Today, I’m talking with Verizon consumer CEO Sowmyanarayan Sampath, who goes by “Sampath.” As CEO of Verizon’s consumer division, Sampath oversees the biggest part of the business: Verizon Wireless, which has more than 115 million wireless connections, and Fios, which adds another 10 million or so consumer broadband connections to the mix. That’s something like a solid third of the entire country in business with Sampath.

As you’ll hear, Sampath is a longtime Verge reader, and in particular he’s been paying attention to my coverage of both 5G and the net neutrality debates for a long time. So we talked very directly about whether the huge 5G investment had actually paid off — whether the “race” we were supposedly in with China was actually worth it, and what kinds of new apps and services actually come to light. You’ll hear Sampath say that network capacity isn’t a small thing, and 5G really helped with the explosion of mobile data — although I was told we’d all be doing robot surgery in driverless cars by now.

The mechanics of all that buildout are fascinating to me, regardless of the hype. Building network infrastructure takes years — everything from figuring out what spectrum to bid on to making deals for towers and running fiber to them is a huge process, and I wanted to ask Sampath how he balances all those very long-term, very tactical decisions against the decisions we all see and feel in very obvious ways, like consumer pricing. Sampath had one of the very clearest answers to the classic Decoder question about decision making here — like I said, you can tell he was very ready to be on this show.

Sampath and I also spent a while talking about the push and pull of integrating content and applications directly into the network. That’s something that all the big wireless carriers flirted with in various ways, especially in and around the peak net neutrality debate. In 2018, AT&T bought HBO and CNN parent Time Warner, right after Verizon bought Yahoo and started something called go90. For one brief extremely weird moment there, Verizon even owned Tumblr.

The goal of all these media plays was to differentiate the various networks from each other, and it basically didn’t work for anyone — even though Verizon in particular fought the various net neutrality regulations that would have prevented this sideshow in the first place. Sampath and I got pretty far into the weeds on that. You’ll hear him say he thinks no one cares about net neutrality, and, well — I disagreed with him.

And while Verizon has always fought tooth and nail against the sort of “normal” regulations like net neutrality, the second Trump administration isn’t nearly as hands-off when it comes to things like holding up deals because of DEI policies — something that’s happening to Verizon right now. So I had to ask Sampath if he was going to push back on that kind of clear government overreach as hard as Verizon has pushed back in the past.

This one is all kinds of wonky, because it’s two telecom nerds going at it, so some notes before we start. You’re going to hear me talk about local loop unbundling, which is when multiple internet providers can use the same fiber to reach consumers. We don’t really have it in the United States — there are not four different ISPs all competing to reach your house on the same mile of cable or fiber — but many other countries do, as a way to increase broadband competition.

The other thing you’ll hear us get into is something called OpenRAN. That’s an initiative the US government committed some money to last year that’s supposed to open up competition in the broadband hardware space and get US providers off of Huawei components. It’s a little complicated, but it’s a big bet in the mobile industry, and I wanted to hear Sampath’s take on it.

Okay: Verizon Consumer CEO Sowmyanarayan Sampath. Here we go.

This interview has been lightly edited for length and clarity.

Sowmyanarayan Sampath, you’re the CEO of Verizon Consumer Group. Welcome to Decoder.

Good to see you, Nilay.

I am very excited to talk to you. It’s a topsy-turvy time in America, but telco stocks are doing great. Everyone needs internet service, and there’s a really interesting dynamic there that I want to explore with you.

But first, we don’t get a lot of telecom CEOs on the show. It’s hard to understand how telecom companies are structured. I want to start with the Decoder question because all of these companies have been built by mergers, spinoffs, sales, and recombinations. You were the CEO of Verizon’s business unit. Now you’re the CEO of the consumer group. Explain how all of this is structured and what the Verizon consumer actually is.

The way we think about it is, we are customer-facing. We have a Verizon Business group, which takes care of all businesses — small businesses, government, large enterprises, anything from Walmart down to Joe’s Pizza. We take care of that.

Then everyone else is a consumer. It’s you and me in our private lives. The wireless service, the fiber service, the broadband to the home — all that’s in consumer. The consumer business is a little north of $105 billion in revenue. I’m responsible for sales, marketing, product offering, customer care, and digital. It’s an end-to-end unit getting things done. Most telcos, at least Verizon, have a common network that serves all of us. We call it the Intelligent Edge Network. It’s a single network that serves different types of underlying networks and businesses, so we don’t multiply costs in it. That’s it. It’s a pretty simple structure. Go to market for business, go to market for consumer, and a common network that is a shared service that serves all of us.

I’ve spent a lot of time covering telecom companies, and in particular, wireless and mobile companies. One thing that always feels opaque to me is the differentiation that you can do to a business or a customer based on the network. This was a big fight in the net neutrality days, that you’d want to differentiate the actual network for different customers in different ways, and that has waned.

When you say you go to market for the customer, are you in charge of things like pricing and promotion, or do you actually get to say, “Here’s what the network can and cannot do?”

Here’s what the network can and cannot do. It’s the whole end-to-end offering, including the product. The way we think about it is that a product is a network. Look, there will always be a good operator and a non-good operator. It’s historically always been the same. For a country as large as America, you’ll always… Look at the end of the day, we and my peers use the same steel to make the towers. We use common tower companies to take care of our tower business. We use pretty similar vendors. We use Ericsson and Samsung. Some may use Nokia. There’s a commonality there. But the secret sauce is in the operations. When you’re running a network this big, how do you take care of outages? How do you plan for it? And even small things like how you design your tower or the specs you put on your tower?

For example, almost all our sites have generator backup. Whenever a hurricane falls or there’s a natural disaster, our sites are up and running because we have backups on our sites. Second, we like resiliency on our sites. There will always be a better operator and not a good operator. We know it. We’ve got 25 years of experience or 30 years, depending on how you count it in the space. There will always be a gap. Now, is the gap narrowing between us and the second network? Probably yes. But there’ll always be a gap, and that’s why people always choose us for the network. If you go and ask people who’s the better network, they’ll always end up saying Verizon. We have to stay on top of it by using more AI in our operations. Self-healing networks. Spectrum is another thing. We have to keep buying a lot of spectrum to put it… There is always a case for a better operator and a better network.

I think that’s the heart of my question. Verizon, as a company, in a very reductive way, has to go to the government, lease a bunch of spectrum capacity, and go to a bunch of tower operators. I always think it’s funny that the biggest one is called TowerCo. It’s just a very literal name that no one except you has to consider. You have to go to a bunch of tower companies and site operators, give them the spectrum so they can operate the towers, and lease the capacity back from them to run your network.

Your competitors might be on the same infrastructure or the same sites in many cases because it’s so hard to build a cell tower in many places. Then, somehow, at the end of that, the consumer perceives Verizon’s network as the best one. That is as much a function of consumer marketing and experience… When I moved to New York in the early 2010s, everybody had a Blackberry and Verizon, and that impression has stuck with me for all this time.

Maybe that’s still the case that Verizon is the best, but what happened was I moved to New York in the 2010s, and Cingular had the best network in Chicago at that time (where I moved from). Then Verizon was the best on the East Coast. I spend a lot of time thinking, “Where does that impression actually come from?” Because no customer holds two phones in their hands and says, “This one has better service.”

There are two ways. One is, is it a better network, and then the other is, how do people perceive it? We talk about national networks. We are the largest network because more customers are on us, but the end of the network is a very local thing. The vast majority of our customers, I would say 60 or 70 percent of our customers, never go outside a five-cell cell site travel zone for the year. Think about it, you may have a great network in Alaska, but 70 percent of your customers are not traveling within five or six-cell sites outside their home or their workplace. It’s a very local conversation about how good you are.

Of course, there are some people who travel a lot. People like me who are on the road 150 days per year. It’s a different equation. But for most people, how good is the network at home? How good is the network at work? How good is the network during their commute? And that’s what makes the experience. You cannot have national standards to do these things. A lot of it is a local game. Getting access to sites locally. Getting access to better network technology at the local level. Having people who know the network topology at the local level. It’s a very local game, this network thing. And that’s why you see Verizon tends to do well when local networks come into play.

How much impact do you have on the capital expenditure of building the network itself? You have a number of customers, and the business side has a number of customers. Do you jointly show up in front of the core network team and say, “You need to build more capacity in Connecticut”?

The consumer group has 85 to 90 percent of all the usage on the network. Maybe it’s a slightly stilted conversation there, but the way it works is… Network is a little bit like scotch. It takes a long time to really become good. Decisions we make today on the network tend to show up three, four, five, or seven years later down the line because there’s a huge lead time to it. You’ve got to procure the fiber, backhaul the sites, and the spectrum to do that. The way we start is by asking, “What is the network I need to deliver my customer’s needs?” And it’s at a local level. In Philadelphia, by markets, what do I know? How much share do I have? What am I going to grow? Then we take that back to the network team and say, “Based on what assets you have, what’s available? Can you deliver this?”

And 9 out of 10 times, they can deliver it. The one out of 10 times they can’t, then we move our marketing plans… If they don’t have capacity, they’re not going to get capacity in this market in the next few years. The second thing is having objective metrics of network performance. For a common person, if a call drops or a network data session goes down, it’s bad. But we track it a million different ways by design. Having that common nomenclature… Back in the day, it was dropped calls. Thinking, “How many drop calls did you have?” We used to measure every single one of them, and then if a drop call happened, we used to go and do root cause analysis. Now it’s data sessions. How many data sessions do you have? When did it drop? What were the speeds? How did the signal flow?

Having really good instrumentation at the local level is really important. We’ve created basically a digital twin for every single customer. Every customer has a digital twin in the network where we measure the network 24/7, and we give them a score. Nilay had a score, it’s out of 10. You can say Nilay had a 9.7 score this morning on his commute from A to B. It’s a good score. If he had a score of five, we would go and understand why the score of five happened. Every single customer is monitored 24/7, from a network performance perspective, and then we track them through the system.

When you look at that kind of data at scale, you obviously have lots of customers who, I’m guessing, have lots of different scores across the board. Where’s the median? Where do you want people to be?

You want people to be in the eights and nines for the vast part of their lives. Now, one of the things with the network is… A bigger issue we have to face, as an industry, is that when things work well, nobody thinks of us. When things don’t… It’s a little bit like parenting sometimes. When the kids are happy, they don’t think of us. Then, when they’re less than happy, they think of us. It’s a little bit like that, because 90-something percent of the time, the network works flawlessly. Whether we’re discussing the fiber in the home or the wireless network. And we are largely forgotten, which is good. But when it doesn’t work, all the emotions come out in front. We are trying to solve for that one percent where the network doesn’t work, or we’ve disappointed the customer. Most of our work is isolating where it doesn’t work. Because where it works, we know it, and we just have to let the machine run its course.

We’ve talked a lot about the mobile network. You also have oversight of Fios, right? The fiber buildout. Verizon has been up and down with Fios over the 15 years that I have covered this company. There are huge investments, huge government-backed investments, disinvestment, and reinvestment. Where are you with Fios now?

We love fiber. Some of it is personal. Look, I built my career on fiber. In every role I’ve had at Verizon, we ended up building more fiber. Verizon was the first. In 2004, we created the Fios network. It was the first, probably in the world, to have fiber. We slowed down a bit in the middle; it was mostly a capital allocation issue for us. We always liked the business. We committed to it at an OPEX level, but we had to invest in 4G. We had to buy the spectrum to invest in 4G. Then we had to buy the spectrum for 5G. Some of the capital that we would’ve put into fiber then, we ended up moving to the wireless side of the business. Now that the 5G is towards the back end of its build, we are back in the fiber business.

We bought Frontier, and we are going to close before the year’s end, or maybe in the first quarter of 2026. That’ll give us another 10 million homes. Nilay, we’re going to have a path to get to 40 million homes of fiber. And fiber is so resilient. The churn is very low. The customer satisfaction is crazy high. Then, when you put the two together, mobile and home, the economics are fabulous. We are really committed to a converged strategy on fiber plus mobility, and that’s the business we are in. We are not distracted as a company. It’s very simple. People ask me, “What do you do?” I say, “I connect things, I connect people.” Full stop. There’s no other business we want to be in, and fiber is part of that story.

Well, I’m on a Fios connection now. So if this drops, it’s on you. Just letting you know.

Nilay, it ain’t dropping. I’m on a Fios connection too, and it ain’t dropping.

Describe the structure of Verizon broadly. You’ve laid it out in a big way, right? There’s the network, there’s business, and there’s consumer. How many people do you have in consumer, and how is that actually structured inside Verizon?

We work with partners, as well as… A lot of our call centers have partners. Some of our stores have partners. We have probably around 30,000-ish people in the consumer unit broadly. I probably end up managing, directly or indirectly, the 30,000 people between our call center partners and our agents to do the work that Verizon consumer is tasked with. The way we are structured internally in Consumer is that we have a gentleman, Frank Boulben, who runs revenue for us. He has all the product offerings. The go-to-market sits with him.

Then we have Kevin [Zavaglia], who runs our field. We are the sixth-largest retailer in America. We have 8,000 exclusive stores between our brands. Kevin runs all of that, plus our relationships on the distribution side. Then we have Nancy Clark, who runs our prepaid business. It’s a fully integrated business. We’ve turned around that business really well. We should touch on that at some point. She runs that. Monica Hammond runs all our customer service, customer care, and customer success organization. And Brian Higgins runs customer experience and digital. A very typical structure that we have. I think most telcos end up doing something quite similar to this.

Is there any innovation in that structure? We have a lot of startup founders come on the show that have big ideas about structure, and then maybe they land in one of two or three different ones. But then you see some companies show up and say, “Okay, there’s actually efficiency from rethinking the structure.” Telcos are big, regulated companies. Obviously, that shapes how you would run a company like Verizon. But do you see any efficiency there compared to your big competitors like AT&T or T-Mobile?

I was a management consultant. I was with a Boston Consulting Group for a very, very long time before this. We had a strong thesis on organization and structure. Then I became a day-to-day operator. I think as long as your structure is not crazy stupid, it works. It all comes down to people and taking friction out of the process. Because any structure will have a seam in it, otherwise the president of the country should be running every single thing on a single day-to-day basis. It doesn’t work. When you break it up into a company, break it up into divisions, you’ll have a seam everywhere. The question is, how do you manage the seam? Because every time you ship a product out, you ship your org structure. We know it. Okay? The question is, where are the seams, and how do you take the friction of those seams out?

For example, customer care and digital. One is the digital channel. One is a human being getting called. When you call up, does the lady in customer care know what you did in digital five minutes ago? There’s a seam in the organization. You have to go and manage those seams, not get too bent about who reports to whom and what structure it is. And I think people tend to change structures a lot, hoping they’re going to get magic out of it. But I think the magic is in friction, seam management, and never ever shipping your org structure on your product. Your product should not look like your org.

The other big Decoder question I ask everybody is about decisions. You have a lot of decisions to make. Everything from going to try to buy spectrum, to customer-facing decisions. How do you make decisions? What’s your framework?

I have a framework for decision-making. It’s a two-by-two matrix. Of course, any recovering consultant has to tell you that. On the X axis, it’s what I call a one-way decision or a two-way decision. One-way decisions for me are really big and very difficult to unwind. It’s things like buying other companies. Even some big leadership decisions, long-term pricing decisions, and big product categories that you want to be in. These are one-way decisions. The cost to unwind is very high, very painful, and will take a toll on the company and its finances. That’s one. Then I have what I call two-way decisions. It’s a promotion. It’s moving a few people around in the markets. It’s trying different things with retailers, trying different compensation things. These are two-way decisions. They can be reversed. Damage can be limited. You could make a bad decision, but you can limit the damage.

Then on the other axis, I have low dollars and high dollars. Low dollar, I have a certain threshold in my head. These are low-dollar decisions. And then you have high-dollar decisions. You never want to spend time with one-way, low-dollar decisions. They’re just bad. You should never even have one of those, but people spend time in companies on that. I work very hard not to make those decisions at all. Because if something is a one-way decision and big enough, it had better have a high impact. If it’s low impact, you shouldn’t do it because there’s risk in it, and it’s not worth it. Anything that is a two-way decision, that is also low-dollar, you should delegate it. Someone else can make the decision, you’ll be fine. And you have to trust the governance model, that there are checks and balances, so nothing bad happens.

I try to spend 60 percent of my time on one-way decisions that are high-dollar value. Where to invest? What categories to do? Who are the big leadership positions I want to fill with talent? And I map every decision down to this framework, and I actively manage it. Just moving from a staff role into a large ops role, the best way to understand how a company runs is to follow the decision-making style. How do decisions get made? Who makes them? Who has the pocket vetoes? You can understand every single thing about a group you’ve inherited by just following decisions. Take five decisions, and track them end-to-end. That’s it. That’s all you need to know. You know about people. You know about pocket vetoes. You know who the naysayers are. You know who’s not ambitious, and you know about talent. For me, decision-making is a science and an art, and one that I practice really aggressively every time I can.

That was the best answer to the decision-making question we’ve ever gotten, but that’s the thesis of the show, right? Inside the question about structure and decision-making, I can tell you about 80 percent of your company. I know it from the jump, and then we can go on to the actual specifics.

But yours is an interesting answer because you’ve devalued the structure question, and you’ve raised the value of the decision question. Most folks do it the other way around. Why make the flip?

Because I think organization is a people construct. When people like each other, when they feel like their work is well taken care of and appreciated, and they’re in a well-supported framework, they’ll give it their best. I think people use structure sometimes to pit people against each other and give conflicting priorities to different people, thinking they’re going to get the best answer. I do not believe in that. I strongly believe… I have extra time sometimes. Rare, but I do have extra time. This afternoon from 4PM to 6PM, I don’t have anything on my calendar. I will go back and say, “Where can I take the friction out of the organization?” Because when my role is… The best use of my time is when I take the friction out. For example, there may be friction between two groups. I go sit down with them, “Tell me, why is there friction? Oh, we have an issue with priorities. Let’s go sort it out.”

The minute we take the friction out, people work well together. Organization gets devalued completely. Then it all comes back to decision-making. And second is… Organization is always a moment-in-time issue because priorities change, people change, and organizations change. You can’t use an org structure to drive performance in the field and priorities. I like decision-making as a more powerful tool to do that. And the second is, how much data is enough to make a decision? This is something, I think, a lot more material should be written about. I promise I will not write a book on this when I retire. People think there are some groups that think they need 100 percent of the data to make a decision. Some make it on a gut. I don’t think both are right or wrong, but we also need a good framework for that as well. Like, “When are you ready to pull the trigger on things?”

It’s difficult in big companies, but it’s also very difficult in startups, and I think it’s a very different construct. Because in my business, I know the market, I know the customer. I have 100 years’ worth of customer experience and customer elasticity models to work with. It’s a very different thing. When in startup mode, you’re creating a category. TAM doesn’t exist. You’re stretching PowerPoint to make your TAM real. You have to rely more on the gut to make those calls. When it comes to people decisions, you can put them through evaluation, but you have to make decisions based on gut [feelings]. But when we’re making a pricing promotion, or product-fit decisions in my world, I think it’s a little more data-driven. That is another area where I think I’ll become a better manager if I practice more of that.

Let’s put some of this into practice. Let’s just walk through an easy one here. Verizon just announced a three-year price lock for customers. This is right on top of the Trump tariffs. I always wonder how these decisions get made. A long, long time ago, I asked another network executive, “Why is this pricing like this?” And the answer — this is a real answer that I received from your other big competitor, but a decade ago — was, “Well, this is how the computer system works. To make it different, we would have to reprogram a bunch of billing systems, and that’s costly. This is how our pricing works.”

It was a very honest answer. It was also somewhat disillusioning, in the moment, that the pricing of the network plan was more related to the structure of the computer system that ran the billing than anything customers could perceive. I’m assuming it’s been a decade. I’m assuming everyone has upgraded their billing systems. But you announced this three-year price lock right on top of the tariffs. The marketing is that you want to give customers predictability. Lots of people have lots of feelings about wireless pricing. I have lots of feelings about wireless pricing. But that decision, you’re going to lock rates for three years, that touches your marketing, revenue, and forecasting. How do you make that kind of decision?

We are in the middle of a large turnaround in the consumer business. Verizon used to take the most share of the growth. We had the lowest churn in ‘21. In 2021 and ‘22, we lost our way a little bit. My chairman, Hans, spoke about it. We got sloppy. We know it. Since then, we have been back on the turnaround track to do that. It’s a three-step turnaround. Step one for me was to fix sales. Fix our stores, fix our compensation system. Become a local company. We came to national… Everything was decided in this building. It didn’t make sense for a company our size. Step two was to take care of a price premium.

We had a 40 percent price premium to the market on our plans. I think that is unsustainable for us. Look, we had a better network, we knew it, but 40 percent was too much for us. So we tweaked it. We are now down to 10 percent, 10 to 15 percent, depending on how we look at it. I think it is sustainable. We are growing our non-connectivity business really well. The third journey for me was to get my churn back in shape. Churn is the number of customers who end up leaving us every single month. It’s still very low. It’s like 0.7 or 0.8. So think about it. Nine percent of our customers leave us every year, which means our average customer stays with us for 11 years, which is crazy, by the way. And in Fios, that number is close to 20 or 30 years, depending on how you look at it.

It’s a loyal base, but you want churn to go down every single time because it’s expensive to acquire customers. We spend a lot of money upfront acquiring them. For me, it was how do you get churn in a much better place? We went and asked customers. For the last four quarters, we polled customers. Nilay, I’ll give you a great tool that we built with AI, where… It’s an internal tool we built. Every day, I get half a million calls coming into my customer care centers. Every single day, they come in. And we take every one of those calls, transcribe them, and put them in a generative AI machine. At 9AM, I get a sheet of paper that tells me the sentiments, key themes, and what happened with all of yesterday’s calls. It is mind-blowing because earlier [before this tool], I had to call people. We had analysts who used to listen to calls, transcribe them, and make notes.

Now every single call is put in a machine, and I get a piece. We picked up two very clear themes when I did that. The first one for me was certainty. People are like, “Everything in my life is uncertain right now. There’s uncertainty about the economy. There’s uncertainty about my job. There’s uncertainty about work from home. There’s just a lot of tension. I want certainty with prices, especially with categories I cannot live without. And phone is one of those categories.” Certainty kept coming up again and again in our reads from the market. The second thing that came up was flexibility. I never want to pay for something I don’t use. That’s one of the reasons why we took out all those inclusions. You get this free, you get that free. And people are like, “I know Disney is good. My kids love Disney, but I don’t watch Disney. Why should I have to pay for it?” Those were the two really clear insights we got from customers. That’s one.

Second, we’ve done a series of price increases over the last couple of years. We had a couple in January. We know elasticity. When we do a price up, we model it. I would say the last price up we did, the elasticity was a little more than what we had modeled. That, coupled with all the insights we got from the customer, we had to respond and tell the customers that we are going to lock prices for three years. Now, it gives customers certainty. It puts a little more pressure back on us because we have to drive revenue in other ways. We have clear ways to do that, but I think it’s an industry first, especially now. Look, we knew the tariffs were coming in the day before yesterday. It was announced… It’s on TV. We are very comfortable doing that because, at a time of uncertainty, we want to give our customers some certainty and some peace of mind.

All right, I’m going to be pretty hard on you now. I have covered the telecom industry for a long time. I noticed yesterday [April 3] that every other stock in the world dropped, and telecoms went up. Because people have to have your service, right? It is a requirement of modernity to be connected to the internet and to have a phone. And it feels like one of the reasons fiber customers stick around for a long time is that there’s nowhere else to go, right? The switching costs are very high, or maybe infinite, because there’s literally no other competitor that can offer the service that Fios offers, in many places. 

That’s true for every residential broadband provider, as competition is generally low. In mobile, the competition is basically you and AT&T, and then T-Mobile is a very strong challenger that’s gotten bigger. But Verizon and AT&T are the biggest players still.

There was supposed to be a fourth player, and the government created this Dish Network deal, which I will ask you about soon. That has gone nowhere, right? There’s not actually a fourth national carrier that competes with you. When you say customers stick around for a long time, what I see is, well, there’s nowhere else to go, right? How much does that come to your thinking that, actually, a Fios customer can’t leave?

A Fios customer can leave tomorrow morning. Every single Fios customer today can move to cable broadband that afternoon. There are very few Fios customers who don’t have a cable competitor in play. I just think Fios is a better solution. It’s a significantly better solution than cable because of the pricing practices they’ve had and the quality of their network. But there is a full-

I hear you, and I’m sorry to interrupt, but I’ve heard that line from many, many, many telco executives in the United States. I look at the United Kingdom, for example, where they’ve done local loop unbundling. And, actually, my colleagues who live in London have 15 fiber providers at much lower prices and much higher speeds because the presence of competition is so much higher, right? It’s not literally switching from fiber to cable, a different kind of network, it’s different providers within the same fiber network offering more competition. We’ve not done that here, but they’re always switching, and you can see the prices come down because of it.

I think it’s a careful balance. Canada has moved to a local loop unbundling scenario as well. The UK has always had it. That’s why European telecom has one of the lowest investments in telecom ever, in any country, almost in any region of the world today, because they just don’t have enough of a profit pool and cash flow to reinvest back in their network. Think about 5G. We are going to get to 90 percent deployment in a country as large and diverse as us. Most European countries are not even at 40 to 50 percent deployment of 5G. I think one of the things, if America wants to be competitive, if America wants to compete with China, having high-quality broadband is super important for that. Whether it’s mobility or whether it’s that. And you want to create profit pools in the system to go back and reinvest.

Think about it. I’m going to invest close to $18.5 billion in just capital. In the last seven years, we’ve invested $200 billion in the network. $200 billion between spectrum and fiber, and the mobility business. I don’t see any European carrier investing $200 billion in seven years. In fact, if they invest a 10th of that, they’re going to get an award. I think it’s super important for a country to be really at the cutting edge of technology, the cutting edge of fiber, the cutting edge of 5G. And look, when 6G comes, we will be the first. I’ll tell you that today. We were the first in 5G. We beat the Koreans by 12 hours.

We will beat whoever else… And most likely it’s going to be China on 6G as well. I think it’s a national security issue. It’s a national competitive issue. We need to keep reinvesting in broadband every single time, and we are not going to stop. You’ll see us investing at the same rate for a very, very long time. And all our investment is in America, in LA. I’m not investing in other countries. I’m 100 percent invested in America, and I think the profit pool lets us do that.

All right, we’re here at the end, right? 5G is done, it’s deployed. Everybody has the 5G phones. What were the stakes of losing the race to 5G? No one ever gave me an answer to this question. If China had somehow won the 5G race, what would they have gotten that we didn’t get?

What does it mean to lose the race? I think the first thing is standards. You want a standard that… You don’t want a bifurcation of standards, and it may happen. It could happen in the chipset, it could happen in AI, or operating systems and handsets. You’re in an interesting phase where you could potentially see a bifurcation of standards between the Western world and China. On 5G, we aligned on a very similar standard, and I think it helped because you’re able to take the price of the technology down to do that. The second is to think of all the applications and the tech economy that you have today. The app ecosystem, the Amazons of the world, the Googles of the world, and the Metas. They rely on the mobility network more than anyone else. Think about Instagram, it is worth zero without the cell phone and the connectivity. In fact-

No, but specifically 5G. Instagram was worth a lot on the LTE network, and most of the 5G networks we have today are not running on millimeter wave 5G or whatever. They’re running on LTE enhanced, right? We’re still running… The core of the network here in the United States is a hybrid of 4G and 5G.

No, if we had stuck with just LTE, we would not be able to take care of the capacity that’s happened. Think about it, the pandemic happened, we all moved to working from home, people… We were able to take care of those networks immediately because we had 5G; we could scale up. So 5G adds capacity in a way 4G could never have added because it comes down to spectral efficiency to do that. And we also ended up adding… Look, at the end of the day, capacity is a very simple thing. You can either add more spectrum, which is expensive, but you can add it. We do it every X years, we do that. Or you can add more cell sites because you can reuse the spectrum, or you can get back to spectral efficiency. That’s the only three ways to add capacity. And 5G helps with one in three.

It puts new spectrum bands into play, and it also helps with spectral efficiency. You got a lot more capacity out of 5G, and also took the cost of capacity down. Every incremental bit that abides, depending on how you put it up in the air, the cost goes down significantly. We saw a 30 to 40 percent reduction in cost per bit when you move from 4G to 5G. Without 5G, we could not have done that. The second is that we created a whole business around fixed wireless access, providing broadband by using wireless. It’s all on 5G. It ain’t on 4G, I’ll tell you that. It’s a multi-billion-dollar business for us. In a few years… Right now, probably 50 to 60 million homes could potentially get that. It’s another competitor to broadband. And what has happened is that cable has not been able to raise prices, and so it’s a lot more beneficial to customers because they have another competitor in the space to do that.

The third thing, which I think has been a little underwhelming for me, personally, is new products and services using some of the 5G capabilities. And you never get this right because you have to get the network, the handset, and the apps all perfectly aligned. I don’t think that happened very well on 5G. I think the third leg of the stool still has more work to do. I think a lot of the AR/VR type new form factors that we are seeing are going to require really low latency, and that will require 5G. You definitely can’t do it on 4G. Sitting where I am, I’m so happy we did 5G. I wish we had done it two years earlier, but that’s always the scenario. And look, we’ll keep building our network, adding more capacity till 6G is up.

I sat in all the demos of all the hyped 5G applications. I went on so many self-driving car rides at CES, powered by 5G, with the data center somewhere that was driving the car for me because the latency was low. I heard a lot about robot surgery — an infinite amount about robot surgery. Those were the applications that were promised. That was the vision of the war: that if we lose the war, China will own this set of industries that is built on the new network. It was not about network capacity in COVID, right? No one foresaw that what we would need is a more efficiently run network.

It was not really about bringing prices down in any substantial way, because the only thing that ever happens is prices go up. That’s the spread, right? I’m just wondering, do you see the gap in how much hype there was for 5G and how oversold it was? And also the consumer reality of it, which is, yeah, everyone’s phones might have a little more network capacity. Yes, it’s great at the concert or the football game to have millimeter wave serving everyone really fast. I agree that that’s a great use case. But it actually didn’t pan out that the big set of applications we were promised, to justify this enormous cost and all of this literal war hype. It never came to pass.

Yeah, I think wireless is one of the categories that has had negative inflation over the last 15 years. Think about it. When you had 300 minutes on your phone, you were allowed… Evenings and nights were free. You remember the time when you got maybe 300 texts.

You were paying more for that than you’re paying today when you get a 5G phone with unlimited data, unlimited text, and satellite messaging included. There are very few categories… In fact, the only other category I can think of that has had negative inflation over the last 15 years is TVs. Flat-screen TVs have really crashed in price. TVs and us are probably the two biggest categories that have gone up. Everything else, prices have gone up significantly. All the way from eggs to others, everything has gone up. Our prices have actually come down for what we provide there. I think the industry is probably one of the most unique industries where we tend to give you a lot more for the same amount or a little less than we did 2, 3, 5, or 10 years ago to do that. The second thing is I do… Technology is interesting. For decades, nothing happens, and then in months, decades happen.

You saw that with AI, right? You and I have been listening to AI ever since we were in college, and then suddenly, two years or a year and a half ago, the whole thing exploded. I think a lot of these special cases, new applications that we are working on, tend to be like that as well. It takes a while to work it. Look, I don’t think autonomous cars or robotic surgery are going to be the ones that are going to break the bank or are going to be the breakthrough services. But I see new form factors emerging. We may be in the last generation of phone form factor as we know it right now. I think you’re going to see a lot more control of the phones using LLM models.

I think it’ll have huge implications for the operating system and huge implications on the app store. But I think this may be the last generation of smartphones that we see. And to power all of the other things, we just need a better network. The second thing is that we are doing a lot more work on 5G in the industrial space. It’s not something you and I end up using a lot, but in my old job, we used to do a lot of private networks, low-latency networks, running robotics, and factory automation based on a wireless network. There’s a lot more happening in that space. When you roll out such transformational technologies, you’ve got to play the long game. Look, Verizon is a company. We are committed to America. Look, we don’t do business in other parts of the world except for small business. We are committed to America, and we want Americans to have the best network, and we’ll keep deploying capital to make that happen.

I think most consumers saw 5G as something that just sort of happened to them, right? They upgraded the phone, the network is there, and then I hear you on negative inflation and value over time, but the main thing that happened is you upgraded your iPhone. Now it’s 5G capable. You need a 5G plan, and that costs more than the plan you had yesterday.

That is not true.

But that was true for a vast majority of people across the vast majority of the industry. The plans did get more expensive.

Actually, it didn’t. If you see phone ARPU, average revenue per user, which every carrier reports in some form or the other. ARPU has moved very little on a per-phone-line basis in the last 20 years that I’ve been in the space. And you’re getting a lot more now than you got many years ago. I think people have this notion that the phone bill is going up, but the phone bill is going up because they’re signing up for new services. For example, in my case, we are going to have a $2 billion business selling streaming services and entertainment services. Now, we give great savings. We give you a 40 percent discount when you buy Disney or Netflix, Max, and all the cool services from us. But the core phone service, as you know it, has really not gone up in price.

And we’ve delivered more than pretty much any other industry for the price point that we do. And I’m very, very comfortable with that equation. It’s my job as the CEO of the business to ensure that we give more to our customers. Satellite messaging is one. It’s something… We found a way, we made it work, and we included it in our plans. You pay zero for it. I think we are adding significant value to customers, which is why our churn is so low.

People can move wireless easily, but because of the value we bring to them every single day, they like us and they stay with us.

One of the things that’s interesting there is the comparison to televisions, which you made. The reason televisions got so cheap is that the industry realized that putting a computer in the TV let them develop huge new revenue models, right? A $500 TCL TV is actually a revenue generator because there’s advertising baked into every layer of that TV now, or effectively a free Roku box. Roku owns the entire ad stack. That’s how they make all their revenue. It’s easy for them to subsidize the TVs in that way. That’s not really the case with mobile handsets, right? It might be the case on maybe the low end of Android phones. There’s some subsidy happening there. With the iPhone 16 Pro, Apple’s not going to let you monetize that device as directly as anybody wants to. The push has always been to go up the stack, right? To go into the application layer. 

A decade ago, we had a fight about net neutrality where the carriers really wanted to be in the application layer and modify the network to prefer some applications over others. That has petered out. I think it burbles up every now and again, but it’s mostly petered out.

Now you’re saying the opportunity is to bundle Disney+, right? To bundle other applications. Is that settled now that that is the strategy? Or is the push into developing custom applications for your network, or the push to prefer bandwidth for some applications? Is that incentive still there?

I think what does Verizon bring to the table? We bring distribution. We bring billing and customer care. We know how to manage subscriptions. We bring a brand that has got unaided awareness of close to 100 percent. Those are the big things we bring. And I think what we’ve decided is that we want to build our non-connectivity. My non-connectivity business is almost $15 billion. It’s the size of Hilton Hotels. My non-connectivity business is a standalone company. And we partner with the best in the world to do that. I partner with Netflix. I partner with Disney. We have an insurance protection product inside that. We partner with other folks to do that piece. I think we want to use our distribution, our ability to scale subscriptions, our ability to lower the cost of acquisition, and partner with the best. But second, the more people use my network, the more excited I get.

We are working very, very closely with the likes of Amazon, Meta, Google, and Apple to ensure that they use our network more to develop more services because we will monetize it at the network layer really well. Then, on top of the network, we’ll find adjacencies to sell in. But if people don’t use my network more, all the cool services I build ain’t going to matter much. I think we’ve pivoted to a strategy, which is single network, scaled network, put as much traffic on it as people can, and we’ll find ways to monetize it in a pricing construct if we do that. On top of that, we have adjacencies. We have six categories we sell into, and there we use our scale distribution, and some integration back with our network to make it real to do that. I think, in a way, we are probably the largest incubator in the world today because we work closely with companies for them to go and innovate services. It’s a pretty cool position to be in. And as we finish deploying 5G, we’ll start seeing more and more of those cool services come out.

One of the things that I think about is that there was just an explosion of big carriers buying media, right? Because you have the distribution, you have the billing. It made sense to vertically integrate all the way up to the top.

At one point, Verizon owned Yahoo, which was very confusing. You started what was like a prototype of TikTok called Go90, which failed almost immediately. It’s a long-running joke in our other show, The Vergecast. You owned HuffPost. You had Tumblr for a second. On the business side, you bought BlueJeans, which is a video conferencing solution.

That notion that you need to actually own the icon that the user is pushing on or the media they’re consuming on the phone. AT&T bought Time Warner, another obvious disaster that collapsed almost immediately. Where did those incentives come from to do that? And do you think that they’ve gone away? Has everyone learned their lesson?

Yeah, and look, it happens in every industry. Look, at the end of the day, you can either grow horizontally or you can go vertically. There are only two ways to grow. We went through a phase… Look, I was on the leadership team of Yahoo and AOL. I was sent from Verizon to help run those assets and eventually sell them. And part of the thesis was that we create data in the telco world, whether it’s clickstream data, location data, or other network data, that, if matched with online data, could create a better targeting engine for advertisers. That was a thesis going in all the way. But then we realized that customers trust us because we don’t share data. And it was a deep, insightful conversation we had within ourselves, and we realized that we should not be in a business sphere, so we sold that business.

But, again, for a company our size, if you look at our enterprise value, it was still a rounding error for us. Fine, it was something we should not have done, but it was still a rounding error for us. At the end of the day, we are focused on our core. Our core tends to be mobility. Our core is fiber, and our core is cloud/AI, enabling those things. That’s it. Those are the only three businesses we are in right now. Carriers around the world have sort of come to a similar place, but we are probably, I would say, the most focused of carriers anywhere in the world right now.

Verizon fights the hardest against any net neutrality rules that appear at the state level and at the federal level. We can set a watch on the Verizon lawsuit against those rules. It happens every time. You’re saying you don’t need those rules, right? If I’m listening to you correctly, you’re saying, “We’ve learned our lesson. We’re not going to monkey with the network. We’re not going to prioritize our own services or someone else’s services. We just want volume,” right? Pure volume is what will drive your growth. Why fight against the rules if you’ve already come to the same conclusion?

I don’t know what net neutrality does. I’ve been in this business for a little north of 20 years. Before that, I was in college, having fun. And I still don’t know what problem we are trying to solve with net neutrality. I actually don’t understand that, because at the end of the day, if there are 300 people using a congested road, and if a firefighter needs to go through, people are going to move aside and let him go through. That’s all we are saying. 

For traffic management purposes, we need to have some controls in the network. Otherwise, when in New York City, when the traffic core or the emergency services want access and it’s congested, I’m going to have to move people off so they get priority service. That’s all it is. And we need traffic management tools. Or you have one person in your town who’s sucking up all the bandwidth, and we have places like that where they use 100 or 200 times the average bandwidth. And because of that, 99 percent of the people are getting a shitty experience. We’ve got to manage the traffic. That’s all we are saying about it. 

I think this was a little bit of drama about nothing. And at the end of the day, we want to build networks and we want people to use them. We just want the ability to manage network management in a way that’s good for the broad set of people. I don’t think we really care about making one group better versus one group worse, which is why I think it got no traction, because I don’t know what problem it was trying to solve.

To be clear, it did get traction. It passed through the FCC under Title II in one administration. It was undone by another administration. I have never encountered a net neutrality rule at the federal or state level that hasn’t allowed for reasonable network management practices, right? That’s always the heart of the rule. It’s really: can you prefer some content and services over others, right? No blocking, no throttling. It’s the heart of the rules, and Verizon has fought against those.

I have seen your competitors do things like sponsor data, right? I’ll let you off the hook for one second. AT&T, when they owned Max, exempted Max from their own data caps, which was just openly unfair, right? It’s just straightforwardly unfair that using Netflix would cost a customer money, and using AT&T’s owned and operated service would not. And that was an advantage that AT&T wanted to give to Max at that time. And that is the sort of thing that isn’t a firefighter on the road, that isn’t one person using all the bandwidth. That is a distortion in the market. That’s what the rule is supposed to correct.

What good did it do for AT&T or for any customer in that scenario? Look, I don’t know what AT&T did, but I think at the end of the day… I don’t think a lot of these folks understand how networks work. Because at the core of these networks is a very complex set of technologies. You have the access network, you have the core, you have peering points, and you have the backbone of the network. Congestion can happen anywhere in the network. For example, if someone doesn’t come and peer with me, they’re going to get congested, but it has nothing to do with me. They’re congested because they chose not to peer with me, and they chose not to peer with me because they didn’t want to pay the $3 or three cents it needed to go and get a port in a particular porting hotel that was needed.

I think the network is a very complex thing. It’s not a single pipe that starts at one end and ends at the other. There are a lot of different interfaces and other things. I think this whole concept that you’re going to unfairly manage someone’s network or give someone an unfair advantage, I don’t think it works that way. The second is an economic incentive piece. Look, in the case of AT&T, again, I’m being an academic here, which is that they bought Max, or whatever it was called back then, and they were willing to give up their dollars because network costs something. They’re able to give that up to drive sales of content. That’s a decision they made. In the end, it didn’t work out that well because it didn’t matter. I think there are economic incentives in place to manage all of this. That’s why I think net neutrality has gone literally nowhere because it doesn’t solve any problem that I think is a problem.

Again, I want to be clear that it’s the law in New York. It’s the law in California. It was the law at the federal level, and it was rolled back. So, it did go somewhere at one point. And the most instructive thing I ever heard about net neutrality was that it was just the arc that Netflix went on. Netflix was once the loudest proponent of these rules. And then several years later, I saw Reed Hastings at the Code conference, he said, I believe it was to Peter Kafka, “Oh, we don’t need to have this fight anymore because we’re so big it doesn’t matter.”

All that indicated to me was they needed the rule and they needed the fight when they were small, to preserve the access. And then they got so big that they didn’t need the rule to give them access because they were so big that taking access away wouldn’t be a problem. That’s the competitive playing field that I think you want at the application layer as well. It’s really hard for people to perceive.

I think they didn’t realize it was a problem because it was not a problem. I think it had nothing to do with size. 

Well, that’s what Reed Hastings said. He said, “I am so big it doesn’t matter.”

I know, but I think he probably said that to cover up something where he made a big to-do about nothing. In the end, it didn’t change anything. If customers want Netflix, they’ll get Netflix. We make shitty content, no one’s going to watch it. I think it’s the core of that issue. He very soon realized… I remember that conference. He realized this was a fight about nothing. At the end of the day, if you make good content, customers want it, they’ll appear, and you’ll get access to it. If you make poor content that nobody wants to watch, nobody’s going to watch it. Irrespective of all the so-called net neutrality rules you have. I think that was the perfect example of how it had nothing to do with size. It just had to do with the fact that they were chasing something that made no sense at that point, and he was the first to admit it, so he could focus his efforts on putting on new shows. By the way, I like their new show, The Residence. It’s a good show.

You live in a different regulatory environment now. You have Brendan Carr, who’s the chair of the FCC. I’m not a fan of Brendan Carr. I think he’s wielding his power irresponsibly, and he knows that I think that.

You mentioned earlier your deal with Frontier, that’s going to go through, and Brendan has been very clear that he will not allow these deals to go through unless you get rid of all your diversity initiatives. And he says there’s an investigation against Verizon on these initiatives. That feels like straight coercion. If Verizon wants to go body up and file lawsuits against the FCC imposing regulatory requirements, this would be where you do it. Are you going to fight that fight, or are you going to roll over?

For Verizon, there are a couple of things. Internally, as a company, certain things are non-negotiable for us. How we treat human beings, human dignity, and ensure that we have the best talent pool when we build and operate the largest consumer telco company in the world. We ain’t walking away from that. I think that’s there. At the end of the day, we have to follow the rules and regulations of the land, and we will work constructively. We have a constructive dialogue with the FCC to ensure that we follow the rules and regulations of the land. That’s exactly the piece we are in right now with them. We got the letter from them. We are working closely with the staffers out there. Whatever the regulations or rules of the land are, we’ll follow them.

Do you think that you’ll fight that as vociferously as you fought net neutrality?

It goes down to the rule of law. Whatever the rule is.

Do you think the FCC commissioner has the authority to tell you how to hire and fire inside your company?

It goes back to the rules and regulations that the administration puts in place. They have certain rules, they work through it, and we have to follow the rules of the way it lays itself out. That’s all we are going to work towards: a constructive dialogue with Chairman Carr on that piece, and we’ll continue to deliver that. But at the end of the day, I want the best talent pool for my company. I will do anything that is needed to get the best operators, the best marketers, and the best network operators in the world for Verizon to do that piece. I’m walking away from that. I’m going to follow that. In the process, I have to follow rules and regulations — we’ll do that. And if there are some changes that have to be made, we will make those changes. But let’s not forget, this is about getting the best talent from us, and we are not walking away one minute from that mission.

Literally, just a few days before we’re talking, I think T-Mobile closed the deal to do a JV on some fiber initiative. But to get there, they had to delete the word diversity from a bunch of its web pages. Is that the level of meddling that you’re willing to accept?

We are still working with the FCC on what the rules are. How we develop the rules. How we execute the rules on that. It’s a work in progress right now. We are still working through that. But look, at the end of the day, we’re going to have to follow the rules of the land. We did that before, we’ve done that. We’re going to do that after that. This is just one of those pieces, we’re going to have to follow the rules that are there.

I want to be very clear, I’m being very explicit about this. When the government passed net neutrality rules, it wasn’t, “We have to follow the rules of the land.” It was, “We are going to file lawsuits for a decade to get out of these rules because we think they’re dumb.” And in this case, you’re saying Brendan Carr, who has been openly censorious, openly chilling of speech, openly hostile to companies because they have diversity initiatives — you’re saying you just have to follow his rules?

We have to follow the rules of the land. I don’t think those are his rules. They are the rules of the land and administration —

Are you going to file a decade’s worth of lawsuits about these rules?

We don’t know. We’re going to work constructively with them to follow the rules that are needed. But at the end of the day… Look, our role is to the stakeholders that we have. My stakeholders are my shareholders, my customers, my employees, and society at large. We have to manage, and we are going to deliver to those stakeholders what’s needed for us, and we will do whatever is needed with the administration to deliver to all the stakeholders. When you run a large company our size, you have to balance the different stakeholders, and we will balance those stakeholders. We’ve done it extremely well in the last 25 years that we’ve been Verizon, and we’ll continue to do that going forward.

Can you put this in your decision matrix? When you look at an administration that shows up and says, “We’re going to start meddling with your deals.” From an agency that traditionally has not done this specific kind of meddling in how you hire and fire, how do you think about that decision in your two-by-two matrix?

This is not the first deal we’ve taken through administrations. We worked with Democrats, we worked with Republicans, we worked with both sets of the administration. We’ve had good success working with both of them on that piece. Every time we take a deal through, there are trade-offs that have to be made. When we bought TracFone, we had trade-offs we had to make with California that were probably not perfect for us as well. I do think there are trade-offs that need to be made in any large transaction. But at the end of the day, you have to follow the rules of the land. I think that’s the most important thing, and we’ve got to manage the stakeholders. And that’s pretty much all we are going to do here.

All right. I want to wrap up by asking you about a very silly thing from the first Trump administration. In the first Trump administration, T-Mobile bought Sprint. And at the time, that set of players in Trump one concocted a deal where Sprint would sell some assets to Dish Network, and we would invest in a thing called Open RAN to get away from Huawei and have more competition at the tower level. These are all good ideas in theory. Have you spent one minute thinking about the competition posed by Dish Network?

Look, we have three large wireless carriers in a very, very competitive market. We have Dish with its Boost network, which is competitive in the prepaid space. They’re a pretty large player in the prepaid space. They’re starting to build postpaid. Then you have the cable companies today, Comcast, Charter, and Cox, all offer wireless services at very competitive prices. In fact, in many cases, the first line is free, and their average ARPU is significantly lower than ours. It is one of the most competitive telco markets in the world today. Three full players, a fourth facility player who is building up, and then a cable company that has 100 percent coverage of the country, which is able to offer wireless. It’s a pretty competitive market that we have in this space right now.

No, but I’m asking you specifically about Dish, right? That was five years ago. I think June or July of 2020. That deal was completed. They had access to T-Mobile’s network for seven years. We’re coming to the end of that. They are supposed to have a next-generation 5G Open RAN that is a national competitor to you by now. Do you think they have a national, competitive Open RAN that causes you to lose any sleep?

They’re a competitor to us. They have, I think, a little less than… The Boost network creates millions and tens of millions of customers —

But that’s just what they bought from Sprint. That was the existing asset they bought from Sprint. That’s not the thing they were meant to build.

They’ve built local networks. And do we compete with them? We compete with them in the space. But think about it, we have three full networks that are building up, and then we have competition. This is one of the most competitive markets. Look, Open RAN is a completely different animal. We should definitely talk about that. Look, over a period of time, Open RAN is a good thing, because it deconstructs the network a little bit. You can have multiple players in a stack. And for folks who are not close to this, today, when we buy equipment from a partner, we end up buying almost the whole stack. The head end that goes on to the tower, all the equipment at the bottom of the tower, is from a single provider. And Open RAN says, “Make those interfaces more standardized so that you can mix and match operators more easily.”

I think philosophically, Open RAN is a good concept. I think more work is needed to get to the performance that makes us comfortable using it at scale. But what we’ve done is we’ve virtualized our networks more than any other carrier in the world. What does that mean? We’ve put generic hardware in our sites, and then the equipment providers have to load their software on it. We think that’s actually a more powerful tool in mix and match, scaling, and loading cost than sometimes de-standardizing or standardizing these interfaces. But over a period of time, it’s important to have more diversity of vendors and have more people play in this space. It’s good for us, it’s good for the economy. I think we are very supportive of O-RAN. We are part of the Alliance. We spend a lot of time, energy, and money on R&D and making those interfaces stable. But I feel it has a little more to go before they’re standardized, and they give you the performance that I get to put on my network.

Well, Sampath, you’ve given us a lot of time. I feel like you should come back and we should talk about Open RAN for a full hour, because I can. I think it is fascinating. We’ll have to have you back soon. Thank you so much for being on Decoder.

Take care, Nilay. Bye.

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

On today’s episode of Decoder, we’re talking tariffs and trade wars. One of the ways I’ve been trying to sort out the chaos of this moment is by talking to the people behind the software that actually runs our global trade systems.

We got a lot of notes from listeners who really appreciated our episode with Altana’s Evan Smith, which laid out how data from his supply chain systems illustrated where manufacturing was moving as globalization starts to break down. So today I wanted to bring back one of my favorite Decoder guests: Flexport CEO Ryan Petersen, whose software manages the logistics of moving things around the world, from the factory to your doorstep. 

Ryan has been on Decoder twice before, once in 2022 and then again in 2023. Check those episodes out; they’re great. What makes Ryan such a fascinating figure is that he’s unique in that he runs a software company in Silicon Valley, but his product is all about the very real and very complicated process of actually shipping things on planes, boats, and trucks. I’ve always enjoyed his perspective to connect the dots between what’s happening on the screen and what’s happening in the real world. In the case of tariffs, that means Ryan has a real-time view of how his customers are dealing with them as products arrive in US ports and go through customs — Flexport is itself a customs broker, so it manages the tariff process pretty directly. 

But nothing about Trump’s tariffs are business as usual, so I really wanted to dig into how they are working in reality — at the highest level, the Trump administration has launched into what feels like a game of chicken with the Chinese government that’s completely upending the stock market and putting industries of all kinds into outright panic. We can all see that, but down on the ground the systems still have to work, good and services still have to move across the ocean, and these tariffs need to actually get paid. You’ll hear Ryan say tariffs are already changing where things come from, and that shipments from Vietnam on Flexport’s platform overtook shipments from China for the first time in the past week.

What you’re also going to hear is that Ryan has basically been in war mode for the past several weeks trying to keep up with all the changes — Flexport is running a literal tariffs live blog for its customers just to try and keep the latest information going out the door. In fact, Trump’s tariffs are so unpredictable that when we sat down to record on Monday, we knew that the numbers we were talking about would change by the time we published this episode on Thursday. And we were right — the tariffs on syringes from China somehow went from 154 percent to 245 percent, and lithium ion batteries shot up to 173 percent.

So instead of specific numbers, Ryan and I really focused on how this system works, how it’s supposed to work, and how it’s working now, if it’s working at all. If you follow him on X, you know that he thinks things are breaking in ways that create nothing but paralysis in the short term and potentially widespread bankruptcies in the long term. 

So I just asked him: is there anything anyone can do right now to pressure the White House to make this make sense on a timetable that saves the US economy from devastation? I’ll let you listen to his answer. Ryan, to his credit, takes a pretty valiant demeanor here — he says he wants to give the people in charge the benefit of the doubt and to keep his cool so he can keep advising his customers. But I think you’ll hear how quickly he’s running out of patience, and he didn’t hold back when it comes to describing the kinds of scary consequences that might happen if things keep escalating at their current rate.

If you’d like to read more on what we talked about in this episode, check out the links below:

• Flexport tariff live blog | Flexport
• US tariffs: how Trump’s tax is hitting Big Tech and beyond | The Verge
• How much will Trump’s tariffs cost US importers? | NYT
• How much are tariffs on Chinese goods? It’s tricky | NYT
• How Trump’s tariff chaos is already changing global trade | Decoder
• Can software simplify the supply chain? Ryan Petersen thinks so | Decoder
• Why Flexport CEO Ryan Petersen took his company back | Decoder
• The US-China decoupling arrives | Axios

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

Today, I’m talking with Vlad Tenev, the cofounder and CEO of Robinhood, which is one of the most well-known consumer finance apps in the world. It started as a way to open up stock trading, but the company’s ambitions have grown over time — and they’re getting even bigger. 

Just a day before Tenev and I talked, Robinhood announced it would soon be offering bank accounts and wealth management services, which would allow Robinhood to be involved with your money at every possible level.

So I was interested to sit down with Tenev and really hash out where Robinhood is going and why he’s so adamant that big ideas, like prediction markets based around everything from sports games to presidential elections, are going to play such a pivotal role in the future of finance. And I really wanted to talk about the responsibilities that come with that role.

Robinhood is on a lot of people’s phones — especially young men — and it’s a quick jump from doing a little bit of casual retail investing to potentially dumping all your money into a bunch of unpredictable, unstable markets. There’s a whole generation out there who might have bought a GameStop share as a joke during the COVID-19 pandemic and are now finding themselves consistently gambling in the crypto and prediction markets.

Tenev and I really dug into some of the complexity around these ideas. For example, you’ll hear him say he thinks of prediction markets as “the news faster” and that there is a meaningful difference between a prediction market guessing if the Lakers will win their next game and just simply placing a bet on DraftKings or FanDuel for the same outcome. You’ll hear him say that prediction markets communicate unique information that reflects reality, rather than just a dressed-up version of gambling that mostly reflects how people feel.

You will also hear my deep skepticism of these ideas. I really pressed Tenev for answers on how he thinks about the risks involved, especially for regular retail investors, and whether the regulatory environment can keep up with that escalating amount of risk. It’s already causing problems: New Jersey and Nevada both ordered Robinhood to halt its prediction markets, and the company’s partner Kalshi launched lawsuits to push back. 

You’ll hear Tenev say that he has no firm idea where any of this might go, but that he fundamentally believes people should get to do whatever they want with their money — and that he wants to position Robinhood as a central destination for all of those transactions. 

That made me curious as to what Tenev sees as Robinhood’s ultimate destination. So I asked him outright: does he think Robinhood is just selling an updated version of the American dream, where you can make the right wager on a prediction, buy the right stock, or invest in the right meme coin to shortcut your way to financial freedom? I won’t spoil it, but I think you’ll find his answer pretty illuminating. 

Like I said, you’re going to hear us disagree quite a bit throughout this episode, but I want to give Tenev credit: he was game to really sit in some of the ambiguity and controversy here and talk about it in ways that many in the crypto and finance worlds simply aren’t. I rather enjoyed this one, and I think you will, too.

Okay: Robinhood CEO Vlad Tenev. Here we go.

This interview has been lightly edited for length and clarity. 

Vlad Tenev, you are the co-founder and CEO of Robinhood. Welcome to Decoder.

Thanks for having me, Nilay.

I have so much to talk to you about. I think I have 900 pages of questions for you. There’s a lot going on in financial services. You all just launched banking services. There’s a lot going on in crypto. I know you’re very interested in prediction markets. I have a million questions about that. So, if you’re game, I’d like to go through the Decoder questions about structure and decision making very fast at the top and then get to the rest. Are you okay with that?

Yeah. I’ll try to be brief.

Because usually I do all this windup, but I think people know what Robinhood is, and I think your ideas about where it’s going are really interesting. Let’s just start with Robinhood. You founded it, then you were the co-CEO for a minute with your co-founder, Baiju Bhatt. He went to become the chief creative officer, you became sole CEO, and then he left the company. Just talk about that set of choices. That’s a pattern we see in startups quite often. How did that all go down?

Yeah, I think we technically started as co-CEOs as well, although in the early stages things are always a little bit murky. And then what happened was in the early days of the company, roles don’t really matter because you’re 10 people and you’re all in one room and you’re doing what’s necessary to make the company win. We both didn’t really have any concrete finance, business, or even technical skills when we started the company.

We met in college. We were actually both physics majors at the time, and we were brought together by our love for physics. We wanted to seek to understand answers to the big questions like what happened before the Big Bang? Or my favorite one: how can we unify general relativity and quantum mechanics? That’s how we became good friends; we were banging our heads against the wall trying to do problem sets in physics and math in college.

Then we learned to build products and build businesses together. We were co-founders of a few companies before Robinhood. When we started Robinhood, we moved to San Francisco together. He became a designer. He literally bought a Wacom tablet and started designing. And I became an engineer. I still remember on the Caltrain ride from San Francisco down to Palo Alto, I would watch Paul Hagerty’s iOS development classes at 2X speed, and that’s how I learned to be an iOS engineer. So that was the division of labor; I was writing code, he was doing the designs. We built a team around that. Then we became managers, and then we became executives. I think at each point, we reassessed what we wanted to be spending our time on, how we could add the most value to the company.

When it became clear that we’re going to have executives and we’re going to be a public company, Baiju made the decision that he didn’t want to be the CEO of a public company — that didn’t feel like how he wanted to spend his time and his energy. So he took the chief creative officer role, and then I think a couple years ago came to the decision that he wanted to go back to the original passion when we met, which was doing things with a more overt math and physics component.

So he ended up starting another interesting company, Aetherflux, which aims to bring solar energy in a more efficient form down to Earth and targeting remote outposts and military locations where energy is sorely needed. So he went off to start that and is still a board member. As the company changes, I think we’re pretty good at reevaluating our roles and what we want to spend time doing, and it was just an organic evolution over time.

So much of our audience is people who build things and people who want to be founders, people who are at different stages of being a founder. I always say that no one ever talks about act two — going from managing people to being executives and managing managers. It seems like you have chosen you’re going to be the eye of the storm of the public company that’s changing how finance is done. Are you comfortable with that now? It’s been several years; you’ve been the CEO since 2020. Have you settled into that role?

Well, I don’t know. You can never be too comfortable. What I’ll tell you is the things that I enjoy, do tend to align pretty well with being a public company and the activities of a public company CEO, like communicating the vision and figuring out how we can distill it in a simple form. I love the idea that institutional power and things that normally would’ve been reserved for institutions are now, through technology, available to individuals. That’s very much part of the genesis of Robinhood. It was all about individual participation in equity markets. 

As a public company, we allow individuals to participate in being shareholders of Robinhood in a much bigger way, and so we’ve been doing all sorts of things to change what being a public company looks like. We were among the first to actually engage retail [investors] in a substantive way in our IPO. I think, to my knowledge, we were one of the biggest, if not the biggest, in terms of retail IPO allocation at the time. We let Robinhood customers through the Robinhood app participate in our own IPO. Then we bought a company called Say Technologies that does shareholder communications. If you’ve listened to Tesla, Palantir, or Robinhood earnings calls, they take questions from retail via our Say Technologies platform.

Last quarter, we did a live video earnings call. That was actually very cool. I think the inspiration was a post-game interview at an NBA game. We were like, “That’s fun, what are the other settings where you have people that just went through something hard talking about it and it’s enjoyable to watch?” And I’m an NBA fan and I actually look forward to the post-game interviews. I want to see what they’re wearing. If it’s a big loss, then I want to hear LeBron James talking about what they could have done better. It’s also fun if they win.

I think news, entertainment, financial services, sports to some degree… these are all merging over time, and they’re part of our collective consciousness. I think retail investing is a big part of that. A big, enduring trend is power and capabilities that were formerly reserved for institutions going to the individual level. I think we can change a lot of things about what it’s like being a public company participating in earnings calls. Traditionally, they were viewed as a chore and nobody really liked doing it, and it was considered B.S. that you have to deal with that was a cost, but I think it as an opportunity. I think if executed appropriately, it’s another opportunity to get the message out. I enjoy doing that.

I’m still learning, and I don’t have it figured out yet, but I think we can discover new things and being a public company definitely provides us opportunities to do different things that we wouldn’t normally be able to do.

I want to come back to that theme of what it’s like to be the founder and CEO of a startup versus the public company CEO of what is now basically a bank. Those are different characters, and I’m going to come back to that theme a few times throughout this conversation. How is Robinhood structured now? You’ve had some changes, you’ve had some layoffs over the past few years. How have you organized the company now?

So we have a parent company, Robinhood Markets, Inc., and that’s publicly traded. We have a number of subsidiaries and we typically organize the subsidiaries by business line, but now we also have international businesses, so we’ve got some international subsidiaries. I’ve got general managers who are basically CEOs in their own right that report to me that manage the business. Steve Quirk, who’s our chief brokerage officer, is also the general manager of our brokerage business. He’s got a couple of broker dealers that he oversees, like Robinhood Financial and Robinhood Securities, which is our clearing firm. We’re essentially a vertically integrated brokerage business under Steve Quirk. 

Then we’ve got Johann Kerbrat, who’s our general manager of crypto. He runs our crypto business. Deepak Rao, who presented at the Gold event and runs Banking now and our credit card, he’s the general manager of money. He’s running his own PnL and business as well. Then we have a couple of GM’s in different areas that are building businesses. JB Mackenzie runs futures and also prediction markets. He’s got international brokerage responsibilities. There are a few others here and there, a few earlier stage efforts, but those are the big businesses. We also try to give full accountability, full responsibility, to our GMs so they have their own product resources. By and large, they have their own engineering resources as well. Although we have central platform engineering that builds the underlying technology for everything.

Design and branding is more so centralized under me because I think it’s important for all of our business lines to have a cohesive design language and also cohesive story when they communicate to the public. Those remain centralized, but most everything sits under the GMs.

That’s unusual for a startup of Robinhood’s age. Most startups at this time are still pretty functional. Everything rolls up to the CEO. Have you structured the divisions in this way because of regulatory concerns, or because it’s more efficient? The app is still the app. You express all of Robinhood to the user as one single app. Why have you broken out into divisions inside of that app?

We do have multiple apps. We’ve got a crypto wallet, which is a separate app, and banking is a separate app. Then there’s the main Robinhood app and all of our trading and investment services are in that one. So we do have three apps, but you’re right in that the GM divisions are not per app. GMs collaborate and share app services and I have a product leader that manages the core app. They’re in charge of navigation and design of the overall experience. We try to solve it that way.

But there’s puts and takes with every org structure. When people are separated in divisions, as you say, then it becomes a little bit harder to find ownership over the core surface areas because that’s not natural. And also, the tension points where they interact or where maybe they have different goals bubble up to me or to the core product team. We’ve tried to engineer around that by making those tension points at least knowable and specific. We know it’s going to be design, core app, and marketing.

But yeah, the flip side is if everything’s functional and you’re rolling out multiple businesses, then nobody owns the PNL (profit-and-loss) of those businesses. If I ask someone, for example, “Why is options market share the way it is or options revenue?” They’re like, “Well, I don’t know. I’ve got the product. The product’s great. The engineering is working according to spec.” But the PnL accountability for all business lines would be me. There’s not one person in that structure besides the person at the top that’s living and breathing and losing sleep over the business results of what they’re working on, which is, I think, a flaw. At least for us, which is a multi-product line, multiple entity business, I think it’s worked really, really well.

We made that transition in 2022, and I think since that point, the results have become apparent. I think all the GMs feel huge accountability and responsibility over their entire businesses. They have more ownership. They love it, and I think we’ve been executing really, really fast. It’s worked very, very well for us, but not without risks that have to be managed. 

I think there is a nice thing about it being aligned with the regulatory structure. We’re regulated. We have lots of regulated businesses. Each of them have different licenses. A lot of the time, the regulations stipulate that the business needs to have a president with actual authority, a chief compliance officer, and so aligning the regulatory structure to how we operate very, very closely just reduces complexity.

I think in a functional structure, you have to fight against that and compensate for that to some degree. Other people have strong feelings about actually hiding the regulatory structure from users and absorbing that complexity as an organization. But I think if you can figure out how to align with it and actually use it effectively, it saves time and streamlines things so that we’ve embraced it rather than fought against it.

The structure question is the big Decoder question. The other one I always ask everybody is how do you make decisions? What’s your framework?

I would say I operate at opposite ends of the spectrum. On the one hand, I’m a math and physics person. I love data, I love numbers. I’m very quantitative. I like digging into the details and breaking things into constituent bits; I’m reductionist in that way. 

One of the values that we have is first principles thinking. I’m generally allergic to thinking by analogy. “Oh, this is the way it worked at E-Trade,” for example. When I hear something like that, I immediately get skeptical. I don’t love reasoning by analogy. I don’t like doing things just because others have done them that way. We have this saying, “We only follow the crowd when they’re right.” I don’t want to be contrarian for the sake of being contrarian either. I think that’s silly. But the ultimate thing is “is it right?” And we’ll follow the crowd if they’re right. If they’re wrong, we’ll gladly go against them. That’s one side.

I think the other side is I’m also incredibly comfortable just doing things based on gut feel. I think you need that in order to actually push design forward, because design by its very nature is not super quantitative. A good design is opinionated, which means you’ll piss people off who disagree with it. I have no problem with that. I’d like to say the endpoints are important. It’s good to make some decisions intuitively based on gut feel, others quantitatively, and generally the middle takes care of itself.

Let’s put some of that into practice. You’re big on prediction markets. I know you have a lot of thoughts there. I’m eager to dig into them. This is the rare Decoder where there’s breaking news on this show that is nominally about org charts. Just before you came on to tape with us here on Friday, March 28th, the New Jersey Department of Gaming Enforcement, which calls itself Nudge (NJDGE), asked you to halt bets in New Jersey, on prediction markets on March Madness specifically. 

I’m just going to read the quote: “This activity constitutes a violation in New Jersey sports wagering acts. Which only permits licensed entities to offer sports wagering to New Jersey residents on collegiate sports events occurring in New Jersey.” It’s similar in Nevada. The quote from the chairman of the Nevada Gaming Control Board: “Every sports pool in Nevada must undergo an extensive investigation prior to licensing.” Massachusetts is also investigating this.

You’ve halted trading in New Jersey, you’re complying with the cease and desist there. It’s a big decision to go launch in these markets when you know that there’s an enforcement authority, particularly in New Jersey and Nevada where they have casinos at scale. Why make that decision? Why not go to them first and say, “Are we in compliance?”

This is sort of new ground. If you think about the history of how this came about, Kalshi had a big lawsuit.

Kalshi’s your partner on prediction markets.

Yeah. For prediction markets we’re operating under the CFTC regime, which is the Commodity Futures Trading Commission. They regulate futures and swaps, and prediction markets falls under that purview. We registered to be a futures commission’s merchant, which is basically the equivalent of a broker, but in CFTC land. And companies like Kalshi, which we partner with for these particular contracts, also ForecasteX, which we partnered with for the presidential election prediction market last year, they’re called designated contract markets (DCMs). They’re like the exchange to our broker.

With all prediction markets, since we’re not a designated contract market, we rely on the DCMs to list contracts. And once it’s listed by a DCM, which again is the exchange in CFTC land, we can list them on our platform. Our view is we want to list all prediction markets. We believe that they have societal value in addition to any value they have as a trading asset for our traders. We think they’re a better source of information.

Obviously the line between prediction markets and what should be federally CFTC regulated and what should be under the purview of states who have gaming — which is regulated and taxed at a state level — that line is going to be debated right now. I think Robinhood’s a big part of that because we believe in prediction markets. That’s the intersection here, particularly with sports. While we believe that these are CFTC regulated products, we also recognize that this issue has to be debated and worked out, and it’s not very, very clear. 

For that reason, we decided to respect the state of New Jersey’s demand to halt operations for its citizens, even though we disagree with it. We’re going to work it out over the next couple of months. We’ll be in conversations. But as you can imagine, there’s a lot of states — there’s a lot of people, a lot of counter parties — that could take issue with various aspects of it. And a lot of established interests are at play here. I think this is going to be an interesting area to watch. But I do believe prediction markets are the future, and they have societal value across all categories.

Were you ready for this? When you launched it, did you know a bunch of states are going to get mad and we might have to geolocate our services or halt them in certain states?

Well, we launched without Nevada, as you know. Yeah, of course we built the capabilities of that, as we have in the traditional non-prediction markets business. Crypto also has a state by state component to it. For example, in New York, we don’t offer crypto transfers. You can’t transfer in and out. There’s differences between the coins that are available on a state level. Up until recently, we weren’t in all 50 states. It’s nothing new to us. There’s a state component to everything that we do.

But were you expecting New Jersey to show up and say, “Turn this off until we deal with it”?

I don’t know if we were expecting New Jersey in particular, but obviously it’s not a surprise that if we’re in this new area where states have vested interests to make sure that it’s state regulated that they would have concerns.

The argument that you’re getting at is the difference between a prediction market and gambling. But straightforwardly, that is what is happening here. The states are saying, “We regulate gambling in our states. You pay taxes, we have revenue. We want to protect our citizens. This looks an awful lot like gambling. Go ahead and stop.”

I would offer you the opposite argument. I know a lot of people who believe the markets are gambling, that merely investing in the stock market, or meme coins and meme stocks, is a kind of gambling that has taken place because it has been democratized by apps like Robinhood and even E-Trade before it. 

I see the difference there very clearly. In the stock market, you should be able to look at the fundamentals of some company or its earning reports and you should be able to derive some secondary value. “I understand what this company is doing, I understand how much money it’s making, how much money it’s losing, where it’s investing, what its opportunity is. I can draw some line to its future stock capability.”

That is the most important thing that undergirds the market. That’s the argument against “the market is gambling.” There’s some mathematical reality there. What is that same argument for a prediction market on sports? Because you can’t go look at the Lakers and say, “Well, LeBron’s there, so they’re definitely going to win every game.” That’s just not how it works.

You’re making a valid point. I think the line between gaming and gambling and finance is a debated thing. There’s people that will go on Twitter and say, “Anytime you’re taking a risk, it’s a form of gambling.” I think the term is not properly defined and specifically defined, which I think adds to the confusion. And in particular when you deal with derivatives markets, which I think prediction markets are a subset of the overall derivatives market space, there’s several types of market participants. There’s folks that are coming into a market to hedge, and if you’re a farmer, then you’re sensitive to crop yields and rain and weather and all sorts of things. 

One of the original use cases was for these derivatives markets to apply to farmers for them to hedge their exposure so they can smooth out their returns and their risk over time. And actually, I think for these types of historical reasons, the CFTC and these derivatives markets are overseen by the Senate Agriculture Committee [specifically the U.S. Senate Agriculture Subcommittee on Commodities, Derivatives, Risk Management, and Trade], which is a weird historical fact. And now crypto being in there has a side effect of Senate Agriculture overseeing crypto. But yeah, it’s the historical reason of futures and derivatives being especially valuable to farmers.

You have hedging, which is one use case, but you also have speculation. Speculation is people just making predictions on what the price is going to be in the future. Without the speculators, it’s not an effective market for hedgers because you can’t just have people taking the opposite view of what’s going on in reality because then it won’t be an effective hedge, so you need the speculators to be in there speculating in order for the market to be liquid. 

Then you also have arbitrageurs. And arbitrageurs, which is how I began my career as a trader, just look at all markets and, using technology, compress the prices. If you have the same thing trading in two different places, you just buy it where it’s cheaper, sell it where it’s most expensive, and eventually the prices converge. These are the three types of participants necessary to make any derivatives market work. And so now your question is if you look at folks that are speculating, is there a difference between speculation and gambling?

Let me make that question more specific just based on your example. My father-in-law is a farmer. I married a farmer’s daughter. The utility of him being able to hedge is very clear. Yes, it’s historical and, yes, now the regulatory scheme has this quirk of the [Senate Agriculture Committees] overseeing the derivatives market, but we all still got to eat, right?

Yep.

The farmers have to stay in business. The utility of that remains exactly the same as it was when these regulations were initially passed. And then you need to do some market making. You need to have the speculators and the people doing arbitrage in order to create the market for the hedging. But the utility of that for the farmer and then downstream of the farmer, us literally filling our plates, is obvious. What is the utility of that for sports? For the Lakers. Are you trying to hedge against the risk of losing for the Los Angeles Lakers?

I’ll tell you that sports is a big industry in the US. There’s lots and lots of different types of businesses that rely on sports and the sports industry economically. And it’s gotten much bigger. It’s not just like fans, but sports betting, since that’s been legalized, has just grown to tens of billions of dollars. And it’s still much smaller than what it would be in Europe.

Do you think that’s good? Do you think sports betting is good?

Do I think it’s good? My view is people should be allowed to do what they want with their money. Yeah, I think that markets are good, generally individual accountability is important. There’s folks that trade very, very actively and process lots of information and actually are quite scientific about taking advantage of mispricings. And as a former arbitrageur, I do think that that has value. I want to get away from actually trying to judge every contract on an individual level because I think you can get into trouble. Of course, maybe I can come and give you examples of contracts that I don’t think are great and I wouldn’t trade personally, but I think prediction markets does have significant societal value.

It’s an evolution of what the newspaper served in the past. You have the front page, which is events that people want information about that are trending right now, then you have the business section, arts and leisure, style, and of course you have sports. And the newspaper obviously had value. People were paying for it after the fact. Prediction markets actually give you that news faster; in some cases before it even happens. I think it certainly has enormous economic value.

I view sports as a subset. It’s one of the categories of information and news that people really, really care about. That’s why it’s so interesting to people and people are going to want to protect their purview over that domain. But yeah, I would distinguish prediction markets from gambling in that way. I have mixed feelings obviously about gambling in general, but prediction markets I’m a big believer in.

When I said I had 900 pages of questions, by the way, that thing you said about information, that’s 850 of those pages. We’re going to get into that. But I just want to stay on this for one more second. What specifically to the user, as expressed in your app, is different from betting on sports versus buying a contract in a prediction market?

Because I looked at Robinhood today, and I understand there’s some difference and there’s some vocabulary differences, but what I saw was I’m looking at March Madness, and if I pay 80 cents for a contract and this team wins, I’ll get $1. And that feels a lot like betting. It was Auburn, by the way. I don’t know if they won or not.

Yeah, they’re the No. 1 seed. Traditional sports betting, let’s say digital sports betting, not even on-premise stuff, there’s a house. That means that when you enter a bet, you’re basically betting against the house, and with that comes all of the negative effects. There’s no market, the house is just giving odds. There’s a line. They’re setting the line, and if you win too much, you get kicked off the platform, which is unfortunate. In most cases, once you’re locked in, you can’t get out of your bet.

Because this is a market, there’s no house. Buyers and sellers are meeting directly in an exchange. We’re crossing orders, which facilitates price discovery. Since there’s no one setting the line, the market sets the line. It becomes a more effective prediction, and from the user standpoint, the spread gets tighter because, for a variety of reasons, price discovery leads to tightening of spreads. I think that’s the major thing. There’s no house. Buyers and sellers meet. You can get out of a position during a game, which at [sports] betting platforms is not a commonly offered feature. It’s very similar. You get all the benefits and the power and the rigor of financial markets.

But just at the base level, some 20-year-old kid downloads this app and they want to wager on March Madness, the technical implementation of “put in some money and get some money out if the team you’ve predicted to win, wins” is different. And I think the regulatory approach you’re taking is different. You’re saying these are effectively derivatives contracts and you should be regulated differently because it’s not traditional gambling. But the effect on the user is the same. 

The reason we regulate gambling is because it has bad collective effects in society. People can get addicted to it; they throw their savings away. There’s a lot of reasons outside of the technical implementation of “the house sets a line and can move against you.” There’s reasons that we regulate it. Do you think those reasons are applicable to what you’re doing with derivatives contracts? Because I look at it from the user pushing a button, and the button says, “If they win, you get money,” and the technical implementation of that doesn’t really matter.

I think some of those reasons are applicable. A lot of the origins of the state-by-state regulations come from a world where you actually had physical places where you would go. And so these turning into digital platforms in and of itself, not even CFTC but also state-regulated gambling, are new things. I think the regulations have to evolve either way. But yeah, certainly we want to make sure that suitability and all of those checks are followed through.

And I think actually the traditional financial markets, futures markets are highly regulated. You do KYC (Know Your Customer regulations), you do monitoring and surveillance. There are suitability checks that make sure people really know what they’re getting into and they have to self-certify. And I think that’s a benefit for prediction markets being in the CFTC regime because you have high standards. Saying that these are financial markets isn’t, in my opinion, a lowering of standards in any way, I think it’s a heightening of standards.

Well, the standards don’t exist yet, as applied to this specific thing. The standards exist abstractly for derivatives markets, and now we have to apply them to this behavior. And at least some states are saying, “Actually, this just looks like gambling to us.”

Yeah, but what I’m saying is CFTC-regulated markets are highly regulated. It’s not accurate to say that there are no standards, because we’re following the CFTC standards, which are very rigorous. 

One of the things I worry about with our audience, we have a lot of young men who listen to the show, and who read The Verge, is a sense that these types of markets — whether it’s crypto, the regular stock market, derivatives, prediction markets, or just FanDuel — are a quicker path to riches than regular work. There’s a sense that we’re replacing the American dream with a very financialized secondary market economy. Is that how you want people to perceive Robinhood, that this is the future of the American dream?

Yeah, I did write an article, back in I think it was 2021, about how the American dream itself as a concept has evolved. It used to be very tied to homeownership. You’d buy a house, you’d get a 30-year fixed mortgage, and that was the American dream, and that’s actually not ideal from an investment standpoint. The amount of interest in fees you’re paying on that, if you view it from an investment standpoint, is actually incredibly high. If you’re going into U.S. equities — which are now commission-free and very, very low cost — that the American dream should perhaps evolve towards U.S. equities.

I didn’t make the claim that it should be crypto and derivatives or all of these things. And in fact, a couple of days ago, we actually crossed the bridge from being a purely self-directed platform into offering investment advice with Robinhood Strategies. If you look at the asset allocations of the portfolios there, it’s very much listed equities and ETFs. And then if you actually look at what we incentivize as business, what we’re giving matches for, it’s things like retirement where you get a 3% match on contributions.

I think I would distinguish between what the right way to invest is for the bulk of your money, which I do think for most people that have income and assets should be passively managed. But also, I do think people that have the income and can passively manage a portion of it, I don’t think it should all be passively managed, I think there is a room in your portfolio for every person for it to be actively managed. That could be in things that you have high conviction in, whether it’s individual stocks, cryptocurrencies, or options. If you’re at a startup, you implicitly have high conviction and lots of concentration in the company that you’re actually working for. And if you consider yourself an expert in an industry or even in sports, I think the derivatives markets live in that bucket.

But yeah, I wouldn’t say if you look at Robinhood, the actual mission and the future vision is for us to manage every dollar. I don’t think every dollar should be in derivatives markets; probably a small portion of them. But the reality is people bet on sports, people engage in derivatives trading, and that’s money that’s leaving Robinhood accounts. If we can serve all of those dollars with our platform in a seamless and easy way at the lowest possible cost and the best user experience, then we’ll have full wallet share with our customers across multiple generations. I think we could both add value and build a significant and important company that way.

“Wallet share of our customers across generations,” is an all-time Decoder phrase. I thank you for it. It’s going on the wall. Let me ask about information and risk. What you’re describing is a spectrum of risk. You’ve got your new bank accounts, you’re paying people, what, a 4.5% yield — that’s low risk. That’s just your bank. All the way on the other hand, you’ve got prediction markets for sports, which are maybe the most risky thing you can do. And then in the middle, you’ve got your thesis about information. Prediction markets are this new source of information.

The thing that gets me is when you make prediction markets, the value of the information skyrockets, and then you have a lot of incentive. You’ve created an enormous incentive to affect the outcome. In the best case, you work at a startup, you’ve got stock in the startup. You have a huge incentive to affect the outcome positively. The company will be a success. You’re going to work really hard. You’re going to make a lot of money.

I look at the NFL, for example. I’m a big NFL fan. The amount of time we now spend talking about referees in the NFL officiating because of gambling has gone up. The notion that the league is scripted and that the games are rigged because any individual referee can make one penalty call at the end of the game and shift the outcome is skyrocketed because of the inclusion of gambling by the NFL into the product itself. That feels like a bad outcome to create all of these incentives to shift the outcome without any regard for the quality of the outcome itself. 

How do you manage the prediction markets against that incentive? Because I see that as totally distorting and in most cases negative.

Yeah, I think that’s a great question. And that’s one of the areas where the traditional financial system already has lots and lots of infrastructure because we’ve faced this problem for decades. You have insider trading rules and regulations, and it’s very analogous to a company insider using proprietary information for their own benefit to make money in financial markets. That very much exists.

There’s also general anti-fraud protections that go into place when you’re not dealing with securities. If you remember, Coinbase had a case a couple of years ago with the DOJ and what they found was that some employees used knowledge of forthcoming listings. I think it was some meme coins, they bought those meme coins because they knew that they were going to be listed and made a bunch of money. And of course, this was caught and tracked by surveillance and they got in a lot of trouble. I think generally the same principles apply. 

Sure, but how does that track with your sense that this is the new source of information? Because the information only enters the market if people have it and they begin trading on it. If you want to outlaw insider information, you have to prevent people who have that information from trading on it. How does that get into the prediction market?

Yeah, basically individuals who have proprietary information shouldn’t participate in the prediction markets, and all of the DCMs basically have rules against this. Because we know who’s making the trades, everyone has to be KYC’d (know your customer) at regulated DCM / FCM (futures commission merchant)-regulated prediction markets, we have the capabilities of identifying abuse. And of course all of these rules can evolve over time. If there’s new vectors for abuse as the markets expand, there’s mechanisms for those to be incorporated and to become new rulemaking. 

Hopefully the rules should evolve, as with any system. And if there’s new vectors coming in, then we can evolve the rules to account for those. But I’m actually not sure if there are new vectors that aren’t accounted for by the existing rules. And I think this is one of those things where because the state regulatory regimes haven’t really accounted for this, they may be less well positioned to oversee abuse than the federal level.

Where does the information come from then? If I’m looking at a prediction market on Robinhood and the line moves sharply, that’s the information you’re talking about. This is the new source of information. You’re going to get it before the news gets it, or the traditional media gets it. You’re going to watch that line move and you know something happened before everyone else knows it happened. How do you get from that second order effect — the line moved, people started moving their money against some new information — to the information itself?

In the same way it happens in financial markets. You have sophisticated participants. Some of them are retail, many are institutional that actually make sense of all of the data that’s coming in real time and actually crunch the numbers and see what it means. And a lot of this is happening using automated computing methods. They’re crunching all of the data.

But yeah, you can think of it as, let’s say you’re watching the news on election night and you’re getting all of this polling data and all the early returns from the polls, and they’re telling you Ohio results just came in, and there’s this many voters in Ohio out of this many that are reported. There’s a process by which you take that and actually price what the likely outcome of the election is. And so the people that are really good and fast at doing that have an opportunity for profit.

That opportunity for profit isn’t the information itself, though. That’s what I’m getting at. I’ve heard you say the information line before. You said it to my friends, Casey Newton and Kevin Roose, on the podcast Hard Fork. You said you’re going to get information before it happens. That prediction markets are not just the future of trading, but also information. What prediction markets are “is the news faster.”

But here, you’re saying prediction markets are reacting to the news, they’re reacting to information. If you’re a regular Robinhood investor and you’re looking at the line move, how do you get back to, “Okay, the smart money made some decisions. The smart money is watching Harry Enten say on CNN, ‘Here are where the votes are,’ and now I’m repricing the contract? That’s not the news, that’s a derivative of the news.

Yeah, I guess this is also an area where there’s ambiguity in what we’re saying. I think traditionally what I would say about the news, about the election, is when the news networks call. “Election news” is probably when CNN, MSNBC, NBC and all those networks say, “Okay, Donald Trump’s the winner of the election.” That happened the next day in the vast majority of cases. The news of who won the election hit next day, but the prediction markets priced Trump at 95/5 within a few hours of the polls opening. I would call that knowing the news before it happened. If you were paying attention to the prediction markets, you knew what the outcome of the election was within some band of error. But I’d say 95/5 is pretty good, right? At that point.

Right. The difference for the media outlets is they can’t be wrong. Or at least when they’re wrong, there are consequences for them being wrong. They lose their credibility or they have to issue retractions or, I don’t know, Trump sends them all to El Salvador or whatever he wants to do. There’s not a consequence if you trust the prediction market and it gets it wrong. The consequence is you lose money. How do you think about that accountability?

Well, I think that accountability is only part of the story. I think the other part is that there’s an incentive to keep viewers glued and entertained. And you want them to watch for longer, so you don’t want to just be like, “Oh, the election’s over. Everyone can go back to what they’re doing.” They have an incentive to actually prolong it and say, “Hey, it’s still anyone’s game. Anything can happen. Keep watching.”

And it’s the same thing for sporting events, actually. I remember I was watching the Jake Paul-Mike Tyson fight. And that wasn’t available in the U.S.-based prediction markets, but I opened up Polymarket to just view what was going on. I think by the second round, the prediction markets had it quite clear: it was 90/10 for Jake Paul. But if you listen to the announcers, it was, “Oh, all Mike needs is just one punch,” or, “He just needs to hold out a little bit longer.” It’s like they made it sound like it was 50/50.

You didn’t know the outcome of that fight before it started? I knew the outcome of that fight before it started. I know why lots and lots of people that were betting on it did bet on Paul, and they knew why the fight was stretched out the way it was. And that’s the danger. What you are describing is the danger. Here’s this totally synthetic event that we’re going to make people pay for and then have people bet on, and then everyone will believe at the end that it was rigged. 

The announcers have nothing to do with it, it’s the distorting factor of the gambling that occurred around it that made everyone think, “Oh, this is totally rigged.” Both Mike Tyson and Jake Paul got a lot of money for participating in this thing that most people believe was a sham, and they had a lot of incentive for the outcome to be made a little bit longer, right? Make it feel like this isn’t totally a joke. Everyone knows that that’s what happened. Or at least they perceive that they know. And that’s the distortion that I worry about.

I don’t have any details about that, but my point really was, I think to your question, I don’t think the only incentive is for the media to get it right, I think the incentive is actually to have people watching longer, which sometimes conflicts with giving you the information as quickly as possible. I think you have to agree that that’s the incentive. I think platonically, yes, the media is held to a very high standard; you have to get it right. We’d like to think that everyone is working with that in mind, but there’s also a strong incentive to maximize viewership, to maximize time spent versus other networks. Because if you don’t do that as a media enterprise, eventually you’re drained of resources and you die. And I think sometimes that incentive actually can contradict with getting you the absolute truthful information as quickly as possible.

Yeah, I work in the media, I would say there’s an equal incentive to be first. That is equally damaging. As somebody who runs this race every day, that’s equally distorting. But then on the flip side, I would say it’s also not working, right? The media is losing jobs and it is dying and it is falling in relevance. There’s a lot of reasons for that that I don’t think have anything to do with stretching out the information. 

But what I just keep coming back to is the incentive to change the outcome by allowing people to have a financial stake in it seems very, very distorting, and there might be some rules against it. The NFL does injury reports because getting access to injury reports early was helping gamblers, and the NFL wanted to tamp down on that, and that’s the history of the injury report. I get it. We’re doing transparency. We’re doing regulation around this stuff to try to control it.

But here it seems like the regulatory regime is new where you’re trying to fit a behavior into a regulatory regime that wasn’t built for it from the start and you’ve got a lot of young people taking part in this behavior who believe now that everything is rigged, that the world is a casino of this kind. This is what I was going to ask about. That seems like that’s where you have to make the turn from startup founder with a disruptive idea to, “I am the CEO of a bank.” The responsibility rests with you. And I’m wondering how you’re shouldering that burden. Because once you have the bank accounts and you can move it into the highest risk category, that’s a big deal.

Yeah, I think that is a criticism for sure. I think in a more precise form, it’s basically do you want your betting in the same app as your bank account? Can these things coexist under the same platform? I don’t think that it’s the highest order criticism because if you think about it, all these things are on your phone, and so it’s fairly straightforward. And now everything can link to your bank account anyway. If you had a DraftKings account, you could link your DraftKings to your bank account pretty easily and move money back and forth. We get into this, well, but should they be separate apps? Should they be separate brands? Do we want the same thing? But the reality of it is everything’s on your phone. It’s all your phone. It’s pretty easy to just go to the home screen, tap a button, and go from one service to the other.

I actually don’t think that criticism is higher order. Basically, it’s just a completely meaningless decision. It’s a business decision. It’s like where do customers want it most? Because if they want it in the same app, functionally there’s no difference between it being in the same app and being in a completely different brand elsewhere on your phone. Because the difference between going from one to the other is three seconds instead of two.

There have been some good reasons in the past to separate these at the corporate level. And I hear you that it’s pretty easy to just move money into FanDuel. But there was the Great Depression, and then we did Glass-Steagall, and we said the investments and the banks have to be farther apart. Then we allowed them to get closer together, and we had the 2008 financial crisis, and then we moved them slightly farther apart. And now you’re saying, “Look, because of phones, this distinction is meaningless,” but history suggests that bad things happen when you let investments in banking get closer together.

Well, that’s for safety and soundness reasons. You don’t want the same organization to be over leveraged and to take proprietary risk with their capital when they’re also supposed to be safeguarding that capital. But we’re not taking proprietary risks because we’re just a broker routing orders to a marketplace where buyers and sellers interact, so the safety and soundness concerns that you’d want to separate proprietary trading from retail banking don’t apply in this case.

Don’t the concerns apply to the individual consumer, though, who is less sophisticated than banks, which often get themselves in trouble?

Well, again, that’s the point of the individual has money. They have a bank account. They’re moving things to different accounts. It’s the same thing from their perspective. It doesn’t affect them whether those accounts are at different entities or one, barring safety and soundness concerns that are entity concentration, which don’t apply in this case.

Would you accept a regulation that said when you open the Robinhood banking app that you were prohibited from advertising or marketing the derivatives products in the Robinhood app themselves?

Possibly, yeah. I haven’t thought about that. I don’t think we should accept new regulations lightly in any sense, but I’m certainly a believer in regulation. I run many highly regulated businesses. Yeah, I don’t want to just be glib about it and say, “Oh, that sounds like a good one.” You’d have to look at the pros versus the cons deeply.

I generally think that we get into trouble by having lots of regulations. We don’t have mechanisms to remove them. Once it’s added, you just keep building this giant, thick underbrush of regulations and nobody ever wants to remove them. And then now we’re in a situation where, as a country, we’re looking around and saying, “How did we get here?” We can’t do anything anymore. You can’t build a bridge. You can’t trade crypto. And yeah, I think that’s a problem. I think I generally agree with some variant of for every one you add, you should probably remove one because things change and the regulations, to a large extent, don’t make sense.

I’ll give you another example: accredited investors. This is probably one that maybe you’d agree with even because it’s less controversial. But accredited investor rules basically stipulate that you can’t be investing in OpenAI or SpaceX unless you’re accredited because of some variant of “they’re too risky.” And why are they too risky? Back in the day when these regulations were created, it was hard to get information. There were no prediction markets, there wasn’t the internet. There was a lot of murkiness. And maybe the wealthy folks had access to information and enhanced due diligence, normal people didn’t so we have accredited investor rules. Now we’re in a situation where meme coins are fine. You can put all of your money, anything you want in meme coins, sports betting, whatever have you, but OpenAI or SpaceX, companies like that are too risky. I think those cannot stand.

As you can probably tell from our conversation, I don’t think we should ban trading in meme coins and sports betting; people should generally be allowed to do what they want with their money. And so I think the accredited investor rules need a complete reboot and probably something closer to self-certification and some requirements for brokers and platforms like ourselves to put these things into buckets based on how much disclosure they have. Maybe if you’re an early stage startup and there’s no disclosure, we have to put a skull and crossbones in red and tell you, “This is an incredibly risky thing. You could lose 100 percent of your money.” But yeah, I think the status quo needs a serious reboot.

I just had the idea of Sam Altman doing a meme coin to fund OpenAI’s expansion. It seems more likely than not. Let me ask you about crypto, just to wrap up here. I have another 850 pages of questions about crypto, for sure.

Let’s do it.

Why should anybody sell a bitcoin?

My general philosophy for myself is I’ve only ever had regrets selling investments, pretty much. I think I tell people all the time when I was in college, I bought Nvidia stock. And I thought Nvidia was a great company. They made amazing GPUs for playing computer games. Doom 3 at the time was my favorite computer game, and so I bought it for that reason. And I think I got it at 20 and sold it at 30. And I felt really, really good about myself. I think everyone has those stories of something like that where they exited a little bit early.

I think people should sell. My own philosophy, I’ll sell if I have to. But generally speaking, I’m more of an accumulator. I like to accumulate things and hold onto them for a very, very long time. But people have different needs. Sometimes you need to buy something; you have an expense, and you don’t have that luxury. In that case, the fact that these markets are liquid and you can sell and get a good price is very, very important. And imagine you’re someone who bought Bitcoin in 2011 when you got it at, what was it, $1 or $2 a coin? I think it’s not unreasonable for someone like that to sell at some point along the way.

I asked that question because my thesis is people only care about Bitcoin because of dollars. It’s the value of Bitcoin as expressed in dollars that makes everybody care about it. And if there’s no reason to get rid of one unless you need the money in dollars, then you will never transact in Bitcoin. It will never stabilize to the point where buying an Nvidia GPU in Bitcoin is a better idea than buying it in dollars. And if you can’t get there, then we really do just have a store of value. We just have another thing. We have digital gold. And you run the platform, and I’m wondering if that’s your view as well.

Yeah, I think the properties of Bitcoin right now are much more conducive to it being a store of value than an actual mechanism for transacting and buying things. I think the fees are really high. And it’s very, very easy, including using platforms like Robinhood to take your crypto or any other asset and convert them back to dollars for when you need to transact. But I don’t think that’s unique to Bitcoin. I think stocks can essentially be thought of as a store of value from the retail investor standpoint. Nobody’s sending little bits of stock between each other to pay for things.

But when you buy a stock, you have a thing. You own some insignificant part of a company. You have voting rights. You could fire the CEO, maybe. There’s a tangible value to owning the stock besides the stock itself.

Yeah, yeah, I do, yeah. But what I’m saying is I don’t think that it not being used for payments really puts a ceiling on its value. I think that actually it being a medium, I think the medium of exchange use case is becoming less and less important over time. Pretty much anything can be a medium of exchange. Imagine if everyone had Robinhood accounts. Basically if you wanted to pay someone, you’d only need to convert it to dollars at the moment of paying them, and then the recipient can convert it into an asset that appreciates more. And you would obviously have probably a small portion of your portfolio in actual dollars. I think medium of exchange was more important when fungibility was much more difficult like it was back in the day. But now fungibility has never been easier, and so the store value use case probably dominates.

It’s weird that I have one vision of the economy that’s like everything’s a casino and another one where everything is Disney bucks, and I don’t know how to reconcile those two things.

Well, Disney owns ESPN, so there you go.

We’re all going to be gambling on Disney bucks, it’s going to be great, on ESPN Bet. In order to keep the value of Bitcoin or any of these store value coins high, you need a constant buyer. Do you think that the Trump crypto reserve is a good idea? Because that’s the purpose it seems to be serving.

Well, I’m not sure. We get the question ourselves about should Robinhood as an entity should have some of its balance sheet in Bitcoin. Every other company is doing that. Why aren’t we doing it? And again, I come back to we only follow the crowd when they’re right. We obviously haven’t jumped into crypto on the balance sheet yet, and I think part of the reason why is that it’s not critical to our purpose as an organization. 

Of course, if we incidentally had some crypto as part of serving our customers with various activities, we can and do do that, but we also don’t want to become some vehicle for people that just want exposure to Bitcoin to buy our company. I think generally speaking, with the U.S. government, I think you can make the argument that the U.S. spends money on a lot of things that are worse than Bitcoin, and I think that’s true. There’s just blatant waste out there in things like that.

Boy, do I disagree with you on that. Also, buying Bitcoin is not a service, right?

Yeah.

The government mostly spends on services for the citizens. 

Do you think the Reserve as announced is the right mix of holdings? I’ve heard it called a “Shitcoin Reserve” because it has Solana in it. You trade on all these coins, the platform enables all these coins. Is it the right mix of holdings in the Trump Reserve?

I think that generally what they announced was not selling the coins that they seize through various mechanisms. I think that’s fine. That’s probably the most I would’ve done if this decision was up to me. And I think there’s reasonable arguments to be made for that. Selling something is also a conscious, affirmative decision. You’re moving a market; you’re doing something by selling it. And again, my personal philosophy, I’ve only had regrets when I’ve sold stuff. I actually think what they outlined was reasonable. Now, it depends on the details of what budget-neutral acquisition of Bitcoin could be. I think that’s a fairly high bar for it to be budget-neutral. I think we’ll see. But yeah, probably I think it’s reasonable. It’s a sensible approach. If it was my decision, I don’t know if I would’ve gone further. Probably not.

Part of your new banking service is that you can get physical cash delivered on demand to your doorstop, because you don’t have locations obviously. How on earth does that work? Are you mailing cash to people?

No. We are using on-demand delivery logistics. We haven’t announced our partnerships yet, but we’re not actually doing everything ourselves. Yeah, it’s basically combining the power of on-demand delivery logistics with financial services to bring the retail bank to you. If you look at digital banking before Robinhood, there was always a sacrifice. You had a nice convenience of the digital app, but there were no branches. And so, if you wanted to get cash or even deposit cash, but most likely withdraw, you’d have to go to a 7-Eleven or a CVS or something. And nothing against going to 7-Eleven. I love 7-Eleven, but it’s not quite the private banking experience.

In fact, with my private banking experience, I was a FRB client, First Republic Bank, and they had this amazing feature where they would deliver cash to you. It is a high net worth feature, but of course it would be slow, it would be super expensive, it would be for a lot of cash and it would come in an armored vehicle. We asked ourselves, “How can we make that experience for everyone?” And this is what we came up with. I think we’re excited to roll it out.

Wait, are you going to roll up an armored truck to people’s houses?

Well, probably not for a small amount, but if you think about it people already get iPhones delivered, and an iPhone is an expensive item — $1,000, give or take. It’s a small, expensive item, and a high-ticket charge. My estimate is the average ticket charge for a cash delivery order offered by Robinhood is probably going to be in the low hundreds of dollars. Similar dollar value to a delivery order. Maybe a little bit higher, but probably not quite as high as delivering Apple products. 

I think if you wanted to get a quarter million dollars delivered, you would need an armored truck, and we’d like to facilitate that as well. But for your typical ATM-like transaction, the idea is that that would be a smaller amount, it would get delivered to your house. And we’d like to figure out how to do it outside of your house too, if you’re out at a place that’s in a service area. But that’s a little bit more complicated. But yeah, if it could come to you in 10 or 15 minutes and at low cost, I think there’s real value there. Sixteen percent of payments in the US are still cash payments. Even though we’d all like cash to go away, cash is still very much a giant part of the economy here.

Well, I think you can tell I could talk to you for hours and hours about a wide variety of things. We’re going to have to have you back just to talk about the logistics of locating cash around farmers markets throughout the country, because that’s very exciting for me. Vlad, this has been great. Thank you so much for being on Decoder.

Thanks so much for the time.

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

Today, I’m talking to Rohit Chopra. He was the director of the Consumer Financial Protection Bureau (CFPB) until the end of January, when President Donald Trump fired him and Elon Musk’s Department of Government Efficiency (DOGE) began trying to dismantle the agency.

The CFPB has been around for just under 15 years now. Congress created the agency in the wake of the 2008 financial crisis to protect consumers from various kinds of lending and credit schemes that led to that crash. 

Broadly speaking, the CFPB has been pretty popular! This is the agency that keeps consumers safe from predatory lending practices, limits credit card fees, and investigates various kinds of financial fraud. So effectively shutting it down kicked off a number of controversies, not least of which is: do Trump and Musk even have the power to do this? After all, the CFPB was created by Congress with a law, and the US Constitution says the president is supposed to take care that the laws passed by Congress are faithfully executed and not reinterpreted by an unelected billionaire who is deep into crypto. In fact, I don’t think that came up with the founders at all, though it’s been a minute since I watched Hamilton.

This all led me to ask Chopra several times who actually made the decision to fire him, who is currently responsible for the various policies of our government, and whether any of those things add up to a clear plan for which someone can actually be held accountable.

I ask questions like this on Decoder all the time, and there are usually answers, even from some of the most powerful executives in the world. In fact, especially from some of the most powerful executives in the world. But here, well, you’re just going to hear Chopra say, over and over again, that he doesn’t know. Sincerely, I don’t think he knows — and that should probably be as worrying as anything.

He and I also talked about the clash between the two main factions of the Trump coalition. On one hand, there are the Musk tech libertarians, and then there’s the more populist MAGA wing. Right now, these two factions are having a big fight over tariffs (and it’s still hard to tell what’s going on there), and they seem poised to be potentially even more at odds as Trump reshapes more of the government. 

The CFPB sits right in the middle of that fight; it’s a lot easier for Musk to turn X into an everything app crossed with a crypto payments platform if there’s no regulator on the beat. But the populist wing of the party isn’t exactly in love with big banks and Big Tech gaining even more power. I have no idea how that will play out, so I was curious to see if Chopra had any insight — and what he was most worried about happening without an agency like the CFPB standing guard.

If you’d like to read more about what we discussed in this episode, check out the links below:

• Trump fires CFPB director Rohit Chopra | Associated Press
• Trump orders CFPB to stop work, closes building | Associated Press
• CFPB workers reinstated after court order but still can’t work | The Verge
• Trump admin to appeal order blocking CFPB shutdown | Bloomberg Law
• A shady tech bootcamp may be sneaking back online | The Verge
• CFPB won’t enforce long-awaited payday lending rule | Bloomberg Law
• CFPB seeks to vacate redlining settlement, refund lender | Banking Dive
• CFPB signals it will drop rule regulating BNPL like credit cards | PYMTS
• CFPB drops fraud lawsuit against banks, Zelle | CNBC

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

Today, I’m talking with Daniel Dines, the cofounder and, once again, the CEO of UiPath, a software company that specializes in something called robotic process automation (RPA). We’ve been featuring a lot of what I like to call full-circle Decoder guests on the show lately, and Daniel is a perfect example.

He was first on Decoder back in 2022, right before he moved to a co-CEO arrangement with Rob Enslin, a Google Cloud executive brought on to help steer UiPath after it went public. In January of last year, Daniel stepped down to become chief innovation officer and Rob stepped up to become sole CEO — and then, less than six months later, Rob resigned, and Daniel took his job as sole CEO back.

Founders stepping aside for outside CEOs and then returning as CEO later on is quite a trope in the tech world, and Daniel and I spent a while pulling his version of that story apart. He made some pretty key decisions along the way to relinquishing control of the company he founded — and then some equally important decisions when coming back. If you’re a Decoder listener, you know I’m fascinated by the middle part of these stories that usually gets glossed over, so we really dug in here.

But there’s a lot more going on with UiPath than C-suite shuffles — the company was founded to sell automation software. That entire market is being upended by AI, particularly agentic AI, which is supposed to click around on the internet and do things for you.

The main technology UiPath has been selling for years now is RPA, which has been around since the early 2000s. It aims to solve a pretty big problem that a lot of organizations have. Let’s say you run a hospital with ancient billing software. You could spend millions upgrading that software and the computers it runs on at great risk, or you could just hire UiPath to build an RPA system for you that automates that software and presents a much nicer interface to users. This decreases the risk of upgrading all that software, it makes your users happier because they’re using a much nicer interface, and it might provide you some efficiency by developing new automated workflows along the way. 

UiPath built a fairly successful business doing that basic version of RPA; I encourage you to listen to our episode in 2022 where we unpack it in great detail. But as you might expect, that’s all getting upended by agentic AI systems that promise to automate things in much more powerful ways, with much simpler natural language interfaces. So Daniel has to figure out how UIPath can integrate and deploy AI into its products — or risk being made obsolete.

Daniel and I really got into that, and then I also wanted to push him on the practical economics of the business. The big AI startups like Anthropic and OpenAI don’t have to make any profits right now. They’re just raising mountains of investment and promising massive returns when all of this AI works. 

But UiPath is a public company, and it’s licensing this technology at a cost. So I wanted to know what Daniel thought about the cost of licensing AI tech, selling it to customers, and trying to have all of that make a profit while the underlying economics of the AI industry itself remain pretty unsettled. 

We also talked about what all of this might mean for our experiences at work, and whether a world of robots sending emails to other robots is actually a good goal. This one really goes places — Daniel was game to truly dig in. I think you’ll like it.

Okay, UiPath CEO Daniel Dines. Here we go. 

This interview has been lightly edited for length and clarity.

Daniel Dines, you’re the founder and — once again — the CEO of UiPath. Welcome back to Decoder.

Thank you so much for having me, Nilay.

I’m very excited to talk to you. I love a full circle episode of Decoder. You were last on the show in the spring of 2022. It’s been a little bit of a roller coaster since then. You were just about to have a co-CEO named Rob Enslin. You hired him from Google Cloud. Then, you stepped down a little over a year ago to focus on being the chief innovation officer. Then, Rob was the sole CEO. Then, Rob stepped down, and now you’re CEO again. You’ve made some changes to the company.

Explain what’s going on there, because that’s a lot of decisions. Obviously, we’re a show about decisions, and there’s a lot of AI stuff I want to talk about. But let’s start with that little bit of history. Why step down and why come back?

Well, roller coaster is a good word. Sometimes people exaggerate with it, but in our case, it’s really what happened. Why? Look, I was always trying to do what’s best for this company. This company is, in a way, my baby. I spent almost 20 years [building it]. This year, 2025, is 20 years since I founded UiPath. I thought that if we can get the best talent, and especially with [Enslin’s] background in go-to-market, this is going to help us. And Rob is a nice guy. We got along pretty well. And look, it’s been mostly a good ride. It gave me some time off, so I switched to chief innovation officer. I ran our product and engineering teams.

In 2023, I had my own time for reflection, especially after I moved a lot of my responsibilities to Rob. I spent that summer in reflection mode, honestly, with a bit of soul searching around “what do I want?” I would say that I missed my early 20s craziness, with people having a lot of fun and going on spring break. I had to work. In post-communist Romania, there was a lot of turmoil, so life was not that fun for me at that stage. I thought maybe I will get to experience what it means to take it a little bit easier.

It was important for me because I discovered that UiPath is actually kind of an anchor for me. It gives me a framework of mind, a direction. It’s very hard for me to wake up every day and give myself something to do unless I am in this big machine and this machine is on a trajectory. It forced my mind to be there. And I’m surrounded by great people. I talk to smart investors, analysts, customers, and partners. It’s a living organism. So, I discovered that this is a gift that I have, being in the position to run this company.

Then, things in early 2024 didn’t go well for us, from an overall market perspective. I think the macro was pretty bad for some companies. We had some execution issues. Our initial go-to-market was “land and expand,” and we over-rotated the company to go mostly after big deals. So, our float business suffered, and paired with some of the macro challenges, it created a difficult environment. Rob decided to leave the company in May 2024. In all fairness, at the time, I was ready to take it back. It came faster than I anticipated, but mentally I was prepared after my summer and my time off.

Did you go on a spring break? Did you take a minute? Were you in Palm Beach?

No, no, I didn’t go to Palm Beach, but I spent a few weeks in the Mediterranean on a boat. So maybe close to it.

Spring break is not the same in your 40s as it is in your 20s, is the thing that I’ve discovered.

Yeah, exactly.

I always want to drill into the actual moments of change. I always joke that I watch a lot of music documentaries. There’s act one where everyone’s in the garage, and there’s act three where they’re playing Shea Stadium. And act two, where the actual moments of change happen, are often glossed over. This is one of those moments. You made a decision to come back as CEO, Rob made a decision to leave. What was that conversation like? Did you initiate it? Did he start it? Was he leaving and you already decided that you were coming back? Walk us through it.

It was simple actually. We decided to meet in New York following Q1 2024. He told me that he thought it was better that I take the company back and he resign for personal reasons. Indeed, he needed to take some time off because some members of his family were not well. I told him, “Let’s reflect a bit on this. Let’s think a bit.” But in the end, he was resolute in his decision. 

I also realized after that discussion that there will be many changes in the company. We needed to contract a bit. We oversized the company for this elephant hunting, so there needed to be a few changes. And I realized it’s actually better that I do the changes. It’s going to be a lot of pain, and we’ve already been through some pain. The last three quarters were not easy for us by any metric.

Would you have made the change if he hadn’t volunteered? Was it obvious to you that you were going to come back as CEO?

I realized something. It would be difficult to get an external CEO while I am here. It’s kind of not possible. I would consider growing someone internally rather than bringing in someone externally. It’s really hard to know someone after you talk for a few hours and you go for a dinner, and it affects the culture of a company so much. Even if I have the controlling stake of the company, it’s not like you get someone and you can command them every day, “You do this and this and this.” No, it has a really huge implications. 

I care deeply about the company and the people. Rob had all the best intentions in the world, but seeing the things that sometimes made me uncomfortable, it was not easy, and it’s not easy for anyone. Naturally, there were two camps created — Daniel Camp and Rob Camp — and sometimes they didn’t talk. Again, without our intention, it was a dynamic that didn’t work well. So to me, it was clear that I had to either take back the CEO role and drive the company, or next time I would step down completely.

This is a pretty common problem with founders. Obviously, The Verge is much smaller than UiPath, but I only have a handful of co-founders left. I often tell people that they should be the editor-in-chief, and it’s perceived as a threat. They’re like, “No, we wouldn’t do it if you were here.” Did you have the power as the founder and the controlling stakeholder to say, “I’m just making this decision, I’m coming back?” Was there an approval process? This is one of those moments that seems like it comes up a bunch with founders.

Theoretically, I had the power to do this, but in practical terms, it’s something very difficult to do. Look, we are a public company. It’s board governance. I have a seat on the board. The board should make the decision. So, the board would have to make a collective decision to fire Rob on my pressure. They could have mutinied against me, but it’s not so simple. It’s doable, but–

That’s really the question. We see some of these decisions from the outside. The founder coming back as a CEO seems like a very natural course of events, but then it’s very complicated on the inside, particularly with founders who were the CEO, stepped aside for another CEO, and then came back.

If there is a battle between the founder and the CEO, yes, things could be pretty ugly. In our case, it really was not. Rob really exited under the best conditions. He gave me all the time. He assisted me with the transition. He then took some time off to fix his personal stuff. From this perspective, it was a smooth transition.

You mentioned the company had grown in ways you didn’t want it to. With a new CEO, there are cultural implications with how they would like to run the company. Then as the founder, you come back and want to change it back. You just reported financial results. Things seem to be a little more stable than they were in the past. What changes have you made, either to go in a forward direction or to go back to the way things were when you were CEO?

I wanted to bring back some of our mojo of being customer-centric, working with customers, and doing whatever they required to be successful. We went back largely to “land and expand,” to being customer-centric while still preserving the muscle to do big deals. We need both. Forecasting is kind of difficult in a company that depends only on big deals. The lumpiness in revenue can create issues with forecasting. It’s normal to have both sides of the equation. 

That’s also a thing that I didn’t realize. We’re not a technology that you can go day one and say, “I will sell you a $100 million of automation.” Let’s see a smaller division and see how it works. Then, expand into other divisions, and then company-wide.

So, regardless if you have a good Rolodex, you won’t go to another CEO and say, “Okay my friend, give me this big deal because I’m here for you, and I promise you we’ll do it best.” You need to prove it, and you need to earn your way into a company. That’s why, in our DNA, the essence is to stay extremely customer-centric, work with them, help them find opportunities, help them deliver the value, prove the value, and have them message internally about the benefits of automation. We kind of lost a bit of this muscle. 

And now, we’ve segmented differently. I created an executive accounting program where we have our top 50 diamond accounts with all our executives attached to those accounts, and we are taking it very seriously. We also have a co-innovation program where we build software together. We decentralized our customer success function that was centrally run. It was a bit disjointed from the sales motion, so we decentralized it into the region, and it’s much more aligned with the customer right now. We even changed  the compensation of our sellers and customer success to be closer to the adoption of our software. Regional partners were also moved within the sales teams. I simplified and streamlined the international part of our business into one big region. There have really been a lot of changes.

Were all those changes in your head while you were the chief innovation officer? You were watching the company change and the results, and you were thinking, “This is how I would fix it?” Or did you come to this plan after you retook the CEO role?

I think some of the pain that we were experiencing was known at that point. The changes? Not really so much. It took me a month to understand who the people on my team would be and what kind of changes we were going to make. 

I love having people come back on the show because I get to read their old decision-making frameworks back to them. You left, you took a break, you got to think about who you wanted to be and how you wanted to spend your time. 

The last time you were on the show I asked, “How do you make decisions?” You said, “I’m trying to learn more by listening to people. I have no idea how to run a big company at this stage because I have never been in a similar situation before, but I’m trying to build a close-knit executive team that relies on each other.” Then, you said the thing people say, which is to “[make decisions] fast if they can be reversed, and do them slowly if they’re irreversible.” Is that still your framework? Have you come to a different approach? Are those still the basics?

I think largely, yes. I like to give space to people to delegate. My style is to agree on goals, agree on the plans, and then let the people run. If I find issues, even small issues, my style is to dig around to see if there are signs of potential cancer or things that are completely not working. You discover interesting things. But yes, I think the faith of the company depends extremely on the cohesion of the leadership team. A big difference in how I make hiring decisions compared to 2022 is that I will never trade chemistry for talent. Bringing talent that doesn’t fit into an organization never works, and long term, it creates really big issues.

I asked you about the structure of the company last time, and you had a really interesting answer. You didn’t talk about the structure at all. You talked about the culture and said you want the culture of the company to be “one single word.” The word you picked was “humility,” and you talked about that for a minute. It’s been two years since then. I’ve come to believe that the structure question is really a proxy for a culture question. By describing the structure of the company, you’re describing the culture. Would you still pick “humility” if I asked you to describe the culture of the company?

I think at that time, humility was the most-needed aspect because we rode a very successful IPO, and our stock was very high. Many people made a lot of money. We lost a bit of humility at that point. Right now, we are back to our roots. I think the company has been through pain, and we understand better.

Look, I am not smart enough to learn from successes, and UiPath is not smart enough to learn from successes, but I think we are smart enough to learn from pain and suffering. Humility was in the genesis of our company and it’s an integral part. What we need now more is to be bold and fast. We are making a big push into our agentic automation era, and I see great things happening. It’s a new energy.

Also, we ran RPA (robotic process automation) for seven, eight years. There was a bit of fatigue at the end. We were just perfecting the software and getting into some white spaces, but it was not that exciting. Agentic AI brings a lot of excitement to the table. We pivoted in product and engineering overnight basically, more than half of the organization into building the new agentic products. All the teams are energized because they know. We basically put agentic automation as our number one priority as a company.

We literally changed direction. It’s not the Titanic, but it’s a big boat. I think very few companies have a chance for an act two, and we have this chance. AI and automation are so synergetic. I think more and more people came to that conclusion. Agentic, in essence, is AI plus automation. It’s the fusion of AI and automation. We’re so well-positioned to deliver on this promise. So our product and engineering is going at a breakneck pace, making really bold decisions. From a technology standpoint, we’ve replatformed our workflow engine to a more modern technology. They really embodied being bold and fast. I cannot say yet that this is true for other parts of the company, and this is where I work with our leaders to be completely prepared for our act two. 

I’m going to ask one more question about structure, and then I have a lot of questions about agentic AI and automation. One of the big decisions you made when you took over the role as sole CEO once again is you cut about 400 people. You laid off 10 percent of the company. Did you end up restructuring around that cut? Why make that decision, and what was the goal?

We looked into our central functions at that point. And in all fairness, we over-hired people in those central functions, and we had to streamline the organization. Decisions to fire people are the hardest from an emotional standpoint, from a cultural standpoint, and financially. It’s very hard to make them. Every time we had to do them, it’s been a thorough process. I was never rushing, and I was always fighting more on “do we really need to?” 

And it came at one of the lowest moments for us, along with the CEO changes. I think now, as we put it behind us, we are more prepared. The world is in an interesting, challenging phase right now. Nobody knows where it’s going to go. I think we as a company are a bit more prepared, more streamlined, and agile. We took time to heal the pain, and I think the confidence in the company is restoring. Looking back, I think that was the right thing to do for the company.

I wanted to ask that question as the lead-in to AI. You’re describing making these cuts as a low moment, as something that was very difficult to do. The right decision, but very difficult to do. You pull the thread on AI, and what I hear from our audience is, “This automation is going to come for our companies and we will all be out of a job.” White-collar workers will be out of a job. Software engineers might be out of a job. Lawyers are terrified of being out of a job. Do you see that connection, that if your software is successful you will reorient the economy and a lot of people might lose their jobs?

If we are realistic right now, it’s all a matter of the time of change, not the change itself. Your job and my job have changed over time. Jobs change. It’s a matter of when it’s going to be and how compressed the change is. Right now, I’m not so fearful that it’s going to come so suddenly. If you look at AI and the real use cases, we still have to see widespread adoption. It’s a productivity gain right now, more like an assistant type of AI. I ask something, I get that response, I do my job a bit faster and better. It’s not at the point yet to affect really huge volumes of the population.

I think agentic AI is one of the steps toward deploying AI into more of an enterprise context, and it might accelerate the way jobs are transforming. What do I mean by this? I think a job today is not a simple task. There are very few people whose job you can describe as one single task. So a job is a multitude of operational things, repetitive things, and many ad hoc things. It depends on different environments and businesses.

I think that many repetitive tasks have been solved. We have the technology to basically eliminate many of them from one’s job. Now we also have the technology to help people with more of these ad hoc tasks, like research tasks. I think the jobs will be moved more toward where people make decisions. They’ll analyze what information agents are retrieving and what they’re putting together. Agents plus automation. People will analyze, will make decisions, and then the actions will be carried on by enterprise workflows, robots that we already have. So jobs will transform more into decision-making, inspections, and overseeing from a command plane. 

I think about this all the time. I don’t know that I’m a great editor-in-chief. I feel like you could automate me by just walking into rooms and having a soundboard that says “make it shorter” or “make it longer,” and you just spin the wheel and pick one. But I know when to say those things because I spent years writing blog posts, then stories, and now podcasting. I have all this experience executing the decisions so that I have a high level of confidence in the decisions that I’m making when I make them.

How do you get that if no one is executing the decisions? If that’s all robots? I just want to make a comparison to you. You were the founder, you spent all this time running this company. How would you make good decisions if you didn’t have all of that experience?

The execution experience?

Yeah.

That’s a good question. Eventually, many things will be like a black box. I don’t know why if I press a key on my keyboard it displays on the screen, but I can make the decision to press. In a way, operations will be like a black box for many companies, and decisions will be at a higher level. I think we can still make decisions even if we don’t know how things are cooked behind the scenes.

I’m curious how that plays out. I am of the school that says the best leaders are the ones who spent time on the ground. That’s not always true. I’ve talked to a lot of leaders on the show, but particularly when I talk to founders, that experience at every stage of the company is what informs the confidence to make changes. If operations are a black box, I wonder where that confidence comes from.

I need to reflect more on that. Probably the best people will understand the operations as well. Even if they are carried out by robots and AI, they will understand in order to make better decisions and change the operations. But this is more of an analytical type of person. The types of jobs where there’s more mechanical typing, copying, and pasting are going to disappear.

So the last time you were on the show, I don’t think there was a lot of hype around RPA. I was into it because I’m fascinated by the idea of computers using computers, and when you were on the show in 2022 was sort of the height of that. You were riding high. This is why you said you needed humility. The idea was that instead of upgrading a lot of old computer systems, we would abstract them away with UiPath technology, build new interfaces, and that would allow all kinds of flexibility. That was a big idea. 

I think that has changed. In the AI age, we see a lot of companies promising agentic capabilities. We see a lot of companies saying that they’ll move even farther up the stack, all the way up to decision-making. But when I look back on that conversation and everything that’s happened since, the thing that gets me is that robotic process automation, the idea that you have some old hospital building’s system and UiPath will build a modern way to use it, is deterministic. You knew where all the buttons in that software were, you could program your way through them. Maybe you needed some machine learning to understand the interfaces better or to make it less brittle, but you knew what the inputs and the outputs were. RPA knows the path between those things.

AI is totally not deterministic. The robot’s going to go do something. Is there a connection between the software you were building, the RPA you are selling, and the agentic capabilities you want to build? Because it seems like there’s a fundamental technology shift that has to happen there.

I think you expressed the essence of what we are building when you say deterministic and non-deterministic. These are exactly the terms I use when I am explaining how robots and AI should interact. Look, LLMs are not meant to do deterministic tasks. If you ask an LLM to multiply two numbers, it cannot figure out how to multiply two numbers because it’s not statistical matching. What it would do best is understand, “Ah, I’m required to multiply two numbers. I have a tool that knows how to multiply two numbers, so I will call a tool and I will get the precise answer.” This is how they work. They don’t have the intelligence inside them because it’s a non-deterministic tool. It’s not meant to do a series of deterministic steps.

In the same way, you can think of transactional work that produces side effects in enterprise systems. It should be deterministic. You cannot have a 95 percent chance of succeeding a payment transaction. It has to be 100 percent, and if there is an exception, people should be notified. It cannot be “maybe.” 

Our robots offer this fully deterministic way to do transactions across multiple systems, transactions that create effects on these systems. With LLMs and with technology like OpenAI’s Operator or “computer use” from Anthropic — actually we are users, and we work closely with both of these companies to integrate their technology — you can complement what RPA is doing on parts of the process that you couldn’t automate before. If I have a process that relies on doing research… like if I’m traveling, I want to create a travel agent with AI. This travel agent will do research on available flights and across a multitude of airlines. It’s no big harm if I miss one flight option.

So I can have a non-deterministic tool, go and extract the information, then an agent can make some decisions. It can present to the user, “These are available flights.” But then when I book a flight, I have to use something deterministic. When the money transacts, money changes hands. Basically we can have the best of both worlds. We can extend the reach of deterministic with non-deterministic while accepting the risks of non-deterministic. And there are domains like research or testing an application when we can take more risks. It makes sense. It depends on your level of risk you can accept.

It makes sense to me. I see your competitors and your partners, like OpenAI and Anthropic, and they’ve made their entire technology bet on agentic AI. I assume that their plan is for that to get good enough to do everything. Your approach is that there’s some stuff that traditional RPA, the traditional deterministic computer, needs to do, and that can be layered with an LLM or an AI system. I’m just wondering what the intersection point is. Will there ever be an intersection point when OpenAI says, “Operator can do it all,” and that presents some kind of paradigm shift for your business?

I am absolutely sure that the intersection point is when you can define a task in a deterministic way and know the steps. There is really no point in having an LLM that does this task all the time to rediscover how to do it or to think about every step because it’s impossible to get to 100 percent accuracy. We are testing these LLMs for simple form filling. They can work very well, but think about it. You need to run it hundreds, or even thousands of times to get to 100 percent accuracy. This is not what the technology is for.

What I am saying is that LLMs will eventually create routines that can work 100 percent accurately. But the idea that LLMs will discover a process every time like you would when you see an application or a book for the first time in your life… humans don’t work like this. We learn. You learn an application, and then if you watch yourself, most of the things you’ll do will be on autopilot.

We’ve had other companies come on the show and talk about their agentic software approaches. Actually, they were facsimiles of the agentic software they wanted to build. So, Rabbit came on the show, and its first version of the Rabbit R1 was running testing software in the background. You would ask for a song on Spotify, and it would just click around on the Spotify website in the cloud and then stream the song to you. Its claim was that it actually did build the agent, but it needed to build the first version and have proof of concept. 

But the deterministic system, in one very real way, can act like the thing people want from the AI system. It can almost do it and then it’s brittle, but the AI can make it less brittle by reacting to change or an unexpected outcome. How do you merge those things together? How do you decide which system to use? Because that seems like the technology problem of the moment.

The way we are seeing the adoption of combined agentic AI and automation is by putting a workflow technology on top of it. Our agents are more like data-in, action-out agents — not necessarily conversational agents. We focus on delivering enterprise agents that work in the context of an enterprise process. So to us, the critical piece is this orchestration part. 

Let’s say you have a loan agent that has to approve some loans. A workflow is triggered when the loan application is received. So, you have an enterprise workflow. Then, that workflow will first send the application to a reading agent that is specialized in extracting the information from the application. Then, I can send it to a human user to verify something basic if I’m not confident enough in what I extracted. It can be a more junior person that does this verification. 

Then, the workflow will send it to an agent that will make loan recommendations. That agent can start to call tools like, “Get this person’s credit score.” So this tool is definitely something deterministic. It’s either an API to a credit score agency or you can use an RPA bot. That is clearly deterministic. You are not going to use something like OpenAI’s Operator to just figure out a guy’s credit score. There is absolutely no point. It’s taking too much time and it’s not reliable. 

Already you see it’s a combination. The workflow knows how to direct the fixed paths of a process, and then agents are capable of making recommendations and calling tools that will give the context. Then, after the agent makes a recommendation to approve this loan, it will go to a human user. The workflow will create a task, a human user will get it in their inbox asking them to approve or not. They press a button and approve. The workflow will go back maybe to the last agent and say, “Please compose a nice acceptance message particular to this client.”

It’s a simplistic view, but this is how we believe the world and enterprise customers will adopt agents. Also, they need to have some confidence in the system. You said we are talking about this black box system, a swarm of agents that do their magic and sometimes they make mistakes. Until you accept it, you need to have confidence and you need to see the work. Everybody is more confident when they see the workflow. They can say, “Look, if that happens, it goes like this. If that happens, it goes like this.” So you can trace it, you can understand it, you can reason with it.

One of my takes on the interaction between humans and AI is that for a long time we have to speak the same language. Even when you create an application or an automation, AI actually creates code. AI can eventually work directly with machine code. They don’t have to create Python code, but it’s important that AI creates Python code because humans can reason, change, and accept it. It’s going to be the same in automation applications. AI will use existing platforms, will create artifacts on top of those existing platforms, and people will validate what’s going on there.

On the consumer side, the value of the existing platforms is, I think, under enormous threat. So I call this the “DoorDash problem” on the consumer side. We just had Amazon’s Panos Panay on, and it’s rolling out a new version of Alexa. You’re going to be able to say, “Alexa, buy me a sandwich,” and it will just get DoorDash to send you a sandwich. 

This is a huge problem for DoorDash. Its margins are under significant pressure if their interface gets commoditized in that way. We’re going to have the CEO of DoorDash on the show eventually and I’ll ask him this question. but I can abstractly see the pressure on some of these systems that are going to get commoditized by new kinds of interfaces.

The classic RPA truly depended on those systems existing. You needed the existing loan system that nobody wanted to upgrade so you could build the RPA interface on top of it. You need the credit score interface that might not have a great API, but you can use RPA to go get it from their website. AI changes that because it’s coming to all of those systems as well. There’s some part of the AI industry that’s chasing all of those things at once, not just building this orchestration layer. 

What do you think about the long-term longevity of those systems? I look on the consumer side and I say, “Oh, this is a big problem for DoorDash. This is a big problem for Uber.” I don’t know exactly how it works on the enterprise side.

We’ll see how it evolves. The fact that we still have a lot of mainframes, and our RPA touches a lot of mainframes, shows that the changing of enterprise systems is much more difficult than in the consumer space. If you look at complex enterprise applications like Workday and SAP, I can see people adding a nice layer of voice on top that’s AI-powered. You know, “Change my vacation responder to this.”

But the tablet and mobile phone didn’t make the keyboard or mouse obsolete. I think they will still have to coexist. Many people can work on user interfaces faster with a keyboard than with voice, but voice is going to become a good way to interact with applications. When you need to absorb a lot of information simultaneously, you need the user interface. In many cases, you’ll still need to interact with it. It’s easier than telling the AI, “Please press the okay button.” I will just go and click the button. It’s easier and it’s faster. They have to coexist. 

I was thinking about the DoorDash problem. You’re basically saying that Amazon can build its own DoorDash. If it can control the interface with the client, it doesn’t matter who delivers in the end because– 

It’s not that they will build their own DoorDash. It’s that DoorDash’s opportunities to make additional revenue will go away. It won’t be able to upsell, won’t be able to do deals, won’t be able to have exclusives. The interface will be commoditized and it will just become a service provider with Amazon or whoever’s AI agent being in control of the business. You see that for a lot of these ideas. You need an ecosystem of service providers for the agent to go and address, and that crushes the margins of the service providers.

It’s possible.

I think I see it in the consumer space. You see the back and forth. There’s some amount of, “We don’t want you here. We’re going to block your agents from using our services.” That is already happening on the consumer side. There’s some amount of dealmaking. Then on the enterprise side, it seems like there’s going to be a lot of dealmaking where maybe instead of API access, we’re allowing agentic access or RPA access because the data is what’s valuable there.

To a certain extent, we had the same problem with RPA. Think about the fact that most enterprise or SaaS software was licensed by user seats. With RPA, you needed far fewer user seats. You can have one seat that does the job of hundreds of seats. They found ways to kind of prevent this and create special service accounts to deal with it. Some vendors do not allow it. I’m sure they will find some ways to deal with it because how can Alexa order if DoorDash doesn’t want to receive the order? There has to be something in it for both of them.

I think that’s an enormous technical challenge, and the business challenge is even harder. You have to get a lot of people to agree to fundamentally restructure their businesses in order for any of this to work. Again, on the enterprise side, there’s more dealmaking. You have some instincts, some history, some moves to say, “Okay, here’s how we’re going to structure access to the data.” I have no idea how it’ll play out on the consumer side.

You mentioned a thing about LLMs not having memory, having to rethink the workflow every single time. That’s true. I think the AI companies are working on that. But they’re also pushing the idea of reasoning, that now we’re going to layer LLM approaches over and over again into a simulacrum of human reasoning. I don’t know if that’s correct. They say they can do it. Is that having an impact on what you’re doing? Can you say, “Here’s the decision, here’s the process by which a decision is made”?

The way we are seeing the reasoning part is that it’s more helpful, in our world, for creating automations. We have this Copilot-type of technology where you describe a process and it can create the artifact to execute the process. The smarter an LLM is, the closer to reality the creation gets and the developer has to change it less. So in a way, it’s like creating code, if you want. It’s the same thing. The smarter LLMs will create better code, but that code is still going to be executed by hyperscalers. It’s not LLMs that do that. Think about it. Maybe LLMs will do everything. Why would they generate code at all?

You mentioned hyperscalers. One of the things that I’ve been thinking about a lot is the amount of investment the hyperscalers are doing just to buy Nvidia chips, to build data centers, or to invest in nuclear fusion against the promise that there will be this much demand for AI services.

They have to make money doing this somehow. It’s unclear how the bleeding edge, frontier AI companies are going to make money. I don’t know how OpenAI will ever make a dollar. I don’t know how Anthropic will ever make a dollar except by raising more money, which they’re very good at. That’s on a long-term plan. You’re a public company. You have to make the money. You have to buy the tokens, you have to use them, you have to build the products, you have to charge your market price. Are the rates we’re at now sustainable?

I don’t know if it’s sustainable or not for them, but if I were them, I would do the same. What if this is indeed the biggest revolution of our time? What if all of these GPUs and AI agents will take over the world and I am not there?

But I’m saying you’ve got to charge your customer some price for the use of an AI tool. You’re not running all of your own models. You’re partnered with some of these companies. You’re buying some of their capacity. They’re, in turn, buying capacity from Azure, AWS, or whatever they’re running on. All of these companies need a margin and some of their margins are negative. OpenAI loses money on inference right now, but it’s selling that capacity to you. 

At some point, they’re going to turn the knob and say, “We’ve got to make money.” They’re going to raise prices on you, and you’ll have to pass that cost to your actual customers who are actual businesses trying to automate their companies and raise their own margins. When will it become too expensive? That seems like the correction that’s coming. You’re going to say, “Okay, OpenAI raised our prices. UiPath has raised its prices,” and some customers are going to say no.

If we look at through our lens of the processes we automate, what’s the alternative at this point? Using human labor? I think even if OpenAI increases prices, I still don’t think humans can compete with AI plus automation when it is possible. And long term, the pricing will go down and it’s a lot of competition for the business. I’m not really concerned about this aspect.

Have you structured your technology so that you can swap between AI providers? Are you tied to OpenAI, Anthropic, or is that easily modular?

No, not at all. We actually offer our customers a piece of technology that we call AI Trust Layer, where they can switch between different providers or bring their own on-prem model if they want.

You just bought a company called Peak, which is another AI provider. Why make that bet? Why bring in technology?

We want to get into vertical agents. Peak is a pricing and inventory agent, and it has really solid experience in delivering these dedicated solutions based on agentic AI, and we want to extend that. Of course, we’ll integrate it first into our platform, but we want to come out with more dedicated agents. It makes the entire go-to-market easier. We want it to work a bit like a locomotive for the entire platform because it can create more demand for automation.

How does that technology plug into your existing stack? I understand it has markets you might not have or that you want to get bigger in, but ideally you buy a company and what you’re going to do is sell its existing markets more of your tools.

Definitely. That was on our mind. I think we have really good synergies in our go-to-market, and we can really accelerate its go-to-market, particularly in the manufacturing industries. We have very solid manufacturing practices in the US, Germany, and Japan.

Do you think there’s an opportunity for you to commoditize the AI tools themselves? I just keep thinking about this. You have your AI Trust Layer, you have your own vertical systems that you’re buying that you might deploy. At some point, what matters to companies is the business outcome, not that they have an OpenAI partnership. It feels like the big AI companies are trying to be everything to everyone, and you’re trying to specialize. Do you think at some point you’re going to say, “What we deliver are business outcomes and the technology doesn’t actually matter”?

I think that generative AI is going through this phase. Initially, it was a nice toy. Everybody put budgets to experiment with it, and now we are moving toward the phase where people really want outcomes. Initially, they all used OpenAI, and our strategy was to use OpenAI because it’s the best. If you want to make a proof of concept, why would you use something different? 

But as you go and you specialize it for different types of industries and processes, you can choose whatever is more appropriate. We look at everything from DeepSeek, Llama, to Anthropic. We use all of them in different parts of the business. In the end, we are more of an AI engineering company, and our job is to build nice products that deliver value for customers. Behind the scenes, we use whatever LLMs are best for a particular scenario.

I actually want to ask you about DeepSeek. Was that as shocking of a moment for you? The industry reacted to the idea that you could run the model much more cheaply very harshly — very harshly. Did you see that and say, “This will bring my cost down. This is also a revolution”?

Selfishly, for UiPath, any open-source capable model is a great thing for us and for our customers. My belief is that these dedicated agents will require a combination of fine-tuning and really good prompts. So, if you can have a great model that you can fine-tune and combine with good prompts, that will provide the highest value and the cheapest price. We find you can actually distill it into a smaller model that works very well for a particular domain.

Where do you see the biggest growth for traditional RPA, for AI, and for the hybrid of AI and RPA?

RPA is an established industry right now that grows in the low double-digits. The demand that we are seeing right now for our agentic technology, I have never seen in the RPA world. It really opens all the doors. We get a seat at the table where we are not used to being from the automation perspective. People are really excited about this idea of agentic automation. They get it. The value proposition is kind of simple for us. I can go to my clients and tell them, “Guys, where did you deploy robots? How are people interacting with the robots today? Why are we not reducing the work of people, deploying agents, and creating an enterprise workflow that will connect agents with people and robots?” It’s a no-brainer proposition. It resonates, it’s simple, it creates a lot of excitement.

I want to tell you about my favorite Slack room at Vox Media and get your reaction to it.

We have a room called Finance Support, and in this room, people ask a Slack robot to do stuff: file invoices, give receipts, all this stuff. I look at this room once a week, and it cracks me up every time. I literally fall over and giggle every time because the people who are new to this room type full sentences: “Hi, I need help with this receipt. Can you itemize this thing? I’ve got a flight.” The people who are repeat users have discovered that they just need to scream nouns at a robot.

So they just show up and they just say the word “expenses,” and all of this is in one stack. There are people who are very polite and then people who are just yelling nouns at a robot. You can see this secondary language of human-machine interaction developing: “I’m just going to say keywords to the robot because that’s all it needs from me.”

I look at that and I say, “Oh, that’s a revolution.” First of all, it’s very funny. But this is a revolution in business. You’re going to have some people who are just saying keywords in Slack to get things done for their business to an agent that might just go off and do it, and then you have the people who are used to all of the niceties of business fluffing up their communication. At some point, you’re just going to have robots saying nouns to each other instead of using an API. In many ways, that’s what RPA was. You’re just using the human interface instead of an API. Do you see all of business changing around this as clearly as I do when I look at this Slack room?

Yeah, and even for RPA, this is true. Many people are using RPA by creating a Slack channel that connects directly with a robot that does something. AI just extends the same idea. To me, it’s kind of fascinating how we communicate with bots. I discovered myself — well, maybe it’s just an impression — but if I say, “please,” I think that LLMs come back with better responses. [Laughs]

Here’s something I also worry about. You’re the CEO. You get a lot of emails, you send a lot of emails. Do you ever worry about the loop where you’re responding to an email that was written by AI with another email that’s written by AI and suddenly everyone’s just pushing the summarize button and no one’s actually talking?

I personally write my emails because everybody in the company and clients knows my own tone and my broken English. So I cannot use LLMs. But yes, I’ve seen many instances where it looks like LLMs are talking to each other.

You’re the automation vendor. LLMs talking to each other — there’s something hollow there, right? Is that something you want to achieve with your products, or is it something you’re trying to avoid? 

I think to a certain extent we want to achieve that with our product. We want to facilitate agents talking to each other, but in a more controlled environment.

Daniel, you’ve given us so much time. You’re going to have to come back. I feel like I could just talk about the philosophical repercussions of all of these systems with you for many more hours, but you’ve given us so much time. Thank you for being on Decoder.

Thank you so much. 

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

On today’s episode of Decoder, we’re talking about AI, art, and the controversial collision between the two — a debate that, to be honest, is an absolute mess. If you’ve been on the internet this past week, you undoubtedly know that controversy was just kicked up a notch by the Studio Ghibli memes — pictures cribbing the style of the legendary Japanese film studio. These images, powered by OpenAI’s new image generator, are everywhere; OpenAI CEO Sam Altman has even been posting some examples to his personal X account. And they’ve widened an already pretty stark rift between AI boosters and critics. 

Brian Merchant, a good friend of The Verge and author of the newsletter and book Blood in the Machine, wrote one of the best analyses of the Ghibli trend last week. So I invited him onto the show to discuss this particular situation and also to help me figure out the ongoing AI art debate more broadly as it continues to collide with legal frameworks like copyright. 

Merchant and I tend to agree a lot more than we disagree when it comes to the technology industry. So I did my best to really take the other side here and push on these ideas as hard as I could. Technology and art have always been in a dance with each other; that’s part of the founding ethos of The Verge. So I think it’s important to put AI in that context — not least because we can see the obvious joy regular people find in using some of these tools to express themselves in ways they might not otherwise be able to.

But there’s expressing yourself, and then there’s churning out AI anime slop that is designed to evoke classics like Spirited Away and My Neighbor Totoro in a way that devalues and even outright steals from actual human artists. Add in the way the Trump administration jumped on the trend by “Ghibli-fying” a deportation photo, and it’s not hard to see why a lot of folks perceive this tool as utterly grotesque and offensive.  Or “an insult to life itself,” as Ghibli cofounder Hayao Miyazaki once famously said of an AI demo he witnessed in 2016. 

So you’ll hear Merchant and I really go back and forth, digging in on what this all means — for art and artists, and for a creative economy that has long since transitioned from the world of physical scarcity to one of limitless digital supply. And, most importantly, we spent a lot of time talking about how we should feel using these tools at all when they might pose very real threats to people’s livelihoods and the ongoing climate crisis. 

I’ll warn you: there are no easy answers here, and I don’t think Merchant and I came to a single conclusion. I don’t even think we wanted to. But I think this conversation helped me consider more clearly how to think about AI and art. Let me know what you think.

If you’d like to read more on what we talked about in this episode, check out the links below:

• OpenAI’s Studio Ghibli meme factory is an insult to art itself | Brian Merchant
• Seattle engineer’s Ghibli-style image goes viral | Seattle Times
• OpenAI just raised another $40 billion round from SoftBank | The Verge
• ChatGPT “added one million users in the last hour.” | The Verge
• ChatGPT’s Ghibli filter is political now, but it always was | The Verge
• OpenAI, Google ask the government to let them train on content they don’t own | The Verge
• Studio Ghibli in the age of A.I. reproduction | Max Read
• OpenAI has a Studio Ghibli problem | Vergecast
• AI slop is a brute force attack on the algorithms that control reality | 404 Media
• The New Aesthetics of Fascism | New Socialist

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

I have been recommending people buy whatever Brother laser printer is on sale for three years now, and no one has ever gotten mad at me about it. My own Brother laser printer, whose model number I no longer remember and do not care about, has been operating flawlessly for nearly 10 years now. We use it to print return labels for things we’ve purchased online in a losing effort to dull the pain of modernity, and my wife is a lawyer, a job that requires printing documents and scowling at them several times a day. We have replaced the toner once in that time, and it has never asked me to sign up for a subscription or fallen off the WiFi.

Our newsroom doesn’t have anything to do with Vox Media’s affiliate deals because of our precious ethics policy, but here’s some space I left for the company’s commerce team to put a buy button that kicks them back money if you push it and purchase a printer:

This is the third year in a row that I’ve published a story recommending you just stop thinking about printers and buy whatever random Brother laser printer is on sale, and nothing has happened in the miserably user-hostile printer industry to change my recommendation in that time. …

Read the full story at The Verge.

Today, I’m talking with Matt Bromberg, who is the CEO of Unity, one of the leading video game development platforms. He’s been in the job for less than a year, and in many ways, he is the perfect Decoder guest.

Unity is one of those companies that we love here on Decoder because we interact with its products all the time, but the company itself is somewhat hidden from view. Unity doesn’t make or publish its own games. Instead, its core product, the Unity game engine, is what all kinds of games — especially mobile ones — run on. If you play games on your phone, which you probably do, you’ve used a product built with Unity. It’s not just on phones by a long shot, and you’ll hear us talk about that quite a bit, too, but Unity as a company is what makes a core part of the mobile phone experience work.

That’s a Decoder all by itself, but Bromberg took over the job of CEO in a moment of crisis, and he’s made significant changes to the company since. You’ll hear him describe Unity as being “at war with its customers” before he joined, and he’s not wrong. The former CEO, John Riccitiello, led the company through a major pricing model change, called the runtime fee, that was going to raise costs for many of its customers. The communication around that pricing model change was also, frankly, terrible. Unity developers were in open rebellion, leading to Riccitiello’s resignation and, eventually, the company’s decision to scrap the runtime fee altogether.

So you’ll hear me ask Bromberg about his decision to walk the fee back after he joined, because it was clear to him even before he took the job that he would need to make that choice if there was going to be a Unity left for him to run. And you’ll hear him explain that while deciding to make such a big change was easy for him, working out how to execute that change slowed him down a bit.

Bromberg’s in charge of Unity during a moment of contraction for the game industry overall — studios are closing, and some big bets on things like the Metaverse and live service games haven’t paid off. So we talked about all that and where he sees growth ahead: Unity isn’t just a game engine provider, but the platform for everything running those big live services and the monetization on top of that.

One note before we start, because it really grabbed me: Bromberg’s one of the first CEOs to come on this show and disagree with me about the importance of structure. I tend to think of a company’s org chart as a proxy for its culture and its tradeoffs. I’m always saying that if you tell me your org chart, I can guess 80 percent of your problems, and so we joke that Decoder is really a show about org charts. But Bromberg told me he thinks culture isn’t enabled by structure. You’ll hear what he means.

Unity CEO Matt Bromberg. Here we go.

This interview has been lightly edited for length and clarity.

Matt Bromberg, you are the new CEO of Unity. Welcome to Decoder.

Thank you so much for having me. It’s really my pleasure.

I am very excited to talk to you. This is a perfect episode of Decoder: You are less than a year on the job as the new CEO. You took over, you made a bunch of changes to the company, you restructured the company, you hired a new C-suite, you used to work for the old CEO — and then, Unity itself is a very core piece of technology that I think is under-covered. So, it’s like a perfect episode. Thanks for joining us.

You bet.

Unity really is kind of under-covered. It’s on everyone’s phone: Everyone has games on their phone, and Unity is almost certainly powering those games. What is Unity? What are game engines?

Game engines are effectively the thing that developers use to build games. And maybe that’s obvious, but the technology is pretty deep and pretty complex. Unity is a platform where developers can aggregate all the tools, technologies, and content they need to build interactive applications. And yeah, to your point, it’s a really important tool in our world. About seven out of every 10 mobile games in the world are built on Unity. About 30 percent of the [top 1000] PC games in the world. Seven out of 10 best AR games were built in Unity. Last night, or two nights ago, at GDC, eight out of 10 of the top independent games that won awards were built on Unity. We’re an important partner in that ecosystem.

There’s a lot of history in game engines and investments in game engine development. A lot of companies that made popular games and then thought they could peel the engine out of those games and turn that into a business; that’s worked or, mostly, not worked in various ways. Unity is just an engine company. You’re not trying to make consumer-facing games.

That is correct. We do a little bit. It’s funny you should say that. Generally speaking, that’s absolutely true. We are not a game developer and a publisher. We have recently been doing more development ourselves on a couple of projects. But only to ensure that our tools are production-ready and dogfooding what we’re doing. We have an expert team that does some of that, but it’s really about feeding back into our development process more than trying to be a games company. I was the chief operating officer at Zynga for a lot of years —I know what it is to run a games company. I think it can be very hard and unnecessary to try to run a games company and an engine company.

I made that comparison explicitly because your biggest competitor is Unreal Engine, which is Epic Games. Epic is a games company, they’re the Fortnite company. They’re doing both, and then they’re expanding their engine into Hollywood, into automobile design, all this other stuff. You seem pretty narrowly focused on games still. What do you think about that competition?

There’s a lot to unpack there. First, there are increasingly lots of applications of Unity outside of gaming. And in fact, what we call industries is the fastest-growing part of our subscription business.

But we are, to your point, much more focused now than we were before. We were chasing sort of everything: Hollywood, architects, digital twins for building nuclear reactors, and all sorts of things. Right now, we’re really just focused on a few core verticals.

For example, most of the in-dash experiences in cars are built on top of Unity. If you think about the little computer that’s in the dashboard of your car, it feels a lot like a phone or some other small device. And that’s really our strength to be able to design and develop interactive experiences for those things. [We work in] virtual, retail, and lots of applications on the manufacturing floor, so we do a fair bit outside of gaming.

I guess the other thing to understand is that Unity as a platform is focused on making games, but we are the only company in the world that helps developers through the whole lifecycle of development. From prototyping through building, operating in life service, and then crucially to acquiring new players, monetizing, and inventorying your games. The big second piece for us is the advertising and marketing piece. It’s a piece that we think grows really organically out of our core business, which is about deeply understanding consumers, and helping developers build for those consumers. And then when they turn to that next phase where they need to acquire players, we’re also there to help them.

I kind of understand why all the car makers use Unity. They under-spec the processors in their cars, and the closest comparison to that is a mobile phone, which is power constrained. I don’t think a car is power constrained; I think they’re just cheap, but there they are. They’re just using under-specced mobile processors to drive their dashboards. Unity is a great engine for mobile phones. It makes sense that you just expand in that direction.

But that’s not games, right? That’s just another set of constraints that looks familiar, so you end up there. When you talk about games, we’ve just gone through an apocalypse of live service games recently. That industry doesn’t seem as predictable. It doesn’t seem like the games industry knows where its next set of growth will be, or how many people it should employ quite broadly. There’s some big change there that it feels like I can’t draw the line as quickly as I can to the expansion into cars, right?

Yeah, listen, there is a lot of consternation about the video game industry. I’m someone who shares that. I’ve been around long enough to know that towards the end of hardware cycles, everybody’s throwing up their hands and saying, “Gee, I don’t think gaming’s doing so well.”

I’ve been through this cycle several times. Here’s what happens: Creative people innovate, and they make new experiences. Those experiences explode into hits, and that drives growth. That has always happened. I think the pain that we’re feeling right now in the industry is principally a pain born of creative destruction, so I accept that there’s been a fair bit of pain. There’ve been a lot of layoffs and other really unfortunate things. I think that’s about folks reimagining themselves for the future. The explosion of new devices and new experiences is going to be the gateway to that innovation.

The thing about Unity is…  It’s not so much that we’re optimized to run on low-spec phones, it’s that when you build any experience in Unity it can operate at a really high level of performance fidelity anywhere. And that’s where consumers are moving. We used to be in this world where, and I remember this… I worked at Electronic Arts for many years. We were in this world where folks believed that visual fidelity and high-quality visuals were the primary things that video game consumers wanted, and they were obsessed with it. And I think what we’ve realized now is: that’s not what’s driving the consumer. I don’t think it ever has been what’s driving the consumer. We’re really proud of the quality of our visuals, but what drives consumer adoption is gameplay innovation. And all that is what’s going to drive this industry forward. I’m hugely optimistic about it.

Keep in mind, again, web games, AR, and VR. We’re going to get a new Nintendo platform. We’re going to get new Sony and Microsoft platforms. Phones are going to become ever more powerful. You’ve got dual-screen phones. The form of games will change, and the nature of them will change, but the appetite for interactive experiences, to my eye, has not lessened in any way.

Just in the context of what a game engine is, what you’re describing here is: Let’s say I have an idea for a game. I see a market that’s changing. Unity is going to provide me with the physics engine that lets me build the game, with the ability to render the game in high fidelity across a number of platforms. And that number of platforms is important because a game needs to be everywhere now. You’ll operate the live service that lets people connect and play multiplayer, and then you’ll help me monetize the game by putting ads in it. Which part is the hardest right now? Because it seems like the games industry is “every part is the hardest,” and it seems like you have a much clearer view. What do you think is actually the hardest part?

My experience of making video games for many, many years — which is what I did before I came to Unity — is that generally speaking, especially for game publishers of any scale, there’s a really simple equation. At the end of every year, they look at their P&L and say, “How much money can I spend next year making new things?” They usually have a big life service or many, and that’s what’s kind of driving forward the business on a day-to-day basis. But the video games world is about hits and about new experiences, and you have to make investments. You also have to control and modulate those investments. The most important equation is people and time. If it takes me, if it’s 300 people and three years to make a game, versus 50 people and a year, that equation is very, very different.

The amount of innovation I can afford, the amount of new starts I can afford, and the amount of marketing I can afford, are very different in both equations. To me, when I think about Unity’s place in the world and where we can be the most help, it’s there. We’re a proud tools provider. We help video game developers make games, hopefully, as efficiently, quickly, and easily as possible so they can spend more time creating innovation. And if we’re really good at our job, they can have more starts, because they’re moving more quickly and they can do it more efficiently. And that’s the role I’d love to see us play there. And then to your point, as many new starts as we can get out there, great. Now we’ve got to go out and find new players. So, that’s the flywheel you want to see.

Let me ask one more kind of existential question. I want to get into the changes you’ve been making here. I feel like every industry, but in particular, the games industry got fundamentally confused by the combo platter of the pandemic and Fortnite. Meta was like, “We’re all going to wear headsets all day long in our house.” And I don’t know, it doesn’t seem like that’s happening. A bunch of companies thought, “Okay, we can see all these people playing Fortnite. The future of live music is going to happen in video games, and we’re going to build these huge live service multi-experience games, and everyone’s going to spend all their time in them.” And all that’s left from what I can tell, that’s sustainable, is Fortnite. There’s not another Fortnite, and that drives a lot of your competitor’s business. That drives a lot of Epic’s business. Do you see that as a mistake that the industry is correcting, or do you see other games being able to reach that level and then go even one step farther into, “Okay, we’re having true metaverse experiences”? Because I don’t see that anywhere right now.

I don’t think about it exactly that way. Listen, I think that the pandemic more than anything pulled demand forward. It pulled about a couple of years of consumer demand forward, and then we did see some natural slackening after having pulled all that forward. And I think that part of the dynamic is obvious.

Meta renamed the company, and they were like, “We’re not Facebook anymore. We’re Meta now.” And I don’t know if that’s paid off for them.

I was never a massive believer in the metaverse during that period of time. I’ll tell you why: Because, as a game maker, I experienced all those new platforms and just thought they were garbage. And I thought, “This looks like the games we tried to make 15 years ago. There is no way that’s a sustainable consumer experience.” All sorts of metaverse companies — I was completely confused by them. Having said that, I actually continue to believe, and I see the vibrancy of big live service platforms and communities as being fundamental to the games business. I mean, yeah, Fortnite, but Roblox, FIFA, and every game… Every major company has many enormous live services with millions or tens of millions of players. And in some ways that is the fundamental feature of the video game business right now — 80 percent of the people are deeply invested in this experience that they’ve been playing for years.

And I think one of the challenges now is, “Okay, how do I get those folks to move to something new?” I think that’s the push-pull that we’ve had. But I see platforms, communities, and interactive communities of gaming continuing to dominate and explode. And I do think we’ll see some new things. I think we’ll see innovation, some lighter experiences that will begin to challenge those. But I wouldn’t confuse the failure of the metaverse with some lack of sustainability in major life service gaming.

That’s a connection that has always been the most interesting to me. If you believe that the world is going to be big live service games, it’s not that much of a jump to say, “And we’ll all wear headsets all day long.” We’re spending time virtually which will just get more immersive. That’s a straight line. But it seems like we didn’t hit the first spot.

No, I think some of that has to do with — again, put aside the idiocy of some of the metaverse stuff… The future, to some degree, is going to be tied to massive consumer adoption of peripherals, or maybe that’s an old-fashioned word for it, but new devices. I’m an enormous believer in AR, for example. I think that one of the things that happened during the course of the pandemic and the couple of years since is we’re starting to get closer and closer with every rev to something that’s real here. I have no doubt that a couple of years from now everybody’s going to be wearing AR glasses. And how long has it been since Google Glass came out? But the combination of AI and voice, which enables really easy interactions with the form factor, the battery life that is now possible, and the ability to overlay information and services in front of your eyes, to me, is obviously going to explode.

And we’re going to look back and think about the time when we kept reaching into our pockets to pull out this thing for everything, it’s going to seem quaint. But it takes a long time to get true mass consumer adoption of these devices because it all has to be perfect. But once it hits, it explodes.

You’re a glasses wearer. I can see you wearing your glasses. I wear contacts. My theory is that you have to deliver far more value than is required to care for whatever you mount to your face. And glasses deliver an extraordinary amount of value. I’m horrible at maintaining my glasses; they’re very scratched, they’re very dirty. I wear them just to go to bed. I don’t have to care for them, but they deliver. I’m allowed to see, that’s a huge value amount.

But you go outside in the sun and you put on sunglasses, right?

But I only get value at that time. I don’t come back in and keep wearing my sunglasses, which is what I think Meta wants me to do. I think they want us to believe that transition lenses are going to be cool, and maybe they’ll pull that off. Who knows? Mark Zuckerberg’s wearing a chain. He’s got a haircut. Maybe he’s going to pull this off.

Well, we will see.

My point is, right now the most compelling AR experience is way down on the curve, right? It’s actually the Vision Pro, which is a huge thing that has an external battery and limited app support. Do you see, when you put on the Vision Pro, “Okay, I can get from here to there if the hardware improves on some curve?” Because when I reviewed it my conclusion was, “We’re racing down a dead end. This is the best pass-through VR that I’ve ever seen, and we should not keep doing this. It’s never going to get better than this.”

We partnered with Apple in the creation of that device. And I would say this, I would not… History is littered with first-generation devices that seemed absurd and then in another form became ubiquitous. And I do fundamentally believe that when the device form factors get right, consumers will adopt them. I do think it’ll be a mix though. I think one of the things we learned is that there’s a pretty high bar for folks wanting to isolate themselves completely from the world. And that’s why. When I think about AR, I think that the first mass-adoption devices we’ll see will be those that allow people to continue to be social and interact in the world while overlaying or enhancing that experience. Completely absorptive experiences have… I think folks will engage in those things, but there’ll be a time and a place for it. So, my view of that device was that it was a first swing. And also, by the way, Apple has a very long history of that as well. The first generation of devices is not necessarily the thing that takes off. So, I continue to be hugely encouraged by that.

Are you of the view, I think the cliche was, that this is a simulator for the thing they want to actually build?

I don’t know. I mean, I guess maybe that would be something you’d say after you’ve shipped it, maybe it didn’t go quite as well as you’d want to. But I applaud their willingness to put a stake in the ground, get out there with something, and then go back to the drawing board and continue to invest. At the end of the day, all of us making technology products need to do that. So, pushing the envelope for me is a good thing.

I ask this, because again, with the games industry as it is currently configured, we are at the end of a console generation. It appears to be restructuring, reconsolidating in some ways, and falling apart in other ways. It was just GDC. There’s just a lot of turmoil when you get the whole games industry in one conference center and talk about the future. And the big bets, like “We’re all going to play a bunch of AR games,” while Pokemon Go just got sold. A lot of the big bets seem to be structurally changing.

Unity obviously plays a part in that. But the sugar rush of the pandemic led Unity itself to make big decisions about how much demand there would be that fundamentally you’ve had to undo. The decision to charge the runtime fee feels like your predecessor as CEO was saying, “Okay, there’s going to be this much more demand for all this stuff. I can now extract a toll.” Do you see that in that framework?

I don’t. I don’t. By the way, I also don’t, fundamentally… I believe in the vibrancy and growth of the video game business, which, by the way, is still growing nicely and continues to be larger than streaming media, streaming music, and theatrical distribution combined. And this you will grow in high single digits, or maybe low double digits. I see this business as really vibrant. I don’t think the things that Unity has done in the past were sort of a … And frankly, the other folks were a direct response to demand… I accept it maybe as folks staffed up across the industry and then demand slackened a little bit. I don’t think that’s so much this year, but maybe in the year before — you had some dislocation.

But I think for us, the thing about Unity which is spectacular and fascinating, is that we sit at the intersection of so many powerful forces: 3D, digital advertising, interactive entertainment, AI, I could go on. And you could convince yourself that you have a meaningful role to play in all of that. Because we have this platform and this engine, this tool that is connected deeply to all those powerful forces. But you have to make choices in the world if you want to execute at an exceptionally high level. And the way we’re thinking about this business now is that we are going to make some really focused choices, hit the ball out of the park around those things, kind of re-earn the trust of customers, be better partners, and deliver at a much higher level. And that all good things are going to come from that. Which doesn’t… It’s great to be expansive, but if you want to deliver, you have to make choices.

I think this is a perfect transition to what I think of as the Decoder questions. I want to start with a quote. You gave an interview to IGN where you said, “We want to be a fundamentally different and better company. It is what we want. We have to have a fundamentally different relationship with our customers and our community. We want to develop and deliver products in a fundamentally different way. And that starts with you thinking about it differently.”

One of my ideas for the show is that structure is a proxy for culture. You have to build a corporate structure that enables the culture you want. You’re basically talking about totally transforming the company, in that quote. Everything has to be different, down from how you deliver the products to what your relationship with the customer is. You’ve been there less than a year. How have you restructured Unity to get to the culture you want?

It’s funny. Culture is a really interesting thing. And the way I think about culture is that I think it’s actually fundamentally backward-looking in a way. What I mean by that is, when we think about our culture in the context of a company, we usually, what we’re usually doing is saying, “Hey, we got together to do this thing and it went pretty well.” Now, let’s stop for a second and look backward and ask ourselves, “How were we behaving? What were our values while we were doing that? Because it went pretty well. Let’s write those down, and we’ll put them on a mug or a T-shirt,” and those are going to be … that’s how we’re going to talk about our culture. That’s not a criticism from my perspective. I just think that’s usually this thing where we look back to understand how we’re behaving. And because it’s worked.

And so, the way I think about transforming culture is about setting really clear expectations and communicating more than anything an approach and an expectation about how we’re going to behave with one another. What’s going to count as the way we do what we do? How do you know you work here? Because every great company feels that way. Like, “Hey, this is how we do.” When you unpack that idea, folks get clear on that, having a common view of it. But having that view be authentic to the actual experience is important. For example, on my first day there was a lot of pressure to do the kinds of things that you’d expect in a big company. “Hey, we should send out new values.”

I don’t do that. Because how do I know? I just got here, I haven’t met anybody, we haven’t done anything yet. How about let’s work together, let’s get to know each other. Let’s figure out how. And then we go back in an authentic way that’s actually connected, in a way that isn’t about mugs and T-shirts. Can we get to a short list of what we expect of one another? And that’s sort of how I think about culture.

It’s interesting you say that. It reminds me that when we started The Verge 13 years ago, we used to argue endlessly about what a Verge story would be. And now our audience knows, right? And that’s a backward-looking reflection of how we made some definitions and held onto them, communicated them. But you have also restructured the company. You have a new COO, you have a new CTO, you have a new CFO — that’s an entirely new C-suite. You’ve rethought how the company should work. What are the changes that you’ve made?

The first and most important thing for us was to reinvest in the relationship we had with customers and to be a better partner. We were at war with our customers, effectively, when I arrived. Folks were boycotting us. They were really unhappy with how we were charging them and how we spoke. And you can’t have a business where we’re your customers. That’s insane. The first thing was to take a breath and reorient ourselves around being good partners, around authentically trying to deliver, to listen to people, and authentically trying to deliver what they need as the primary touchstone of what drives our behavior. And I know that, in a way, that sounds obvious and like a cliche. But some organizations do that and some don’t. The process of coming to understand in an authentic, but also accurate, way what signals to listen to and respond to is more complicated than it seems. But the first piece was to be better partners.

The second piece, to your point, was I wanted us to lead differently. I wanted our organization to be different and to have different values. And in my experience, the fastest way to do that is to make changes at the top. Especially, by the way, if you think you might later have to change the rest of the organization, what is soul destroying for people is when somebody comes into a company that’s having a little bit of difficulty and they start restructuring the company and they leave the management layer intact and they think, “Well, hang on a second, who is responsible for all this stuff? How do we skip over the folks who were in charge?” And suddenly like, “I’m taking it in the neck. How did that happen?” If you want different outcomes, if you want different approaches, some of it’s changing culture, but a lot of it is changing people.

And especially if you want to do it quickly, you can’t be afraid, because you need alignment. You need instant and immediate alignment if you want really fast results. So yeah, we changed a lot of leadership. Then, I think about the other things we did… We did some important things in our business, which we can talk about if you want to, just in terms of we needed to make certain investments, especially in our advertising business and other things. But putting those aside for a second, the third big thing from a culture and a company perspective was that I think we went about trying to redefine how we do what we do. And honestly, that is my obsession. I am someone who believes that the what is a lot less important than the how.

If you have the how right, all the what’s will go better. And you have to ask yourself when you come into any organization, if that organization is not organically producing really high-quality answers all the time… What you should not do is run around trying to tell people what to do and ask them to do it differently. There’s no leverage in that. There’s no leadership in that. There’s no inspiration in that. What you need to do is ask yourself, “Why is that? How come this organism doesn’t produce good answers?” And start to get into the how and start to fix the how and then you start getting better answers. That’s the key to transformation.

Those ideas are in a little bit of tension. Let me just push you on that.

There’s the notion that you came into the company, you didn’t want to impose culture from the top down. You wanted to wait and see and diagnose the problem of how the company made choices. And then there’s a need to move fast in taking out all this old management, putting in new people, and you have to have a vision in order to go hire all those new people and change fast. Those ideas are in a little bit of opposition.

You’re right, you’re right. Totally fair. You have to listen. The truth is you have to do both. And you have to do both in a way that is authentic and effective. I am generally very, very clear and we as a team are very, very clear about what we want to do and what we think the right direction is. But here’s where they’re not in conflict: It’s the process through which you get clear about what you want to do, which is where all the action is. So, it’s not, “Hey, here I am, here’s what we got to do.” It’s much more about, “Hey, here I am, help explain to me how we got here.”

What do you do? What are the challenges? How do you guys see this? You fly all over the world, you talk to customers, you meet with everyone you can inside the company. You have hundreds of conversations, hundreds. Because in the beginning when you get here, you have no idea. If you’re being honest with yourself, you ought to have no idea what to do. Where you get that inspiration is through the coalescing of the hundreds and hundreds of conversations.

And at least what I experience is, I begin to understand where the areas of energy are and where the areas of opportunity are. And so then, when you show back up, and you stand up in front of people and you say, “Hey, here’s the direction I think we ought to go in.” It doesn’t feel like something you came up with by yourself in a room with a legal pad. It’s a reflection and an echo of what you heard from all the assembled people. It’s organic to what you’re doing, and it’s a byproduct of listening as opposed to stuffing it on top of people. And that’s why if you do it that way, it ought to be better.

You are the new CEO, but you also used to work for the old CEO: John Riccitiello was the CEO of Electronic Arts when you worked at Electronic Arts. Then he became the CEO of Unity, and you replaced him. What did you learn about what not to do or what to do from that experience working for the old boss?

John actually hired me in my first job in video gaming, back in — well, it wasn’t my first job in video gaming. He hired me at Electronic Arts in 2012.  I learned a lot of what I learned about this industry from him, and have an enormous amount of respect and admiration for him, so I have really no inclination to sort of pick apart what he did or didn’t do.

The world is a new place now than it was four or five years ago. Unity is a new company and we’re pursuing a different direction.

All right, it was worth a shot! So, you made some changes to the top. I think this is the big Decoder question: How is Unity structured now?

Here’s the thing, and maybe you and I just won’t agree about this. I’m much less interested in structure than I am in behavior. And I’ll tell you why. I’ve worked in big companies, a lot of them over the years. My experience is that what people want to do, especially if they think they want to improve things, is they immediately go to an org chart. They go and they start moving things around. The truth is, if people are not properly oriented, if they’re not properly situated, if they’re not spiritually in the right place to do what you need to do, I don’t care what the structure is, it’s not going to work. And there is no right answer to structure. People think, “Hey, if I could just get it in the right shape, it’s going to be okay.” The truth is that some structures accelerate some outcomes and create different tensions, and other structures accelerate other outcomes and create other tensions. I tend not to think first about structure. I try to think first about the landing approach and behavior.

Once we get oriented around how we’re going to do what we’re doing and people start behaving differently, different things happen. If somebody comes to me, for example, and says, “You know the reason we can’t do this, it’s because this thing is over in this division and this thing is over in that division and they can’t talk to each other; they should be in the same division.” I have no time for that conversation. What do you mean? Do you have a phone for that person? Do you guys sit down and have a conversation about that? If we’re properly motivated, what difference does it make where it sits? Look, of course, I’ve restructured things and I’ve changed things, so again, I don’t want to, but in terms of orientation, I want to land one before the other because one I think is more substantive and more likely to fundamentally change things than the other, if that makes sense?

It does. My joke on the show is that I always ask that question, because to your point, if you tell me your structure, I can abstractly tell you 80 percent of your problems. I can identify the trade-offs that come with various structures. And it seems like you are making different trade-offs. You changed the structure, which means there’s something you wanted to change in the trade-offs you were making.

Yeah, I’d say principally the change in the structure… The only meaningful changes we made were around reducing layers and making sure that everybody was communicating, and that we could quickly escalate any issues, make decisions, and move on. We always had the right folks in the room to operate in a unified way. And if ever there were issues, we insisted that they get escalated and resolved quickly. So if there was some organizational thing that made it impossible for us all to sit together, or if there were too many layers and I couldn’t see, we made some of those changes. And I do think those have been important, but again, not as important as the behavioral changes.

Obviously, part of restructuring is shrinking. You’ve had a few rounds of layoffs now; last year, in January, 25 percent of the company was let go. You just had another [round]. Is that all to get smaller to reduce those middle layers, or is there something else going on there?

I wasn’t at the company for that really, really big one, but I will offer my strategy and my approach to these things. And I can already anticipate you’re going to tell me there’s a tension in them.

Welcome to Decoder. The producers, by the way, joke that that is what the tagline is.

Yeah. And by the way, that’s life, right? Nothing is black and white where it’s a Zen thing. But I never… We as a management team do not think about trying to get financial leverage by doing layoffs. I hate that. I think it’s nonsensical. What you have to do is make decisions and try to be really disciplined about the very few things you think you can do really, really well, and be a little bit ruthless about stopping the other things. And so for me, those kinds of restructurings are about putting your money where your mouth is as it relates to investing in the things that are important and ceasing the other things. 

The thing I think that most leaders miss is, when they stand up and they talk about things that are important, what they don’t realize is that all they’re doing is adding another thing to people’s to-do lists. Because last week you said something else was important. The week before you said something else was important, and folks drown in those priorities. And nobody ever tells you what to stop doing, because that requires courage, and you probably wouldn’t be doing them if it wasn’t a pretty good idea at some point. 

There’s a reason why people do things, but if you want organizations to perform really well, you have to be really, really clear about what you’re not doing, and then you have to put your money where your mouth is with respect to that. And sometimes there’s real pain associated with that, and it’s the unfortunate part about it, but in my experience, you come out the other end in much better shape.

Here’s the other big Decoder question: How do you make decisions? What’s your framework?

I’m a big believer in full transparency, context, sharing with people, and having them share with me everything they understand about something. And then holding that up and making a call. And so, for me, it’s about engaging in a really deep conversation, being sure we’re talking about the same thing that we understand at the same level. I expose all the thinking I have on a topic, I’d like you to do the same thing. And we balance the pros and cons and then we make a call and we move on. For me, it’s much more important that we make a decision, than it is what that decision is. And it goes back to what we were talking about before. I believe in the how more than the what. If it turns out that decision was wrong, we’ll make a different decision. But the thing that kills companies, especially large companies, is a lack of courage and clarity around making decisions. So, my thing about decisions is, mostly make them.

That one comes up a lot. You are an interesting person to ask that question, because you came into the company and one decision that you had to make was obvious, I think, right? You said you came into a company that was at war with its customers, and that is because of the runtime fee that your predecessor had announced, which has shifted the pricing model of Unity to a per-install fee. Every time someone installed the game, the developer would have to pay, which would skyrocket costs.

I don’t know why that fee was instituted. It was obviously a problem. Here you are, you’re the new CEO. You knew you were going to walk that back, right? But you’re saying the how was really important. Walk me through how you made this decision, where the outcome seems so clear.

You’re absolutely right, and I appreciate you asking the question because I do think it’s a good example of what we’re talking about. I absolutely knew I was going to roll it back before I even took the job. The important part was not that I was going to roll it back, and not even so much how we were going to roll it back, but in the way in which we’re going to roll it back. For me, the most important thing, as I said, was, “Hey, I wanted to get back into a place where we were being partners with our customers.” And so, what I did was get on an airplane and started flying around, meeting with customers, and saying, “Hey, between you and me, we’re going to repeal this fee. I have some ideas about how, what do you think of them? Here’s what they are. If you want to vent about how upset you are about the old thing, that’s why I’m here too.” But I prefer to say, “If I did this, how would that strike you?”

And we started an open conversation. You have to select the folks you talk to because it is confidential. It’s important, you have to trust people. But we consulted and what we heard again and again was, “Hey, I accept that we’re not paying you enough. I actually think you’re delivering more value. I just hate the way you ask for the money. I don’t like the structure, I don’t like the tone. And by the way, if you’re going to raise prices, I need you to deliver better and execute better, because the two together strike me as aggravating.” And so, by the time… It took a little while, but it wasn’t a very long time. It was a few months and people were saying, “How come it took you so long?”

But it’s because I wanted to be sure we heard people and I wanted to be sure they understood that we were going to do things differently. We’re going to be disciplined, we’re going to communicate, we’re going to execute at a high level, and we’re going to listen. What came out of it was pretty sharp increases that, by and large, and listen, nobody likes to have prices raised… But we came out of it with a stronger bond with our customers, and the way we talked about it internally was equally important. Because to your point, we made this decision which was really ruinous for the company. You have to ask yourself, “How did that happen?” And so, how we make the new decision, how we have that conversation, who’s at the table? What kind of analysis do we do? What kind of feedback do we take? Getting into that and using that exercise to fix the organism and fix the process was equally important. It’s like, again, it was less about the thing and much more about using the thing to get healthy.

Can you just, in a sentence, diagnose how the decision to impose the runtime fee was made since you looked into it? And then how the decision to pull it back was made since you made that one differently?

I see Unity as a product company, and I’m a big believer that the way we create value is by delivering product value, and then folks will pay us for the value that we create. The thing about the runtime fee was that it was a business model. We were thinking about dynamics and how you can get people to do this versus that. If we raise one price here, then they’ll be forced to do that. That has nothing to do with creating value. That’s a trick. It’s a business model trick. It’s a hack. I don’t believe in that. I believe you create value by building better products. And so, the runtime fee decision was about, “Hey, we’ve got an engine business and we’ve got this ad business and we can’t exactly figure out how those two things would leverage one another and work together really well.” Okay, here’s a way. “I think we can structure a fee that brings them together and sort of encourages (which would be a kind word) folks to spend in one area if they spend in another.” I think customers found it more coercive than encouraging, but whatever. 

My point is that, no… The way those two things come together is that both the developer part of our business and the ad part of the business depend on a deep understanding of consumer behavior. And just as folks who are building games need to understand how consumers are behaving in those games, in order to effectively acquire users, they also need to understand how consumers behave and that we can create product linkages and value there. That’s how we’ll drive value, and that’s substantive. I think that was sort of the distinction between the two approaches if that makes sense.

When you rolled back the runtime fee, you increased prices on some of the tiers of Unity. It seems like those could go up independently if you think you’re delivering more value there. But they did go up pretty steeply as you mentioned. And then there’s a monetization side where you’re making investments in ad companies to help you deliver more monetization. There’s a tension there. You can see that increasing ad monetization helps you get more money if there are more installs, so you might want to lower the price on the engine to get more market share and compete there. But if you increase the price on the engine, you might just make more money upfront against this ad thing that you have to keep developing. How are you reconciling those things?

I agree that there is a tension there. 

We’re going to make a Decoder game in Unity called There’s a Tension There.

Any ad business at a high level thrives by opening the top of the funnel as broadly as possible. That’s what drives so that you’re left with something at the bottom of the funnel. That’s what ad-driven businesses do. And you’re right, there is some tension if you charge for the product, then presumably fewer people will use it, and then that’s not good. I think the important thing to understand about our structure is that Unity is free for everybody if you make less than a couple of 100,000 dollars in revenue. The top of the funnel is really big and broad. And there are millions of developers of Unity, and that’s why in the aggregate we have something like eight billion DAU (daily active users). It’s not deduped data. Cut it in half, say four billion people who are touching that Unity runtime in some way. So, it is hugely distributed. 

I accept that if it was completely free, maybe you’d have more people using it, but again, keep in mind that you’re only paying us once your game is really successful. And as a percentage of what you’re making it’s not… it ought not to negatively impact you.

That’s interesting because I’ve seen that dynamic play out on other platforms, particularly the app stores. Where you’re free to a point and then you tick over into the next tier of success and suddenly you’re paying a penalty — or what developers perceive as a penalty — for success, right? You’ve tipped over into the amount of revenue it would cause you to pay. And you actually have an incentive to stay just under a certain level of success in order to stop paying more money, in order to prevent paying more money. Do you see that dynamic in Unity?

Listen, I think that, in general, our customers and the community understand that in order to invest in the engine, it’s important for us to be able to charge folks. And engines are very complex beasts and there’s a lot of engineering there, and there’s a lot of investment we make every year. And in general, I think there’s an understanding about that.

The hard part of the ads business, particularly on mobile, is that you’re up against the platform vendors. Google is particularly interested in ads in various ways. Apple is not interested in ads in various ways. The mobile ad business, particularly in iOS, took a big hit from app tracking transparency.

Apple is interested in their own ad business. Just to be clear.

They’re definitely interested in their own ad business. They’re not interested in anyone else’s business. I think that’s a good call out. 

But they made the ad business on iOS much harder with app tracking transparency. I think I would argue that they made the correct trade-off. They made it harder to track people across apps and platforms, the web, and mobile. I don’t know that anyone perceives that that has gotten better. I still feel awfully tracked. But it had an impact on, for example, Meta’s business to the tune of billions of dollars. Is that a challenge that the ad industry has overcome? Is that a challenge you’re still working on?

I think it’s a challenge that largely the industry has overcome. It certainly had a meaningful impact. Different folks could be more or less cynical about what Apple’s motivations were or not. I don’t know. But I think now it’s largely in the rear-view mirror. And that in general, the systems have adapted.

You mentioned Apple is interested in its own ad business. The other thing they’re particularly interested in is in-app purchases. Their services revenue line is the one that’s growing as iPhone sales taper or even decline. I look at a show like Severance or Ted Lasso and I say, okay, that’s the shiny face on a huge revenue line that is mostly in-app purchases and games, right? Candy Crush whales are the growing business, and we just get Ted Lasso for free because we don’t want to talk about in-app purchases and games in that way.

I think 60 percent of the app economy is actually the game economy.

How much pressure do the platform vendors put on that? Because across the non-game iOS app ecosystem, the developers we talk to are basically just in full outrage mode, right? Apple calls them, they call it a shakedown. You can see various developers argue loudly for and against Apple. Sometimes the rules change. Sometimes all of Europe’s issues or regulations change the app store. The games industry has not been as loud, right? The game makers… In-app purchases make everybody money, and it seems to be left alone. Do you see that same conflict or that same angst there?

Well, I think Tim Sweeney at Epic has been exceptionally loud, in fairness.

That’s true.

But listen, I think that there’s a tension there, Nilay.

Are you all just following behind him? Are you just letting him take the heat?

Apple and Google are incredibly important partners to the industry. And they’re incredibly important partners of ours. They’re platform partners. They’re partners to our customers, and so there’s a tension. I think it’s probably a brilliant flash of the obvious to learn that folks would rather pay less. But I think what you’re seeing in the industry is tension between folks who are relatively small in the scheme of the world, have important partnerships, and probably don’t have a lot of interest or incentive to be at war with their biggest platform partners.

Right, that’s why I’m asking about Tim. I was going to come to that. Most game makers are not willing to be at war. He is, because he has Fortnite, and that’s just a weapon that he can wield or shield, I suppose, from attacks he might face. No one else has that as far as I can tell. Is the industry just waiting to see what happens with Epic, or is there anything else concerted going on? Because I would say the pressure from the platforms is coming in different ways, particularly as different countries around the world regulate their fees and how open they have to be. They’re going to seek that revenue and games are already the largest portion of the revenue.

I think your description of the landscape is accurate. There are clearly pressures, significant pressures, regulatory pressures in Europe, regulatory pressures in the U.S., lawsuits, alternative app stores, and new devices. These are all pressures. And I do think that that is a feature of the world we’re in right now, and I’m sure Apple and Google are feeling some pressure around those things.

At the top of the episode, you were saying you perceive that there’s a lot of growth yet to come in the industry. You mentioned the Switch 2. There’s a new console generation, there are new form factors, and potentially there’s new hardware that’s coming.

All of that is up against those pressures. That’s the escape valve. Apple can keep squeezing gaming on the iPhone for fees, but if you can escape and the business is actually on the Switch 2, that’ll relieve the pressure. If the business Ray-Ban AR games — let’s say that works — maybe relieves the pressure, you get some competitive relief there. What’s the number one relief valve that you see as being viable in the near term?

Well, one of the things that’s really interesting in the most recent version of Unity we shipped has support for web GPU. I think web gaming is a really fascinating, really fast-growing part of our world, and it’s maybe a little bit back to the future. But I think that’s a great example of what you’re talking about.

I see the web as an out for a lot of app store problems, and then there’s a lot of pressure on, are the browsers good enough? Are they being held back by the mobile vendors in particular? Do you see the browsers being held back? Is there something you want to do in the browser that you can’t do on a mobile phone?

I think that as it relates to tools and standards there’s progress we need to make. But I’m optimistic about that, and again, I see web gaming exploding. I think we’ve largely got what we need, and you’re going to see more of it.

Do you think that web gaming is on a phone, or on a desktop?

I think principally you’ll be on your phone.

So, you see people playing more games on mobile Safari.

Why not?

I think Apple at least thinks it’s “Why not?” That’s the problem.

They might, well… I just think that the consumer will ultimately push people and partners where they have to be pushed. I think that. And I don’t know exactly how that’ll play out, but I don’t see Apple and Google as preventing web gaming.

I’m curious because the other thing I see from our audience is, boy, people love building gaming PCs, and actually, that’s a relief valve in a way. People love the Steam Deck; that’s another relief that way. But it doesn’t seem like those are flickers of growth that would give you some relief from the pressure.

I am much more optimistic in a generalized sense about the industry than you are, so I wouldn’t say flickers and you’re making me feel sad about it. I don’t see it that way.

I think the Steam Deck is real, but it’s not —

It’s real. It sort of seems to be, it is what it is. It’s a sizable audience. I don’t know that it’s going to suddenly become more massive, a place of more massive consumer adoption. I think that we have… Gaming happens on your TV. It happens on your phone and happens on your PC. There’s overlap in some of those places. But we continue to be, I think, more of a lean-back society in the living room and a more on-the-go society than we are a sit-down-at-your-desktop society. I think you’ll see those two probably grow more quickly than the third.

I want to be very clear for you and the listener, The Verge as an organization is very high on the Steam Deck. I feel like if I didn’t assign other stories, we would only write about the Steam Deck. That might be the only thing that happens in our newsroom, but I also know who we are.

We’re usually supportive of PC gaming too. As a tool, what we mostly think about is how can we ensure that whatever the new platforms are, we’re supporting them on day one so that developers can push to those platforms. We don’t have favorites. We don’t choose. We want to be everywhere, and that’s mostly how we think about that.

The one thing that would change that is another idea that I think flashed really brightly, particularly during the pandemic, and then has maybe come back to Earth, which is the idea that we would stream the games. Unity would run on big data centers operated by Netflix or Microsoft or someone, and really what we’re doing is live-streaming video, which maybe requires a total re-architecture of the entire industry. Everyone loved this idea. This idea has not, I would say, panned out in any realistic way. Do you think it has a future?

Not in the near term.

Why is that?

I don’t see it. I think the hurdles were a combination of technical, really hard problems to solve there. And I think it was a consumer solution looking for a problem that people didn’t have. It was like, “Well, I don’t know. I’m playing this game on my PlayStation or my PC. Why do I need to use my phone as a controller or get some other controller, because you want to delivery the game? I mean, okay, I downloaded the game last month. What are we talking about here?” I just don’t think consumers cared.

I think there was some desire to play the games everywhere. I agree with you, it obviously didn’t happen. But the thing that really prevented it was that you couldn’t install a game streaming app on a mobile phone, which would have been, I think at least one thing to market as opposed to however the platforms wanted you to do it. Do you think if they opened that up it would become more viable?

I don’t think, I don’t. Again, I think that the input device is maybe more important that people understand in this context. I don’t want to play Call of Duty on my phone with an external controller. And I want to pick a mobile version of Call of Duty, which is a different experience. I think different platforms should be developed as unique experiences. So, this idea that it was going to be this one version, which, I get it at a high level, at a technical level, but I don’t think consumers wanted it.

I feel like I’m required by law to ask you about AI to wrap up this conversation. I have to say it’s been a relief to not be mostly talking about AI with a CEO. I appreciate that.

I would never force you to talk about AI if you don’t want to.

I appreciate it. If you listen to some other episodes of Decoder, you can see, I’m just fighting it for an hour.

The games industry, like almost every creative industry, is in turmoil because of AI, right? The creatives are upset. We see video game voice actors putting out Instagram Reels about having their voices cloned by AI for AI characters. You see entire games being made in AI. Actually, there’s a very funny video where someone claims a game is AI, but it’s actually just Elden Ring. That’s deeply amusing to me.

But you see this happening. I generated some simplistic game using AI. I vibe coded it here. I shipped it, we’re ready to go. And then you see the creatives in the industry really worried that their jobs will be taken over by AI. I know Unity is somewhat agnostic to this. I think that’s actually the phrase you used. “The Unity platform is agnostic to the derivation of the 3D asset,” but then there’s the actual making of the game. Then there’s who gets paid. There are layoffs across this industry for a variety of reasons and rethinking of cost structures. How do you see AI playing out?

Yeah, this is a big topic and we could probably spend longer than you’d want to talk about it. I would say this, first, you’re right. We are agnostic with respect to where and how 3D assets get created because Unity is an assembly point. And it’s always been a place where you ingest or build the assets you want to build. You assemble the tools you want to assemble, and then you build these deep systems around them, which become fully formed games. I actually don’t think that any of the layoffs we’re seeing in gaming are driven by AI. I don’t think that. I think what video game developers and publishers will do, as we talked about a little bit earlier, is if they gain efficiencies in the making of games, I think they’ll make more games. And that’s where they’ll use that extra wiggle.

What we sort of believe will happen at the end of, maybe not the end, later in this process, is that there’ll be human creators sitting in the middle of the process of creation of these games, using a platform to create them. And they will have vertically oriented agents that are helping them do so. An agent that specializes in physics, an agent that specializes in sound, an agent that specializes in I.O., whatever it is. And there’ll be multiple people in that loop, and they will have those tools at their disposal, and they will need to bring them together in an interactive experience. We would like to continue to be at the center of that, and we think there is a need for a platform to be there. I think that it’s sort of a slightly Grognardian, boring point to make, but it’s critical. The biggest problem with either the creation of assets or even the full imagining of games, this hallucination of full game experiences, is that there’s no connection to workflows.

Games are massively deep systems, successful games, that need to be built around some of these assets and even some experiences. I think AI will be a really important part of putting those things together, but the systems, workflows, connections, how you manage large teams of people doing those things, and how we output and distribute those things — those are likely to be a part of this process for a really long time.

We have talked about a lot of things that are pressure on the industry, and you’ve pointed out that you’re more optimistic. What are the things that are making you excited about the future of games right now?

There almost isn’t anything that’s not making me excited about the future of games. Again, I see the difficulty we’re experiencing as this moment of creative destruction. One where the industry is going to have to reimagine itself to some degree, and that reimagination is going to take the form of innovation and gameplay, innovation and distribution, innovation and business models. And just we’re going to rethink. Just to take one example, because I think we haven’t talked about it, think about how meaningful game-related and actual property has become in the broader entertainment industry. That, by the way, was never true, until very recently, right? There had been a million unsuccessful movies and things that related to games, very few good ones. Now, suddenly there’s this whole other life. We’ve talked about some of the new possible distribution platforms, the shifts in the business models, and platform partners. All these things are opportunities. And yeah, we as companies… You feel pain and then you respond to that pain, and I think the history of the industry is that we innovate through those things.

Do you agree with me that Uncharted would’ve been better as a long-running series instead of a movie?

I do.

This is my thesis. Adaptations got good when we started making TV shows out of them, not movies.

It’s easier. Yeah. I mean, even though I think The Witcher was good as a series, at least initially.

Spicy take.

I know. I know. I’m sorry. I enjoyed it.

Fallout.

Fallout was great. Listen, I think there are a lot of things that are better in that form than a full movie form.

All right, well, Matt, we’re going to have to have you back and only talk about video game adaptations for a full hour. That’s my next pitch to you. This was great. Thank you so much for coming on Decoder.

It was my pleasure. Thank you for having me.

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

On today’s episode, we’re talking about bird flu, but in a pretty Decoder way. Lauren Leffer, who recently wrote a piece for The Verge about bird flu and how it’s becoming a forever war, is joining me on the show, and we’re going to talk about the systems, structure, and culture that might control bird flu — and those that might make it worse. 

You’ve most likely heard about bird flu in the context of egg prices. Eggs are expensive and hard to find right now because the bird flu has kicked off an old fashioned supply and demand problem: the virus kills almost every single chicken it infects; fewer chickens means fewer eggs, and fewer eggs means higher prices. You’ll hear Lauren say it’s a little more complicated than that, but that’s the basic shape of it.

But the bird flu is actually quite a bit more serious than just the egg supply. The first confirmed human death linked to the virus happened in January. Now, we’re facing some uncomfortable questions with no easy answers. Like how rapidly the new strains of bird flu, known collectively as H5N1, might be spreading from birds to dairy cows and other animals, as well as just how deadly those new strains are among both birds and mammals, including human beings. 

If you’re like me, you’re starting to feel some uncomfortably familiar feelings, and you’re not wrong to feel them.  There are some real echoes of COVID-19 here — and the Trump administration, with Robert Kennedy Jr. as the head of the Department of Health and Human Services, is not at all equipped to handle a fast-moving pathogenic virus. Kennedy himself is an anti-vaccine activist who peddles junk science with abandon.

Just last week, Kennedy went on Fox News and suggested one way to deal with ongoing bird flu outbreaks was to simply let it spread and kill all the birds, which is a purely stupid idea for a variety of reasons you’ll hear Lauren discuss. Certainly not helping the matter is that on Thursday, just a few days after Lauren and I talked for this episode, Kennedy announced a major workforce reduction of HHS, with layoffs hitting 10,000 employees in a move Kennedy said would a “a painful period” for the department that would involve doing “more with less.”

On top of all that, the messy aftermath of the pandemic and the damage done to institutional trust across the board would make the job difficult for even a credible leader of our public health services.

But like I said, I wanted to talk about this in a Decoder framework, and there’s one big difference from COVID-19 that I wanted to push on here: industrial agriculture in the US is a big business, and the bird flu is a threat to that business. Millions of dead birds represent millions in lost revenue, and if the virus becomes more deadly to dairy cows, that’s even more money on the line. So if the government isn’t capable of mounting a response, is the industry capable of seeing and solving this problem before it becomes even worse?

The answer, as you might expect, is a resounding question mark, but as you’ll hear Lauren explain, different parts of the market are reacting differently. And right now, there still appears to be time: scientists are not yet raising the kinds of alarms that we started to see in the months leading up to March 2020, when covid changed all of our lives forever.

But one thing is very clear: the bird flu is not just going to go away, no matter how much Kennedy might wish that was how it worked. We need to learn how to live with it and manage it. And so long as we can get everyone on the same page, we might even be able to fight it.

If you’d like to read more on what we talked about in this episode, check out the links below:

• We’ve entered a forever war with bird flu | Verge
• Kennedy’s alarming prescription for bird flu on poultry farms | NYT
• First bird flu death in US reported in Louisiana | NYT
• Bird flu found in sheep in UK, a world first | NYT
• Shell shocked: how small eateries are dealing with record egg prices | NYT
• Animal Farm: eggflation’s monopoly problem | The Lever
• At the ‘Wall Street of Eggs,’ Demand Is Surging | WSJ
• How to protect your pets from bird flu | Popular Science
• What to know about the bird flu outbreak in wild birds | AP
• Bird flu continues to spread as Trump experts are MIA | Ars Technica

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

Today, I’m talking with Kakul Srivastava, CEO of music creation platform Splice. I don’t think I need to really introduce Splice, actually — I just need to play this clip:

Disco Guitar sample from Oliver: Power Tools Sample Pack III | Splice Sounds

If you exist on planet Earth, you know that as the guitar loop from Sabrina Carpenter’s “Espresso,” which is an inescapable pop music phenomenon. You can check out the sample in full right here in the “Espresso” chorus.

That loop is part of a sample pack on Splice — in fact, most of “Espresso” is part of a sample pack on Splice, which is one of the biggest marketplaces for loops and samples around. You can just sign up, pay the money, download the loops, and try to make pop hits all day long. This is a part of making music now, and it has been ever since Rihanna’s monster hit “Umbrella” was built around a GarageBand loop called “Vintage Funk 03” in 2007.

Now, if you’re a Decoder listener, you know that some of my favorite conversations are with people building technology products for creatives and that I am obsessed with how technology changes the music industry, because it feels like whatever happens to music happens to everything else five years later. So this one was really interesting because Splice is all wrapped in all that — and some of its new products, including AI tools, might change how music is made all over again.

Srivastava joined Splice as its CEO three years ago. Before that, she was at Adobe, so she has a lot of experience working at a company that makes tools for a creative user base that’s threatened by things like automation and AI. But if you’ve listened to any of our Adobe episodes, you know that the flip side of that is people actually using these tools at high rates, because they’re fun to play with and make some parts of the creative process easier.

So I really wanted to dig into that with Srivastava, not only to understand where Splice stands, but also to see how the broader music industry can try and make sense of this technology and what it could do to music. I also wanted to talk about how the company navigates the incredibly complex minefield of copyright law and attribution on the internet — something that’s only getting more complicated with AI and the increasing number of copyright lawsuits filed against big AI companies.

There’s a lot in this one — and Srivastava was willing to fall pretty deep down some of these rabbit holes with me. Let me know what you think.

This interview has been edited for length and clarity.

Kakul Srivastava, you’re the CEO of Splice. Welcome to Decoder.

I have wanted to have this conversation forever, so I’m glad I’m here.

Yeah, we ran into each other at the Code Conference last year and we just were off to the races talking about music and technology and AI, and I’m glad you’re finally here because so much has changed since then. But all of the issues are kind of still there and still working towards resolution.

Yeah, a lot has changed. A lot is going to keep changing, and you’re right, some of the core issues are still the core issues.

And along the way, at least one massive hit single has been made using loops from Splice, so there’s that.

Oh, come on. Not just one, not just one.

I think “Espresso.” There’s “Espresso” and there’s a lot of other ones.

I love “Espresso.” It’s awesome, but a very large proportion of top music everywhere uses Splice.

Let’s start with the very basics for people who maybe aren’t familiar with how music is made today or with Splice, what is Splice? What do you do for folks?

Splice is a music creation platform that is used by music creators and musicians. That’s our focus — who are our creators? And what we provide to them, we have this tagline “starts with sound,” so we do start with sound and we provide them with probably the world’s most diverse, most high-quality sonic palette. We send people all over the world. In fact, right before this, I was looking at a report from our team that just came back from Brazil, and we’re recording sounds.

We’re meeting artists on the ground, so we’re capturing the sounds of the world, and we make that available through our platform. We also provide AI-based creative tools that help you start with a sound, but make it your own. We have compositional AI. We just launched something brand new at SXSW called Splice Mic, which allows you to hum an idea, start with a musical idea right in your phone, and we’ll help you compose around that by putting the right samples next to it, to help you get you to your final track.

So there’s a lot here, which is “here are the foundational pieces of making a song,” right? “We’re going to do loops and samples, we’re going to have this library of audio,” and then there’s this turn, which I see a lot of companies that make creative software starting to make, which is, “We’re going to do it for you.”

You used to work at Adobe. Adobe I think is the paradigmatic example of this right now. You can just push generate to fill in Photoshop and it just does a bunch of stuff for you. You can prompt Photoshop now in various ways and it does stuff for you. Are you all the way there with Splice and where you’re going, where you can say, “Write me a country song,” and Splice will just do it for you?

Actually, we’re totally not. That’s totally not what we’re trying to do, and I’m so glad you asked this question because I really want to put this idea out there. Our creatives, our musicians, our artists, the people who we think about all day long, the last thing they want is someone to make the song for them. In fact, one of the things that we learned when we launched Create, out of the gate people were like, “Oh, this feels like cheating. This feels too easy. I need more controls.”

So right from day one, we’ve been adding more sophistication, more technology, more customization, more personalization for users. Because for our users, it’s really about the creative process, and how is that interesting? How are they able to get the tools to capture what’s happening inside and turn it into a song, turn it into a vibration, turn it into something that they can share with other people? So it is absolutely not push-button creation. That’s not fun.

This is a long argument in music. It goes back decades. It stretches to before AI hit the scene. You must at the company have some sense of how people perceive building music out of sample packs. And even before “Espresso,” like “Umbrella” by Rihanna was [made with] GarageBand, which I think is just a moment in music that should belong in the history books.

Totally.

How have you dealt with that? Okay, music is now just assembling a bunch of pre-made samples and that’s good or bad. People have a lot of feelings about that. Is that a framework you’re using as you enter the AI generation era?

So, I’m going to take exception with what you just said.

Sure.

I don’t think music making today is putting just a bunch of samples together. I think that using samples to create music is a really profound creative process. 

By the way, I will concede that that is a very reductive criticism, but it is a criticism.

Yeah, it’s a process that’s been developed over decades that it’s really powerful. So I think of samples as the building blocks for how modern music is made, and it used to be a hip-hop thing. It used to be very specific genres, and now it’s in every genre. One of our largest growing genres is country music, which I never thought would happen, but it is. You’re using samples to make country music. 

I think the artistry of using samples to make music is that you start with a sample, you start with the sound, but then you make it your own. One, how you assemble it, but how you change the sound, how you vary it, what you do inside the digital audio workstation, which is the primary creative canvas. I think that’s really important. Sorry, I got really heated about this. This is important to me. 

I asked that question somewhat to provoke that response, right? Because like I said, it is a reductive criticism, but it’s a criticism that has existed. And I guess I’m curious, you have that response to the criticism of sample usage. Is that informing how you’re thinking about the criticism of AI usage?

Fundamentally, it’s about, what is the creative process? And I personally spend a ton of time with creators, and what they are telling me over and over and over again is, “I want better tools.” And when I was at Adobe, this is also something that we heard from people. “I want better tools.” And so the work for us, the work for any company that’s wanting to really meet the needs of this growing and large market is, how do you build better tools in this era of AI? 

It’s not going to be, “Oh, let me type a bunch of prompts and I get a song out at the end,” but what happens next? How do I edit that? How do I change this particular part of the song and get it to sound a certain different way? How do I take this sample and make it into something else? How do I get my musical idea?

You saw this with the Splice Mic launch as well. A lot of it is: how do we get more of you into the music creation process as quickly as possible? So that’s fundamental, whether we’re talking about using a synthesizer to make music, or samples to make music, or AI to make music, how do you make sure the creative process is respected throughout those different transitions in music innovation?

That’s a lot of incoming about what your product should look like. You’re getting feedback from artists, from musicians, from other creators. There’s another side of the puzzle, particularly music, which is copyright holders and labels. Now there are these huge private equity companies that own huge catalogs that want to assert their rights in various ways. There’s the distributors themselves now, like Spotify and YouTube. Do they have a point of view that’s informing how you’re using AI, or how you’re thinking about sample licensing? Because that seems like the most complicated part of your business.

Yes and no. We are aligned across the industry — whether it’s with Universal Music or any of the other key, high-quality players in the industry — we are very aligned that the rights of the creator have to be respected. And again, our position is super simple. We’re going to focus on the creators and what creators want, and we’re going to try to meet their needs. So the rights of the creators have to be respected. 

On the Splice side of things, we take this pretty seriously and we take it seriously throughout the entirety of our process, starting with ingestion. How does a sample producer or sample pack creator come into the Splice platform? We have an entire organization that does the intake, does the quality control, and checks the provenance. If you’re telling us you’re the creator, do we know that you’re really the creator? So that process of ingesting is something we take seriously.

Is it tagged effectively? Is it tagged appropriately all the way through getting onto the platform? And all the quality stuff: Is the sound clear? Is the recording nice? All of that. On the other end of it, what is the experience of someone who is downloading a sample from Splice and able to use it? We want to make sure that every single download that you do on Splice lets you download the PDF that says, “You’ve got royalty free, you’ve got full rights to this material to use it for any kind of creation.” So that’s something that’s a basic part of what we do, and it’s been that way for a long time. 

On top of that is the AI story, and that’s the big story that everyone’s talking about, and I think there, it’s really simple as well. It should be, which is if you’re going to use content to train, you should train on content that you have rights to. It’s not okay to disrespect the rights of creators, and I think, again, most players in this space are pretty aligned on that.

It occurs to me just as you described that, that you are a creator platform for creators.

We are, yeah. Both sides of it.

There are people who sit around making sample packs and then they might make money uploading sample packs to Splice, and then on the other end, you’ve got artists who are downloading the sample packs, paying you money to go use some in other songs. That’s a unique situation.

It really is. Can I just add, go back to your sample thing?

Yeah.

Because you really poked me on that, and I wanted to come back to that for one second. That’s what’s really magical about using samples to make music. It’s not just a random sound that you got on Splice. There’s an artist at the other end of that. Just for the Sao Paulo team, we work with some of the people who are really at the forefront of funk, and what that means and what that sound is, and how it’s evolving. 

So when you’re using a sample pack from Splice, you’re collaborating with those people, and it’s a collaboration. It’s a storytelling between those two different artists coming together. I think that’s really fun. It’s a neat part of Splice.

Are there creators who make their entire living just making sample packs for you?

There are some, yeah.

Is that a viable approach to being a professional musician?

I think it is for some people. For some people, they make hundreds of thousands of dollars. For some people, they’re building their own musical career and this is part of what they’re doing, so we see a big range of people. I will say that the revenue that we’ve shared with the artists on our platform over time, it’s at an all-time high. So it’s nice to be able to feel good about that, too.

I’ve been spending a lot of time just thinking about the economics of creator platforms, whether they’re sustainable over time. On sort of the big consumer platforms, you see that creators have to augment their income. They all have to do brand deals, they all have to do brand expansions or sponsored content or whatever.

You can’t really do that on a Splice. Is there a ceiling to how successful you can be on a Splice? I’ll just use “Espresso” as an example. Sabrina Carpenter makes “Espresso.” I’m guessing the person who made that sample pack did not get paid more money because that song was a hit, just based on how your licenses work.

Yeah, that’s the pro and the con of being royalty-free. We are royalty-free in that, what that means for the creators is they don’t have to get stressed about it. You can use the sample. It’s clear. You don’t have to worry about clearing the rights. The downside is you don’t get to share in the sort of upside when something big like that happens. We’re really here to make sure that as many people can create as possible, and that’s part of how our model works.

How does your revenue work? Where does Splice take the money?

We are a subscription platform. So people buy a subscription to Splice, and that gives them access to this unlimited library of sounds, along with the creative tools that we’re investing in heavily for the future. You get a certain number of credits per month, and use those credits to download sounds that you can then use as you want.

And is growth just getting more and more artists to use Splice on both sides, as creators and as people who are subscribers?

So that is growth and both is, it’s been an interesting journey over the last three years while I’ve been here, but growth is really good.

We should talk about that. I think that brings us to the Decoder questions. You’re a newish CEO. You’re three years in, you were at Adobe before. I think you had two different stints at Adobe?

I did. Right out of business school and when it was a perpetual business and more recently, well on its subscription journey.

Adobe is the creative software company. They have a very, I would say, back-and-forth relationship with creatives. We had [Adobe CEO] Shantanu Narayen on the show. We got feedback on that episode of Decoder like nothing else we’ve ever experienced.

I’ve heard.

People have a lot of feelings about Adobe, what that software represents, what that subscription is worth, what AI means to Adobe as a company and its user base. You obviously have some of that experience. As you’ve come into Splice, how have you thought about applying those lessons to what is well on its way to being one of those companies for the musical community?

Adobe has been a really important part of my career journey. I learned a ton of great things at Adobe, both good and bad. I was also one of the early people at Flickr, the photo-sharing site, which I don’t know if you ever used, but a lot of people loved it.

There’s some Flickr users right now who are writing us emails. I’m just letting you know, they still love it. [Laughs]

I was also head of product and marketing at GitHub. So I’ve had a chance to see creator tools in multiple different places, and all of that has really informed what I’m bringing here to Splice. The journey for me has been a little bit around pattern recognition. One thing that I’ve seen at Flickr, at GitHub, some parts of Adobe, that I see here at Splice is that you have a business that’s centered around content, and you have a lot of rich metadata around that content, and you have lots and lots of impressions around that content so that users are giving you information about it. 

So at Splice, we have about a million songs that are samples that are sounds that are downloaded today. That’s a lot. We have 28 million stacks that have been created using our AI tools. So we have a lot of impressions of what sounds people are listening to, how they’re creating things together, and what sounds go well together. That’s been a really interesting thing. Once you have that data, once you have that metadata, you can use that to build rich experiences on top, which is what we’re doing now with the creative tools, the AI-based creative tools. That feels very familiar to bring to Splice, to bring to the music industry where I’ve seen it at GitHub, I’ve seen it at Flickr, I’ve seen it at these other places.

That turn to, “We are going to make the tools that actually help you create the music.” You can look at it in a slightly more abstract way, right? In an early version of Splice, you downloaded some sample packs, you would open Logic or Pro Tools, and you’re off to the races. Splice doesn’t see what you’re doing in those apps, but those are the dominant music creation apps. To this day, they’re the dominant music creation apps. 

You’re suggesting with something like Splice Mic or Splice Create, you’re going to take some of that creation. Particularly on a phone, I think there’s a lot of opportunity to reinvent how we make music. It’s still fairly cumbersome. Phone screens are small. The features you launched at SXSW are interesting, because they use AI to make that a little bit faster, more seamless, more sketchy. You can sketch an idea very quickly on a phone now.Is that the extension — “We’re going to take some of Pro Tools market share. We’re going to go take some of Logic’s market share”? 

So I think that word “take” suggests a zero-sum game. This is not a zero-sum game, right? It’s about expanding and exploring the creative process. Many of our users use Splice Mic, or use our mobile app as an adjunct part of their process that they will ultimately finish inside a digital audio workstation (DAW), and I love to see that. So one of our super, super top-end producers has worked with many of the big names that you would recognize. He’ll tell me, “I’ll get into my Uber, I’ll start playing the Splice app. I’ll generate a bunch of stacks so that by the time I get to the studio, I’ve got a bunch of ideas that I can show the artist right away to say, ‘Do you want to go this way or do you want to go that way?’” And that’s a really core part of his creative process.

I was just at my kid’s school where they have a digital music production class. And for them, listening to sounds on Splice is a really core part of learning, “What does this genre sound like? What does this genre sound like? What does it mean to create a Bollywood hit? What does it mean to create something that’s a K-pop sound?” And I think that’s a different way to use this experience. So for us, it’s not that we’re going to take [market share] away from this place or this place, but how do we expand how much we’re part of the creative journey in different ways?

But the idea that you’re going to start and finish a song in a Pro Tools, you’re not looking at that.

Do we think that we’re going to directly compete with Pro Tools? No, I don’t think so. I think Pro Tools has its place, just like Photoshop has its place. There are people who tell us every single day, “You will take Ableton out of my cold dead hands. It’s not going to happen.” And there are a lot of other parts of the creative process that are painful. 

So for example, when I sit down with one of our creators, inevitably there will be a situation where they’re like, “Oh, we need to find a certain kind of kick drum.” And they’ll find a folder and they’ll do a subfolder and they’ll do a sub-folder, and then they’ll finally find the sub-sub-sub folder that has 20 kick drum sounds that they have saved.

You just go through and you listen to these sounds. That is painful. That is a painful process, and it shouldn’t be that hard. And so, we’ve just done this new experience that we launched in October last year where we integrated with Studio One, which is one of the top DAWs, and there’s a Splice integrated search with sound experience. So we listen to what you’re creating inside Studio One, and we’ll suggest the samples that go with it right there integrated as part of your creative workflow. 

Do I think I’m going to replace Studio One? Absolutely not. Can I make the Studio One experience a lot better, because Splice is there and Splice is smart with AI? 100 percent, all day long.

How do those conversations work with all those digital audio workstation providers? They’re all very different. The companies that make them are all very different. Some of them are very quirky. Some of them are Apple, which…

All of the music tech industry is very quirky.

It’s all very quirky. There’s a lot of, I would say, eccentric Europeans floating around this industry in particular. It’s great. It’s one of my favorite parts of the tech industry to cover. And then you have a company like Apple, which is, they’re just going to do whatever it wants to do. That’s just how they work. Splice has to integrate with all of it. Some of them are expanding into your zone. They are adding sample packs, libraries, and subscription features. A lot of them are adding AI tools. How does that competition and cooperation work?

Generally speaking, it’s a very strong cooperation. I’ve actually been really impressed at how collaborative the industry really is. So the conversations with Studio One was very, very positive, and we’re working with other partners as well to bring that integration, and it’s been very, very positive. I think there’s generally a recognition that we’re good at what we do, the kind of work that we can do in terms of bringing these sample packs to the world, the global coverage, the high quality, or consistent high-quality, control process. It’s not something that they want to replicate. They want to make great experiences inside Ableton, the next feature, this is not what they want to do.

I think the AI stuff is new to a lot of people in the industry. I come from a core tech background. A lot of the team that I’ve brought into Splice over the last few years comes from a core tech background. So, we have a lot of expertise around that, which is unique in some ways for the music tech space. So I think there’s a lot of respect around that. I think there’s an attractiveness to a subscription business model that has been difficult for this industry to adopt. And so, I think there’s a lot of curiosity about that. Could we use a content business model to get more recurring revenue? But I think many people have found that it’s not as easy as it looks, and they’ve struggled with it.

One of the things you say about bringing people who have a core tech background is that helps you innovate in things like AI, I’m sure, right? Where you just need to be on the cutting edge of the technology. Tech and music in particular have always just crashed into each other. The thing I say on the show over and over again is if you pay attention to the music industry and what tech is doing to the music industry, you have the view into what tech will do to everything else five years out.

How are you thinking about that dynamic right now? Inside of Splice, you’ve said, “I need to hire more tech people.” Is it just for AI or is there something else you’re trying to accomplish with the addition of that talent?

Innovation is really important, and when I look at the music creation process, especially as an outsider, I feel like these music creators have been underserved with great innovative experiences, and I think it’s important to focus on the creative workflow and provide people better tools over time. When I think about the collision between tech and music, it’s weird because there’s actually more similarity than dissimilarity in Splice. Inside Splice, we have some really great software developers who love music, and are music creators in their own right. We have a whole bunch of musicians and artists who think in that same weird mathy way that great software developers think. So, there’s a lot of similarity. Surprisingly, there’s a lot of similarity.

I also think that there’s this mindset out there that musicians are scared of technology, scared of innovation. I actually think that musicians love hacking. They love trying new things. Again, there was all this threat around synthesizers and all of that stuff, and then Stevie Wonder took it to a totally different, magical new place. I think artists love innovation, and it allows them better tools to get to the other place. What they don’t love is push-button creation. I think if you stay away from that, if you stay close to the creative process, you will find the right ways to bring technology innovation here. I think there’s something else that you’re pushing on here that I think is important, and maybe it’s one of your Decoder questions, around how do you bring the cultural mindset from the tech industry and meld it with the music industry? And is that a difference? Is that a challenge?

Yes, that’s definitely where I’m going. I might as well ask you the Decoder questions. You’ve been the CEO for three years. How is Splice structured today? How have you changed it?

This is the tech part. We are fundamentally a product company first. So my largest organization at Splice is the product development organization, and that’s product managers, engineers, designers, and [customer support]. And what’s neat about that is, I do keep CX very close to product. Because I think that tight loop is super important. 

Wait, I just want to make sure. CX is customer experience? 

Customer support.

Customer support, okay.

So support, design, engineers, [product managers], they’re all in one org and that’s product development, and it’s our largest org. Our second largest org is our content team. And those are the people that they’re going to Brazil, they’re going to South Korea, they’re going to India, they’re recording these sounds. It’s our quality control department. It’s our data and ingesting, and metadata tagging groups. So that’s the content org. 

Maybe the third thing that I’ll point out that’s really important to me and how I structure the org. Is we have a very strong central data organization that reports directly to me. So a lot of people put that inside product dev, but for me, data’s important for content, data’s important for marketing. Data’s obviously important for finance and how we run the business. It’s really important for product. So I have that as a central organization, and again, it reports directly to me.

How big is Splice now? How many people is it?

We’re about 200 people, a little bit less than 200 people.

And how is it split between those three groups?

So product dev is our largest org, maybe 80, 60, somewhere between there. And then content is the next biggest, and it’s somewhere between 40 and 60.

One of the really interesting things here, again, it’s a creator platform for creators, which is just an interesting dynamic. Other creator platforms at scale, they say they have investments in content teams, but they really just hope the scale carries them forward, right? Instagram does not have some huge content team that is traveling the world to get content. They just wait for people to come to them.

It’s the same with YouTube or TikTok or whoever else. They might manage some of their top influencers, but really the volume of content comes to them. Is that a tipping point that you think Splice can reach, or do you want to maintain control over the library?

It’s really important for us to make sure our library is the highest quality that it can be. So, it’s not going to be a free-for-all where anyone is uploading anything they want, because we need to maintain that high quality, especially in the age of AI, right? There’s all kinds of stuff that’s being uploaded to all of these big platforms, so it’s never going to be that way for us.

And so, that’s just a core investment, right? It’s a core piece of, I think, your cost model. How was Splice organized before? Again, you’re three years into it, how have you changed that structure? Is it still largely the same, or have you reoriented the company?

So I think the biggest change has been around, I would say, three big ideas Nilay, which are core to how I run a business. The first is data. I’ve brought in a lot more data people. It’s very, very critical. The reason that’s important for me is because I need to understand what our customers actually care about. So, how are they voting with their clicks as opposed to whatever opinions everybody else has. So, that’s a big investment.

The second is design, and that is really where data and the math and the science turns into something else, which is a real experience that people can feel. It’s where the art becomes magic. The reason that’s important is because we’re serving creative people, and that’s what creative people do as well — they take all of these inputs and they turn it into something new. 

So building a strong design team that is either made up of music creators themselves or people who spend a lot of time with music creators is really important. And the third thing that I really brought in that’s important is that we build our products with the customers. So everything that we launch, there are tools that we built in to allow people to give us feedback. In fact, when we launched Create, the biggest button in the Create experience was the feedback button. It was weird, but it was important for us. 

Every single time someone typed in something to give us feedback, it comes into a Slack channel that’s with all the designers and the engineers and the product managers. So we’re actively talking about the feedback from the customers as it’s coming in, and responding to it for the next version. I love that. I absolutely love that we build product that way. I think everyone should build product that way.

One of the reasons I always ask about structure on the show is that it’s a proxy for culture. You kind of get what you get. You make some big choices about how things are organized, and that leads to a culture. You’re in an interesting spot because you took over for co-founders. One co-founder left, he was a CTO in 2019. The other co-founder, Steve Martocci, he’s the executive chairman now, but he’s off doing another startup. How have you thought about changing the culture, inheriting the culture, and the balance between the two?

The reason I love your question around structure is because I do see that it’s a proxy for values, and that’s why I answered it the way I did around data, design, and building with customers. Those are fundamental values that I want to bring and inculcate into the company. There’s something else that we also did that was around building culture. I spent a lot of time listening to the team, trying to learn what made this culture unique, and then I reflected back to the organization, “Hey, these are the values that I’m hearing from you all. Do you think this captures it?”

And we came up with something that we call our DISCO values: direct, inclusive, Spliced together, creator-centric, and optimistic. And even though these are new values that we came up with after I joined, they have felt so authentic to the culture that we have that’s existed for a long time, but it’s given voice to it. So DISCO is something we talk about a lot. Every single new employee that comes on talks about which DISCO value they resonate with most. We use it in performance reviews. We do use it for shout-outs. It’s a core part of who we are.

The second Decoder question, which is also in many ways a proxy for culture and values, is about decisions. How do you make decisions? What’s your framework?

This is something that I’m working on. I’ve always been a very math and science kind of person. I’ve always been someone who’s very analytical. I use a lot of data. I have a framework for decision-making. I study all the different tools for decision-making, but as the decision sets that come to me become more complex, and as we operate in an increasingly more complex world, fires, politics, etc., I have found myself relying more and more on intuition, and I think balancing those two.

So, I would say that my decision-making process is, I will drown myself in data. I will really get deep. People know in my team that I spend a lot of time on our dashboards. I will spend a lot of time watching research videos and understanding how people are using our tools. I will spend a lot of time personally talking to different customers. I’m talking to customers all the time, and once I’ve kind of drowned myself in all this information, I’ll just try to listen deeply, and usually the answer is very clear.

All right. We’re going to put this into practice because the “making creative software for creative people in the age of AI” is about as tense as it gets in the balance between what the numbers are telling us and how the people feel. And what I mean specifically is the numbers are telling everyone that people are using the AI tools. Just down the line, every software maker I’ve talked to has introduced AI tools with any meaningful value, says the users are using them, they’re clicking the buttons, they’re doing generative fill all day long. I’m sure you see that in your numbers, too.

Then what you hear from the creatives on social media or online, or in letters to Congress is, “Get this out of my face. They stole everything from me.” And that is about as big of a divide in tech, in culture, in creativity as I have ever experienced. I think that is challenging a lot of how everyone is going to make decisions. So I’m going to read you a quote from one of your ostensible competitors, and it tracks with everything you’re saying, but I suspect you are going to disagree with this quote, and I just want to sit with that for a minute.

Yeah, let’s do it.

So you have said, “Right, creators just want to create, they want all this stuff to get out of their way.” So here’s the CEO of Suno, Mikey Shulman. Suno is just “push a button, it makes you a song,” right? You say country song, it just spits out a country song at you. And here’s what he recently said: “It takes a lot of time, a lot of practice. You have to get really good at an instrument or really good at a piece of production software. I think the majority of people don’t enjoy the majority of time they spend making music. It is not really enjoyable to make music now.”

Now I’ve made a lot of music. I have no idea what Mikey Shulman is talking about. I think it’s quite fun to make music, but that does track with what you’re saying, that you just want to get the software out of the way. You want to get the creators creating. But he spun the knob all the way to “just prompt me for a song.” And a lot of people reacted to this quote very strongly.

Yeah, I heard a lot of feedback. 

How do you sit in the middle of that to say, “There’s a line and I’m going to enforce the line, and we’re not just going to prompt it all the way to a song?”Also, do you think he’s right? Do you think people don’t enjoy making music?

Here’s what I have learned by serving creative people for most of my career: the creative process is essential for people who create. It’s not essential for everybody. For the people who create, it is a sacred experience. It is a core part of who they are. They can’t not do it. And there is a struggle, but the struggle is to authentically translate what is inside you into something else. And sometimes, your tools will help you —will enable you to do that — and other times your tools will get in the way. Understanding the distinction between those two is the whole ball game, but it’s really about allowing the struggle to come to life. 

Giving birth to something new is hard, but it’s profoundly important, and to dismiss it by this push-button set of tools, it’s insulting, it’s dismissive, it’s reductive. And, I think the creative process and creative people deserve better. They deserve better technology that enables them, as opposed to reducing this profound activity to a button.

So this is where I think the line is inherently qualitative, right? “Well, here’s what we’re going to do and here’s what we’re not going to do.” And the tension of, “It’s not really enjoyable to make music now,” you can describe that as using the software sucks, or I just want to have an idea and hear it as fast as I can. And then you can describe it the way you’re saying, which is there’s some parts of the struggle that are the creative process, which make the art compelling. 

It’s profoundly important.

If the data tells you that people really want to just click the button and make the music, are your values strong enough to not send you all the way down the road?

I think it depends on which people you’re listening to. We are really clear about the people that we’re listening to. We are listening to creative people who love the process of music creation, that it’s essential for them. Yes, are there challenges? But the challenge is the creative process, right? That is the challenge. And for those people, the signals are really clear. They do not want push button creativity. In fact, like I was sharing before, when we gave them Create for the first time, they’re like, “This is too simple. I don’t want this. I want something that gives me more creative freedom, more creative control.” And so for us, the signal, the people we’re listening to are super clear and the signals they’re giving us, there’s no confusion in what they want. They want more creative process. They want more creative control.

The other side of this marketplace is consumers. We see consumers and fans all the time now react very strongly to AI generated imagery. In particular, you make a movie poster and it’s got a bunch of AI in it. The fans are going to–

It’s not just Photoshop.

That movie poster is coming down. It’s maybe different in AI. It’s not in your face. You can’t see that the characters in the movie poster have 12 fingers and their hair bleeds into the skyscraper behind them. It’s not as obvious, but it’s there. Do you perceive that kind of consumer or fan backlash to AI in music the same way that we see it in visual art?

We haven’t seen it yet. Here’s what I have seen. I have seen really clear signals from our customers that they are not really interested in computer-generated samples, and it’s clearly not our strategy. We are in fact investing in more human-created samples, human-curated samples. This is why we’re sending people out to the sort of subgenre locations, talking to authentic artists, getting their voices. It’s really important for our strategy to continue to do that, because people want to connect with the stories of the real artists on the other side of the sample. So that’s really, really important, and really clear for us.

I think what an end user who’s listening to Sabrina Carpenter today and will listen to somebody else’s music tomorrow, what they can hear is going to be interesting, is going to evolve over time. I love that Kendrick Lamar won the Pulitzer Prize for music, and the people who won the Pulitzer Prize for music, when that award was first introduced, it was a completely different sound. So what art is and what is acceptable changes over time. I would expect it to continue to change over time, so I don’t want to reflect on that. I know that artists will use different tools, and they will use AI-based tools. Absolutely.

At least on the visual imagery side, there’s a lot of endless, sort of futile, discussion about watermarks and encryption, and letting people know when images were edited by AI or created by AI. I would not say that’s come to anything, and I would say there are some deep and meaningful challenges with even making that technology work consistently. There’s not anything quite like that on the music side.

No, there isn’t.

Do you think there should be?

I think it’s going to be really hard to disambiguate around sound, and around images, around video. You’ve had some really great conversations about this topic on your podcast. I’ve listened to them. I think it’s a really important debate and discussion to have. There is going to be a bunch of bad AI-generated content out there. It’s already happening. It’s going to happen with music. I think that the toothpaste is out of the tube. 

I think as an industry, we have to do the right thing around respecting the rights of creators, and doing the right thing with respect to training data, respecting credits. This is work that has to happen. I don’t think it’s solved yet. Maybe some of these cases that are open will help us get to the right answer, but I don’t think it’s going to come out of watermarking.

You talked about the flood of AI content that’s coming. We can all see it. The big consumer platforms are embracing it. I think to some extent, Mark Zuckerberg would love it if all the content on Facebook was AI and he was paying zero out to creators. I think to some extent, YouTube is really leaning into the idea that you should interact with your favorite creators through AI avatars, and that they should make even more videos or AI should help them make even more videos to increase the volume of content that appears. 

That’s all very complicated. I don’t know exactly how it’s going to play out, but I understand the incentives for those platforms to make those choices to say, “Actually, what we want always is more content because that will create more attention and we can serve more ads, and we’re in this finite zero-sum intention game.” You’re not in that game specifically. You don’t have those incentives, and you do allow artists to make music with AI using your tools. Do you allow AI-generated samples to enter your library?

We do not. We do not.

Draw that distinction. Why is it okay to make music with AI but not to have it in the sample library?

I think it’s what users are coming to Splice for today. They are coming to find those authentic sounds made by humans. That’s not to say that people aren’t using AI to master sounds or things like that. You’re using AI to master your audio and video probably here. I think those are tools, and that’s fine as long as there’s an authentic artist’s artistic vision and voice behind it. So, that’s super important for us to continue to be focused there. With respect to these social platforms that you’re talking about, I think that’s a really important insight. And inasmuch as these social platforms are important for our creators as a way to share their output, to share their musical idea, they’re really important for us.

But these social platforms have grown because they allow people to have emotional connection with each other. “I’m really angry about this particular issue,” or “I’m reaching out for support for these fires in LA,” or these connections that we make, and finding support around this very specific cancer that I have that I can’t find other people to connect with online. If we erode that, if we erode those actual emotional connections between people in order to save a buck in paying out creators, I think the value of these platforms will diminish over time, and maybe that’s okay. Maybe we shouldn’t spend so much time on TikTok. Maybe we should spend more time creating music on our own.

So I think these are really interesting evolutions that are going to happen in the industry. And as a mom, I care a lot about where some of this stuff goes. For Splice and as the CEO of Splice, my focus is going to stay the same, which is I’m focused on creators, I’m focused on what they need, and so many things our users create just to hang out on their desktop because it was just for the joy of creating. And some of it goes on and becomes a Billboard top 100 hit. Great. I’m happy that that happens, but I’m just as happy that someone is spending time creating, and it just hangs out on their desktop.

Let me ask that again in just a different frame. I want to push on it. It feels important to me.

Yeah.

We’ve talked a lot about active creation and what the tools are for and the fact that your customers, artists, do not want ready-made, push-button songs. They want controls. They want to add something to what the computer-generated product is giving them, right? They want to add something to the AI, they want to add something to the samples, and that process of addition creates additional value. Some very important songs have been made that way using Splice and other tools. 

But you’re saying that is not a good enough argument to get AI-generated audio into the sample library, and I’m just wondering why the difference, because you could make the same argument. I use Splice to generate some samples. I tweak them, I filter them. I made a bunch of different things. If it’s good enough for me to send to a major label and play on the radio, shouldn’t it be good enough to get into the Splice sample library?

So I think that the distinction in my mind, and I think for many of our creators is that, is it AI-generated, or was AI used as a tool to bring a human creator’s idea to life? Do people use technology to create the samples that end up on Splice? Absolutely. People are using Pro Tools. People are using synthesizers. People are using lots of tools and technology, and like I said, some of those tools might be AI-based, like mastering tools or mixing tools, things like that. 

That is really different from, “I’ve created an algorithm to pump out a whole bunch of samples that are computer-generated for the mass market.” Those are not going to end up on Splice, I will guarantee, but if there’s an authentic user, the Stevie Wonder of the AI age, who is creating art that they care deeply about, and they’re using AI tools as part of that process, absolutely. That distinction is very important.

I agree it’s important. I just don’t know how to write it down in a way that can be consistently enforced across all the geographies that you’re operating in, with all of your teams going out in the world, or in a way that’s understandable to artists who might want to be part of Splice. Is there a definition you have of where the line is? For how much AI is too much?

For me, again, I will take it down to something very simple. There’s a human being who we have a relationship with on both sides of our platform, and so on the side of the platform where we are working with a musician, an artist, an instrumentalist who wants to provide a sample to Splice. We actually have a relationship with them, and we talk to them about what they’re trying to do, what the idea behind their label is, what is their artistic vision, and we will work with them. What is your tool set? How are you doing it? How many sample packs do we need every quarter? All of those kinds of things. And some of those people are, there’s a Japanese potter who is making handmade percussion instruments that he then records, that end up on Splice. That’s a really cool part of the process. 

And then, we’ve got crazy kids making all kinds of super electronic, super grungy, super sharp technical sounds, and they’ve got a different tool set that they’re using as part of their process. We’re not going to tell them, “Oh, you can’t use this tool because it’s AI-generated or not,” but do you have that authentic vision for what you’re creating? And it’s not that difficult to tell the difference between a person who is creating that way, and a person who is like, “I typed in a bunch of prompts and I got a whole plethora of computer-generated sounds.”

The other extremely challenging piece of the puzzle with AI-generated content is when you veer into impersonation. We’ve seen this in the hip-hop industry a lot recently. We’ve seen it with OpenAI and Scarlett Johansson’s voice. There’s a lawsuit. The voice got pulled. Who knows how that’s going to play out? We see there’s the Elvis Act, in Tennessee where impersonation is illegal, and I don’t think there’s a great answer for whether Elvis impersonators themselves are now illegal. Are you playing in that space where you’re letting people use artist voices or sound-alikes?

We’re not. I think there are lots of people who are playing in that space or interested in that space. We are focused on creative people, and creative people are actually really clear with us. They are coming to Splice because they want to find their authentic sound, and so we work really hard at the very other end of that, which is how do we allow our users to authentically find their own vibe?

Voices is one thing, right? They’re pretty recognizable. The fake Drake song set the industry ablaze. It was just very obviously a fake Drake song, or Drake’s voice. There’s not a great legal system for saying, “That’s Drake’s voice. You can’t use it.” We’ll get there. It seems like we’re on our way to understanding how to get there. 

Then there’s kind of the existing mess of music copyright. We talk about the “Blurred Lines” case on this show a lot. I think more than any other podcast we’ve talked about “Blurred Lines,” a song which came and went and whose moment is over, but it continues to come up on Decoder maybe once a month, right? That lawsuit is “you guys stole a vibe from Marvin Gaye, not notes, not chords, not anything direct,” but the jury was like, “These vibes are too close. Robin Thicke and Pharrell have to pay the money.”

That’s something you could very easily see a user of Splice wandering into, right? “We’re going to prompt for a beat. We’re going to do a stack. We’re going to layer some samples and we’re going to get to a vibe that’s too close to another artist.” Is that something you worry about? Is that something you try to protect users from? It feels like in the age of AI, it’s ever more of a danger.

It is, and it’s also been a core part of how music evolves over time. There’s this whole conversation around reheated nachos and what that means, and I think artists and musicians build upon each other’s work. They’re influenced by each other, and this conversation’s been around since the beginning of sampling, which is “what am I referring to when I use this sample, and what’s the story that I’m trying to tell?”

You could argue that it’s derivative, or you could argue that it’s an homage, or you could argue that it’s building on a shared piece of work that’s a community piece of work that continues to evolve over time. I think that that’s what makes art and music in particular super fascinating. I love that you guys have this whole debate around that particular song. I think it’s fascinating. I think it’s going to continue to grow, and what’s right and wrong should be defined by the artists.

But the idea that you would accidentally boost too much of an existing song by using an AI tool, which is trained on bits and pieces of existing songs. That’s a new danger, right? I mean, the cycle you’re talking about with music, I agree has existed since music. We’re all building on one another. We’re all lifting bits and pieces. Great artists steal. Everybody kind of understands it, and along the way, there has been a lot of litigation. That’s the other part of the cycle.

The push and pull is people being very unhappy about the money, and now we’re at a place where it’s easier than ever to be derivative, and the money is absolutely not clear — that artists are very upset about their work being trained on, maybe not in your tools, but certainly in other tools. The labels are suing Suno and Udio, its competitor, for training on their data. Do you see that resolving? Because it seems like the problem is going to get worse faster than the legal system will even comprehend the technology.

Most of these problems get worse before the legal system catches up. I mean, we know this. We know this for privacy, we know this for many areas. Technology outpaces how quickly legislative action catches up. I think in the music industry, we’re doing a lot of work to try to create standards within, so we’re a part of a coalition of, again, great companies in the music space that are saying, “We’ve got to support ethical AI. We have to support the rights of creators. We have to make sure our training data is clean.” 

So, I think there are a lot of companies that are trying to do the right thing. Is there one standard that has won out amongst all the others? No, but I know that a lot of people are working really hard on this problem, and we are too. We care deeply about the rights of creators, so that’s going to stay really important for us.

How do you feel about the labels suing Suno and Udio? Is that something that’s a warning sign for you? Is that something you support? Do you think that that is going to get resolved?

I think what the labels are trying to do is support the rights of the creators, and we are a creator-centric company, so we absolutely support the rights of the creators. Do I take one side or the other? No. Ultimately, it’s always going to be about the creators first, and I know my customers deeply care about the fact that they have rights to the content they create using Splice. 

That’s why we allow people to download the rights PDF. It matters to people, even if they’re not putting their song up on Spotify or trying to make a billion dollars from it, they want to know that they have the ability to do that. So that’s what governs our decisions around clean training data, ethical AI, etc.

If I wanted to sign up for a Splice account, download a bunch of tracks, and then train my own AI on them, is that allowed in your license?

It’s not allowed.

And you spell that out. You say you can’t train AI on these tracks?

Yeah.

I bring this up and you probably don’t know, I’m going to pre-apologize to you for this question because I know you haven’t seen this document, but just go with it. Basically what I’m saying, I’m asking a question just today, I’m sure you haven’t seen it, but Google filed a letter with the government basically saying, “Look, you need to make an exception of fair use to allow us to train on everything.” OpenAI filed a similar letter in the past few days.

There’s a big push from the AI companies to say, “Look, we just need this stuff. Give it to us. We don’t want to pay for this. It has to be fair use. This is going to slow us down too much.” At the same time, you’re saying, “Here in our license we’re saying you can’t do that.” Do you think that can get resolved? That seems like a big problem where if you steal enough of it, you get to write a letter to the government saying, “Write us an exception.” And if you steal a little of it, you might end up in court. And I don’t know how to resolve that.

Yeah, I don’t either. It’s such an important issue. And the scale of the Internet, the scale of content on the internet is so vast that — What is fair use? What is not fair use? What is public consumption? What is public record? What is public ownership? We are in uncharted territory, and we’re going to be watching it just like you are.

How would you write a fairer system if you were clean sheeting this? How would you write a fairer system that makes creators feel valued, gets them paid, and still allows people to build these AI systems that a lot of people are getting some value out of?

I would love to say that I’m the expert who could write something like that. I have a much more straightforward problem to look after, which is, how do I help creative people be creative and get the ideas from their hearts and minds out there? Yeah, I’m going to leave that problem to people way smarter than me, who are legal minds who are working really hard on this.

Well, if I get anyone on the show who has an answer, I’ll let you know.

You’re a lawyer, right?

I just talk for a living. I haven’t done anything useful in a long time. Kakul, you’ve given us so much time. What’s next for Splice? What should people be looking out for?

So, what’s next for Splice is that we’re going to keep going deeper into the creative process. I’ve been really public about this with my blog posts and all of that. Users keep telling us, “I love Splice. I want it deeper in my creative process.” So whether it’s these partnerships that we’re doing with DAWs, thinking through how we build more creative flexibility for users on our own platform, whether it’s with Create or Splice Mic, there’s a lot for us still to do, and we’ll keep going down that path.

All right. We’ll have to have you back soon as some of these issues play out. Thank you so much for coming by.

I would love to. I had such an enjoyable conversation. Thank you so much, Nilay.

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

Today we’re talking about the protests against Tesla, which have been branded on social media under the hashtag Tesla Takedown. The protests are, of course, a reaction to Elon Musk, who has managed to install himself as basically a not-so-shadow president who is tearing the federal government apart, leaving confusion and destruction in his wake. 

A lot of people are deeply unhappy with this state of affairs. And because Musk is not an elected official but is, on some level, a car salesman, many of those frustrated and angry people are taking aim at the most visible and accessible symbol of his power and wealth: Tesla itself.

There are protests at Tesla showrooms and charging stations every weekend now, and they’re only getting bigger — and attracting police presence and threats of prosecution from the government. President Donald Trump and Attorney General Pam Bondi have said that any violence aimed at Tesla vehicles or property would be prosecuted as domestic terrorism. This is very Trump, in that there is not actually a specific federal domestic terrorism law, but you know… Trump.

In any case, these protests seem like they’re working; Tesla’s stock price has been sinking, new car registrations are down, and merely owning one has become uncomfortably political for a lot of people. We’re also seeing the hype around the company fading in ways that are frankly surprising — popular YouTubers are pointing out that vehicles like the Cybertruck simply aren’t very good and that Tesla’s missed ship dates and big claims about the capabilities of its self-driving tech don’t really hold up. And if the hype fades enough, it’s possible to break Tesla itself — it’s sky-high stock price is entirely based on hype, after all.

To help pull this all apart, I asked Ed Niedermeyer to join me on the show. Niedermeyer published a book about Tesla in 2019 called Ludicrous: The Unvarnished Story of Tesla Motors, and he hosts a podcast called Autonocast, about autonomous vehicle technology. His sizable following on Bluesky has also made him one of the leading organizers of the Tesla protests in Portland, Oregon, where he lives. So I wanted to bring him on the show to talk me through what he sees with both the big picture of Tesla as a company and on the ground with the ongoing protest movement.

There’s a lot happening and the story is changing fast, even since Niedermeyer and I talked at the beginning of the week. But that bigger picture — the disconnect between what he calls the “myth” of Tesla and the reality, and how that reality is now manifesting into a global protest movement — is shaping up to be the defining story of Tesla’s next chapter.

If you’d like to read more about what we talked about in this episode, check out the links below:

• Is Tesla cooked? | The Verge
• Elon Musk has become too toxic for YouTube | New York Magazine
• ‘Tesla Takedown’ wants to hit Elon Musk where it hurts | The Verge
• The Tesla protests are getting bigger — and rowdier | The Verge
• ‘Tesla Takedown’ protesters planning ‘biggest day of action’ | The Verge
• Tesla registrations — and public opinion — are in a free fall | The Verge
• Multiple Teslas set on fire in Las Vegas and Kansas City | The Verge
• Mark Rober’s Tesla video was more than a little weird | The Verge
• Tesla sales fell year-over-year for the first time | The Verge
• The Cybertruck isn’t all it’s cracked up to be | The Verge
• Tesla autopilot, FSD linked to hundreds of crashes, dozens of deaths | The Verge
• Tesla crash victims’ families worried about Musk influence on investigations | The Verge

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

Today’s episode of Decoder is a little different, and I think it’s one of the more illuminating conversations I’ve had in a while. I’m talking to Evan Smith, the cofounder and CEO of Altana, a company that makes software to track and manage the global supply chain. 

Smith started Altana in 2019 because he predicted that the first wave of globalized manufacturing and free trade would come to an end, and companies would need powerful tools to adapt their supply chains as borders, tariffs, and tensions got more complicated. Here in 2025, that looks like it was a pretty good bet, even if the way it’s playing out is a little more stressful and chaotic than anyone really wants. 

The easiest way to think about Altana’s product is that it’s a map of the global supply chain — you’ll hear us call it that several times in this conversation. What that means is that Smith has a front-row seat to how things like Trump’s tariffs and isolationist trade policy are playing out in real time. So while we talked about Altana as a company a little bit, we really spent our time talking about where things like the iPhone are made and where they might be made in the future. 

You’ll also hear us talk a lot about China and how the Trump administration is focused on reducing our dependency on Chinese manufacturing — a goal that may or may not be possible on the kinds of timelines that matter to politicians, especially ones like Trump.

Smith has a lot of insight here, and he’s very even-keeled about what Altana’s map of the world is showing us about trade policy and conflict between major powers like the United States and China. There are some big, unsettling ideas here, but talking about them directly and with clarity at least made me feel like I had a framework to understand the endless on-again, off-again tariff news cycle.

Oh, two notes before we start. You’ll hear us mention John Mearsheimer, who really was my international relations professor at the University of Chicago in the early 2000s. Mearsheimer is a famous proponent of what’s called realism, a philosophy that says great-power competition dominates world affairs. 

The other note is that you’ll hear us talking about China being granted membership in the World Trade Organization in 2001. That probably seems like a bit of trivia now, but it was a major political decision at the time — a decision that you’ll hear Smith characterize as the end of the first wave of globalization and the end of the post-WWII world order. You can certainly argue with that characterization, but it’s worth calling out because it was indeed a major, controversial decision at the time. There’s a lot in this one; I’m very eager to hear what you all think about it. 

Okay: Altana CEO Evan Smith. Here we go.

This interview has been lightly edited for length and clarity.

Evan Smith, you’re the co-founder and CEO of Altana. Welcome to Decoder.

Thank you for having me.

I am excited to talk to you. Altana builds software that helps people think about supply chains and logistics. There’s a lot to talk about there.

I was reading the company’s manifesto, which you wrote in 2022. You founded the company in 2019 on the thesis that globalization as we knew it would break down and change, and Altana could build some software to help people in a new era of globalization. That bet seems very prescient right now — but also maybe a little bit destabilized, given how globalization might change. Explain what you mean by that manifesto, that founding statement that globalization is breaking down.

The premise was that the side effects and the unintended consequences of globalization since World War II were creating its demise. That breaks down into a few dimensions, principally around geopolitics. So it’s the US and China playing out a great power competition principally in the supply chain and economic theater. We can go back to that—the other one’s around climate. So you had this race to the bottom, where everything was outsourced to the lowest cost, just-in-time manufacturing locations, and the world’s physical production became dirtier and dirtier and dirtier. And you have these increasing climate dislocations and the policies, and the corporate actions, and the capital market actions that are following as a result.

The third one was around the destabilization of the middle class of the West. So if you have a rapid sort of outsourcing and a collapse of manufacturing, then the point of view was you’re going to have a populist response. And I think we’ve seen that play out. The fourth thing we said was that these just-in-time supply chains are only economically efficient under conditions of stability. Because of reasons one through three, our point of view was that the world was getting a whole lot less stable. Therefore, the supply chain itself — the fabric of the world’s physical economy — was going to break down. And I think that’s borne out since COVID. We didn’t literally predict COVID-19, but we did predict the fragility in the system.

I am the editor-in-chief of The Verge; I often think about things in terms of the iPhone. I’m a kid of the ’80s and ’90s. I remember Ross Perot saying that NAFTA, the North American Free Trade Agreement, would lead to a giant sucking sound of labor being moved to Mexico. I think a lot of people believe that it played out exactly that way. 

But at the same time, you also get the ability to make a product like the iPhone, which is just inherently global. It’s manufactured in several countries around the world, principally in China, but lots of parts flow back into China so Apple can make that product. They make it at an enormous scale. But then you can look at the Tim Cook era, and you can say from one perspective, “He increased Apple’s stock price.” But from another perspective, you can look at it and say, “Oh, the iPhone has actually prevented the great power conflict,” right?

Everyone wants an iPhone. If the Chinese government somehow makes it impossible for people to buy iPhones in China, that would be bad. If the US government makes it impossible for US customers to buy the iPhone, that would be bad, and both economies would crash. And that’s going to keep us from being in a war.

And again, I’m a kid of the ’80s and ’90s. When I was in college, John Mearsheimer wrote The Tragedy of Great Power Politics, which is a big book that predicted the United States and China would go to war, and the next day was September 11th, and then that whole thing got sent to the side. Now maybe we’re back to it, but I just look at all of that, and your thesis, which feels correct, like it’s being borne out, and I’m like, “Oh, this thing is actually being held together by some consumer products.” 

Making those consumer products and making our economies interdependent is the thing that keeps us from going to World War III. Are you all the way there? Are you building software to be like, “Okay, I’m going to try to hold this together a little bit longer?”

I’m building software and AI to help the public and private sectors navigate that dislocation and hopefully build resilience and some amount of decoupling without it resulting in a hot war.

But look, the version that you described of the economic ties that bind, I too kind of had that economic worldview. I got an economics degree from Yale in 2007, so at the absolute apex of the neoliberal and kind of neoconservative worldview, where… I literally went to school with Thomas Friedman’s daughter, and I read all of his books. His famous thing was the Golden Arches theory, right? Countries that have McDonald’s don’t go to war with each other, which was true then. So, the premise was that free trade and economic interdependence result in peace between nations. And as I’ve gotten older, I’ve kind of watched the world progress, studied more of history, and read some Mearsheimer, and I’ve actually kind of come to the exact opposite view — meaning peace gives rise to free trade and not the other way around.

When there is an equilibrium in the whole global geopolitical great power competition, then what you see through history is that trade progresses between nations; and then in the moments of disequilibrium and big shifts in power, those become more violent. Then trade collapses. And so, my fear right now is that we’re in a disequilibrium. You have a rising great power. You have multipolarity with Russia and the EU asserting themselves in different ways, and the US, at least on a relative basis as a declining power, is no longer the single global hegemony. Over the last 15 years, if you look at the world, it’s in a state of disequilibrium, and I see the tariff issue. I see all this trade policy and volatility. I see all the economic security policies and industrial policies following as a result. So it’s all sort of nested within that big geopolitical circumstance.

When you think about the thesis of Altana, in which you identified that globalization is running one version of its course, and there’s going to be a next version… I’m reading your manifesto. Your idea here is: We’re going to need a lot more data so we can build supply chains across our partners that we trust, across actual suppliers that we trust, so we can see the social costs in other countries of outsourcing or the social cost of outsourcing labor to other countries. I look at that, and what I see is, “Oh, we should make globalization better.”

It’s changing. There are some great power dynamics, but we are still going to have this. I look at the Trump administration and the imposition of tariffs, and the idea that we’re going to annex Canada, and we should stop it, right? The response from basically every member of the Trump administration to the question “Are tariffs going to cross prices to go up?” Is, “Well, then you should manufacture it here.” That seems wholly unworkable to me. 

Again, I just go back to the iPhone. I’ve listened to every explanation from every Apple executive, from Steve Jobs on down, about why the iPhone can’t be built in the United States of America. And they seem correct. I take them at their word that if they could solve the problem, they would. And yet the Trump administration seems to think that just by turning the knob on costs they will be able to bring manufacturing back to the United States. Do you think that’s possible?

I think it’s certainly possible to bring some manufacturing back to the United States. I also think it’s possible to bring a lot more manufacturing into North America. So, in other words, we’ll do some things more and better, and you have both the labor cost environment and the national security, or sorry, the natural resources, starting in southern Mexico and going north. You have that stratification of low-cost, relatively low-skilled labor going to higher-skilled, higher-cost labor, and you can manufacture most of everything that the United States consumes in our backyard, so to speak.

So that’s just sort of a fact of the market around us. Will that happen entirely? No. And will that happen immediately? No. So then, what’s the policy objective of the Trump administration? I’m not in the Trump administration, so I won’t speak to it with total authority, but I have spoken with people who are in the Trump administration, and I think there are two key themes. One is yes, absolutely, there’s a motivation to bring manufacturing back to the United States and back within the USMCA, which is the US, Mexico, Canada Free Trade Agreement that replaced NAFTA under the first Trump administration.

Which might be again replaced by the annexation of Canada. Even that doesn’t seem like a stable foundation, but okay, continue. We’ll take it, but some people in the Trump administration published that data.

Objective number one. Right. So, theatrics aside, policy objective number one is that we want to bring industry back generally, and we want to rebuild a middle class. It’s an industrialization initiative, and they want to do that by creating trade barriers. The Biden administration had a similar objective, but they did that through very large-scale subsidies and industrial policy initiatives to direct capital toward those industries they wanted to see in America. So that’s thing one.

The second thing is there’s a very clear objective, which he’s given to his policy team, to be strategically decoupled from China in critical industries by the end of his term. Critical industries — a kind of policy-speak for things that matter a lot, like food, telecom, aerospace, defense, and advanced electronics. The things that we couldn’t do without. Everyone’s less concerned about toys and apparel and more concerned about the things that are critical to our economic and national security. He’s told his team that within four years, we need to bend the supply chain such that the United States no longer has these fundamental dependencies on China. So that’s going to impact everything from critical minerals to chips to everything downstream of those, like pharmaceuticals. It’s a huge, huge economic policy objective with sweeping consequences for the private sector.

I want to hold on discussing China for one second because I want to come back to it and talk about it in depth.

Altana builds software that helps people see these supply chains. Describe how you build the software and what it does for your customers because I think understanding your viewpoint and your visibility into these supply chains will help me understand how you can see the shift from China in real time.

The whole world’s a supply chain. At the base level, what we do is model the whole physical world of companies making things and buying and selling them from other companies. So it’s like a bottom-up, Google Maps-style view of all the companies — where they operate, what they make and sell, and the connections between them. In the same way that Facebook built a social graph that connects everybody within six degrees of Kevin Bacon, and LinkedIn is our professional relationship graph which shows how you’re connected to all your colleagues and counterparts… We’re doing the same thing but at the scale of the whole global supply chain network. So that’s the foundation. And the way that our customers primarily use that is to understand and manage what are called value chains. This matters in the context of the globalization discussion we were just having.

So, what’s a value chain? A value chain for any product is the whole network of production all the way back to the soil. Let’s take the iPhone: The silicon comes out of the ground, it turns into chips, and it turns into all the logic in your phone. The graphite comes out of the ground, the aluminum comes out of the ground. The whole system of raw materials through the intermediate goods, through the final assembly, and the sale and end use of the product is called the product value chain. We make it possible to know and manage that whole network for all the products in the world, and it sits on top of that map of the world that we’ve built. So why is that novel? Because of globalization and outsourcing. 

There was a deliberate move to outsource that whole value chain. The whole theory was that we should only do our comparative advantage. We should only do the most specialized version of that within our borders, and everything else got outsourced and outsourced and outsourced and outsourced. So it became impossible to really know your supplier’s supplier or your supplier’s, supplier’s supplier, and even in the other direction. Western electronics companies keep seeing their stuff get into Russian weapons systems. Well, they’re mostly not selling to Russia, so it’s that customer’s, customer’s, customer that’s receiving the electronics, right? That’s the value chain. 

We make it possible to know that and to map it, manage it, and actually collaborate with those value chain partners. The story I like to tell is in the 1950s, Ford Motor Company used to own the entire value chain of its cars. They owned rubber plantations in Brazil to make tires, and now they just do the final assembly. So that’s the point. Boeing doesn’t own its value chains anymore. They can’t even make an airplane. Ford is in the final assembly business. Apple doesn’t even manufacture anything; they just create the software and the designs, and everybody else manufactures it. So what we’re doing at our core is we’re kind of making it possible to do what the industry used to do, which is to see and connect and have some control over that whole network. 

In our case, you don’t have to own the whole network. You can illuminate it to see the compliance, security, and resilience dimensions of it. And we can actually give our customers the ability to map, manage, and collaborate across that whole network.

So you collect a bunch of data from across these value chains from different suppliers. Why do these companies give you that data?

They don’t share it in the broadest sense. What they’re doing is they’re connecting to a network that we provide. And the thing we built the company around, the core invention of Altana, was that everybody could kind of keep their data sovereign, siloed, private, and connected to the network, and we’d sort of solve for both at the same time. What that means practically is that it’s like a hub-and-spoke federated data model. So we provide the shared map of the world and all the software and AI systems in the hub, and our customers keep their data in a spoke, and we bring the platform down to their data and not the other way around.

There’s a Maersk spoke, there’s an LL Bean spoke, there’s a General Atomic spoke. None of those customers are sharing all their raw data with each other. They’re not allowing us to pool it all centrally, but what they are doing is subjecting their data to the network; they’re connecting their data to our platform, and they’re seeing how they fit within the broader network. We learned from that. We build out the network connections, and we train AI systems, and that all adds up to a model where every customer that joins adds to the visibility, connectivity, and intelligence for all the other parties.

Put that into practice for me. Let’s just use LL Bean. LL Bean wants to manufacture some more hats for the stores for ski season. What does Altana help them do?

LL Bean tends to know not just the garment manufacturer it buys from and has a direct relationship with, it will also sometimes know the textile manufacturers that feed into those garment facilities. What it doesn’t know is everything upstream of that. It’s like, where does the cotton come from? Where does the polyester come from? Where does the zipper come from? At the base layer, we just help them answer the question: What’s the whole network of the value chain for every one of my goods, all the way back to the soil? They’re motivated by a few things.

One is ensuring that their goods are compliant and free of forced labor. So speaking of trade barriers, that was a big one that happened in 2022. The US banned all imports of goods into the United States that have any of that upstream value chain content coming from Shenzhen, China. So it’s a rebuttable presumption that the Uyghur forced labor issue was sort of infecting all those goods. So that’s a big one. 

You literally can’t import to the United States, and they’ve had detentions; everybody in the industries had detentions. And so, navigating through that, making sure the value chains are healthy and compliant, and that you don’t have these big multi-million dollar, tens of millions of dollars, hundreds of millions of dollars of border interruptions and delays.

So that’s one. And now they’re expanding that to look at all things sustainability in their value chain. So how do you use Altana to understand your carbon footprint? How do you understand the workers’ rights and ecological impacts of production, all the way through the chain, and on and on and on? We become that sort of foundational network that they use to not just see the whole global value chain network but then engage it and make it better over time.

I’m still just thinking about the incentives to expose any data to your network. I get it at the top level, right? A car company or a shipping company at the top level wants to see all this data, but then you need information from the lowest levels of that from the soil. Why does the big industrial producer in China want anyone to know where its products are coming from?

I’m not sure they do in all cases, but what’s interesting is they now have to in more and more cases. The big arc of policy here, whether it’s tariffs, forced labor bans, or a carbon border adjustment in Europe… I can paint a big picture, but in the last two years alone, there have been 1,200 net new trade barriers added. And basically, what it says is that if you want the downstream goods, the car, the iPhone, or the clothes to make their way into the UK, the US, Europe, Australia, or Canada, then the large upstream manufacturer has to disclose all this information about your goods and your production practices.

That’s always been a little bit true in the context of moving goods across borders. There are certificates of origin and customs classifications that depend on the nature of the goods, and that’s how you pay your tariffs and duties. And you give the Food and Drug Administration the certificates and permits. That all is kind of business as usual. What’s so different over the last few years is they’ve created so many new and more stringent requirements on that whole upstream value chain, saying, “You have to share this information with us, you have to make these declarations, and it’s a matter of market access.” So your customers and your customers’ customers can’t get their goods into the biggest markets in the world without it.

When you look at what Altana provides in that, that’s a problem to be solved. You can say, “Okay, well, add some software to make it easier to share data better.”

You’re also talking about adding AI to it. What kind of insights are you generating from collecting all this data? It seems like you have an enterprise data provisioning problem, but then you also, generally with AI, want to see all the data and collect an insight [from it]. It seems like there’s a tension between wanting to see the information across the whole network and actually generating insights.

Where do you get the data, and what kind of insights do you generate?

We get the data across both the publicly available and commercially purchasable universe of data. I mean, we spend a lot of money bringing all the data to us that we can get our hands on. And then, as I mentioned earlier, every customer that joins the platform is connecting their data to that model of the world, and then it’s subjecting their data to that whole framework of visibility and shared learning. So we’re now at the place where roughly half of all that supply chain connectivity, all the supply chain visibility, is coming from our customer network itself. That’s another way of saying that we’ve doubled what you can theoretically do by going out and scraping the internet, getting publicly available data, and licensing data from commercial vendors. So we’ve exhausted what you can do in the public domain, and we’ve now about doubled that, and the lead keeps compounding.

We use AI to connect all that data together. So it’s billions and billions of data points. It’s really messy stuff. It’s in Chinese, Cyrillic, and Spanish, so it’s in many different languages. You don’t have the benefit of unique IDs, like you do in personal data, where an email address or a telephone number can help you join all the records together, like Nilay or Evan. In the world of businesses and transactions, purchase orders, and shipments, you just have this messy, semi-structured text. You need AI to process all that data and connect it into one clean, unified representation of the world. We’ve been building our own homegrown AI systems since day one to do that. Then, you need AI to map out those value chains. So that’s where we’ve invested really heavily over the last few years.

We can almost instantaneously illuminate the inputs and outputs of anybody’s products across the whole global network with very, very high fidelity. So that’s always a jaw-dropping wow moment with customers, where it’s like, how could you possibly know this, and how could you possibly draw these connections? And then, on top of that foundation, we have AI systems that are doing everything from detecting single points of failure and business interruption risks. We can look at the whole network and say, “Here are upstream suppliers that are at risk of failing financially, just by sensing what’s happening in the supply chain.” That’s been one of our crazy success stories over the last year, just seeing the model perform on that.

We have AI systems that automate all of the cross-border tariff and customs compliance work and all the logic associated with that. So that’s a $150 billion per year industry — just humans reading text about the nature of the goods and then assigning them the codes when they come across a border that say, “Here’s the tariffs and duties, and certificates, and permits.” So we have an agentic AI system that’s doing all of that. You can put the customer’s broker in the loop and give them ultimate control, but that’s become wildly performant. That’s not the whole list, but the point is that it’s AI constructing the map. It’s AI to show the customers how they fit inside of the map, what are the networks that are relevant to them, and then also AI that generates insight and automates entire workflows.

Let me ask the two Decoder questions, and then I want to put all this into practice with what’s happening in trade right now. How big is Altana right now, and how have you structured the company?

We’re about 240 people and our two primary lines of business are enterprise and government. We then have financial services partners and logistics partners through whom we service the enterprise. So, they connect to our platform to better serve their customers and connect them to the network.

How do you make decisions? What’s your framework?

This has evolved pretty meaningfully in the last few months. So, where I have little depth or relatively little depth, depending on the team around me, I try to speak last and ask questions, and then decide when I need to break ties or move the group forward. And then, the net new thing over the last couple of months is permitting myself to be very decisive when I’m deep in the details and driving the company forward that way. So I’ve actually leaned in pretty aggressively on product management and some big audacious product bets we’re making right now, and a lot of the network architecture we’re building to connect the public and private sectors into this whole global value chain network. And so, on those two dimensions, I’m highly opinionated, and I’m driving very, very fast decision-making by doing it myself and bringing groups along with me.

Do you think that’s reflective of the fact that the foundation of the entire global economy seems to be changing very quickly?

It’s a big part of it. We started the conversation where you’re like, “Hey, all these things seem to be coming true,” and that’s true. Our lived experience is that these things we kind of expected would be relevant products that could exist on top of our platform… Over time, we’ll do this, and then we’ll do this, and we’ll do this. There’s demand for most of it right now. And so, the big scramble in Altana is like, how do we launch new software? How do we meet the demand? How do we get out there, and how do we do this from a unified base? I’m finding myself in wartime mode and calling a lot of shots to get through that stuff with velocity.

How real-time is the insight you get? You mentioned being able to detect when an upstream provider might be failing, right? Is that in real time? Is that… You wake up, and there’s a map, and there’s a blinking red light, or is it on a trailing basis?

It’s actually a spectrum. So some of the data we process is — I think of it as negative latency. You see the intent of a transaction before it occurs in real life through what are called booking requests and shipping instructions. That could be a thing that will happen five weeks from now. And then, we get the shipment itself, and then we get the customs entry associated with the shipment. So you see this life cycle of the supply chain activity. We also have real-time event feeds through a partnership with Dataminr, and we can overlay that on the whole map of the world and contextualize that for our customers. You see some things that kind of develop over time — in that, you need more pieces of data, pieces of signal to kind of build up a broader picture you feel good about. So it’s really a spectrum. There are some things that you can see far ahead of the curve. It’s totally predictive. Some of it is more… there’s something sort of taking shape in the network, and then some of it is sort of post-factum.

All right, let me put all of this into a blender. I think I’ve got a sense of the company and what you can see. Trump announces tariffs; 24 hours later, he walks back the tariffs, and then he announces them again. Maybe they’re on. Howard Lutnick, the commerce secretary, shows up on CNBC and says they’re going to go away. Do you see that reflected in the data? Do you see that across the network?

What does that look like? Do you see the aftershocks of it? Do you see the supply chain reconfiguring?

Yes and yes. And this goes back years because the first Trump administration was the first to go hard with tariffs, and Biden continued it. So yeah, you do see a couple of things. The obvious one is for the goods that have tariffs applied. You see a redistribution of imports associated with the tariffed goods from the tariffed countries, and then this is kind of obvious, but it’s either going to be manufactured locally, or it’s going to come from other places where the tariffs don’t apply. The more interesting thing that you see is that the ownership structures associated with these big global supply chain networks change, and so does the nature of the value chains. So what happened in Trump’s first term, when there was a bunch of Chinese economic policy and trade barriers that were put up, was you saw a big shift in Chinese supply chains going to Mexico and Canada, where two things happened.

In one case, they would just do final assembly, so they’d ship all the refrigerator parts to Mexico, turn it into a refrigerator, and then import it into the United States under USMCA and not pay tariffs. The other version of it was through what’s called the e-commerce de minimis threshold, where the premise was like, if the goods are under $800, we don’t need to do a whole customs entry, pay tariffs, and put a trade compliance process in place. Let’s just rip it into the country and get consumers what they want.

So then, unsurprisingly, you had these mega warehouses built up on the Mexican and Canadian borders, where they were drop-shipping items under $800, or even components that were sort of disassembled so that each of them was under $800 and then imported into the United States through that loophole. Both of those things are coming to an end, and it’s obviously a little chaotic, but I see the whole USMCA tariff thing through those lenses. And the other thing I should say is that it’s really interesting when you see the ownership networks because we map those too. You’re seeing all these Chinese companies open Mexican and Canadian subsidiaries to import goods from themselves and then bring them into the United States, subject to those exemptions I just spoke to.

Is that what’s driving all of the posturing towards Canada and Mexico? Is it an underlying sense that China’s actually taking advantage of those agreements?

Yeah. I have to get a little wonky to argue the case. So yes is the answer, and if you look at where the tariff thing landed, and again, by the time this airs, who knows where we are, but where it landed three or four days ago was they said, “Okay, we’re going to do the 25 percent Mexico and Canada tariffs for any goods coming into the United States that are not registered under the USMCA.”

That was fine print in the announcement. That is pretty consequential. What that means is that companies that bother to prove that the material origin of their goods is within the value chain definitions of the USMCA — again, this is the free trade agreement that replaced NAFTA — those companies and products are not subject to the tariffs. That’s roughly 40 percent of Mexican and Canadian trade with the US. So the 60 percent that hasn’t done the work, to say the raw materials and the intermediate goods are all eligible as per the USMCA — and they’re part of our free trade world — those 60 percent of goods are still subject to the 25 percent tariffs. At the same time, they hit China with extra tariffs. So they’ve now done two, a 10 percent and another 10 percent on all Chinese imports. The dotted line through all those things is that the Trump administration is trying to bring Mexico and Canada into tighter alignment with the United States and its trade policy vis-à-vis China.

And all of that is to say, we want more manufacturing here in the North American continent and to get away from China.

You’ve got a big map of the world. You see the value chains, you see where things are manufactured. If you just look at where all the puzzle pieces are today, can you actually accomplish that in four years? Can you decouple the United States from China at the end of Trump’s term? You said that was the goal, and I don’t know how to take those Jenga blocks and make that tower. I know how to knock the tower down; that seems very obvious to me. But I don’t know how to make the tower or make a new tower.

What I believe is that they’re going to try, and then in some cases, you’re talking about billions of dollars, tens of billions of dollars of capital expenditure (CapEx) in the ground in China. Some of the biggest chemical companies in the world have manufacturing locations in China that feed into all of our pharmaceuticals and advanced manufacturing processes. Is that kind of CapEx possible to replicate elsewhere within a four-year timeframe? That’s going to be pretty tough.

Do we even have the capital to do it?

There’s a lot of capital sloshing around the world.

Yeah.

Yeah. So, should the US government invest in that, and will it? Certainly not in this administration. It will not do that, right?

I suppose the operative part of my question there was the “we.” There is a lot of capital, but do we in the United States have the ability, the will, or the coordination to do it? That seems up in the air.

I do think, philosophically, what this administration is saying is rather than the Biden approach of using mostly incentives and capital to build the things that we want here… So the IRA, the CHIPS Act… those deployed a lot of capital and loan guarantees and other ways to kind of subsidize and motivate the manufacturing of those things here. What this administration is doing is saying, “We’re going to use trade barriers to restrict the market access of competitive products,” and a lot of it’s aimed at China obviously, “and we’re going to let the market figure out the rest.” So we’re going to create the barriers. We’re going to distort supply and demand and then let capitalism figure it out from there.

On what timeframe do you think capitalism figures that out? Again, I’m just looking at Altana’s product, right? You have a view of this system and how it reacts to different kinds of shocks. What timeframe do you think that we can figure this out?

It is just a spectrum. I mean, a lot of these things are already in motion, so you’re just accelerating the near-shoring of advanced electronics. You’re accelerating the near-shoring of automotive supply chains. The pharmaceutical CEOs I’ve talked to in the last six months have already been talking about creating new manufacturing locations outside of China for all their key stuff. Some of these things are going to move a lot more quickly, and some of these things are going to be pretty tough. I think for those examples where you just have massive CapEx in China, it’s really hard. And I think the most intractable one, honestly, is what are called critical minerals. 

It is rare earths, which we all have heard of, and then the other metals that go into all of our most important electronics. So, think nickel. Nickel’s not rare, but it’s critical. And China has a virtual lock on all of the world’s critical mineral processing. They don’t have a lock on the raw material itself. The raw materials are pretty pervasive. They’re abundant. China, over the last 25 years, was very deliberate about building a stranglehold on the refining of the raw ores. And this is nasty stuff, right? These are typically big open pit mines where they’re doing cyanide leaching to separate these ores from other metals and refine them. So it’s stuff we don’t actually want in our backyard, which is all well and good if there’s a free trade environment. But in a world of geopolitical and economic competition, China now has between 60 percent and 98 percent of the world’s critical minerals on lockdown, and they use that market position now to kill any emergent refining in a country or a company that they don’t control.

So a new one pops up, and we’re refining lithium, and we’re doing Lithium oxides. Well, you’re going to see, and this has happened repeatedly, China floods the market, depresses the price, and puts that company out of business. They either go away, or the Chinese company buys them. And so, the reason this really matters is that you can’t make missiles, airplanes, guns, telephones, or GPS [without these minerals]. Our everyday modern economy depends on all these critical minerals, and they all run through China. So, can that be fixed in four years? No, not without a radical sort of World War II-style change in our regulatory stance, our willingness to invest, and a massive cross-border partnership with our allies. I don’t see that happening.

When you describe that, I think about directed industrial policy, right? China made a decision. They made a huge trade-off against their environmental policy, the air quality that their people breathe and the labor abuses their workers might suffer, but they said, “It’s worth it, because over some timeframe we will have this stranglehold on the market and that will create peace, create prosperity,” some set of goals that industrial policy will deliver.

We can look at Taiwan, for example. On this show, we’ve talked a lot about TSMC and Taiwan’s massive investment in the chip industry because it saw the future and said, “We’re going to do that here. We’re going to make an industrial policy from the top down that says all the chips are going to be made in Taiwan. TSMC is our national champion; here we go.”

Those things came true. Again, the trade-offs are very real, and very clear. We don’t have a directed industrial policy in the United States. To the extent that the Biden administration was trying, they were doing it in the most Freakonomics neoliberal way you could do it, right? They created a bunch of incentives, and they hoped people would not game them. You can see how that played out for them.

The Trump administration is doing the exact opposite, right? They’re creating a bunch of impediments and saying, “Our market is so valuable that if you want the n+1 American consumer to buy your car, you have to do a bunch of stuff to get to our market.”

That seems just as naive. It’s still not directed industrial policy. It’s negative incentives instead of positive incentives, and you still might game the system in some way. But it doesn’t seem like that plan of, we have to actually say out loud, “We’re going to do nickel refining to decouple from China.” They haven’t said that out loud. 

They’ve said, “Here’s a bunch of structural impediments that might result in some capital directed at nickel refining in the United States.”

And that to me seems again, for all the criticisms of the Biden programs, like you run into the same problems. You’re not actually saying what you want. You’re hoping that the negative incentives will create momentum for the market to deliver the result.

I would add one thing to the picture you painted, though, just to be fair about it. The other thing that the Trump administration is trying to do is massively deregulate. So, why isn’t there nickel refining in the United States? Because it’s impossible to get it permitted. It’ll never happen, right?

But again, that’s a trade-off, right? There’s a reason we don’t want nickel refining.

I’m not making a normative judgment. I’m just describing the landscape I see. I think the Trump thesis here is, “Let’s remove the barriers to building, let’s create the barriers to what they view as unfair trade and economic abuses by adversaries. Let’s let the market figure out the rest.”

And you’re asking, is there a third way? In general, my perspective is that the United States can’t possibly out-centralize China. It can’t possibly out-execute top-down economic control better than China does. Our comparative advantage, our right to win, is through innovation. So create a rule of law and a risk capital environment, and a labor environment where the best ideas can win, people can take risks, and that can flush through the system — that has been our comparative advantage. I’m more sympathetic to an economic security framework that plays to our right to win than one that doesn’t. I think the Biden kind of Freakonomics thing you described was very much like, “Okay, let’s sort of take other examples of central planning and industrial policy, and make these big billion, a hundred billion dollar, trillion dollar injections into sectors of the economy and specific recipients of aid, and then hope it works out.”

That’s what China does. I’m not so optimistic about that working for us. So I do think there’s a big question of, “Is there a third way, and what should the economic security policy framework of the United States be?” That’s a conversation I’m getting into more and more in the academic and think tank worlds. I don’t think we have a coherent answer to that as policymakers, and I don’t even think we have the language for it in a lot of cases.

A lot of what you’re talking about with the comparative advantage of the United States historically being risk-taking innovation, I agree with that. The foundation of that has been the relative stability of the United States after World War II. I’m not saying it’s been totally stable — it hasn’t. No one listening to the show looks back across the last 20 years and thinks, “That’s been stable.”

But fundamentally, politically, we’re stable. Economically, there is a sense of stability. Data-wise, we have not, until recently, had the United States government deciding that it will redefine how GDP is calculated. There’s sort of an essential policy and economic stability to how the United States works.

Your company relies on some amount of stability in the system, right? Even to just look at a system and say, “We can see how things are moving,” implies that all the actors in the system are rational or predictable in some set of ways, or at least participating. That is what feels destabilized right now. 

The tariffs are on, the tariffs are off. We’re going to export bourbon to Canada. We’re not going to do that anymore. Trump is going to say, “We’re going to cut our own lumber in the United States,” but all the trees are in Canada. That seems destabilizing. We’re going to decouple from China in four years. Inherently destabilizing. Apple is going to announce $500 billion worth of investment that they already planned for, and somehow, that’s going to make it okay. All of that seems new and destabilizing in a way that, even in Trump’s first term, didn’t exist. Do you see that? Is that reflected? Do you worry about that?

I mean, back to the beginning of the conversation, we bet our careers on it. I just see it as an inevitability. I am a macro guy. I like these big systems problems, and I’ve seen, for a long time now, the gears of history cranking. It’s pretty obvious to me that the system I grew up in, and I think the orthodoxies of our time, have come to an end. And so, yeah, it’s destabilizing. The moment in time is one of instability and destabilization. I see these trends playing out almost irrespective of who’s in office here or there. Like, look at Brexit, look at what’s happening in the EU. Look at all the climate dislocation and the waves of immigration across borders that then result in nationalism and populism. And you kind of add all these mega-trends up. The rise of China, the relative decline of the United States, the global debt burden and how that’s playing out in economic policy, and wealth inequality are these asset bubbles inflated and reflated. All these things, to me, signal a breakdown of the system.

When you look at your map and you see China in the center of it, and almost all these value chains end up flowing through or around China, can that be reallocated? Can you shift things around to a world where you have the United States and its allies’ value chain and China and its allies’ value chain?

I mostly see it going there, at least for the more sensitive critical value chains. I don’t see it playing out any other way. And this isn’t just a sort of the US doing its thing. China publishes its economic intentions every five years, and then they reify them and the national addresses. This is their explicit strategy: to create economic resilience and dominance in the technologies that they believe will be the commanding heights of the 21st century. And to ensure that the value chains associated with them as either critical imports or as exports where they know they need to deploy their production capacity (because they’re an export-driven economy) — they are explicitly trying to secure both. And they write it down, and they do what they say. To some extent, how did we get here? We had a rules-based free trade order that the United States architected, and everything made a lot of sense economically, middle class impacts aside.

The admission of China to the World Trade Organization in 2001, I think, was the end. That was the beginning of the end. And you had a very large and growing economy that was exploiting the free trade system systematically, stealing trillions of dollars of intellectual property to feed their industrial engine and subsidizing their exports at the expense of everybody else’s markets. And so, 25 years later, we are where we are. And so, it was sort of bigger than its neighbor at an unprecedented scale, and I think the system is responding to that fact, right? That was the catalyst. It wasn’t that somebody here or there was elected and did some policy actions. So I view things in sort of this multi-decade lens, and the policy environment, the political environment right now is a consequence of that as opposed to the primary driver of it, if that makes sense.

I wrote my law review article in law school on China’s use of open-source software to get itself into the WTO. And their approach to intellectual property was, “Oh, well, we’ll do copyright law,” because that is a foundation for open-source. And now we have Linux. That was a fascinating rhetorical move. I don’t know if my law review article was any good, and I don’t know if anybody understood what I was talking about at the time, but I was like, “Pay attention to Red Flag Linux.”

The other end of that, the positive end of that, is one where we start, the iPhone. The iPhone exists at scale, at a massive scale. The idea that we manufacture as many iPhones as we do with three nanometer TSMC chips in them is mind-boggling.

There are many criticisms of Tim Cook one could issue, but… Did you maintain the rules-based world trade order for as long as you did because the iPhone is such a successful product, and you managed all of those supply chains? You managed both Trump’s first term and the Biden administration through COVID-19. That’s a remarkable fact of American industrial policy, right? Apple exists, and it held the world together for a minute.

Do you think that there are still those kinds of products or those kinds of companies that will keep this thing going even as everything else gets destabilized? Because that is the success. There has not been a great power conflict or a hot war between the United States and China in all this time. Even as things frayed, even as both sides felt taken advantage of or felt hostile towards each other, or sent fighter jets flying at each other for no reason, just to have a good time.

Honestly, I hope that I can look back at the whole arc of Altana and say that we are one of those companies that bent the arc of history in that way.

How’s that?

I think we’re making it possible to be resilient and to navigate this change for both the public and private sectors. I mean, we work with governments of the West and allies to enforce this stuff, to navigate it, to find some of the dependencies and vulnerabilities in these networks, and to mitigate them. We also help them simulate policies and see how they might play out. What are the direct impacts for the second, third, and fourth order impacts so that better decisions get made? And then we work with the private sector to navigate all that stuff. And so, if we succeed at the scale that we’re trying to, our mission is to fix globalization.

Then, I think you can envision a world where business in the West and business in the East continue to be transnational, and you have these miracles of productivity, and you have the economic ties that bind. At the same time, you de-risk the most critical infrastructure and the most critical supply chains. Through that de-risking, you prevent accidents from happening, and you prevent the worst from happening.

So, paint a picture like, let’s say China moves on Taiwan by the end of the decade. And I’ve been in a war game with US military leadership, where one of their assumptions is that GPS goes off, and all of the critical water and electricity facilities in America are shut down. Our cars can no longer navigate because of all the logic in our cars. And those are all through supply chain exploits that have been built into the systems by China over the last two decades and known vulnerabilities, known state-sponsored cyber attacks that create those exploits…

In that scenario, is America more or less likely to go to war? I actually think it’s more likely to go to war. I think there’s a higher probability that escalation occurs and tensions fray. There’s more popular support for conflict. And so, in a world where we’re more and more de-risked, where you don’t need to go to war over an island, where you don’t need to defend a specific supply chain or you don’t need to react to a vulnerability that your adversary exploited and shut down all the critical infrastructure in the country, that’s a world that, to me, is less likely to go to war.

I feel like some people will take issue with your characterization of Taiwan as an island, but I get it. Is there enough world? This is the biggest question that I have for you. Your software builds you a picture of the value chains across the world.

I don’t know if there’s enough capacity. I mean, it’s a big world. I am not saying all of it has been industrialized. We’re producing everything everywhere. But it does seem like with the past 25 years of globalization and maybe the next wave of globalization that you’re helping bring about, it has reached an equilibrium. And destabilizing that equilibrium just implies there will be more capacity, that we’re going to move the chips out of Taiwan and make them in India or something. I’m not sure that there’s enough capacity to reorient everything and maintain an equilibrium or to change it without something even more destabilizing happening.

I’m wondering if that’s where you’re at or if you think you can actually make the shift. Because if you’re like, “We need to move to the next phase of globalization,” my view is that Trump may kind of want it to stop, right? And it still feels like tension.

I think there’s enormous production capacity at basically every stage of the value chain. So I don’t think there’s any kind of fundamental limit that we’re bumping up against, either in terms of labor price, labor skill, raw material availability, or manufacturing sophistication. And I’m also reasonably optimistic. I’m not maximally optimistic about automation and software-defined manufacturing and where a lot of this is going. So I think there’s an enormous amount of capital out there chasing good returns, and there’s an enormous wave of productivity gains that are already underway, and it’s going to keep scaling with AI and robotics. All those things add up to the means to refactor.

You’ve got the big view; you’ve got the software. If you’re someone listening to this, you’re an entrepreneur, you’re a builder — we have a lot of those listening to the show — where would you go build? Based on your map, what seems the most stable right now?

Geographically or in terms of industries?

Both.

So, tariffs notwithstanding, if I were starting from scratch and not doing Altana, the thing I’d be most excited about is building a software-enabled and AI-driven manufacturing value chain for advanced electronics in North America. I’d be focused on producing in Mexico, sourcing most of my components from North America, serving US and Canadian markets, and taking advantage of the free trade agreement that exists.

And again, the way that the tariffs just played out only applied to anything that wasn’t certified under the USMCA. As you know, North American origin got hit with tariffs, but everything that’s certified is still sort of free trade. I would go long on the Mexican demographics. I would go long on the Mexican economy, generally. I would go long on North American demand for advanced electronics components in everything from automobiles to aerospace, to defense, to the whole picture. And I would try to build a Shenzhen of North America. One big, integrated, and co-located value chain where you could really do mass-scale drone manufacturing, optics, the whole thing.

The Chinese government also made a huge investment in labor, right? When Obama asked Steve Jobs, “Why can’t you make the iPhone here?” His answer was, “There’s just not enough manufacturing engineers. It’s not going to happen.” 

Do you see that labor pool existing in the United States, Canada, or Mexico? Do you have to create it? Is the Mexican government going to be better at that than we are?

Mexico’s a lot better at it. We’re terrible. And I think some interesting experiments were begun with junior colleges and community colleges and some of those initiatives. But we’re nowhere close to having the skilled labor necessary to execute the industrialization objectives of either the Biden administration or the Trump administration. So, there is a massive gap in both of those worldviews and approaches.

Do you think that can be reconciled?

Partially, through automation. Not fully.

Evan, this has been an incredible conversation. I obviously could talk to you for another hour. You’re going to have to come back.

I want to know more about your law school thesis. I think China’s doing the same thing with open source and AI that they did with open source back then. But anyway, maybe that’s for another one.

Thank you for being the only person, including the editor of the Wisconsin International Law Journal, who has ever said, “I want to know more about this thesis.” I appreciate you for that. It was a good suck up to end the show. We’ll have you back. Well, here’s what we’ll do. When tariffs go on, we won’t have you. When tariffs go off, we’ll have you. So, every few weeks, we’ll have you back on the show.

All right, I’m going to be a regular.

Yeah, that’ll be great. Evan, thank you so much. We’ll talk to you soon.

Thanks for having me.

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

On today’s episode of Decoder, I’m talking to Verge policy editor Adi Robertson about the Take It Down Act, which is part of a long line of bills that would make it illegal to distribute non-consensual intimate imagery, or NCII. That’s a broad term that encompasses what people used to call revenge porn, but which now includes things like deepfaked nudes.

The bill was sponsored by Sen. Amy Klobuchar (D-MN) and Sen. Ted Cruz (R-TX), and it just passed the Senate. It would create criminal penalties for people who share NCII, including AI-generated imagery, and also force platforms to take it down within 48 hours of a report or face financial penalties.

NCII is a real and devastating problem on the internet — it ruins a lot of people’s lives, and AI is just making it worse. There are a lot of good reasons you’d want to pass a bill like this, but Adi just wrote a long piece arguing that giving the Trump administration new powers over speech in this way would be a mistake. Specifically, she wrote that it would be handing Trump a “weapon” with which to attack speech and speech platforms he doesn’t like.

At a high level, her argument is that Trump is much more likely to wield a law like this against his enemies — which means pretty much anyone he doesn’t personally like or agree with — and much more likely to shield the people and companies he considers friends from the consequences. And we know who his friends are: it’s Elon Musk, who now works as part of the Trump administration while at the same time running X, which is full of NCII.

Now, Adi and I have been covering online speech and how it’s regulated for about as long as The Verge has existed. We have gone back and forth on where the lines should be drawn and who should draw them as many times as two people can over the years. But that conversation has always presupposed a stable, rational system of policymaking that’s based on the equal application of law.

Here in 2025, Trump has made it clear that he can and will selectively enforce the law, and that changes everything. Once you break the equal application of law, you break a lot of things — and there’s just no evidence Trump is interested in the equal application of law. You’ll hear us really wrestle with this here. The problem doesn’t go away just because the solutions are getting worse, or that the people entrusted with enforcing the law are getting more chaotic.

So in this episode, Adi and I really get into the details of the Take It Down Act, how it might be weaponized, and why we ultimately can’t trust anything the Trump administration says about protecting the victims of this abuse.

• The Take It Down Act isn’t a law, it’s a weapon | The Verge
• A bill combatting the spread of AI deepfakes just passed the Senate | The Verge
• Welcome to the era of gangster tech regulation | The Verge
• FTC workers are getting terminated | The Verge
• Bluesky deletes AI protest video of Trump sucking Musk’s toes | 404 Media
• Trump supports Take It Down Act so he can silence critics | EFF
• Scarlett Johansson calls for deepfake ban after AI video goes viral | The Verge
• The FCC is a weapon in Trump’s war on free speech | Decoder
• Trolls have flooded X with graphic Taylor Swift AI fakes | The Verge
• Teen girls confront an epidemic of deepfake nudes in schools | NYT

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

Today, I’m talking with Almar Latour, who is the publisher of The Wall Street Journal and CEO of its parent company Dow Jones, which you can think of as a huge research and data provider for companies of all sizes. Dow Jones itself is part of Rupert Murdoch’s News Corp. 

Latour is a fascinating guy. He started as a news assistant at the Journal in the ’90s, spent time as a tech reporter, and eventually rose through the ranks to become CEO in 2020, putting him in charge of how everything makes money.

And if you’ve been paying attention, you know it’s a tough time to be making money in the news business, especially the paid news business. In addition to competing with the flood of free content on various social and streaming platforms, the industry is also facing new challenges like fights about copyright and AI and battles with the Trump administration, which has been pushing hard to shut down critical reporting and limit press freedom.

Latour has insight into all of that: On the news side, he’s made a big content-use deal with OpenAI while also suing Perplexity for training without permission. On the other side, he’s pushing to build his own AI data products for Dow Jones customers. 

He is a fierce defender of press freedom who fought to have Journal reporter Evan Gershkovich released from Russia after being imprisoned for more than a year. But at the same time, he works at News Corp, whose chairman, Rupert Murdoch, has deep ties to President Donald Trump and who has overseen a vastly polarized and politicized era of news media.

So I asked Latour about all of that and really pushed him on a few of his answers — even right at the top of the conversation, when I asked him about the Journal cutting a huge chunk of its tech reporting team literally the day before we recorded. To his credit, Latour was game and he hung in there, but you’ll hear him congratulate me for almost getting him to slip up. I did my best; I think many of you will have thoughts about it.

Okay: Almar Latour, CEO of Dow Jones and publisher of The Wall Street Journal. Here we go.

This interview has been lightly edited for length and clarity. 

Almar Latour, you are CEO of Dow Jones and the publisher of The Wall Street Journal. Welcome to Decoder.

Great to be here. Thank you.

I have a lot to talk about with you. There’s an entire set of complicated AI questions that might be existential for the media industry, but you’re heavily invested in building some of that technology and some of those services, which I think is interesting. There’s the general state of the press in 2025, which I want to talk to you about. I know you’re very interested in press freedom. 

But it happens that I have you on the day after the news, so I want to start with the news. Just last night, The Wall Street Journal, of which Dow Jones is the publisher, restructured how it covers tech and media. That involved cutting about 10 or 15 editors and reporters. Obviously, I’m very interested in how you structure a newsroom to cover tech. Why make that decision? Why get smaller?

First, this is a newsroom decision, so this is squarely in the terrain of [editor-in-chief] Emma Tucker, who’s a new editor, relatively speaking. She’s in year two, moving into year three. Emma was hired with a remit of helping to increase engagement with our existing readers, to expand our readership, and to maintain and enhance the quality of our coverage. She has set out over the past two years to rethink how she wants to offer news with The Wall Street Journal newsroom. 

Her consistent message, and this is one that I subscribe to, is that distinctive journalism is what makes the difference. It’s about knowing the interesting story, the story behind the story, and to have exclusive journalism and insights. That I say as a preface because Emma has been making changes to nearly every part of The Wall Street Journal and continues to do that.

So what happened yesterday was a continuation of that. Generally, when you look at the changes that she’s brought in — I’m not speaking specifically about the San Francisco bureau but tech —  it’s been new talent. She has an antenna for what she thinks works there. And one of the areas where that’s been very pronounced in recent months is in Washington, D.C. That has gone through several cycles of changes.

She’s brought in new people, and that has had consequences. The marching orders are slightly different when there is a closer connection to the center of the newsroom, where the decisions about news can be made in the context of a broader story that’s happening around the world, rather than in isolation around a certain topic. So that’s the context for yesterday. 

I’m not going to comment on specific individuals or Emma’s specific plans. I’ll leave that to her. I’ve worked with quite a few of those people. As you know, I was a tech reporter myself. But overall, what Emma has focused on, and what the Journal and Dow Jones are focusing on, is going deeper and having more exclusivity, more “proprietary content.” Moments like yesterday are absolutely never easy.

I think I see a thesis of Dow Jones as a company, and we’ll get to the big picture in a second here. The idea is you’re going to give a bunch of people in the business world an information edge, whether it’s with the Dow Jones Information Services, some of the AI tools, or with The Wall Street Journal, which gets a bunch of scoops and tells people stuff they didn’t know before. 

I’m just looking at this broadly. I’m looking at Dow Jones as part of News Corp. I’m looking at News Corp’s financials from last month. Revenue at Dow Jones is up to $600 million. That’s up 3 percent. Your earnings are up 7 percent to $174 million. But the cuts are in tech, which is dominating the world. I’m just wondering about that resource allocation because that’s the role of the publisher.

Well, it is in the sense that the newsroom has a budget, and we support quality journalism. Frankly, we are so successful at this moment as a company that all of our investments are in enhancing the quality of our journalism, data, analytics, etc. Getting better news, getting better information is the mission, and we are investing in that.

So, I just want to correct one simplification that sometimes comes to the surface at moments like this. I don’t think you intended to do that, but I think it’s important to make a distinction. It’s super hard when you go through what happened yesterday and many other times in journalism. But this shift, like any other shift that Emma has gone through, is not to eke out more profit by having fewer resources. There is an overall climate inside our company. You see our earnings, revenue, and subscription base growing.

We will invest wherever there’s a good business case to be made. Tech and the cross section of tech with policy, politics, global trade, and society is one of the top priorities for Dow Jones, and one of the top stories in the world. So, don’t take a snapshot and say, “Okay, we’re going to stop there.” This is a top priority. Tech will permeate everything — it is permeating everything right now. Look at Washington, D.C. or at the announcements from President Emmanuel Macron in France. So, don’t take the snapshot, that is where I was headed. This is a moment in time. We can talk again in a year, and our tech coverage should be broader, deeper and probably have a larger following.

I run a tech publication —  what is nominally a tech publication — and we are heavily invested in covering policy. One of the lines we’ve always used, a cliche even, is that, “The Verge can cover everything because everything is now a tech story.”

That was a way 10 years ago of maintaining a broad focus, and now, it’s very real. Elon Musk is at the State of the Union. You can see the tech giants fighting tooth and nail against the Digital Services Act in the EU, and that is now a part of American foreign policy. Do you see that as, “Okay, maybe all of The Wall Street Journal is about tech in that way?” Is that getting more expansive for you?

There is a current of tech that runs through every story at The Wall Street Journal. It runs through everything at Dow Jones in two ways: as a story and as technology. There’s not a part of The Wall Street Journal or Dow Jones where tech does not feature. Having people cover tech in isolation is one way of covering this. In addition, tech is a core part of many other beats and areas that we cover. Tech is a horizontal and a vertical at the same time.

We’ll come back to the AI deals you’ve struck. I want to talk about them expansively. Just in this context, like so many publishers, you’ve struck a deal with OpenAI and other AI companies. Are those providing enough revenue for you to invest against in the newsroom, or are you still in wait-and-see mode with those deals?

I don’t think we should tether our investments in AI to any individual AI deal. That’s not how I look at it. It’s not like, “Oh, AI brings in this much money, and now I can invest this much in AI.” You could, I guess, rationalize it that way.

Or the newsroom.

AI in the newsroom or in the newsroom [itself]. We have our investment priorities, and we’re following a game plan that we’ve been following for a while. It meanders every once in a while, but the goal is pretty clear. We intend on growing in three ways. One is by going deeper, so investing in the depth of our content, our data, and our analytics by growing wider. 

We do this  by adding new areas of expertise. Just last week, we announced our agreement to acquire Oxford Analytica. Going deep into geopolitics is an addition. But in future investments, once it’s a part of Dow Jones and The Wall Street Journal, we will probably invest in going deeper into that area of geopolitics. 

Then, the third wave is by connecting everything that we have. There should be easier access to the data that underlies everything at Dow Jones and everything that we do at The Wall Street Journal.

Now, to your question, investing in the newsroom is a goal in itself. We are a successful, leading, subscription-driven news organization, and we grow by investing in our journalism, not by shrinking our journalism. Sometimes there’s a temptation to be in austerity mode and to just take away in order to eke out a profit. That’s not us, and that’s not how we’re growing whatsoever.

I think it’s more than a temptation for most media businesses right now. That is the reality of the situation, right? It costs more to make the information than most people can return on it.

Yeah, but if your answer to that time and time again is, “Okay, I’ll cut in order to make ends meet,” then you’re not addressing something at the base. You’re not addressing something in your model. You might not be addressing something correctly in the way you’re organized or where you’re focused. That, to me, was never an acceptable method to grow or create great journalism. I understand that sometimes you have your back against the wall as a company in any industry. You may have to cut in order to make ends meet, but that’s not a strategy.

People get that wrong. I think you’re absolutely right. That might be a prevailing tendency, but I don’t think it should be. The prevailing focus should be about making a difference in the news and information that you offer. How do you make that distinctive? How do you add value? How do you allow people to make decisions based on that? How do you convince people that they should recognize the value of the information that you offer? 

People in the past used to say, “You’re a subscription business. You’re The Wall Street Journal. You’re all about business, and therefore, that doesn’t apply to anything else.” I don’t think that is true. I think people recognize the value of a lot of different types of information. It doesn’t just have to be about business. So, I actually think there’s a lot of opportunity in shifting from this austerity mode to creation and building mode. Easier said than done, and sometimes, you have to step away from things that just aren’t working.

I think that recognition of value is very challenging. I understand why it happens in the business community. I even understand why it happens for us in the tech press because it’s often tradable. You can pay a high rate to The Wall Street Journal if you are a Wall Street trader, an investor, or some other kind of business professional because the information has a clear value that you can use to trade upon to make a deal or an investment, to buy or sell a stock.

I think that information is not tradeable for the average consumer. They just open TikTok. Maybe there’s some influencer reading The Wall Street Journal to them for free. That elimination of scarcity, I think, has been the fundamental challenge. It was the challenge when we went to social media, the challenge when we went to these social video platforms. It feels like the challenge again for AI, right? The AI platforms are going to take all of the world’s information and completely eliminate even the scarcity of having to click. They’re just going to tell you what the models have read on the internet. Do you perceive that as existential a challenge as some of your peers in the media do?

I think you’re absolutely right in that the bar on being distinctive with the content, news, or information that you create has gone way up. So, you have to be more discerning where you focus. You ask about the existential threat around AI, and that goes back to the recognition of value. I want to make sure that the industry — certainly Dow Jones, The Wall Street Journal, and all of our publications — don’t fall into a trap that is similar to two decades ago when all information had to be free, people took scraps from search engines,. and then found out over time that they’ve effectively lost, that we seeded that market. So, that’s why we are investing time and resources right now into making sure that large players in the AI space recognize that value.

Our push is to make sure that there is a commercial agreement around that. And I think that is the preferred outcome in many cases on both sides of that fence. Where we can’t reach a commercial agreement, where there are fundamental differences of opinion, we are prepared, in some cases, to say, “Then we’ll fight it out in court.” So, we’ve walked both paths, with a preference for the first. 

I’m still answering your question as to what’s existential here and is there an existential threat. You spun it forward and we can get through that, what it means for the end user and how consumers respond to that. But I think first, we have to get the starting blocks. These are the companies that are providing generative AI, UX, and new interactions for consumers. They have gotten to that point by using information. We need an acknowledgement that the information that has value, certainly our information, and that if you want to use that information on an ongoing basis, that some of your generative AI-produced content answers to queries are current and reliable. You’ll have to pay us for access if you value that. That part we cannot skip over. There’s a whole other part of this where we don’t yet exactly know how the user is going to interact, but we see the trending there. We have to get that part right. We’re in the middle of that, or maybe we’re at the first part of that still.

I’m going to ask you one more very existential philosophical question, and I need to get back to having you explain the company — 

Yeah. All my favorite topics.

There’s that line, that a lie goes around the world six times when the truth is putting on shoes. It feels like we might be describing a world where regular people are awash in a sea of free lies that come to them on social media. The social media companies are all giving up on fact checks. Various billionaires say whatever they want on podcasts with no pushback. And what you are providing is a very expensive source of truth, or hopefully what I am providing is an affordable source of truth. That’s a big discrepancy, right?

I would say we’re as affordable as a cup of coffee a day. Everyone drinks coffee. I think it’s a myth that access to reliable information is unaffordable. It is an individual choice. I realize it’s a hard choice to make if you don’t have significant disposable income, and that’s what you’re indicating, but I do believe that there’s a choice to be made.

I understand why people would buy The Wall Street Journal or some of Dow Jones’ products, which we should come to. But that’s the bigger picture, right? It’s convincing the next consumer that they should pay for information as opposed to picking a filter bubble on social media. Whether that’s AI, social media, or if it’s just algorithms, that seems like the challenge the media faces.

The challenge is to convince someone that it’s to their benefit to invest in having access to reliable information. Whether that’s for making decisions in the realm of investments, technology, policy, or whether it’s hyperlocal and I actually want to understand what’s going on in my community, some of that I might get from AI. But some of that I might not. I might want to invest a small amount of money into understanding what’s happening in my community. There’s some examples of that popping up.

In response to this huge question that you’re asking, I think we’ll see innovation in journalism and a lot of creativity — already and in years to come — where, undeniably, there’s this flood of information of mixed quality.There’s also, undeniably, a huge demand for reliable information. People are craving it more than ever before. In fact, the more noise there is, the more people are confused and the more they are asking, “Hey, tell me what this means.” 

We see this in our data. We see this when there are moments of friction in society, in business, or in geopolitics with any market. We see a spike and people coming to us for free, but we also see a spike in subscriptions. Demand for reliable information, I think, has gone up as the pool of unreliable or uncertain information has grown. Some of it might be reliable, some of it might not be.

I think you cast that as an existential risk. I can cast that as an opportunity. I like to be aware of the existential risk and take the precautions there, but mainly focus on the opportunity and meeting that demand. I don’t think we have met that demand by any stretch. I think there’s a huge opportunity for us and for other publications in lying ahead to meet that demand. I think that demand will actually only grow.

This is a good place to back up a little. We’ve talked a lot about The Wall Street Journal. I think people know about The Wall Street Journal. I think Decoder listeners also probably know about Rupert Murdoch and News Corp, which is the parent company of The Wall Street Journal. Describe how all of that fits together in your role as CEO of Dow Jones.

In its most simplified form, I think of Dow Jones as a Rubik’s Cube, and inside is all of our premium journalism, exclusives, or explanations of what’s happening right now. There’s our proprietary data that we have on many different sectors in the global economy. There’s Factiva with thousands and thousands of sources from around the world sitting inside that Rubik’s Cube. 

Each tile on that Rubik’s Cube is a way to get out of Dow Jones what is important and relevant to you. Maybe I want a couple of different tiles or maybe I want the whole thing, but that’s fundamentally a way of thinking about how we operate and how, ultimately, we’re organized. AI is actually helping a great deal with this and it’s accelerating with generative AI. AI and automation has been with Dow Jones for a long, long time.

Let’s then make that a little bit more complex. If one of those tiles is about bond trading, I ought to be able to get all information: premium information and live information, but also analytics and forecasting to help me in my job. We’ll get to AI in a moment, but the cards I was dealt when I took over almost five years ago as CEO was to look at Dow Jones as a platform of verticals. 

You’ve got business news, which is both a horizontal and a vertical, but that’s The Wall Street Journal, as you say. It needs no huge explanation there. There’s wealth and investing, where we have Barron’s, MarketWatch, Financial News, which is a title in the U.K., and Private Equity News. We’ve got a compliance arm that has data on compliance. It helps companies discern, “Should I do business with this person or not? Are they on a sanctions list or not?” We’ve since added an energy arm and within that, commodities and petrochemicals. We’ve added a leadership arm, organically, looking at what it takes to be a modern leader. 

Each of these verticals, if you will, is successful when it does four things. You have to have leading news in that particular vertical, in that particular area of concentration. So pick your industry. Unless you have that, you’re not relevant. Second, you have to have proprietary data. Some of that can come out of the news, some of it you have to build, some of it you have to buy, and we’ve done that. If you have those two things, you can do proper analytics. AI helps with that to some degree. If you sell products that can do forecasting and analytics, that’s where the value keeps going up. Then, the fourth factor is convening power — that is bringing people in that industry, in that sector, in that vertical together.

If you have those four, you actually get a mini-network effect inside that industry. We’ve seen that happen, and it happened by design. We’ve executed this with our manager, Joel Lange, who runs our risk and compliance business. If you have all those four parts, then you see the leaders in that industry come together. Take risk and compliance. You see compliance officers coming together and our [Chief Compliance Officer Council] at Davos and in other places. So, now you have thought leaders there.

Well, the procurement officers who buy our data products see, “Oh, yeah, the people that are leading my department or leading my company are talking on a Dow Jones platform about these big themes.” So it anchors Dow Jones more deeply, whether that’s just an observation and anecdotal or whether that actually translates into business. Often it translates into business. That has the same effect on our analytical products.

Then if you add news to that risk vertical — we have mentioned internally that we will have a risk journal —  you will have risk industry folks, compliance officers and the like, tap into that news product to actually start their workday and understand what’s happening. Now, we are present in your workflow in that industry end to end. By the way, since we’re The Wall Street Journal, we’re probably also still with you in another way when you go home.

So that’s the view of the company and our operating model. Build verticals that have these four parts at a minimum. Go deeper and make that exclusive. This is why what Emma is doing in the The Wall Street Journal newsroom really matters. More exclusives and more distinctive journalism helps with the news obviously. 

We’ve built, bought, and created more proprietary data. We’ve spent well in excess of $1 billion on getting companies that are specialized in that added to our roster. We’re doing more analytics and even a little bit of consulting. We’re never going to be a consulting company, but that’s an outflow of analytics. Then,there’s convening power. I think for a long time, the media sort of misread that as just events, but it’s something far bigger than that. I think it’s a subscription business.It’s a recurring revenue business, it should be. It doesn’t mean that you can’t have sponsorship for it, but it is fundamental to anchoring the decision-makers of a certain industry in your platform.

When we have all four parts, we call it a full stack. So we’ve got a full stack in risk. We’re building a full stack in energy and subsets of energy, so we’re going deeper there. And it’s a scalable model for the future because we now understand how to build these verticals. Either we can mine The Wall Street Journal for new verticals because we see what people gravitate to and where we have the expertise, and then build on that. We can inorganically add as well, or we can organically start outside The Wall Street Journal newsroom.

So going deeper, going wider, and then connecting everything. Let’s go back to the Rubik’s Cube, if you want to buy that whole Rubik’s Cube, we will also make that possible for you. So we’re making sure that these magnificent data pools are going to be available. Some of this is, of course, still in the works but going to be available to our journalists so that they can do exclusive work with it.

You’re describing Dow Jones as something that makes really high-quality, rigorous information across newspapers, magazines, and data products and that sits within News Corp. How often do you hang out with [Founder] Rupert Murdoch?

I wouldn’t put it in the category of hanging out whatsoever. There’s healthy, friendly interaction. He’s chairman emeritus, he’s built the caricature on the side. He’s built enormous media success stories over time. And so, from a business point of view, there was a lot to learn from him, certainly in my early days, like, “How do you create businesses?” But from News Corp, there’s been nothing but support for our growth story.

So I’m very thankful for that. It’s Rupert, it’s [Chair] Lachlan [Murdoch],  and also [CEO] Robert Thompson, that’s a whole apparatus. But we also have access to capital as Dow Jones. Five years ago, we hadn’t done an acquisition for well over a decade. Now we’ve done billions in acquisitions. That support is a vote of confidence for the direction that we’re taking, for the strategy that I outlined to you: growing deeper, growing wider, connecting things. There’s also a deep respect for the independence of The Wall Street Journal and the value that comes with that. I’ve seen that consistently applied. So yeah, that’s my answer to your hangout question.

Rupert Murdoch plays on both sides of the information crisis. You can watch his other properties create whatever reality is politically expedient for Donald Trump. And then I see The Wall Street Journal rigorously cover the impact of tariffs all the way down to the opinion pages, which sometimes say the tariffs are bad.

Oh no, the opinion pages don’t just say tariffs are bad. Prime Minister of Canada Justin Trudeau said in a press conference — I’m paraphrasing here — “I don’t often quote The Wall Street Journal, but they said that this trade war is the stupidest trade war in history.” Something along those lines. We don’t hold back. Our opinion pages don’t hold back, and their assessment is based on well-established principles of free markets and free people. The independence to make that judgment is core to who we are. For my part, I’m focused on the Dow Jones part, of making sure that sings.

But do you see that contradiction? Do you think your team sees that contradiction, that you’re trying to sell really high-quality information while another part of the structure that has the same ownership is contributing to an information crisis?

Listen, I can’t comment because of the shared ownership structure. I’m not going to comment on my colleagues at Fox. We had protesters outside in Fox Square, and people are protesting while the reporters at The Wall Street Journal and Dow Jones are doing their jobs. So, there is an awareness of the perception of Fox and what we’re focused on. I think there’s also a great awareness amongst our staff, and has been for over 15 years, that those two things are separate — without specifically commenting on how you characterize that, because I am just not going to get into that.

That makes sense. I want to ask the last Decoder question, and I want to end by talking about AI at length. You’ve had to make a lot of decisions. You’ve changed the way the company works a lot. You obviously have a way of thinking about the company that’s very specific. Although I will say that Rubik’s Cubes are meant to be solved. I think you want them to remix the cube, not solve it. But what’s your framework there? How do you make decisions?

I take in information a lot. I like to be, probably to a fault, familiar with the facts on the ground, so I am very consultative, take in expertise, and absorb. So, I feel like I have a sufficient level of mastery to then make a decision. So that’s one part. I’m very inquisitive. The strength of good journalists is that they know how to ask questions, and they’re driven by curiosity. As a company, we’re driven by curiosity. As an executive, I’m driven by curiosity. I want to figure out how things interact, to understand the nuances. That’s for my part.

I also, at the same time, believe in letting managers manage or creators create, and making sure that what I’m doing is to help make those managers or those creators successful. It depends on where we’re focused. If we’re focused on something that is important for the whole company, I will be informed and I will make the decision by consulting my management team.

Within the framework of the strategy that I’ve outlined and the intricacies of that, I am a firm believer in letting the managers make decisions. So, I have a very flat structure where there is a lot of autonomy to, within that framework, make decisions. I think that allows people to move faster and allows the company to move faster. It allows us to experiment without going through a central clearinghouse constantly. So, I want both. On the one hand, I want the expertise so I’m informed. There’s a limit to that because you can’t do that when you are thinking about everything’s macro issues. I have to be judicious in how I do that, where I focus that thirst. On the other hand, we have that flat structure and empower people.

Let’s apply that framework to AI. I think it’s useful to have the framework where we talk about this next set of big shifts because that’s a lot of decisions. You mentioned Factiva earlier. That’s a data platform. You’re promising some generative AI tools there. You want to offer more of those tools to your customers across the board. You’ve made some deals with OpenAI, but at the same time, you’re suing Perplexity.

You think it’s taking information away without compensating you. You made a deal with OpenAI at some rate, we can talk about whether that rate is enough. And then you’re offering the tools to the users. What is the shape of the AI opportunity to you? When you look at that whole set, the number one question to me is, “How big of a business is this really?” Because I don’t know if anyone’s making more money from AI than they’re spending on building the tools right now.

We are in a very, very early stage in that, so it’s hard to say. What I can say is that on the product side, some of our products outperform versus what we had planned for them. We have a product in our risk business called Integrity Check. It’s more of a self-serve model where you don’t have to wait for Dow Jones to get back to you and do its computation — using AI ourselves but out of the view of the customer. Instead, we’re having the customer do some basic research on risk and compliance themselves, assessing who they can do business with, and getting that to an 80 percent reliability. From there, they use that as a springboard to say, “Now, with the remaining 20 percent, I need help from the company.” 

That product is fairly young, and that’s outperforming expectations. On the whole, I think this will be a net positive, but we’ve got to unpack what we’re talking about first. We’ve got to get the foundational elements of this right. If we don’t have proprietary information that is truly proprietary, then we’re going to lose this game. You see us engage in building products, building a marketplace with Factiva, and deploying tools internally, but all of that has to happen in tandem with solving that foundational question.

Last summer, the deal with OpenAI, the reporting was that it’s worth about $250 million. Is that correct?

Read The Wall Street Journal.

That’s The Wall Street Journal’s side, so I believe them. I had Nicholas Thompson, the CEO of The Atlantic on the show a few months ago. He told me one of the reasons that he made a deal with OpenAI was to set the market rate, which is useful for fair-use litigation and for other kinds of deals. Do you think $250 million is enough of a rate to set the market?

I’m not going to talk about specific amounts, but you’ve got to ask what the amount is for. What I am personally less interested in is a single amount and more in an operating model and a business model for how you do business over a long period of time. How does that operate? When information is used, is the value of that information recognized along the way, and is there a mechanism that helps in realizing that valuation?

Set the dollars aside, inside of the OpenAI deal, what are the signals you’re looking for that indicate whether the deal was a success or a failure?

Simply because of the way we’ve set up that deal. I’m not going to talk specifically about it. That was a super nice try because I almost bit. 

I’m going to try it again, don’t worry.

I’m sure you will. I can talk more broadly. How can you tell the generative AI tools that you’re deploying or the models that you’re deploying within your own business are successful? That’s by usage. Is it generating revenue on a consistent basis? Is it just a blurb? Like, ” It’s a novelty factor, and now we move on?” We’re pretty early in that process, so I don’t know yet what’s a head fake and what’s real. Some of the products, I can tell, are real. Some of it requires a shift in user experience and user requirements. This is going to have to be table stakes, like offering a UX that is built around generative AI because the customer expects that. So, you’re getting at the trickiness of establishing the value writ large.

But overall, I think of generative AI as an accelerant for the strategy that we have. It will allow us to go deeper in our verticals faster and more efficiently, and in ways that we couldn’t even imagine. We’ve all talked about the wonders of generative AI, in doing research in ways that we couldn’t do before, human and otherwise. Within Dow Jones, we talk about authentic intelligence. That’s the combination of generative AI and human guidance. We find that that’s a sweet spot for certain B2B products that we’re building.

It’s accelerating, going deeper. It will accelerate going wider, i.e., scaling our vertical strategy because we can stand up verticals much faster. If it’s a geographic vertical, we can now say, “All right, we can launch in this language, and it’s so reliable and it’s a lot cheaper.” And then connecting everything, generative AI is a massive accelerant because now with a thin layer on top, we can extract data from all these different data pools.

Do the tools work well enough for you to trust it?

On a case-by-case basis, when it’s very specific and we are answering a question from a customer. If it’s a co-creation where we are solving a certain problem and we have very narrow parameters, then I think it works. When you go wide, you get a wide answer. So, our strategy is built around being specific and being focused on verticals. AI fits nicely with that, and in fact, it allows us to go much deeper, be much more specific, and be more discerning. Under each vertical, you can create sub-verticals using a much larger data pool.

You’re describing something that happens within Dow Jones and its products. More broadly, News Corp has been pretty harsh about platforms and work. News Corp CEO Robert Thompson is famously critical of Google because the company was behind the laws in Australia that require platforms to pay publishers for linking. AI represents that opportunity as well, or that challenge as well. Instead of using your tools, someone might use a ChatGPT or a Google Gemini, and just receive an answer. Do you think that these deals you’re making are hedges against that outcome? Are they investments in that outcome? 

I’ll just give the example of the millennial digital media startup boom, which was predicated on “we will just be the most viral thing on Facebook, and Facebook will pay us that money.” That obviously did not pan out. I think people are very wary of making that same mistake with AI. But you have one of these deals.

This is why I said at the very start that I see those deals in a separate category. It’s foundational, it’s about principles. The money to be made in AI, it’s really on us to make sure–

So it’s on your proprietary tools?

It’s on us that what our generative AI spits out is relevant to our customer in a way that some other provider with maybe a more general offer is not. We have to make sure that when we combine our proprietary journalism, our proprietary data, and our convening power with generative AI and LLMs, that the outcome to a query is A, reliable, and B, something that you can’t find somewhere else or be at the scale of where you can find it somewhere else within that vertical. I think there’s a distinction between establishing the principles and getting value for that, getting forward value if you are being used. Then, there is a separate category of efficiency tools that we use in the company.

Yet, there’s another category where we say, “Here’s where we build products that have to answer a certain question that exists in the market in any industry.” We’re going to give a superior answer, and you’re going to need that answer in order to be more successful than the next person working on solving that problem in a certain industry. So, if that’s about forecasting energy prices, we want to be the most reliable on that. We want to be the leading voice on what’s happening in the chemical industry, and generative AI should be one way in which you get that out of us, but in a proprietary sense. 

That should be, hopefully, very different from going to any chatbot and asking that same question. Maybe you get an approximation, but it might not be as reliable. Hopefully, there’ll be sufficient proprietary data in our answer that will make that competition uneven to our advantage. That I think is the task. I feel very strongly that we cannot go into this new era with a view of, “Well, this is what these companies have to do for us.” We have to agree on the principles and the value, but then it’s really up to us to create superb products and answers to complex questions in a very complex world, to realize the value that these new tools offer. Both those things have to exist.

I’ve talked to a lot of publishers and media CEOs over the past several years about where the traffic comes from, how the payments work, where the value is going. Setting aside AI for a minute, it feels like the nuclear question everyone is asking is if Google is just indexing our sites and taking the data, eventually we will have to block Google in a way that many publishers were comfortable using their robots.txt file to block OpenAI and other crawlers. Have you ever considered going that far?

I’m not going to speak specifically to Google. We’re partners and we have lots of things that we do together. There’s also things that we disagree on.

News Corp has famously been the most outspoken on this. That’s why I was comfortable asking the question.

Absolutely. This is not on my radar in the way that you express that, but that’s the short answer to that. Taking your question in a different way, we have to emphasize owned-and-operated. We have to make sure that being in our world — in our universe, in an individual vertical, in one of our broader products, or in the entire Rubik’s Cube — you have an experience that you cannot have somewhere else. That’s on us.  How far do you go? It’s about putting a wall around that. Yeah, we’ll see over time.

You are in litigation against Perplexity. It’s taken some data. I’m guessing by the fact that the lawsuit was filed, you didn’t like that. If you win that case, The New York Times Company wins its case against OpenAI, or, I don’t know, Sheryl Crow wins her case, that will upend the market as we understand it. There will be some new fair-use precedent that is created. How does that change how you think about building and deploying your own AI tools?

Again, I put this in a separate box. We’ve got to build our–

Well wait, let me challenge you on that for one second, just to get it into the right framework. Right now, it feels like the entire industry is just assuming that, win or lose in these cases, the money will be sorted out and labeled to build at the same rate we’ve been building.  Maybe the rates go up, and it’s just more expensive to do what OpenAI is doing because it has to pay all the singer-songwriters in the world. It also seems to me that potentially, the rates are so high that the entire structure of the industry changes.

The industry, in this case, is AI.

The AI industry changes. Suddenly, the compliance cost of making sure all of our data is licensed before we feed it into the model for training skyrockets because the penalties are high under copyright law. That feels like an under-considered risk. These lawsuits are just going to play out and something will happen.

The way that I think Google was able to roll over the Viacom lawsuit when YouTube started or the Google Books lawsuit was because it was sort of the plucky upstart. The value of those tools was so high that it got to win a bunch of lawsuits. I don’t think the AI companies feel like plucky upstarts. I don’t think that public sentiment is with a bunch of giant tech company billionaires anymore. It feels like those lawsuits might go the other way. At that point, some of the tools you are using to build with or some of the partners you have, their cost structures might change so dramatically–

That that is going to stop us from that development.

That everyone’s strategy has to change. I’m just wondering how much you’re considering that.

I see where you’re going with it. It’s a little bit of the blind man and the elephant. There are different patches of fair use that different legal cases are pursuing. One case might reverberate, but it’s not going to necessarily be absolute. I hate to say this as an answer to any question, but it’s a big wait and see. 

At the moment, I have to go on the assumption that as a technology, generative AI is present in my world, is going to be present even more, and is going to be an expectation from consumers, whether they’re corporate or consumers out in the wild. So, we cannot continue to build and then think at the same time about how it may all just disappear. And by the way, we might be the culprit because we’re applying that, which is… it’s an interesting scenario. I don’t think it will play out that way. 

But you’re one of the litigants. That’s what I mean. It’s interesting. 

But I don’t know that it will be debilitating. I don’t think that the commercial agreement we have with OpenAI — with the value of which I can’t say but you cited — has stopped it from developing. I believe in a market mechanism, and I think that’s where we will end up, that there will be a gravitation to that rather than stopping the industry in its tracks.

OpenAI famously has not made $1 in profit. It has to build a business that’s valuable enough to support these deals. 

Yeah, but Amazon didn’t for a long time either.

I feel like we’ve brought up Jeff Bezos in a variety of ways on this episode. [Laughs] I’m very curious to see how your lawsuit plays out with Perplexity and how those businesses develop. We’ll have to have you back as that progresses. There’s something there that feels almost invisible. You’re right, the blind man and the elephant. It’s there, it’s very big, and I think this next year, we’ll see how it shapes the business.

I want to end by talking about press freedom. It’s something you care about a lot, you’ve talked about it a lot. You’re a publisher of The Wall Street Journal. You famously had Evan Gershkovich detained in Russia in March 2023. You worked very hard across a number of administrations to bring him back. This is a very challenging time for press freedom, both abroad and it feels like in the United States. What’s your view of the landscape right now?

We live in a time of tremendous change, of polarization, and that makes covering the news trickier than ever before. It also, I think, increases the value and the contribution that we bring to society as a free press. With all the changes and challenges that we’re seeing, including against the media, this is a time that any journalist should be made for. If your heart is in explaining complexity to the world, then there’s never been a time when we’ve had this to grapple with.

On the one hand, we can offer enormous value, and on the other hand, it’s become a lot harder to do that. The statistics around the world don’t lie. There are well over 300 people who were killed last year doing journalism, and people being put in prison. There is a harsh dialogue in society that makes it, under many circumstances, less comfortable to go after a story. Sometimes I measure whether we did a story well by how much I got in terms of complaints from the left and the right after certain stories. So, the temperature is high.

Let me push you on that too. That’s an old chestnut in journalism, right? If everyone’s unhappy, you’re doing your job right.

It’s also my daily existence, honestly. So, it’s a very young chestnut for me. It’s there every day, but yes, I know where you want to go.

This is a pretty asymmetric information landscape right now. One side is vastly more willing to lie. One side is vastly more willing to even change the data. The Trump administration is making noise that it’ll take government spending out of GDP, which would dramatically change almost everything The Wall Street Journal does, right? 

At the most fundamental level, we might not be able to trust government data anymore. That’s a threat to press freedom, and it will spin it as a good thing. Elon Musk is out there trying to spin this as a good thing. You don’t see the left playing that kind of game with the data to metaphysically create political outcomes. There’s not as much trying to tweet things into reality that Elon is doing.

I’m not going to left-right things in this conversation. What I can say is how you respond to an information ecosystem, or maybe in an asymmetrical manner —

I’m saying, how do you respond to an information ecosystem where Donald Trump has threatened to sue pollsters in Iowa because he didn’t like the results of their poll? Or where Brendan Carr, the Chairman of the FCC, is potentially holding up the CBS Skydance merger over his investigation of 60 Minutes’ editorial content?

There’s a very clear answer to that. It’s not an easy answer, but the first answer is stick to your principles. In our case, we believe in reporting the facts in the newsroom. We believe in free markets and free people on the opinion side. You stick to that and you do not let go. And you double down on that. That’s our contribution to the information ecosystem, and we’re going to do more of that. 

That’s a demand-driven thing as well, but we’re talking about press freedom. This is an answer to that. If you start making concessions in your reporting and omitting facts that you know to be true or start self-censoring, then that game is lost.

Second, you’ve got to keep a cool head. We live in an environment where taunting and provocation is the norm. You can take that bait or you cannot. Then, you have to then, in my view, not be hysterical in response to every little provocation that might exist. In fact, you might get more respect if you do not respond to every provocation.

Then, you have to recognize the moments when principles are at stake, when you have to fight or express your disagreement. So, keep on doing what you’re doing. Do even more of it, in our case. Create reliable information. It’s going to be good for society, it’s going to be good for you as an organization. Keep a cool head and stick to your principles.

That last part is also non-negotiable. All these three, in fact, are non-negotiable. You’ve got to stick to your principles. If you start shifting and making certain concessions at the wrong moment, there will be a very high price to pay for that.

You have colleagues in similar positions across the media that are making concessions. ABC settled its case with the Trump administration. Looks like CBS might settle the 60 Minutes case because the threat of the Skydance deal being blocked hangs over it in some way. Are you saying you would not make those concessions? Are you saying they should not?

I am not saying either one of those because I’m not commenting on their individual situations. I just think that there are moments where, as an organization, you’re going to have to evaluate. The AP just went through this. Is this a moment where I speak out and where I stick to my guns or not? I think in those moments, you better choose carefully. You better have a clear view of your principles, understand what you actually stand for, and understand the ramifications. People will point back to certain moments, and you want to make sure that you were on the right side.

I’m going to ask you this more directly because I think I just need to hear it therapeutically, but I think your reporters probably need to hear it too. If the pressure comes to you from the Trump administration, are you saying that you’ll fight when it feels like a lot of other big media companies are choosing to cave?

The question is just to load it in and make it very specific. We have fought for our principles for decades. We have stood up for our reporting for decades. We have a legal team that is incredibly strong, that has fought for press freedom and for our journalism again and again. If we make a mistake, we correct it. We own up to that. That is absolutely part of the value structure. We stand up for principles, period.

What happens if the generative AI makes a mistake?

It depends on what mistake it is. Actually, in building and co-creating some of these products that answer very narrow questions, we were sometimes surprised at mistakes that snuck in. We want to make sure we don’t release products without having screened for that. But you’re going to have to correct.

Do you think that an information environment where some of the tools and the institutions are less reliable, or perhaps even openly hostile to the press, is something you will be able to chart as Dow Jones alone? Or do you think that’s an industry-wide effort? Because it does not feel like there’s a lot of coordination across the industry right now.

I think the industry should shoulder a lot of this together in a loosely formed coalition or through solidarity. I’m focused on Dow Jones’ success, but I certainly share our findings with colleagues all the time, and there’s a very active dialogue amongst media leaders that A, I want to foster and B, I want to participate in. If you look at very difficult moments like the Evan case or getting people out of Afghanistan during the U.S. forces’ rapid withdrawal, we worked very, very closely together. There’s very close contacts among a lot of those leaders. I think that’s a healthy thing, and I would like to see more of it.

Almar, you’ve given us so much time. Tell us what’s next for Dow Jones.

Well, it’s tomorrow’s news, so I want you to tune in. Definitely come to The Wall Street Journal every day. For us, you’ll see us focus on international rebalancing. We’re rebalancing our portfolio to make sure that we are as strong outside the borders of the U.S. as we are here. We’re focused on video, focused on deeper data products. Overall, we will continue to be focused on what we have been focused on for the entirety of our existence, and that’s providing reliable information.

Amazing. Thank you so much for being on Decoder.

Thank you so much for having me.

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

I was on a flight back to New York the other day, and something really odd happened when we landed: the passengers burst into applause. I’ve only ever seen that happen after a flight with really bad turbulence or a big delay. But this was a totally boring flight that took off and landed on time with no drama whatsoever. So why the clapping?

Well, you know why: there have been a lot of plane crashes, or near crashes lately — and it’s all against the backdrop of the Trump administration and Elon Musk firing Federal Aviation Administration (FAA) employees and talking about upgrading everything with Starlink, or whatever they’re posting about on social media today.

Just this week, news reports have detailed exactly how Musk has wedged his way into the agency to force SpaceX into the conversation — including a move to take over an existing Verizon contract and even threatening FAA employees with termination if they don’t get fully onboard.

So the reason it feels less safe to fly — the reason people are clapping when the plane lands — isn’t just the tragic accidents. It’s that the system we took for granted to keep us safe and solve problems when they occur is being destabilized right in front of our eyes, and actually improving that system takes more than posts and bravado or conflicts of interest so intense it causes a constitutional crisis. It’s actually a complicated dance of people, technology, and policy — you know, Decoder stuff.

So today I’m talking to Andy Hawkins, The Verge’s transportation editor, about what’s going on in the skies. He just edited a big piece for us by writer Darryl Campbell that helps put a lot of what’s happening in air travel right now into perspective. There’s some very reassuring data about how safe it is to actually fly, but there are also some big questions about what we need to do next to regain our confidence in air travel. 

He and I talked about how safe it really is to fly right now — extremely safe — and how the current air traffic systems might change for the better and the worse. And, of course, we talked about Musk. 

If you’d like to read more on what we talked about in this episode, check out the links below:

• What’s the deal with all these airplane crashes? | The Verge
• How Elon Musk muscled his way into the FAA | Bloomberg
• Elon Musk says upgrade of FAA’s air traffic control system is failing | CNN
• FAA targeting Verizon contract in favor of Musk’s Starlink, sources say | Washington Post
• FAA officials ordered staff to find funding for Elon Musk’s Starlink | Rolling Stone
• FAA announces ‘hiring supercharge’ for air traffic controllers | Forbes
• Air traffic control trainees to get raise, in nod to cost of living | NYT
• Some of the 400 jobs that were cut at the FAA helped support air safety | AP
• DC plane crash marks first major commercial crash in US since 2009 | ABC
• What the ATC controller sees | Flight Training Central

Today, I’m talking with Panos Panay, who’s in charge of devices and services at Amazon. That includes everything like Alexa, Ring security cameras, Eero Wi-Fi routers, and the Project Kuiper satellite internet service that’s meant to compete with Starlink. 

Panos and I talked the day after he announced Alexa Plus, the new AI-powered version of Amazon’s famous voice assistant, and this episode gets pretty deep into the weeds of how all this works and how Panay thinks about running his teams to make it happen. 

This is actually another one of those full circle Decoder episodes — I talked to Panay’s predecessor, Dave Limp, on the show in 2021.  If you’re following executive shuffles, you know that Limp left Amazon to go work for Jeff Bezos as CEO of Blue Origin in 2023. Panay was hired as his replacement from Microsoft, where he was running Surface and Windows. It’s safe to say that the two have very different approaches to running this team and its products, so I was excited to dig into what changes Panay had made in order to make the new Alexa Plus happen.

Now, I’ve known Panay for a long time — if you’re a tech fan, you know that he was the Microsoft exec who really brought the Windows hardware market back to life by introducing the Surface line of tablets and laptops, and he eventually ended up overseeing Windows itself. You’ll hear Panay say that the idea of infusing Alexa with AI really drew him to Amazon — like so many folks in tech, he sees AI as a platform shift that will change the way we use computers, and Amazon has a big advantage with the enormous number of Alexa devices that are already being used globally. Just making them a bit smarter and more capable with AI sounds easy, but actually doing it is fairly hard, and we sat in the weeds of the execution for a while. 

There’s a lot here, and a lot of different parts of Amazon that needed to work together in new ways — that’s pure Decoder bait, and Panay was game to really get into it. It even got a little emotional there.

One note before we start: Panay talks about “experts” a lot, and in this context he means the individual services that power different parts of the Alexa Plus experience, kind of like apps on a smartphone. You’ll hear what I mean, but if it gets confusing, just think “app” and it’ll click into place.

Okay: Panos Panay, head of products and services at Amazon. Here we go.

This interview has been lightly edited for length and clarity.

Panos Panay, you told me that you don’t care about your title, but technically it’s SVP of devices and services at Amazon. Welcome to Decoder.

Good to see you, man. I love being here.

I’m really excited to talk to you. I was sitting in the audience yesterday as you were announcing Alexa Plus. I have a lot of questions about how it works, the feature set, where do you think it’s going. But it occurred to me, as I was sitting there watching you present it, and then later as I was watching some of the demos of it working, that to make it happen had to have required some big structure and culture rethinks inside of Amazon itself.

You joined about a year and a half ago. Decoder is all about structure and culture rethinks. So there’s a lot here. There’s a product to talk about, but then there’s the path of getting to that product. Is that how you see it? That you had to reset some parts of Amazon to get to Alexa Plus?

I don’t think resetting Amazon; Amazon’s incredibly ambitious in so many ways. Always learning, changing. I mean it’s pretty powerful. I think resetting the devices team a little bit, yeah. First off, we hadn’t really had a large-scale event, as I understand it — obviously, I wasn’t there — since pre-pandemic.

The events under your predecessor, Dave Limp, they were entertaining in a way. It was, here’s a firehose of stuff with Alexa in it. Microwave, a coffee maker. We would count, maybe, like, they announced 45 products.

Yesterday, you announced one new product, Alexa Plus and no new hardware, and that’s a pretty big difference.

I think that was important. So yeah, I guess that’s a change for sure from what it’s been. What we did yesterday as a team, it was a little bit of a reset. The team was pumped to do it, excited. We were never going to announce hardware. It wasn’t a goal. We need to reset Alexa for the world, and bring Alexa Plus forward. That is a bit of a cultural shift. We’re just going to focus on the service and what it’s going to be.

Great products are coming. We already have great products in market. We launched stuff at the holiday. And the team, they rallied. The company rallied. It’s pretty awesome. Having [Amazon CEO] Andy [Jassy] there is fantastic. And you can feel a vibe in that room for sure. I hope you did. I mean, you made your snarky comment about the music when you got in there. Man, we check every detail. I think I missed, I may have missed, I don’t know…

The chiptune rave music? It was pretty good. I always wonder who sets the playlist, ’cause you can do a lot with music in the pre-show.

Every single part of that show after the moment the mic starts has been very, very well thought through. Yesterday’s event was the highest risk event I’ve ever done. I mean, bar none.

I mean, I watched you reintroduce laptops at Microsoft in competition with your partners.

It doesn’t compare.

Really? Why so risky?

Because when you’re basically doing hardware, you have fallbacks. The demos aren’t, they’re not not live, but you always can just go to the hardware. When you’re reinventing or re-architecting an entire service, there’s no backup. It was the product. I think the only product video we had, like actual video, was the kids’ portion. Because, honestly, you’re not kids in the audience. So sharing a kids feature without some emotion is a waste of time. It’s like, here’s a kid feature, please write about it. So putting a little bit of emotion and storytelling in it.

Those were all real demos. That all really happened. That was one of the principles of the event. It wasn’t like, let’s go make up a fake story and we’ll just put film. That was the one area where it was just, it wasn’t a vision piece, it was the product, but it was the only area that wasn’t live. And so there was a lot of trepidation. This was the hardest kind of event we’ve put together, risk profile-wise.

Let’s talk about Alexa Plus for just one second and get a sense of it, and then I want to talk about how you made it happen.

So I think there’s a part that seems very obvious to people. You see an LLM, you see it interact with you. You’re like, this thing is great at natural language input and output, maybe it’s going to lead us to AGI and maybe it’s not, whatever, but the core piece of it is, the computer can talk to you in a non-deterministic way. Everyone saw that and said, okay, Siri should work like this. Alexa should work like this. Google Assistant should work like this. And then the actual implementation of it has taken everybody a really long time.

What’s the gap there?

It’s not just an LLM. I think that it seems easy. Put a voice to the LLM, let the LLM talk, or [text-to-speech], bring it out, bring out the voice. Or if it’s speech-to-speech, it doesn’t matter which tech, but if you want the elements of connecting to thousands of —I’m speaking for Alexa. You asked a broader question, but let me just talk about Alexa.

You want the element of connecting to thousands and thousands of APIs, partners that have been connected to Alexa forever. You’re trying to manage hundreds of millions of customers who already have the product. You want to update as many of those devices as you possibly can, meaning you don’t want to leave a customer behind. And there will be some devices that are eight, nine years old that won’t work. But everything else, most things will, relative to what’s used in the market today.

So you’ve got to carry forward all that history because people still love Alexa. We’re still growing. We still have usage that’s higher than you would expect, and we can’t leave those customers behind. That’s the worst thing. We focus on not doing that. So there’s that element. Sitting on top of an LLM, you’re now going, okay, just talking is just not that interesting. Although, awesome. Like having ambient conversation, I think it’s a superpower moving forward for Alexa. It’s different today on Alexa. It’s like point, shoot, ask the question. Hope to get the answer.

Yeah. You guys call it Alexa Speak.

Yeah. I do. Like with my team a year ago, we’d be in meetings and product meetings and we’d be talking and people would say, “Let me show you the new Alexa with a demo.” And they would Alexa Speak to it. And it was like, nope. Speak normally. Go to natural conversation. Don’t adjust your speech for Alexa. That’s exactly what you don’t want if you want natural conversation.

It’s hard, though. You’ve been training people, we’ve been training ourselves for 10 years. Calling a timer is, “Can you set a timer for eight minutes?” Calling a timer on the new Alexa is, “I’m making a ramen egg.” “Gotcha. I’ll set a timer for eight minutes,” where she just proactively comes back and sets it. I didn’t demo that yesterday because I didn’t want the timer headline, but it’s a really badass experience. It’s really cool. And so there’s a level of that transformation where — I’m off-topic, let me go back.

At the end of the day, the LLM needs to be able to, now it’s the base layer. Then you’ve got the next layer, which is just a series of different models. Picking the right model to do the job. And then that model is basically picking the right expert. And so the LLM plays a role, especially in the natural side of it, but as it makes it through the stack, it narrows down for accuracy. It narrows down for speed. It then narrows down for holding memory and personalizing it. And now you just have a series of experts basically sitting on top and one of them is conversational.

And so, that’s not just an LLM, that’s a series of… if you look at one of these other products, they’re not just LLMs, they’re basically, they’re mainly, I don’t know, overstating it, understating it but, so not to be rude, but they’re chatbots. And they’re pretty good. They’re damn good. And then when you start typing long form and rewriting and dropping in summaries, that’s very powerful. Creating videos, creating photos, isolated but powerful. But the idea that these experts all sit on top of the stack and basically kind of, there’s a runtime that orchestrates and says, okay, call these experts. These two experts have to work together. Got it. And then it operates. That’s just not simple. And the first thing I was asked when I got there was, I don’t know, I actually don’t…

It’s like 18? Something like that?

Yeah. I don’t know. It doesn’t matter. But it doesn’t feel like anything short. That’s for sure. Hey, why don’t you just change the brain with an LLM and everything will be fine?

Yeah, I think I probably asked that question the first time when we first spoke.

You might have. Yeah, I mean it’s the first question. And I’m like, well which one? And it won’t work. All you’ll do is talk, and it’ll be super verbose, and it’ll sound like you’re talking to the internet, and it’s just not that. It doesn’t work. And then everything else breaks. Which is the hardest thing. I don’t think anyone else is doing what we’re doing. We’ve got thousands of APIs now that we’re able to call. You’re able to get these, if you will, experts or agents, whatever you want to call them. It’s not a real word, it’s just being able to talk to each other at the right time. And then try, the invocation is like there’s something invoked and now the LLM at the bottom is arbitrating like, oh, what’s she trying to do? What’s he trying to do? Got it.

Route it to the right model. Route it to the right expert. Got it. This expert needs to talk to that expert. I’ll give you an example if you want it. But that level of complication — there’s nothing simple about it. It’s why you haven’t seen it. It’s why it doesn’t exist outside of videos. So the biggest thing I needed was to not do a demo, but to use the product live. Meaning you can code a demo just to be a demo. It’s code. But the principle was very, very clear. And this hasn’t changed at Amazon, to be clear. The team’s all in like we are going to show the product. And that’s what you saw.

One of the questions I have is just about that orchestration layer. We’ve seen other companies try to build it. Even when Microsoft launched Bing with ChatGPT several years ago, they were talking about orchestration at that time. Is that something that’s evolving in the same way in different places? Do you have a unique approach?

Yeah, I think we do.

Is that competitive?

I think it is. I think it’s hugely competitive. It’s pretty easy to invoke a single API off — I mean not easy, I don’t want to discount anything but, orchestrate to a grounding, let’s say the expert is a grounding expert. I’m going to ground the local info. We’re in New York. I know everything about New York. I’m going to make sure this conversation stays within New York. Calling one API, make sure you’re grounded to that local info.

Is “expert” a term of art within Amazon?

It’s just my term. As a team, we talk this way. I don’t want to overstate it. I think some people call them agents, some people call them APIs, some people call them, I don’t know, grounding to a certain experience, maybe? Our challenge was, that’s not enough. We already have that. I mean it’s deterministic today with Alexa, but we already have it. And so, meaning you can call a single API at a time, but then you get frustrated ’cause you’re like, I needed more than that. 

Let me give you an example. It’s a simple one. Let’s call “photos” an agent or an expert or just an app. I mean app’s a bad word, ’cause you’re not opening an app. But let’s just say the photos expert, and the music expert are both very important to this next example. The other day, I’m leaving the house. And I go, I have Alexa Plus, obviously. And I go, Alexa, do me a favor. Find all the photos of Mary’s… Start a slideshow and put music behind it.

Okay. I just did a search command. I did a photos expert command. And they have to talk to each other. He’s looking for Mary, slideshow, got it. And then that expert has to call the music expert and basically say, play the music. All right. It does a phenomenal job. It does it in under two seconds, and I get a slideshow. It’s pretty cool. Music’s playing. I’m about to leave the house. It automatically chose music and some playlist. And then I just said, change the music to, in turn without reinvoking Alexa, which I think you saw yesterday if you were watching, it’s very small. And I just said, put something on that Mary would like. And then it switched it and I’m perfect. And I just walked out the door. Okay, that’s an emotional moment. It’s one of my favorite parts of the product. If you said, P, what’s one of the things? I’m like, that’s it.

You’re pulling emotion out of the things that matter most to you. Mary wakes up, she comes in the kitchen, there’s a slideshow playing and it’s got music. She texts me, do you know Alexa’s on right now? I don’t know what’s happening. And I’m like, well, do you like it? She’s like, it’s fun. I’m not turning it off. I’m like, well, I left it. It was a message I left for you. Now the next step of that is to, Alexa, leave a message for Mary when you see her. And she will. But these are all, they’re multi-turn conversations, but they’re also “and” statements. So when you have, basically these conjunctions coming together, the continuation of a statement, ’cause I just want to talk in natural language. To invoke all of that in one place is, I think it’s beyond, it’s incredible what Alexa can do. I don’t see that anywhere else. It’s quite powerful.

So even in that example, and this is what I was saying at the top — it’s complicated.

It is super complicated, but you’re like, a slideshow, what’s the big deal, P? I’m like, well, I’ll be clear, on that screen, it’s emotional, it’s ambient. It was natural. Yeah. But it is somewhat simple in the way you talk about it.

Well, right, the outcome is simple. This is a thing that I want. But I’m looking at, okay, to make that actually happen, my photos need to be in Amazon’s photo service.

Correct.

I need to be in Amazon’s music service.

Correct. Well, no, Spotify would’ve worked there too. But yes, you need to have a music service.

That’s compatible.

But I would like it to be Amazon.

Yeah. Those divisions inside of Amazon all need to talk to each other in a common framework that Alexa can address.

Correct. Yeah. I happen to be responsible for photo service, so I’ve got that. It’s a blessing.

But I look at Amazon, I look at Amazon’s structure. Again, a lot of Decoder is like you can describe Amazon, can you describe other companies the same way? Okay, then. 

Amazon specifically has a language, how it describes how it’s organized. So famously, it’s single-threaded owners, right, like single-threaded leaders?

When you came in, obviously from a different management culture at Microsoft, how did you say, “Okay, I need everybody to participate,” because that seems like the thing in particular that Amazon has not been great at? And to make Alexa work the way you want it to, Amazon has to be great at it.

I think it’s a good question. At the end of the day, first off, all of Amazon’s rallying around Alexa. It’s crazy. It’s so cool. It comes down to a few things.

Actually, can I ask about even that, is that instinctual? Is that you got them to do it? Is it, Andy Jassy sent an email that said get on board?

Yeah, I think Andy’s been a huge part of it. I have a role. I mean, I came in with a vision that I think Alexa is a thing that we can anchor and change the world with.

Is that what drew you, this is one of my other questions, is that what drew you from Microsoft to Amazon is Alexa Plus?

Yeah, of course. Yeah, 100%. I don’t know if it was Alexa Plus, I’m not going to say that. It was the advent of where we can take AI and, yeah, I’ve got two questions in my head now, man. I need to compartmentalize both, but I’ll go there. You can see the turning point, I was there, I was in the middle of it, and it is just awesome moments, what Amazon brings relative to just even what I’m responsible for and how they can all connect magically through AI.

I fully believe this transformation’s happening, and Amazon’s the leader in ambient AI, period, end of story, and in the home, if we can connect all these things. A year and a half ago, when I was talking to Andy about joining Amazon, he was just so ambitious about it. He’s like, “Look, come in and do it. Let’s do it.” And so that is the tipping point. There’s a lot of nuance in that, but that was the tipping point, like, “Let’s go. We can change the world. You can think of the scale, the relative level of investment, the ambition, the patience that Amazon brings, but happy to talk about it.”

But yeah, the answer to the first question is sure, I come in, lay down a vision, kind of re-architect the team a little bit, get the explicit focus on, first thing we’ve got to do is get Alexa right. Once we do that, we’ll bring the hardware together. And to get Alexa right, it takes music, photos, shopping, and these are — you know, photos, of course, is under me, but you have across the company, you have music, video, shopping. We’ll just use those three as huge tenets for the product, and those leaders are exceptional. There’s no “we’re not going to work together,” it’s the opposite.

At Amazon, we set goals, and they are cross-company goals. And so the goals are set out from Amazon Nova, which is one of the anchoring points of the product, to what music needs to be on the product. Sure, the expert is kind of a joint thing, the music expert, but ultimately that music service has to be perfect and the music team’s killing it right now. Shopping, all in, how to make it great. We didn’t do a lot of shopping yesterday just because it would’ve been like a meme, you know, of course shopping, like oh, yeah, it’s going to be amazing. And then video, same, and there’s other areas, but we align and we go.

But it does start with a commitment from me for sure, you know, I’m in, I’m all in, I’m going to re-architect it. It’s not going to be easy. It’s going to take time. Andy’s patience, I would say the company’s patience to get it right for the customer is extraordinary, like extraordinary. I mean, Andy was pushing me. He wants urgency, of course, like you would expect from an Andy Jassy, but he also wants the right thing for the customer. And when you talk about customer obsession, let’s get it right. Let’s do it right and get it right. And we didn’t move slow. Even though you asked what’s taken so long, I don’t see that, you know what I mean, from where I’m sitting. I know it feels late because there’s been a lot of announcements, but I think we’re here at the right time.

You have a big team, and you talked about re-architecting. I think this brings me to the Decoder question. You oversee everything from Ring and Link to the photo service to the satellite service, Project Kuiper.

You took over what, October of ’23? November of ’23, you cut some folks. How have you restructured your group?

We refocused on Alexa, we really did. It was in a lot of different places, and so we just made it super clear. I had an Alexa platform team and an Alexa product team. It’s not a platform team, maybe that’s not the right way to say it, but just an engineering going across and then a product team vertically, is the way I look at it, and that AI stack going across. And so once you get that focus and that clear ownership, that leadership, you quickly see speed change.

That was the biggest shift, I think. Also I made some shifts as a team where a lot of the core horizontal functions are, if you think about the lowest level of the OS or the stack as a horizontal or hardware or supply chain, we’re kind of intermixed with the product verticals. So I’ve shifted that around too, just to get more product focus. One of the number one tenets is we’re going to make great products. I’d like to just start there.

I heard a rumor that at one of your first meetings you said that there were not great aspirational products, and that’s what you needed to do. Is that true?

Yeah, I mean, look, I don’t know exactly what was said, but at the end of the day, I immediately started pushing the team to have amazing pride in their products. We have to, because that pride shows up for our customers, and yeah, we want to push for it. That is a little bit of a, it’s just, let’s be super clear, these products have to be great. We’re not making tradeoffs if they’re not.

One of the things about Alexa is, again, in a previous administration, we would see Alexa coffee makers and microwaves, and the idea was we would just push microphones and speakers out everywhere and you would build this ambient platform, everything is sort of listening, everything is sort of aware of you. That was the big dream of ambient computing, that the computer would vanish into many different devices. You’re laying out something a little bit different, right, that there’s going to be a focal point in a piece of hardware. Yesterday was a lot about screens.

It was.

There’s a lot of multimodal interaction where you’re talking and touching a screen at the same time. That’s different, right, to say, okay, there’s going to be a place where you interact with Alexa?

Yeah.

That implies you’re going to cut down this giant ecosystem of ambient devices. How are you seeing that roadmap?

I think you’ve got to focus the roadmap. I think there’s no doubt. What you need is products that people want in their home, but also need, so I don’t think that history is broken. Obviously, the more endpoints the better, but they’ve got to be the right ones and they’ve got to be the ones that people want to use.

At one point, I think I saw a smoke detector with an Alexa microphone in it. I was like, we’re getting a little far afield here.

Here’s what I will say. The go-forward is: focus on making great products and the right ones. I don’t think you’re going to see thousands of products a year coming out, that’s not the goal at all. What I want is some attention to detail, making sure the right products for the customers are there, the things that fit into your home, the things that fit on your eyes, things that fit in your ears, so you can take Alexa with you, and just narrow the experiences that are great that way. And I have to tell you, the focal point, yeah, it is a screen on an Echo device in the home that can run your home. You don’t need it. With Alexa Plus, you actually don’t need it, it’s just a better experience.

And so when I’m asked, because there is a little, I mean, I’m treading a little bit here on some hallowed ground, like there’s a little bit of… Look, we’re going to light up all your Echo devices, but it’s just going to be awesome if you have a screen. And so when somebody says, “So, do you recommend a screen?” My answer is, “Yeah.” Do you have to have a screen? No? Well, you’re still going to have a great experience. Remember, you have a screen in your pocket. It’s called a phone. That phone has an incredible new Alexa Plus app on it, and so you have a screen, but you don’t need it to operate it. But let’s say you start a conversation with your voice and you just want to remember what that conversation was, you’re going to go to your phone to just capture it or you can send something to your phone. I think we cut it just for time in the demo yesterday, but anything you’re doing, you can send to phone because it’s like a longer form I want on my phone.

We’re also launching Alexa.com, so you’re going to use it on your PC, so it’ll be in the right places, but at the end of the day, if focal point is to control your home, which by the way, hundreds of millions of customers, that’s really the focal point today, you put a screen there, it’s emotional, it’s informative, it’s useful, and it’ll make a difference. It’ll make a difference.

So you come in, you restructure, you obviously want to get more focus on the products. All of that feels like we’re trying to change the culture, right? The structure is really a proxy for culture, in many ways.

Yeah.

That brings me to the other big Decoder question. Amazon has a famous decision-making culture, one-way doors, two-way doors. You can write books about it. You’re writing the press release before you write the product. You have a long history at Microsoft, you’re obviously trying to change some of that culture, how are you making decisions there? What’s your framework? Are you inheriting all the Amazon approaches or are you bringing your own riff to it?

I often get accused of making the final decision only when I have to. It doesn’t mean I’m not making decisions. When I was studying up to come to Amazon and making that decision, that was a life decision for me, it was a big one, and I was so inspired talking to Jeff, talking to Andy, just inspired, no doubt. I also love Microsoft, so I’m inspired where I was sitting, so there’s all these conflicts. Those are personal to me, oh my gosh. But when I started reading, let’s go to decision-making, and then I just watched a few stories that Jeff had told, talked to Andy about it, it basically, from a leadership principle standpoint and from some of the things you hear about on decision-making principles, like one-way, two-way doors, it’s hard to explain this, but it’s so aligned to the way I was running my team. That’s how I’d operate. It was weird. I was just reading the LPs and I’m like, I used to have a culture box.

The LPs are leadership principles?

Oh, sorry, right, leadership principles at Amazon. You should check them out. You can go to Amazon and find them. They’re rad. They’re inspiring, and they’re almost, sometimes they’re just obvious, not all, not all of them. And they’re hard to believe, like big bets, is that real? I’m like, yeah, it’s pretty damn real, it’s pretty incredible. Leaders, they do, they dive deep. Yeah, they do, they get into everything. And I think those are real, but in the spirit of when I started reading them and then the way I made decisions, Nilay, they were aligned. I mean, I’m not, no BS. They were just, it felt right. I had a culture box when I was running my team at Surface and Windows, and that culture box had five cultural principles. They were basically five of the LPs, but that’s how I ran it, and so it was so connected.

And when I got to Amazon, it was almost — what a team! I found this team that was not only hungry, but unbelievably talented, massive and capable, knows how to ship, knows how to invent, and it’s just a little bit of direction, that’s all. My job’s to give that direction, and so making sure I lay out the vision, making sure everyone knows where we’re going, what are the highest priorities, but when it came to decision-making, to answer your question, is I fully operate in the values of, all right, let’s make this call today, but no. And I think one of the strongest points of a leader, without any doubt, and I learned this from one of my colleagues in the past who I worked for, he used to teach me. He’d go, “Hey, Pete, when you’ve made a decision, the best leaders in the world are willing to be wrong. Now, you’ve got to be right a lot, but you’re willing to be wrong.”

This is simple to say, but it’s a powerful concept. What does willing to be wrong mean? It means you’ve got to put your ego aside, you’ve got to be vulnerable. Do you know how hard that is, in front of a team of thousands of people? Just, “Yep, I was wrong.” What does being wrong mean? It’s not like this dramatic, “I’m wrong, I’m sorry.” That’s not it. When being wrong, it’s not necessarily the wrong statement, it is the you got new information a week later? Then use the information. And if it was a two-way door decision, guess what? Make the right decision. But if you’re not a great leader, you don’t change that decision because you’re like, “I already made the call, sorry,” but you knew it wasn’t right for the customer or for the business or whatever the reason. It’s just a fail.

And this was very early in my career. It’s very similar to the two-way door, one-way door. Once you’ve made the hard call and you’re past the point of return, that’s it, you made the call. And you have to make decisions sometimes, man, and those are hard. You lose sleep over it. When I made the decision to have the event, “We’re doing it.” And they’re like, “Well, the product’s not 100% done.” I go, “It doesn’t matter. I’m at 90% usage. We’re going.” And everyone’s like, “You realize that,” and that was a two-way door decision until I send out the invites. And so we checked the information a day before we sent out the invites, and like, “We’re going.” The minute you send the invites, that’s a one-way door decision. There’s no pulling back. It didn’t matter how sick I was, it didn’t matter who couldn’t make it, none of it mattered.

And then we’re lining it up, we had the venue booked, and we’re like, okay, that wasn’t a one-way door decision, you can always cancel the venue, not cool, but if you had to. And you kind of go through it, and then you get to that point, you know, that’s it. There’s no new information that was going to change it. And so great leaders, they’ll make those decisions, but they’ll always be willing, they’ll always be willing to check themselves, and not just check themselves, but be willing then, when they have new information, if the right decision is in front of them, you’ve got to change it, and I always live by it.

And so when you come to this world of, when you say this culture, Nilay, the Amazon culture is incredible. You have no idea how empowering that is. It’s a two-way door decision, all right, let’s make the call. If we’re wrong, let’s deal with it, but then we move, and we move. And I get accused a lot of, you know, like to make a call and like, “Do you have all the info?” “Probably not, but we’re moving.”

Yeah, we’ve got to try something.

Yeah, and it’s been pretty fun that way.

Let’s put this into practice. I want to talk about Alexa Plus in great detail now. I think I have a sense of how you got the team to get the product so you could have an event. The big announce, the last thing you announced, was the pricing, and you started with, it was big reveal, well done, well played, you said it’s $20 a month, and-

Credit to Andy. That’s Andy, that wasn’t me.

And it’s free with Prime. This is a big decision, right? Pricing is maybe the most important decision.

Yeah. I think the exact words were, “$19.99 per month, but free with Prime.”

I will note that Prime itself costs $15 a month. You’re pricing the service $5 more than Prime. Are you subsidizing Alexa Plus with Prime?

I don’t think I understand.

Does it cost you more to run than you’re getting inside of that membership?

I want customers to understand that the service is better with Prime. At the end of the day, if you have Prime Video, Prime Shopping, Amazon Prime, you fundamentally get the best music experience. You get photos, unlimited photos. That just makes the Alexa experience better. You don’t need to have it, it’s a great experience without it, but it’s just better. And so we talked about it, we want people on Prime. If you’re on all those services, it comes together and as a collection on your product, it just makes the personalization so much stronger, it makes the invocation of services so much easier.

Was this an obvious decision, from day one this is going to be part of Prime?

No.

How’d you make that call?

Just a series of events. I think back to two-way door decisions, that definitely, I don’t think it was the first decision, there were different ways to think about it. It costs more to run the service, that’s all there is to it. You’re going to invoke an LM, you have many models working, there’s a lot of inference, that’s true. Then you heard Andy talk about how much cost is coming down with Trainium2 and you just see the efficiencies, if you will, that are coming through, those are plumbed through the plan. We have an incredible opportunity in front of us. And so it wasn’t about how much you’re spending, how much you’re making, it is about making a great product. And once we were like, we want to make sure people have the best product possible, that is the anchor. And so we’re like, all right, it’s got to be with Prime, that’s the best way to get customers there. And that’s it.

I think people want it to be more complicated, because I’ve been asked this question a bunch of times. I generally haven’t answered it. I’d be like, oh, you have a choice. You can pay 15 or 20, it’s your choice. Just choose. But not to be, I’m not trying to be pompous or whatever. I think if you’re on Prime, you’re going to love it, so I inverted the equation.

The other piece of that I see, the other way to think about it that I was curious about, you mentioned this, Alexa has distribution, you have a huge installed base of devices. This is I think the first at scale non-phone AI product. I can’t think of any others. 

Yeah, it might be. I have to think about it.

There’s Google Assistant, but they haven’t launched the way that you’ve launched this product yet. Gemini isn’t doing all this stuff yet. There’s Homepods but Siri doesn’t do it yet. I don’t think the Humane Pin was keeping you up at night, and now it’s gone.

Well, it’s not gone, I think. Went into HP, right?

No, they shut it down.

Oh, they did?

Yeah, it’s gone. They won’t work anymore in a couple of weeks. It’s a real thing, we’ve been breaking news to you here on the show.

Wow, that’s huge. You’re so informed.

Sadly, that’s my only job, is to be informed. Make no decisions, just know everything.

[Laughs] I don’t see it that way. 

But that is the scale. If it’s not a phone, you need something else. There’s been a lot of excitement about what something else could be because you have a new user interface paradigm with voice, with natural language. But you already have it, you have the installed base.

And saying it’s going to be with Prime means you’re just going to deploy it to that installed base, because I’m guessing people with Alexa and people with Prime has a pretty massive overlap.

Yeah, there is.

So you’re just going to launch it to that whole service. Is that going to be a flywheel?  Because the promise of Alexa 10 years ago was this will compete with your phone. I don’t think that actually happened. Do you think that this will help you compete with the phone in that way?

I think it’s more of a compliment now than it’s ever been. You need the phone, we send things to the phone, we want you on it as well. I want you on the Alexa app on your phone, it’s an awesome experience. We can play with it if you want after, but I think it’s a compliment to the phone, I think it does replace a lot of things. I’ll tell this, I say it to my team all the time, look, our customers are going to find the easiest path to something. They just will, it’s innate. It saves time, it’s about speed, it’s about efficiency. The only time that’s not true is when you’re getting more joy, and a lot of times joy comes from speed or happiness comes from being able to complete a task quicker. And so let me go back to the point of ambient. One of the core tenets when we started Alexa Plus and the vision for it was we have the largest install base in homes on the planet.

I think that’s a pretty definitive statement, I think it’s true. I probably have to check with the lawyers to say something like that, so maybe I’m wrong, so let me qualify it. We might have the largest install base on the planet, and it’s incredible. The way Alexa Plus is designed is it’s meant to be ambient, it’s meant to be a conversation, and it will replace tasks you do on your phone. It’s going to happen. And so does it replace the phone? Absolutely not. But does it replace certain things? I think I told you the story before, let me tell you again. When I was building laptops 12 years ago, when I’d first started on Surface, people came to me and said, there were a few people that were like, “You’ve lost the plot, P. You’re going after this thing and the laptop is dead.” Why? Because phones are replacing the laptop, and I mean you’re using a laptop 12 years later and it’s pretty important to you.

Probably more important now than it was 12 years ago. So what had happened was jobs moved to the phone that were really important, shopping, social media, your photos, I don’t know, pick communication. But what happened was the things that didn’t move to the phone only got stronger on the PC over that time, and so they essentially became compliments to one another. If you’re going to sit down and write a long story, you’re going to do it with a keyboard. You want to be snackable information, you’re going to pick up your phone. And then one got better at one of them and the other got better at the other, and incredibly so. It actually strengthened them both.

I see this as very similar. I think as Alexa Plus comes into market, I think it’s going to be better at a lot of things and it’s going to move jobs to it. I believe that. I think there’ll be more emotion to be pulled out of something that’s conversational, knows you well, is personal to you. You can have a conversation, it knows your calendar, it can get some stuff done in a simple way. You might not always do [the task] on it. I don’t know, it doesn’t matter to me where you do it. I just want to give you the shot, and if it’s the easiest way to do it. Can I give you just a fun example? I was sitting on the couch last week with Costas, my son. He’s 24. I don’t know, he’s 24 ish.

Those are pretty fuzzy ages.

I think maybe 24. He was born in… Yeah, 24. And so we were hanging out and we were talking about the Clippers and he had asked me a few questions, and I’m a fan of the Clippers growing up, and then of course since Steve [Ballmer, former Microsoft CEO] bought them, I just love the team. And I asked “Costas, did the Clippers win last night?” He goes, is Kawhi even playing?” This is, I think, a week and a half ago. I don’t remember the day. And now we have Alexa Plus in the house everywhere, and my son works on AI now, he’s blown away by it. He had to sign an NDA that he can’t talk about what he sees. And I realized right at that moment — Nilay, I was going to lose him, because you know what happens? You pick up your phone, you open it, now you see your notifications, you know that feeling, and you’re like, oh, I’m going to check my notifications, or I’m going to jump on TikTok, or whatever it is that you love about your phone.

He’s going to go get the information, answer it, and I’m going to lose my kid to his phone. And now all of a sudden we went from this moment hanging out to him on the phone, it happens all the time, and it blew my mind. He goes, “I don’t know. Alexa, did the Clippers win last night?” And Alexa goes, “The Clippers did win last night.” And then his score and blah blah, Kawhi Leonard scored so many. And he’s like, “Is Kawhi playing?” “Yeah, Kawhi’s been back for several weeks.” And he now started having a conversation, the three of us are having a conversation, the job moved. He would’ve never done that.

So this was the promise of the original Alexa, right? There’s celebrity ads during the Super Bowl, people are just hanging out with their Alexas.

It was a great ad by the way.

It was a great ad.

Oh my gosh, what a great ad.

But it couldn’t do it. A decade later we have trained a generation of consumers to believe that these products are limited and that we should use them to play music and set timers. How are you going to teach everybody that it can — actually, a more important question: can it do it?

It can do it. I think we’re resetting the next 10 years right now.

Are LLMs durable enough as a technology to build all the things you want them to do?

Not just the LLM. It’s not just the LLM.

I understand that it’s not just the LLM, but it is the enabling technology that’s making all this go.

They’re durable, but they’re going to continue to evolve at a rapid pace, and they have to. They are. But you have to be smart about how you build on top of it. I mean, obviously everyone’s doing a great job, I’m sure. I think the promise is there. I’m not going to understate it, I won’t overstate it, I can’t, I believe the promise is there.

I’m here at Amazon because I believe it’s going to change the world how people engage AI, and it’s going to be easier because your device is there and ready for you, and we’re going to make beautiful devices. And so all this will come together in a way where there’s a team that’s going to connect all these experiences. You saw a little bit of Fire TV and Ring, that all of a sudden these natural moments are going to happen and you’re not going to have to guess, you’re not going to wonder.

If it can do it, because it’s not deterministic, you’re not issuing these Boolean commands.

Correct. Exactly, right. And so hopefully everyone understands that concept, but since it’s not deterministic and now you’re going to ask a question, even if Alexa doesn’t do it, she’s going to talk about what you’re trying to solve and you’re going to actually get to an answer. As opposed to, “I don’t know.”

One of the things that I think is really interesting about the product, you talked about the kid’s demo where it was telling a story to a kid. I’ve had my kid talk to ChatGPT in that way, I think it’s fascinating to see that interaction develop. Then there’s simple stuff. Yesterday I sat in one of the smart home demos and they turned the lights from blue and green to a warm yellow and I was like, that’s a lot of data center to turn a light from one color to another. So you can see inside of the orchestration you’re describing, there’s the most expensive thing, to have this real time creative story. Then there’s “turn the light off,” which should be simpler and cheaper. I’m assuming the orchestration is picking what model to use when.

That’s exactly right. And some will do it on the edge too. You don’t have to do it all. If it’s a point and shoot command, we’ll do it in a simpler way.

But then I ran into Mike Krieger from Anthropic, who was at the event. Anthropic is one of your models, and he said the most interesting thing to me that I heard yesterday. He said, “Sometimes when I talk to Alexa, I can tell when it’s Anthropic because I know our model so well.” And he’s like, “No one else will be able to tell.” But he was like, “Sometimes I talk to it and I say, oh, that’s my boy,” which was incredible.

A product person knows their product and maybe they’re seeing ghosts in the machine, but it was just incredible. How are you picking between Nova and Anthropic? How are you picking the cost of these different models that you have to invoke? What are they better at? How are you making that determination?

Actually, the orchestrator picks the model that’s right for the job. The how, I won’t get into the details, but there’s some awesomeness here. One of the things that inspired most people is that we’re using a multi-model approach, which I think is a little bit novel. But at the end of the day, it depends on what the task is, it depends on what’s being asked for. I think right now you’re seeing 70% of the utterances running through Amazon Nova, 30% running through Anthropic, something at that rate. It changes, it just depends on how you use the product and what you’re using it for. It is also non-deterministic. Basically, there’s a model that’s like, what’s the best model to pick? And then you’re looking for accuracy and speed. First understanding, then accuracy, then speed, and you target. Then you move it, you pick the right model and then you fire to the expert, and there’s a small model and the expert if you will sometimes, and then those all orchestrate together and that’s how it works.

Inside of that is the way that you talk to your partners.

Slightly different than all of that.

I think you just did an API-driven one where you asked for an Uber, and Uber’s got a bunch of APIs and you just talk to them.

Uber’s been awesome. Uber, OpenTable, Grubhub, these things that you use every day, they’re just in-depth connected. That’s like opening an app on your phone, at the end of the day.

We understand how computers work. You call an API, it delivers a result. You call another API, great, the Uber’s booked. Then there’s the more agentic stuff that you were showing off. It wasn’t quite ready yet, but a lot of people have this idea. I believe the example was we’re going to book a stove repair, and it was a Miele stove.

He was going to choose last minute depending on how the demos went. I think he did, did he do a Miele dishwasher?

I know it was Miele because I was like, oh, those are expensive to fix. That’s what I knew in my head.

[Laughs] That’s what he said. It was pretty funny.

And then he went on to Thumbtack, which is a partner, so he had permission, but what it was doing was it was looking at the Thumbtack website and clicking around and reading that back to you. And even with permission, I think of that as why wouldn’t you just get an API? If you have the permission, why not do it deterministically?

Yeah, then the partner just has to do the work.

Right, so this is basically cutting down the amount of work a partner is doing.

Yeah, you don’t want to do the work, no problem. It’s just a couple of different ways to engage it. From an SDK perspective, this is just basically permissions, and we have to work on authorization and payment at the end of that, which is the trickiest part. I’m not going to get into how, but that’s the trickiest part. And so completing the task is the trick, getting almost there, it’s not that hard, but completing it. And so that’s where you need the partner to be like, yeah, sure, we want this traffic and we’re going to go create the service and send it through. Great. If you don’t, no problem.

But the answer on why not do the API is just these relationships are different, partners want to work in different ways. One of the things we are trying to do, and I’m really re-engaging Alexa, is we want to open SDKs. Basically, we want to open the product up for developers to come in and do what they want, come make it great. And if somebody asks to fix something in their house, we got it, we have a way to get you there.

So that implies a lot of things. Having tried to get a Miele dishwasher fixed in my life, it is expensive. 

The repair person has to actually be on Thumbtack, they have to actually be using that service to actually book their appointments and take payments. That is not necessarily true, they might just be marketing there, but there’s a lot of things you have to know that you’re depending on that ecosystem to provide you to make Alexa just book a repair service professional for you. That’s the part where every time I talk to anybody about agentic systems I’m like, oh, this is where it falls in, payment is the other one. And the thing I’ve been calling it is just the DoorDash problem.

If you say, “order me some food,” and it goes and uses DoorDash for you or GrubHub or whatever, you’ve commoditized those service providers and you’ve started to crush their margins. And after a while, you might not want to be… Because they can’t upsell you anymore, they can’t sell you their subscription credits or whatever else they want to do. They can’t put advertising in front of you because the robot’s looking at their website, not a person. And I don’t know why they would participate in that unless you have actually solved this payment problem, to make that valuable to them.

I think the partnerships are unique for sure. I think it’s quite different. Remember, you always go back to your phone, the information’s there, it’s in the app. It’s not like we’re doing something on the side and doing it anonymously and you don’t have the customer info, I think is one thing. The second thing is when you have those challenging… Let’s use a Thumbtack example, let’s stick there for a minute. If you don’t have a Thumbtack account, the first time you do it’ll just pop a QR code and say, here, connect, authorize, go. And then forever then you’re going to fix things and Thumbtack’s going to push you through it. There are just some simple things that you can do that make the customer journey simple and gets you to those connection points. And once you do that, which is everything, God, you understand this, setup is everything, removing that barrier to entry. To make Alexa Plus great, you’ve got to share your contacts, you’re going to want to add your photos.

I think you’ve got to share your contacts. You’re going to want to add your photos. You’re going to want to connect your service providers. It’s a one-time kind of low barrier to entry go, and then you’re all in. And the partners, we don’t talk about the deals with the partners or anything like that, but there’s benefit on both sides. But at the end of the day, it’s the right thing for the customer. And I think there’s a lot of partners out there that believe in that same philosophy. Let’s get our customer to the endgame.

But if you run one — say you run food delivery service A. I won’t name names to keep them out of it. But if I run food delivery service A and I have a deal with you, and food delivery service B shows up and signs a deal with you, and I just ask Alexa to order some food, suddenly Alexa is in control of a lot of revenue.

Yeah. But you have preferences, customers have preferences, they know what to say.

Why would they have a preference over where the sandwich comes from, like what intermediary brings you the sandwich?

That’s their choice. You can’t speak for that. You can’t speak for it for the customer, but I would say they just have a choice, and they’ll get a choice.

And you’re going to express that choice on a screen?

I’m going to keep partners out of it for this, so I won’t give you the examples, but there will be simple ways to make it clear to the customer what they want.

The other part of this, which is equally complicated is partnerships, and that’s agentic stuff. And usually when I talk to people at agentic services, it’s to open the ecosystem to say, “Okay, we can browse the web for you. Now we have access to everything.” You are doing that in a much tighter way. You’re saying, “This is how we’re going to bring partners in.”

Why make that decision? Why not say, “We can just go browse the web and do whatever”?

I just think it’s right. It’s their business. And so, we’re seeing a lot of participation. There’s a lot of partners.

They’re excited, from what I see. Not all; I can’t speak for all of them. I’m not trying to talk in absolutes. But you have this moment where you’re like — the promise of Alexa is here. Ambient is here forever. They’ve all made skills in the past or they’ve done something that they didn’t get invoked. And it’s hard because the customer had to point and shoot as opposed to just speak in natural language; they had to know exactly what they were asking for. But at the end of the day, now you have a truth in: just speak, and something comes up. And now partners are like, “Well, if they’re looking for something from me, I’m in.” But I think it’s right to be partnering and not doing it another way.

Which I’m pumped about. We have a great biz dev team, it’s what they do.

So that’s asking Alexa to do something, and it goes off and does something in the world, right? It schedules a person or orders some food, it books a flight, great. Then there’s the stuff in your home, which Alexa has historically been very good at. 

Turn the lights on and off, make a routine. I’m very intrigued by the idea of automating routine creation with natural language. Right? Make a bedtime routine for me. That is as messy as it gets, right?

No.

That’s not even partnerships. That’s Matter and Z-Wave and all.

We do it all before then. This one’s different. We already have partners that work with Alexa. If you already work with Alexa, you get the magic.

That’s it. It’s awesome. You saw it yesterday. There was no new code written on the partner side.

Really?

Nothing. I have my Govee lights at home right now that I put on the house. I’m just talking to them to change the color. That’s it. I would’ve never opened the app to change the color on my lights.

It just seems like the promise of the smart home forever, and this is what you’re describing, is that it will get more invisible.

This is what’s awesome, dude. Right?

It’s going to get more invisible.

You have to understand this is freaking awesome. 

But I’m looking at the last five years, like, “Oh, this is more visible than ever.”

You have no idea how badass my team is. This team, now I’m talking Eero, Ring, Blink, Fire TV. This team, including Alexa, Kuiper, they’re incredible, man. They’re so damn capable. I’ve not seen invention like this. Now how we get it to the customer, we refine a little bit of that. But I’ve got to tell you, and this is a great example, because this works with the Alexa program and the thousands and thousands and thousands, dare I say, hundreds of thousands of things that work with Alexa. That is one of the largest connective tissues on the planet. It’s crazy. And they’ve set it up so well that now when Alexa Plus shows up, your routines are by voice done, like 100%, Nilay.

It’s so damn cool. The other day Mary was so frustrated with me, and I don’t have a smart home at my house in the Seattle area, but I use it in another area. And she was so frustrated with me. She’s like, “The lights are on all the time.” I just grabbed my app. I’m like, “Alexa, every night just turn off the lights outside at 10:00 PM and don’t turn them on again until 7:00 PM the next day. That was it.

The promise of some of the smart home standards that have made this messier, like Matter or Thread, is that you will be able to control these devices device-agnostic, right?

Yep. We’ll take advantage of those as well. Yep.

For example, everyone talks about the smart home only in the context of their own lived experiences.

Well, how do you not? What are you going to do?

It’s hard to be on track.

What story are you going to tell? I’ve got plenty of customer stories.

But my joke is that if a thing doesn’t show up in control center on my wife’s iPhone, it doesn’t exist. She’s not going to open an app. She’s going to swipe down and see that panel and that’s how we’re doing it. So you’ve got to bridge into that. The promise of something like Matter is, we’re going to see it across all of these surfaces. It’s all going to work together. Are you thinking that far ahead? Because where does the logic of my smart home live?

Especially if you’re talking about putting hardware with a screen centrally in your home. Okay, now you’ve got a little computer running your house. And everything should talk to that, and that’s where the logic should live.

In theory, but we also have the cloud to arbitrate. We have so many different methods in. You can use Matter, you can use Bluetooth LE sometimes. You can use Zigbee, but you can also —

Ring famously runs on Z-Wave all the time.

You can use Z-Wave. You can fundamentally use Works with Alexa, just plug them right in. There’s no limitation for us to connect these things, because basically we can orchestrate to it. The team has thought through it from every way to Sunday, but they’ve also been working on it for 10 years.

It’s phenomenal. It’s probably one of the things I’m most excited about, because you basically democratize the smart home, a hundred percent. Yes. It won’t work unless you gave someone a button on their phone today, but we just talked about this. You know where the job’s better? Just say what you want.

It’s a much better job to be done. I tried to do it with the music demo yesterday. I’m not sure it landed this point, which is like, just plug them in. The speakers were there. I’m going to move music to the speakers. I’m going to do it nuanced. I think one time I said “Move. I want to move the music. I want to hear the music. I want you to bring the music here.” I used different language so it wasn’t continuous. That was all real working. Probably those little nuances get lost on the natural language as if I had a direct command. I didn’t. It could have been any of those. Or play, which I try to stay away from. And so, it’s the same concept. You just think it and say it, think it and say it. It’s very powerful. And on smart home, it comes to life amazingly. And this is credit to an incredible team. They’ve thought it through.

Do you think that we’ll see more of an explosion of consumer smart? There’s big investments people.

I think so. I think this is the tipping point.

You’ve got to put a bunch of light switches in or buy all new light bulbs.

I think so. Tipping point, because you don’t have to be an expert. Just plug it in, that’s it, and then say something.

I want to believe you, but I’ve been burned so many times.

I don’t care if you believe me or not at this point.

I’m just saying.

When you get after it, man.

I’m ready to get the products. I’m ready to try.

You go get after it. It’s pretty fascinating. This is what an LM is great at. And then, the expert that we have to go rationalize and so it doesn’t have to be deterministic. And so, it’s pretty interesting.

By the way, it has to learn as well, so if you go, “Turn on that light.” “Which light?” “That one over there.” “Oh, you mean the one in the living room?” “Yeah.” “Okay.” Now that’s not a good example, because you’re up against a switch, which takes, is just go touch the switch. But how fast the system learns, that’ll never happen. It’ll never happen again. It’ll be like, “Oh, I know what he needs. He’s asking on this device and I got it. I know I’m turning on the light.”

What’s one thing you want Alexa Plus to do that it can’t do today?

I’ve shown you everything, but I’ll tell you, and if I can touch back to my Mary example, I want these moments to connect not only the home but the family. And it’s got some pretty amazing attributes. The idea that I can leave the house and leave a message and walk out the door, and then when Anastasia shows up downstairs, she gets the message, and it’s a lovely note from her dad with maybe a direction of what to do. The fact that it’s this totally natural language moment feels magical. Alexa is being proactive on your command, not intrusive, but you are asking her to be. When you start seeing those things, that’s the thing. That’s the thing I want it to be. Because you’re just going to connect deeper into people’s lives in a way that makes it better, that you know me well enough. I want you to use these products and tell me your life is better.

But there’s not a specific thing where you’re like, “I need the next turn of capability here.”

Look, I have a vision for where this thing goes. I can’t take you there. We’ve already revealed everything, and we’re going to preview in a month. And it’s like, I’m sure we tipped over a few carts yesterday, and so I’ve just got to be careful how far I take it. There’s so much for the future. But I showed you a few of my favorites. And that’s what we did. We narrowed it down. There’s thousands of things it does now.

Try to narrow it down to the ones that both told the story but are also most emotional to me, because that matters, what I’m presenting and I think sharing. And the biggest thing, you want the team to have pride in the best stuff they’ve created. And those moments are pride moments for the team.

There’s something I’m really curious about. I’ve asked basically everybody who has had something to do with Alexa about this for a decade. Amazon always calls Alexa she. For some reason this robot has a gender, and it’s a she, and it’s always a she. Why is Alexa gendered in this way?

There’s eight voices with Alexa Plus. I don’t think we talked about it yesterday. It’s in the blog post that we wrote. Not the blog post, the About Amazon post. I’m told I’m a dork when I say blog.

I run a blog. You can say blog.

Okay.

It just depends what you’re using. Pick your voice. But the default, the default, I use the default voice. I love the new voice.

You can use the old voice. I love the new voice. It’s the default. And then, you can pick a male voice or another voice, and you can call it what you want.

I just wondered. For a decade, you gendered this robot pretty real, honestly.

Yeah. It is. This voice, the more we’re using, I called her she yesterday. I understood that. I had a couple of people ask me, I’m like, “Well, that’s kind of how I was thinking about it.”

I don’t think it’s more complicated than that.

That makes sense. I don’t even mean to… I understand it’s a loaded time in American history we’re asking this question, but I actually don’t even mean it in that context. I just mean it’s a robot. It doesn’t actually have one of those. It’s only what we assign to it.

Yeah. I think look, it is, but it is getting more personal. It’s going to be more meaningful in your life.

For sure. Do you want people to think about it as a person in that way?

You don’t want to go all the way there, but yeah, I think it’s okay that you think you have another set of ears when you want them, another set of thinking if you need it. I think it’s quite powerful.

All right, last question. This is rolling out soon to some devices. I think it’s the screens, the Echo Show 15 and 21.

Yeah.

When is it going to hit everywhere?

Actually, the 8, 10, 15 and 21.

Eight, 10, 15, so the screens.

It’s rolling out next month starting with those devices, and it’ll be a gradual rollout. And then, it’ll roll out to all devices. If you want to be in first, my push is I want people using screen devices, for sure. We’re rolling it out there first because it’s such a great experience. You go get a device and you’re on the list, you’ll be first to get it. That’s basically it. If you already have a 10, a 15, a 21 and you subscribe, then we’ll get it out to you as well. That’s where we’re starting.

And it’ll light up the whole house, by the way.

Oh, the other Alexas?

Yeah. 

So you have a screen, and it comes to your screen.

Yeah. Let’s say you have five Echoes at home right now, and you just go get a screen and it’ll light up your whole house.

Do you think you’ll drive a hardware cycle of people trying to buy screens to get Alexa?

I hope so. I think they should. And not because I want to sell another device, but I want people to have that experience. I think it’s a miss not to have it. It’s a miss.

I want to drive a cycle in the spirit of not trying to be sales, not my thing. But I will say if you want the best experience, go get a screen device. We’re pleased already. I didn’t expect… Pleased just seeing the reaction from yesterday. It’s nice to see. But I think in a month, people will get it in their hands, we’ll start the preview. Most features will be done, most. There’ll be a few that are coming later, for sure. And then, we’ll roll it out to everybody when it’s the right time.

Last question, you’ve said you’ve got a vision.

This is your third last question.

I know, but that’s how I do it.

I love this.

This is why I’m good at this. It’s tricky. Really, last question: You’ve laid out a vision for where you want to go. You’ve talked about the big opportunity here. I’ve asked you if you think LLMs are durable enough to pull all this off. You said they are.

Do you see this as a platform shift the way that other people have talked about it as a platform shift? Do you think we’re going to actually reconsider how we interact with computers at the biggest level, the way that touch screens did it, the way that mice and keyboards did it?

Not to be too cliche, I think 10 years ago was a magnificent moment when Alexa launched, 10 years and a couple months, but what a moment. It really was a reset. I think right now, 10 years later, I actually do think this is that next moment. But this one is, to your point, that promise. I think this is the shift. I think this is that time. It’s going to take years. This is not like, don’t worry, you’re not going to miss out. Somebody’s like, “Well, why are you so late?” I’m like, “Late? Do you know we’re just at the beginning?”

And by the way, our Roadmap is awesome. And I believe in this team, in their invention, and the company’s patience for invention, and its ability to make the big bet and stick with it. It not only creates an incredible future opportunity, but with that opportunity and bet and invention, you also have the moment right now is just starting. It’s literally just starting, dude. It’s just starting.

It’s a great future. It’s fantastic. And I think the home transforms forever starting now. But it takes time. It takes time. And I would say patience is one of the strongest qualities of Amazon. I had once heard infamously on a great leader, and I don’t know the quote, but our best overnight invention took seven years. It takes time. But right now, we’re here. 10 years later, here we are. And it’s the beginning of that next gen. I think it is a shift. Right this moment.

All right. No better place to end it, Panos. Thank you so much for being on Decoder.

Great to see you, Nilay.

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

Today, I’m talking with Vimeo CEO Philip Moyer. You probably know Vimeo from its beginnings as an artsier, more creative competitor to YouTube. But over the last few years, and especially after it went public in 2021, Vimeo has really turned itself into an enterprise software company, selling video hosting services to companies of all sizes.

And this episode is a particularly fun full-circle Decoder moment — I interviewed Philip’s predecessor, Anjali Sud, both when she was the CEO of Vimeo and again more recently in her new gig as CEO of Tubi. So it was fascinating to close the loop and see how Philip is changing Vimeo after taking over, especially as the entire ecosystem of online video is shifting so rapidly. 

Philip is pushing to make Vimeo a different kind of YouTube competitor, one that can support everything from independent creators to huge corporations. It’s a shift from the strategy Anjali used to reset the company and take it public, and there’s a lot of interesting nuance to it. Everyone wants to put videos on the internet, it turns out, but only some of those people want them to be ingested by YouTube’s advertising and recommendation systems. 

Philip himself has a ton of big tech experience: he’s worked at Amazon, Google, and Microsoft, and he’s deep in the weeds on both the tech and the business. You’ll hear us talk about Google’s business and YouTube in particular quite a bit in this one, but we also get into TikTok and what it means to have the incentives of algorithmic video platforms drastically influence both creators worldwide and the culture we all consume.

Of course, we talked about AI, too, and how it’s upending every platform in different ways. Vimeo has been marketing itself as an “AI-powered video platform” lately, so I wanted to know what Philip is thinking about Vimeo’s creator-focused mission colliding with the pitfalls of AI-generated video. 

We also spent some time on the simple supply-and-demand math problem that seems like it will change the creator economy drastically in the years to come: that is, if the amount of video on the internet explodes because of AI while the total amount of time we can spend watching video remains relatively fixed, how is anyone going to make any money at all?

This is a fun one; I think you can really tell that Philip and I could have kept talking for a very long time.

Okay, Vimeo CEO Philip Moyer. Here we go.

This transcript has been lightly edited for length and clarity. 

Philip Moyer, you are the CEO of Vimeo. Welcome to Decoder.

Thank you, Nilay. It’s great to be here.

I am very excited to talk to you. I have got to tell you, just structurally, this is one of the very first truly full-circle Decoder episodes. Because we had your predecessor Anjali on as CEO of Vimeo, and then we had her on in her new job as CEO of Tubi, and now we have you on as her replacement.

Wow, it’s full circle.

It’s full circle for a show that’s about structure, decision-making, and organizational culture — this is as good as it gets for me. So thank you so much for joining us.

That’s fantastic. I have a lot of respect for Anjali and have to say a big thank you to her for all she built here, so it’s wonderful to hear.

I’m very curious. Anjali was in the middle of executing a big pivot when we talked about Vimeo. People know Vimeo as what it started as — a consumer video service, kind of the art-ier competitor to YouTube. She pivoted it to a [software-as-a-service] business. She was very open: “This is now a SaaS company, we’re doing a lot of enterprise work, we’re a video hosting provider to a lot of people, we are a video creation tool for small businesses.” That was several years ago. I talked to her three years ago at Vimeo. You’re the new CEO. What do you think of Vimeo today?

I think in a lot of ways, I mean a little bit of what she started, and a lot of what we’re continuing here, is going back to the roots of what Vimeo is known for. People came to us because, at the time, 20 years ago, it was hard to upload video. I kind of call that… I’ll say the second or third epoch of video was right around that time period when video files were getting bigger. It was hard to stream video. They were no longer just a single file.

We were starting to get multiple formats. People came to us because of the quality of our transcoding and then, ultimately, our ability to be able to serve it directly to the person that was most important that they wanted to provide that video. And a lot of what we’re doing right now, and what we’ve taken that Anjali had started and extended pretty significantly is… It turns out that lots and lots of organizations and individuals around the world want to be able to keep a video private or they want to be able to serve it only to an intended audience.

They don’t want to have someone collect data, they don’t want to have an algorithm that tries to send them down a rabbit hole. Instead, they want to just literally be able to provide that video. And so we’re seeing that video makes up about 82 percent of the world’s internet today and we’re seeing that same concept arrive with private individuals, doctors, educators, and large organizations. And so yes, I would say that what we’re really becoming now is… Our goal is to become the largest private video distribution platform in the world. We think that there’s an increasing demand for video that isn’t public or algorithm-driven, but instead, it could be very personal and delivered at the right moment, at the right time, to the right individual.

Let me try to just understand that in context. There was the beginning of [shared video], right? You’re saying there are several epochs of [online] video. There’s the beginning, which was on iOS as a “send to YouTube” button at the operating system level because no one thought that public videos were a big deal. You just needed someplace to watch videos and YouTube was that. 

And then YouTube grew into the social media juggernaut it is now. TikTok exists, all that’s there, and the initial reaction was that it’s very hard to compete with that. “We’ll be for enterprises, business customers have video needs, we’ll service them.” Are you saying there’s some kind of middle ground, in the middle of the spectrum between big algorithmic consumer video platforms and enterprise, where just a lot of regular people want private video sharing?

We have hundreds of thousands of school districts that use us. Teachers want students to be able to upload video assignments. Especially in a world of ChatGPT and AI, teachers want to see that the student is actually doing the assignment. We’ve got tens of thousands of medical professionals who want to be able to send a video to a patient without having some algorithm capture their disease state or the question that they have. We’ve got lots and lots of marketing organizations that want to be able to serve their video over to an individual inside of a big company, or to their client, and they don’t want it to be public. And then what we’re also finding is, on the YouTube front, organizations that were hosting entire YouTube sections on their website… YouTube is redirecting its customers over to some other place.

I was meeting with a really large financial services company’s CEO and the chief economist. I took them to the website and I said, “Let me show you your chief economist speaking.” I clicked on the video, and before the video played, we had to watch some advertising on some Bitcoin or something. And on the right-hand side, it said, “Here’s how the top three credit rating organizations are trying to control the world. This is the video you should watch next.” And we’re just trying to watch the economist’s video. So organizations are getting tired of being redirected or having their data captured. 

A lot of organizations are starting to bypass what I call the big walled gardens, and in a lot of ways we’re back in the MSN and AOL era for websites. That’s where we are with video right now and people want to be able to go directly to their consumers. They want to be able to serve the message in an unfettered way, and then ultimately, they just want to be able to ensure the highest quality and the most personalized experience. We’re seeing tremendous demand for those kinds of scenarios. Again, it’s everything from the school teacher, the fitness instructor, and in some cases, the faith institution, all the way up to some of the biggest companies in the world — and I do mean that literally the biggest companies in the world are becoming customers.

I think the thing that I’m keying on there is… there’s the consumer-facing video platforms and then you have a bunch of enterprise video needs and you’re in some way describing a bunch of core enterprise customers, right? Large school districts, and companies big enough to have a chief economist. I think those are classic enterprise customers and then somewhere in the middle, those companies actually want to reach consumers without participating in algorithmic media, and that consumer surface is what Vimeo has gotten away from. Are you saying you’re pushing back towards it?

We get about 100 billion views of video a year on Vimeo, and only 20 percent of it is on Vimeo.com. We show up on e-commerce platforms. I was meeting with a physician, who is an independent practitioner, a fertility doctor, and he records a video before his patients come in and he sends that video over to his patients in the fertility clinic. So we have individual proprietors, people who want to be able to share a family video or something that they learned if they had been at a doctor’s visit or otherwise — so we have lots and lots of people that are using [the platform as] individuals. 

Now on the filmmaker side… What is most amazing to me right now is the sheer number of filmmakers that I have coming to me, and this has really started kicking up since I’ve become CEO. I would tell you it’s been a trend for about the past six months. I have so many filmmakers that are coming to me saying, “We don’t like the deal… We don’t like the deal that we have with the big studios. We don’t like the fact that if we go to YouTube, as an example, they take 45 cents of every advertising dollar. Or if we want to go onto one of the big platforms, they’ll take as much as 50 cents of every dollar. Is there a way for me to be able to sell tickets to my audience?”

In some cases, some of these filmmakers that come to us have audiences that are bigger than they might get on one of those platforms and so we’re finding people who want to go direct. Our streaming business is like that. When you post on some of these big platforms… I really encourage people to look at the terms of service of the major consumer-based video platforms. It says in their terms of service that they’re able to monetize your content any way they want to. They can reuse your content, they can serve the content, and quite frankly, they’re capturing 45 cents of every dollar in that process. So a lot of these organizations want to be able to bypass those kinds of economics. 

So is that you building a consumer interface? You’re saying it’s only a small percentage that’s coming to Vimeo.com. Where are they finding an audience? Is it all on their own websites? Is it on other people’s platforms? Where is that audience actually going?

It’s been interesting because there is this trend among streamers in particular where they’ll go to the large platforms, they’ll get some following, and then when they want to be able to serve premium content. That’s when they’ll come over and say, “We want to put SVOD. We want to be able to put a gate in front of this content.” 

Or they may want to go live, they may want to go asynchronously, and so they’ll come to us and say, “Look, we want to be able to have a common library so that people can see past live events.” Plus, they want to be able to serve new content. In some cases, we’re getting organizations that want to do more interactive content — so like clickable videos, as an example. It’s a whole variety of creators that are saying, “Look, we might not want to adhere to the… let’s call it compliance requirements, the economic requirements, or the IP requirements of the big platforms. Give us an environment that we can control ourselves.”

And that environment lives on their websites?

Exactly. Sidemen is a great example of a group. They have one of the biggest followerships in the UK. They sold out Wembley Stadium in two hours and they famously played a soccer match where, when a yellow card was shown, one of the Sidemen held up an Uno reverse e-card to the ref and it blew up the internet over in the UK. They serve on Vimeo. They have both content that they put on YouTube or they put on Instagram, but then some of the more extended content they actually put on Vimeo, and then that library lives on us as well for a lot of things. Dropout’s is another great example of that. Try Guys is a great example of that. Zeus Networks, Martha Stewart — where they want a little bit more control over the content and they want control over the monetization more so than what the traditional platforms give you.

I’m going to ask you a question, and you’re just going to have to bear with me on the mathematical nature of this question. Hopefully, it makes sense. I have a lot of CEOs of web hosting companies on the show because I’m very curious about the web in the age of platforms and where the audience comes from. The last great referrer of web traffic is, as everyone knows, Google Search; Google Search is undergoing some sort of gigantic AI-powered identity crisis. Who knows what’s going over there, but it’s changing. 

So I have the CEO of Squarespace or the CEO of Wix, or whatever other hosting providers, and I say, “Why does anyone build a website? Why would you do that instead of starting a TikTok channel now?” And they all say, “Well, it’s to do e-commerce, right?” Embedded in that is some sense that, “Okay, you’ve built a following on some platform, now you want to sell something to your audience. And you have to sell the spoons somewhere, so you’re going to start Spoons.com and that’s going to be hosted on Squarespace, and that’s the way it goes.”

You’re describing the content itself as being valuable, and being more valuable when it’s hosted on Vimeo. Maybe you’re selling it, maybe you’re doing subscriptions, whatever you’re doing. But that’s happening on a website because you can’t transact that way on YouTube or TikTok. You can’t make the content valuable, but you’re still stuck with how you get the audience to come to the website that’s still just some fraction of a search audience or some fraction of conversions from one of the social video platforms. And that — this is the math — seems like the upper bound of your growth. 

Because some number of people have to come to the website, some number of people have to choose to transact on a video from the Try Guys, and that can only grow insofar as all of those individual customers can get people to come to the websites. Do you see that the web is the limiter in that way?

No, not at all. I’ve worked for Google, Amazon, and Microsoft in my life. Most recently at Google, I worked in all manners of businesses and data problems, and I have this foundational philosophy that there’s actually more data behind firewalls and paywalls than there is in front. There’s more information behind those firewalls and paywalls than in front. And when I take a look at the enterprise market for video… In the past, you’d have a marketing video and it was hard, or you might have a couple of product videos for e-commerce. Or you might have, for example, the CEO’s message. Video is coming to literally every single element of business. In the same way that it’s 82 percent of the internet, it’s coming in [to business]. And so whether or not it’s that e-newsletter, whether or not it’s for sales… You look at an organization like Seismic or Gong, that records sales calls, and then it helps to coach individuals.

If you watch a video, you’re 67 percent more likely to buy a product. And so we’ve got very large e-commerce customers where they now have millions of videos on us that are serving to every single product page on their website. So what I’m seeing is quite frankly that there is an explosion of video. It’s such an engaging medium. When you watch a video, you have 91 percent better retention than when you read something. A lot of the stuff that’s behind the firewall and behind the paywall is now getting video enabled, and it’s going across every single division inside of an organization, and it actually dwarfs what I’ll call a lot of the content.

We’re going to see video show up in so many different ways and in so many different businesses. People are starting to use video to be able to determine efficiencies inside quick-service restaurants. They’re starting to use video to be able to evaluate what’s on a shelf and whether or not there’s a stock on the shelf. So when I think about this, I don’t think about it just in terms of one segment of our organization. 

Actually, the beauty of Vimeo is that we’re able to live inside and outside the firewall, and YouTube does not live inside the firewall. We’re able to hook in and sign a [business associate agreement] to do HIPAA for a doctor. YouTube’s not going to do that. You think about all the interactions in the healthcare industry that actually can be video-enabled… And so our upper bound of growth is a larger opportunity than what YouTube is focused on right now.

YouTube, right now, is focused on video podcasts. They picked their shot and they’re going to take it. By the way, every time anyone says that stat about video retention, I feel like a dinosaur because I need to read. For as much video podcasting as I literally do, I am a reader. 

I’m an underliner. I have to underline things.

I learned to highlight with five colors in law school. It’s still where I’m at. Maybe the future is video and that’s why we do a video podcast, but I’m still a reader to my core. 

There are three parts of the video business. We’ve talked a lot about distribution, and where you might distribute that video. It sounds like that’s where you think there’s a lot of growth across organizations, even to consumer in some new way. Then there’s monetization, which I want to go to, but the first part is the hardest and I think undergoing the most change in terms of what we expect videos to be, and that’s obviously creation. You need to make a video, you need to distribute it, you need to monetize it. 

The creation of video right now is super interesting because you have not just the young generation, but everybody learning to speak the language of TikTok. TikTok, I think, is most importantly expressed to people as a video editor, not just a scrolling video tool, and it’s a very powerful video editor that you can also use in CapCut. 

Then there’s AI, which is making it a lot easier to make all kinds of videos in all kinds of ways, and then there’s the smaller AI components, like it’s going to write a script for you that you can read and maybe that’s good. Maybe that’s bad, but it’s all just in the mix. Everyone is expecting the tools to guide them. You can see in particular how TikTok tools, challenges, filters, and templates create a kind of culture that builds upon itself. Are you thinking of that component of it? Like “We need to build an enterprise TikTok editor for people just to bring them into the pipeline?”

I think there are a couple of dynamics that are happening right now. This is what gets me so excited about this… One of the biggest things that brought me here is that the barriers to video creation are dropping so dramatically, which leads to that mass proliferation of video, and then the difficulty in being able to manage at that scale. That’s just, foundationally, the market forces that are behind us. 

I always pause for a second and tell people, “I’ll be able to talk to you for a long time about artificial intelligence in a couple of seconds, but let me talk to you about what’s happening in video formats.” You’re one hundred percent right. Right now, we’re in that era when mobile video is becoming much easier. People are becoming more comfortable. Covid really helped us get comfortable with dogs barking in the background, babies being inserted into frames, and basically, I’ll just call it more casualness, in video.

Before it was highly scripted, if you recall. Highly scripted. And so culturally, people are getting much more comfortable shooting video. The proliferation of tools has been extraordinary. Now, we did make some acquisitions in the past. Magisto was an example of this because we really felt that we had to make it easier and easier to be able to create video. Well, I was thrilled with the proliferation of tools. 

We shot a video for our Reframe conference. I shouldn’t say shot a video — we actually created a video using 16 AI tools that didn’t exist 18 months ago. Over $15 billion of venture capital has gone into creating those tools and that’s just one small set. But you’re absolutely right… You’ve got all these tools that are being created and so we are thrilled about that, but simultaneously, the format of video is also proliferating, and so you’ve got traditional like 1068, and you’ve got 4K that’s starting to become more commonplace.

8K is arriving. When you do 8K, it’s roughly about six times the size of a 4K video. Well, 16K and 32K televisions are on the horizon right now. You’ve got wide stream formats, square formats for podcasting, rectangular formats, and then we just recently released Apple Vision Pro support to be able to stream on an Apple Vision Pro — which is 8K per eye, 36 frames per second. Simultaneously, while the tools are proliferating, the format types are also proliferating.  So, your ability to both accept video from any format… In some cases, you accept something that’s an old, old format that you have to get improved, or it’s a super high quality, giant widescreen format that needs to be cut up for all the different areas that you’re going to serve that video.

What I would tell you is that it’s becoming more complex for the creator to choose which tool to use and when, and then how to ensure that the right format gets served at the right moment. So the two simultaneous things that are happening in our business are creative tools and formats, and they are exponentially growing right now. They’re exponentially growing the amount of video that a creator has to deal with.

What’s growing the fastest?

It’s really interesting. As you can imagine, I think the square format is popping up a lot. We are seeing a lot of demand for 4K — 4K in live formats and in serving formats. I think people are starting to demand that format more, which is obviously for us… We have to move more bits, we have to store more bits, we’ve got to transcode more bits, and so I would tell you that’s probably the thing that we’re seeing spike the most in terms of consumption. The traditional mobile stuff is going to be there and it’s going to be constant. I think it’s almost growing at the speed of so-called mobile phones, but I’m actually surprised about how many people are coming to us asking us for 4K.

Why do you think that is? I look at the broader industry and you see the big streamers are pushing everybody to 1080p with ads. That’s sort of the default for Max or Netflix, and then you have to pay extra for 4K. Are you seeing that demand in the same way they do, which is that people will pay extra for it? Or are you seeing that demand as this is now the understood industry norm?

It may be the place that we sit in the industry. As I mentioned at the top of our talk, people have always come to us for quality. So it may just be that because we’ve been known for quality — we’ve been known for the quality of our transcoding, our stream, the service that we do — that people are not finding that kind of support elsewhere and coming to us for it. I think that people are experimenting with those formats. 

I’ve been pleasantly surprised with the sheer number of people who have come to us since we launched the Apple Vision Pro. They are coming to us with really interesting film projects to do 8K per eye, stitching all the camera work together.  You’ll see us talk a lot about this at SXSW, about what we think. I’m seeing some good excitement in those 8K formats as well, that’s all I’ll say, but it might just be the position that we sit in the industry.

It’s so interesting to see the rest of the industry basically insist that consumers don’t care about 4K. You and I are talking two days before the Super Bowl. For all of Fox’s talk about 4K, they’re still producing that in 1080p and then upscaling it. It’s fascinating to see the consumer side of the market land at one standard quality level while you’re saying the enterprise side of the market — the more discerning part of the market — is now not only assuming that 4K will exist, but that you will support 8K per eye, 36 frames per second on the Vision Pro.

And when there’s 16K for TVs out, I think people will be buying them.

That’s a lot of cost, right? I mean you’re talking about moving an enormous amount of data. Are you just getting ahead of it because that’s what the customers expect? You have a background in cloud services and big data. Is that something where you say, “Okay, this is just scalable? We can solve this problem with the tools we have”? Or do you have to build new systems?

It’s a little bit like… I’ll use the corollary of what’s going on with token size inside of AI models,  where everybody knows that the very first version of ChatGPT was maybe, I don’t know, a hundred million tokens and then it popped over to a billion tokens and will be up to a trillion tokens. So the cost to be able to deliver all of that will come down over time.

The cost of storage comes down, the cost of bandwidth comes down, and then even the innovations that are in the televisions, those costs will be coming down. When you think about the quality of the TVs we have now versus even just 10 years ago, it’s so discernibly different and I think that as those costs come down, somebody has to serve that content and our infrastructure… We have the infrastructure to be able to do it, and so for some of us, we have to stay slightly ahead so that we are that place that’s always viewed as quality. So yeah, I guess it has to be part of our DNA that we’re always going to support those cutting-edge formats.

We’ve opened the door to AI and I definitely want to talk about that and in particular, using AI as a creative tool, and how your customers might be thinking about that and their relationship to it. But first, I just want to get to the Decoder questions. We’ve talked a lot about how you’re thinking about growing Vimeo’s business. What would you say is the most tangibly different thing you’re doing compared to your predecessor Anjali?

There are a couple of things. I think Anjali was supporting a lot of different businesses. I’ll say as you went through covid and as you went through when video was hard… I would say it wasn’t as culturally ingrained as it is right now. She had to make a lot of decisions around the business. When I got here, a lot of people asked me this question: “Well, do we serve the consumer or do we serve the enterprise? Do we serve the filmmaker or do we serve the physician?” When I really spent time with who our customer was, I really had to get deep, deep, deep down inside and go, who really uses us? Show me the type of company, show me the names of the companies, and the industries that we’re in. 

It was very clear to me that we serve the creator who is professional, somebody who is using video for their business professionally. It’s not a hobby. It’s actually to get a job done. Being able to consensus around that creative pro, not trying to go and create a YouTube competitor or a low-cost tool for the hobbyist, but truly that we serve that professional creator, and then being able to describe very succinctly the fact that all of that comes together. Sometimes a professional creator wants to serve a single video. Sometimes they need to be able to manage thousands of videos. Sometimes they want to be able to go live, sometimes they need a high quality, and sometimes they need help to divide it, to cut it up into rectangular or square formats. 

I think one of the core things that I did when I got here was really obsess about the customer. Every meeting we start, we start by telling a customer’s story. I learned a lot of this, I would tell you, between Google and Amazon. Amazon is, I would say, renowned for this, but we really tell the stories of who our creators are and then really build the ability to move quickly and listen to those enterprise requirements. When you start a company, and… I started at Microsoft in the very early days of Microsoft. It wasn’t an enterprise company then, believe it or not. Amazon, when I got there, didn’t have a lot of financial services companies using the cloud. I went through that transition. At Google, it wasn’t really known as an enterprise-grade cloud when I first got there. 

So I’ve been through this transition, and when you start a product that is going to serve the enterprise, what I always tell people is that it’s easier to get more complex, but it’s really hard to be complex and get easier. And so we were starting from these roots as a consumer and filmmaker’s product, and a lot of what I’ve focused on is really listening quickly to not just our individual customers, but the entire spectrum, and being able to say yes to those requirements.

I also come with a tremendous amount of experience as you can imagine. You could look at my background — it’s years of enterprise experience, and so, I know what’s required to be able to do HIPAA. I understand how to do General Data Protection Regulation (GDPR). I understand compliance requirements. When we go into things like artificial intelligence, or we go into storage and distribution, I have a lot of instincts around that. Spending the time to explain where we’re going as a company, to be able to serve both inside and outside the firewall, and the requirements that we’re going to have architecturally, and just explaining that to the organization and weaving together… Hey, that filmmaker wants their content protected the same way that the largest retailer in the world or the CEO wants their content protected, then weaving those two messages together and building a product roadmap that’s going to serve both.

I feel like that’s probably the biggest thing that I’ve done since getting here: unifying the vision into a single cohesive vision. And then the second thing is really making sure that all of us are telling the stories. “Hey, did you know that this fertility clinic is using us in this way? Did you know that this school teacher is using us in this way? Did you know that this faith organization uses us in this way? Do you know that the largest retailer uses us this way?” Getting people to tell the stories internally, I think, was important. 

The last thing I’ll tell you is that there was a little bit of a shine that came off the company after the IPO. I think, giving the company confidence and saying, “Let me tell you who’s actually using us.” I don’t think that a lot of people really realized, internally even, the broad array of customers we have. I put this in our shareholder letter: we have eight out of the top eight big box retailers. We have eight out of the top eight media companies that are using us internally. We have huge numbers of insurance and financial services companies, companies that could use anybody, and I said they’re using us for a reason. So getting some confidence back into the company that we actually are an incredibly valuable tool in a world of increasing complexity — I think that gives the company a lot more confidence to be even bolder as we go forward.

You’re talking a lot about culture change, and renewed focus. The thesis of this show is that that comes out of structure. I’m just looking at our notes here from my producers. In the past few months, you’ve named a new chief marketing officer, a new chief information security officer, a new people officer, and a new chief revenue officer. You’re obviously making some changes in the organization. How was Vimeo structured and how are you restructuring?

I think that, in some cases, our technology organizations were incredibly siloed. How we did trust and safety as an example, in some cases how we did data, and who owned which parts of engineering. They were actually broken up in a lot of ways. I think oftentimes our marketing organization had one mandate while the sales organization might’ve had another mandate. A lot of what I would tell you was really important to me… One of my first hires was a chief technology and product officer, Bob Petrocelli. I brought in an individual who unifies product engineering and is able to have single-threaded leadership over top of parts of the business that are that important. They have to be working together well. 

Some of the things that we’re doing in trust and safety, we actually think we can turn around and expose that to our enterprise customers. It turns out they probably don’t want things on their platform the same way that we don’t want certain things on our platform. So getting any internal function to become an external function and getting that kind of view that we all serve the customer in some way, shape, or form is super important inside of our product and technology organizations. One of the most important things I had to do was bring in our Chief Marketing Officer, Charlie Ungashick, who has really extensive experience marketing to individuals and enterprise.

We are going out and talking about how we can protect videos, serve videos, and provide AI to that entire audience. So we had to get an individual who was able to oversee both parts of the business. I’ve also done some recent restructuring where we put an individual completely in charge of what we call our self-service business. To be able to move even faster in that part of the business and obsess on everything from the top of the funnel all the way through when a subscriber comes in, right down to “what are we actually using in the product”… And this is a big element as well.

I will tell you one of the other changes I didn’t mention is that we’re obsessing on use right down to the feature level. I look at those reports on a weekly basis, like how many people are using our edit feature, and how many people are using our live feature. How many people applied permissions to a video? Did that increase week over week? What did we do? And so that self-service leader is really now a single-threaded leader, and we also had a single-threaded leader around our streaming business. We really started seeing some of the results from that inside the company. Giving single-threaded leadership, I will tell you it is talked about a lot, but oh my God, it’s beautiful to actually be able to call up somebody that owns the number, that owns the resources, that worries about it as much as you do every single day.

Single-threaded leadership is an Amazon concept. You’ve worked at all the companies, so I can pick out where the concepts come from. It’s pretty fun for me. That’s a classic Amazon concept.

It is.

You need to have a pretty small team that owns the thing and there’s a leader who is responsible for the whole stack. That is how you get silos. You can look at Amazon’s product and say, “Oh, there’s a bunch of single-threaded leaders here.” This is not necessarily cohesive. Everything is running as fast as it can, but the holistic vision of the Amazon product suffers for it at scale, right? You started out talking about having too many silos and we’re talking about single-threaded leaders. How are you managing that tension?

I was asked one time to give a talk on what it was like to work at Microsoft, Amazon, and Google. I got an opportunity to work directly with Gates, Ballmer, Jassy, and then certainly with Thomas Curry and Sundar Pichai. One of the most important lessons I learned very early on at Microsoft was about really establishing a strong single-sentence vision for the entire company about what we’re trying to do in the future. And you wake up every single morning and I mean… I was there in the early days when the vision was a PC on every desktop in every home. That was extraordinary at the time. Now we have a PC in every pocket, but we all knew that what we were trying to do was unlock information for the world by putting this powerful computing device in someone’s hands.

And so regardless of the divisions or otherwise, it all feathered into a common vision. It was a lot of what I had to do when I got here. I owed the company a strong agreement among everybody in the company about what we’re trying to build. Are we trying to build the best livestream product? Are we trying to build the best marketing platform? Are we trying to build a product for filmmakers? We settled on this common vision and then we’re able to say, “Okay, this is the individual that owns this part of the business.” There’s a huge portion of our individual business where people swipe a credit card and start using us or register for free. A huge number of those customers actually end up as enterprise customers. I called up one of the top retailers and started talking to him about Vimeo and he said, “Well, first of all,” he goes, “You don’t have to tell me who you are.”

He goes, “My son is on Vimeo every weekend. He’s an independent filmmaker.” He goes, “So I know who you are, but why are you calling me?” And I said, “Well, we have 2,600 accounts, self-service accounts that are on [our platform]. We should do an enterprise agreement.” So being able to explain to the organization how the two sides work together and being able to make decisions in a room between where we’re applying more features may be in one part of the business or another… And how those features actually feather, how we might start them for an individual, but they have to grow to work for an entire enterprise — is all really, really important. 

Starting with a strong vision that everybody buys into, that they understand their piece of it, is really critical. And then for each one of those leaders, I expect them to have a strong vision for how they’re going to contribute to the overall vision. That’s another important thing. You can’t let their vision exist in the absence of the rest of the company’s vision, so you have to actively stitch those visions together.

You brought up decisions. That’s the other classic Decoder question. I will warn you: this is a honeypot for former Amazon executives. When you ask Amazon executives how they make decisions, everybody sings chapter and verse, but you’ve worked at a bunch of places. You are now the CEO of this company. How do you make decisions? What’s your framework?

The other company I didn’t talk about that I learned a lot from was Google. One of the things that I would tell you Google gave me was that they managed something like 10 out of the top 11 billion-user products in the world and were really thinking big. Actually giving an organization incredibly lofty goals, and sometimes you only reach 80–90 percent of them. One of the things that I do first and foremost is — that I really am a believer in — that you’ve got to set these very high goals. You need to have this vision and you need to be willing to put yourself out there to set extremely high goals. And then back into that from a decision-making process, I would tell you that we’ve made a number of decisions around which products we focus on, which areas we deprecate.

I come back to the customer. One of the things I really try to hold people accountable to, and I think it’s really important, I learned a lot of this both at Google and at Amazon, but actually explaining the customer problem that we’re trying to solve. And there are all kinds of studies that you can do. There are user studies, data studies, and so forth, but truly being able to assess what that workflow looks like. What are we trying to solve? What is the most challenging thing for the customer? What is actually frustrating the customer most? And really having a strong sense for your customer and the customer anecdotes as well as what we call… At Google, we called it customer empathy, actually putting yourself in the shoes of the customer. One of the things that we ask everyone inside of Vimeo to do is be a user of the product.

So the things that are frustrating us, we are elevating those into our decision-making process. We both bring the voice of the customer in and, we bring our own voice in, and then we also are saying, “Okay, well what’s going to help us grow? What’s going to grow the next million users for us or what’s going to grow us to 10X?” So I can’t just tell you it’s one thing. It’s a little bit of a framework of the customer, making sure we’re tethered to big ideas and really making sure that we’re being innovative enough in how we push the team.

Well, I applaud you for being the only former Amazon executive to not talk about one-way and two-way doors when I ask that. You’ve done it, you’ve achieved escape velocity. I mean, I appreciate the one-way and two-way doors. I’m just saying.

No, I get it. I don’t know if I’d buy that. The thing I do love, I mean, at Google, I think they’ve proven that there’s not a lot of one-way doors.

Fair enough. Google’s interesting. We’ve talked a lot about YouTube during this conversation. Vimeo has come out against YouTube. You have entire blog posts about how your search capabilities are better than YouTube’s search capabilities or how Vimeo is a better platform. You’ve talked, even in this episode, about wanting things to be private, not being part of the algorithmic ecosystem or the advertising ecosystem on YouTube. 

Google is big. They think so big that sometimes they let opportunities just slide away because they think at such a massive scale. How are you thinking about competing with YouTube at that scale when they seem to own so much of the attention, space, and video?

I don’t think that a lot of product companies love the fact that you have to go to YouTube to get some customer support for one of their products. Meanwhile, one of their competitors could be rolling right next to you. And I think that when I look at YouTube… I was at Google and spent a lot of time with customers, and I really foundationally believe… I love YouTube, I’ll watch YouTube as much as the next person. I think that what they’re doing for, I’ll call it the attention economy, for what they’re doing around content, for democratizing access to more and more content — I think it’s absolutely wonderful. And quite frankly, as I said, a lot of our customers are great YouTube customers as well. People will house their videos on Vimeo and post on YouTube in a lot of ways. But I really do think that there is… In the same way that you don’t do a lot of your business on Facebook or you don’t do it on LinkedIn, you kind of do it behind closed walls.

I think that a lot of the economy runs behind firewalls and paywalls. So I think that we can go directly at that. The other thing I’m going to say is… Think about what happened in content and why some of these platforms rose. Think about, and again, I’m old enough to remember MSN and AOL — the reason why we went there was because news had to be consolidated. It was hard to create websites. It was difficult to find information, it had to be curated. Well, Netflix and YouTube were born in an era where it was really hard to categorize content to say, “Hey, this video is about a cat,” or “This video is about how to plug an HDMI cord into the back of your LG TV.” And so there was categorization that had to take place. There was standardization of the data, the metadata, and then recommendations engines. I don’t know if you remember, but Netflix famously paid a million dollars to be able to write their recommendation engine.

They went out and said, “Whoever builds the best recommendation engine will win a million dollars.” Well, with an AI model, I can categorize content in seconds now. With a recommendation engine, I can buy recommendation engines off the shelf. And quite frankly, the metadata that we can produce now out of a video is extraordinarily more detailed than a human being can even write. I can tell you precisely when the purse left the beach, who was carrying the purse, and what brand of shoes the individual was wearing. All stuff that may be missed when a human being has to enter all that metadata. And so what I see is that there is going to be a democratization of content classification, content recommendation, and context discoverability. I think that there are searches… There is a single search place that you go to be able to get your content, videos, and maps, and pick your favorite things.

But I think that there are billions of dollars going into discovering other ways to find and interact with information. And so I think that Vimeo can serve that kind of information outside of a traditional Google search in a lot of ways, whether or not you’re on an intranet inside of a company, whether or not you’re inside an AI model and you don’t want to leave the ecosystem of the AI model. We can provide an answer to a question as well. I think YouTube is fantastic, I love it, but I think that for the discoverability, accessibility, indexing, and recommendations, there’s a whole new era coming and we intend to be a part of it.

There’s another piece of that dynamic that I’m wondering if you are considering, which is that Google is a huge company that is under an enormous amount of pressure right now. Maybe it’s so much pressure that it will be hard for the company to execute. There’s an antitrust trial in this country that resulted in the government suggesting Google break itself up and sell Chrome. There’s Donald Trump in the mix, who may or may not make some sort of deal. There’s a Donald Trump in the mix who’s done a tariffs regime with China that resulted in a Chinese antitrust investigation of Google — which is amazing because Google doesn’t really operate in that country.

Europe exists, much to the chagrin of many of our tech companies. There’s just a lot going on. There’s a lot of pressure on Google to not flex that dominance. And then there’s competitive pressure from the AI products, like ChatGPT, SearchGPT, and Bing — to whatever extent that Microsoft believes that Bing is a real competitor to Google. Does that create an opening for you? Do you see that as a real opening or is that just well, if those doors open, you’ll be ready?

I have a lot of friends at Google. I really enjoyed my time there. I don’t wish them ill in any way, and I really hope that they sail through this era of challenge for them in a really great way. Clearly, I came here from Google because I saw the opportunity. I really did see the opportunity that… We’re about to go through a seismic shift in the accessibility of information with new ways to go and access it. Whether or not you’re using ChatGPT, Anthropic, or Mistral, there are so many different ways to be able to discover information. The notion of the common crawl on the web, the ability to be able to crawl the whole web, index it, and then to be able to ingest it into these models shows that it’s democratizing access to that information discovery. 

Video is a very important element of video, and I think you’d agree. You can’t just imagine only one platform is going to serve all the video answers in the world, and so that’s where I see it’s such a super opportunity for us at Vimeo.

Let’s talk about AI. I want to start by asking a mathematical question. One of my theses for the year is that the creator economy is under an enormous amount of pressure. Not just from AI, but also from what you’re describing: this huge shift to video. You can see that there’s just an exponential increase in the supply of video on all these platforms. More and more kids are making videos. More and more people are choosing to communicate in video-first ways. More enterprises are doing it. And then you have AI, which is just making it easier and easier to produce a massive amount of video. So the platforms are getting flooded with supply. There’s not as much ad revenue as there is an increase in video supply, so you do the division and you’re like, well, the ad rates are going to go down, and then attention is sort of fixed. There are only so many people with so many hours in a day, and presumably, people do have to eat and do productive work. 

So attention is kind of fixed, right? It’s just like a fixed number that you can capture. That all just seems like it’s a bubble that’s going to pop. You flood an ecosystem that has been pretty stable for a few years with an enormous amount of supply, the ad rates go down, and attention stays fixed. Something happens in there, and it seems like, to me, AI is the most important component of that because it’s the thing that can change the economics of the supply the fastest. 

You just say, “Make 50 videos about my product,” and now we have 50 videos about the product on whatever platform. Is that an opportunity for you? That this whole creator economy, or the video creator economy as we know it, seems like it’s going to have a pretty basic shift in its economics?

That’s a huge question. 

To me, it is the question of 2025 — if you asked me, “What is The Verge doing in 2025?” There’s Elon Musk and DOGE, and then there’s what happens to the creator economy. 

I think with the creator economy, we are reaching saturation. I mean, think of your own experience. I don’t know when we’re going to get to the post-mobile phone era, but this is not a way we’re going to live for the rest of eternity as human beings. And so on the creator’s side, yes, I think that there is a saturation point, but I also think that people are looking for a little bit of a higher quality experience. 

I think people are getting tired of the doom scrolling. I think the mere fact that we name it, the fact that we are now acknowledging that we get sent down rabbit holes… I do think people will like storytelling. And I do think there’s going to be really different opportunities. I get asked all the time, “When will AI be able to take my favorite book and turn it into a movie?” And now, think about that. Think about how wonderful that would be. Think about being able to take your child’s favorite book and turn it into a video for them that has an extended storyline. I think storytelling is as old as humanity and it’s going to continue forward and so I do —

Can I just stop you there for one second? I know the Decoder audience fairly well. A lot of people just started screaming at you in their car because they think that’s a bad outcome.

Do you think that’s a bad outcome?

I have a young child. The idea that we’re going to read The Wild Robot and then some AI tool is going to make the movie The Wild Robot instead of the beautiful actual movie made by people, The Wild Robot — I would argue that that’s a bad outcome.

I think it is for a class. Here’s what bugs me the most right now. Last year the big six studios only put out 88 movies. 88.

Right, because the economics of video have collapsed on them. They don’t have a distribution monopoly.

Exactly, and I think there are so many stories to be told. If the creator of The Wild Robot gets paid for having a movie and is able to be monetized in some shape or form, in a really beautiful way, I actually think we’re supporting storytellers in a foundational way. I think that that’s a decade away. I would say maybe five to seven years away. And so first and foremost, I do think that AI is going to help people create more stories. I think they are. I think they’re going to be able to illustrate more stories, let’s put it that way. 

I talk to a lot of creative types who tell me, “Look, AI is fairly disjointed right now. It’s indeterministic. I don’t know what I’m going to get out of it.” Human curation of AI creation is going to be a necessity, in the same way that shooting on a green screen and then being able to put in a background for a movie is indeterministic until the human being decides what’s on that green screen. What I’m saying is that I do think that longer-form stories are going to be more compelling. I think people are going to want to stay inside of a story a little bit longer. That doesn’t mean that the creator is going away. It doesn’t mean human curation is going away. I just think that we’re going to be able to tell more beautiful stories in more ways. So I’ll park there because we are pro-creator, we are pro-filmmaker, and we serve a lot of them. They’re not going away. We’re going to uplift them and make them faster.

I think this is a real tension, and I see it expressed all the time. I’ve heard it from your peers on the show when they tell me about the tools that people use in Photoshop, right? Generative fill in Photoshop, according to Adobe CEO Shantanu Narayen, is like a hundred percent usage rate. But then, everyone yells about generative fill existing, and there’s a real mismatch between consumer expectations and how people feel about AI,  and then about the creatives actually using the tools at high rates. I get all that. I also think there’s a mismatch between you saying you’re for filmmakers and how marketers want to use AI.

I think we’re careening towards a world of basically custom creative being shown to individual users. [A world] where some brand uploads their assets to a video platform and ads get assembled for you in AI. For you, a specific user, [to get] ads targeted to your interests. We’re headed there and the big platform’s already talking about it. But those really commercial uses of AI — “we’re going to make a whole bunch of ads and we’re going to do some of the most creative filmmaking that exists” — they don’t seem like they’re happening at the same rate or with the same level of acceptance, or even like they should happen with the same tools, and you have every piece of the puzzle in front of you. Where do you see the biggest growth and where do you see the biggest pushback?

One of our creators, Jake Oleson, recently shot Currents for Apple Vision Pro. And when you shoot for an Apple Vision Pro, you have to maintain perfect stillness in the camera. You’ll shoot four to six cameras and by hand, you have to stitch all these things together. If you get an opportunity to watch Currents, it’s absolutely stunning. You kind of look at it and go, “Oh my God, I’ve seen the future of filmmaking. I truly have.” And I do think that this is where I say that I think we’re going to get into a post-mobile phone era for watching content and creativity. And I think that we’ll experience film in new ways. We’ll experience stories in new ways. And I think that I’m seeing the best creators blend together the content and AI usage with traditional techniques to make something incredible.

Most filmmakers that I talk to start with something they’ve shot and then enhance it with AI. One of the things that is most interesting to me is that, in the marketing world, the thing that’s taking off the most right now are not avatars (and avatars, I like to call them…  like nobody wants to talk to a robot), it’s actually people that are sitting there going, “Hey, I just bought this piece of furniture. This looks really cool. Let me show you what it looks like inside of my house.” Actually, authenticity in a world of robots I think is actually… I’m already seeing it. We offered to a number of our customers, “Hey, would you like us to do some avatars?” And then we also offered to them, “Hey, we have this super down-and-dirty create tool where you can record, we can put a teleprompter up in front of you so you can do your own script. And we can either make the avatar look perfect or you can be sitting in your living room and do this quick thing about your product or about your service.” And inevitably, all of them go to the real human being doing this. 

I’m just going to tell you point blank: I’m not seeing the robots take off. I’m not seeing it. And we’ve tried to serve both. I think that humans have always risen above. They’ve always brought authenticity. They’ve always brought through how you kind of know when you’re getting something and when you’re not. Even in the chatbot world, how many people get frustrated when they’re talking to a chatbot online? They quickly want to talk to a human being. I don’t know how to say it to you, but we sense that there are no ghosts in the machine. So I don’t know how to say it to you any other way. I studied artificial intelligence for a long time and I’m very confident in the beauty of the human soul in the context of creativity.

I feel like I’m more cynical than you, but I spend more time on social platforms, it feels like. And the problem, generally, is you can sense it. Some people can feel it, and a lot of people cannot, right? Or they just let it wash over them and then you end up in sort of interminable fights about metadata or labeling. Google just rolled out SynthID for images that you edit in Google Photos. 

None of that stuff seems to have landed. It has certainly not landed in a universal way. Vimeo has some labeling features. You have some ideas about how you might show people AI-generated content or expose that metadata to people. Do you think that’s working? Is that something you’re going to continue? Is that something that you think needs to be expanded?

When I was at Google, there were about 42 different regulatory bodies that were working on AI legislation. The last we checked, there are over one thousand on a worldwide basis. And I’m raising that to say that when you do translations in certain states inside of the United States, like Illinois or Texas, you can’t actually modify people’s lips and put words in their mouths, so that’s just one of the regulations. Over in Europe, you actually do have to identify that something’s been AI-enhanced or modified. So I do think that we as humanity are wrestling with it when we want to know that something’s not real.

The mere fact that that’s coming from all over the world, that you’re seeing the desire to know when something’s not real, I can’t say whether or not that’s good or bad, but I can tell you that it’s actually a human desire. And so, getting something done like changing the credit card on my telephone bill, I’ll deal with a bot to do that. But if I’m really having a problem, or my elderly father-in-law is having a problem, I actually do want a human being to pick up the phone and just talk him through it. I guess what I’d say to you is… I think in filmmaking as well, we’ve always used tools to tell our story. It’s been the invention of so many tools to help us tell stories. AI is just another tool to help us tell the stories. I would certainly like to know when characters aren’t real.

I guess this is the hardest question. This is an existentially philosophical question, but: Where do you draw the line? Where do you personally think the line should be for when you have to put the label on something?

I’ll give you two examples. I think when I stabilize a video with AI, that does not require it. I think if I’m talking to a deepfake avatar, or a creator that is photorealistic, that is just lying to me, that situation probably requires it, right? I’ve replaced products in this movie with other products and that may require it. Where do you think the line is?

I think that I haven’t been asked this question before, but as I reflect on it, I would like to know that a particular character, animal, or something that’s in the film is actually not real — that it’s completely made up. Now you can tell that in animation, but in a real film, if I know that a certain character or a certain scene is actually completely fabricated with a single individual in it, I probably would like to know that. Or when there’s dialogue involved where something’s talking back to me that’s not for real, I probably want to know about it. When I look at some of the Marvel movies, clearly you start crossing the line. Well, does the fox in Guardians of the Galaxy do it? We all know the fox is-

The latest Marvel movie they just announced is Fantastic Four, and they made the poster with AI, and there’s fan backlash to it. So I’m just wondering if you see the norms moving faster than the technology, or slower? You are shipping these products and you have such a direct line to creators, so it feels like you’re caught up right in the middle of where we put the labels. 

The thing that really does bother me right now is the influencers that don’t even exist. If I’m looking at something that’s clearly animated, I’m okay with it. I would love to know that somebody’s voice was actually used for real by that individual. So it’s going to be complex and I wonder out loud, will we stop caring, and at what point? Will we become comfortable that basically the whole thing is simply animated because that’s really what we’re talking about? We’re just creating animation that’s higher and higher fidelity in a lot of ways. 

But I think it should probably be noted that, at some point, the human doesn’t exist. I feel like that’s probably where I’d cross the line, or that dog doesn’t actually exist, especially if the dog’s a main character. So you might end up doing it based on the classification of the importance of the character and whether or not there’s actual, true existence there. And how much modification was done to the individual based on the class of character in the story?

I asked that question three different ways and pushed on it again because it feels like the pressure on the creator economy and the social platforms is just going to go up. With Meta, Mark Zuckerberg is out there openly saying, “We will have AI content in people’s social feeds on Facebook and Instagram.” How they choose to label it, whether or not the mean dogs have labels — I don’t know. I don’t know what Mark Zuckerberg is going to do. No one can see into his soul, but it’s pretty obvious that if he could get a bunch more cute cat videos on Facebook using AI, he’s going to do it. It’s obvious why he would do that. 

Does that create an opportunity for you to say Vimeo is for real content or real people? Is that something you would lean into? Because you do have the AI tools. When you open the website today, it says you’re an AI-powered video platform. There’s a lot of conflation between “we can do better classification and with better recommendations” and “we can do better marketing videos and we’re going to steal everyone’s data and make cat videos for Facebook attention spans.”

I think as I sit here and think about protecting the Vimeo brand, I aspire to actually be the place that people trust. When I first got here, we were approached to basically crawl our content and it was like as many companies were approached… We talked to the creators and the creators said, “Listen, don’t replace us. Just uplift us and protect us. Make sure that Vimeo is always a place where we’re protected and that’s why we come to you and that’s why we want to continue to come to you.” 

So we made a decision not to allow that crawling, and then shortly thereafter we had to say, “Anytime you use AI on Vimeo, we’re going to actually guarantee that none of the AI models will improve based on your usage of it.” Unless you say, “Improve based on my usage, like understanding my storyline, my filters, my dialogue, or my style of dialogue,” we’ll do that for you as an individual creator. We’ll create your own private AI.

We were actually approached as well to say, “Listen, we need you.” A number of the creators said, “We want you to help us identify when content has been generated by AI.” We do a lot of business over in the EU, and so we said, “Yes, we’re going to do that as well.” The thing that I describe about AI… Back in the era of the production line, humans stopped being able to keep up in a lot of ways, and so we started creating robots. We started creating machines that turned on screws and so forth. And next year at this time, there’s going to be more information created in the next year than there has in the history of mankind coming up until this point. Humans are struggling to keep up with a production line of information. So we’ve invented these machines.

I also think AI is going to help us identify these things. It will help us filter and help us be able to say, “This is AI generated,” or “This content cannot be verified from the source.” We have to do some work around what we call KYC or know your creator. In some jurisdictions that we operate in, we have to actually say, “Yes, this is a human being that created this, this is the company that created this.” I actually think we have an opportunity to serve as that — like “Yes, this was created by a real human,” which actually stands out in a world of robots. So I think there are a lot of opportunities to protect the viewer and the creator, as well as serve them in helping produce stories faster.

I want to ask one more foundational question about AI. There’s a lot of talk about cost right now in the AI world. There’s DeepSeek, which might have brought down the cost of training. There’s an argument about whether the cost of inference will drop. At the same time, Sam Altman is saying he’s going to build $500 billion worth of data centers all around the world with SoftBank. 

You use these tools, right? You’re deploying these tools against some large data sets in video, which is where the costs tend to go up the fastest. Where do you think that is? Is that working out for you? Are you making more money on the use of AI than you’re spending on it right now?

The short answer is yes. I would tell you that I expect the cost of inference to drop dramatically. We were experimenting with some of the exact same things that DeepSeek claimed to do, to be able to use really low-cost chips to be able to do inference. And I do expect inference… the cost of inference is going to go through the floor. I used to joke to say, “If I need to order a Frosty and a double burger at Wendy’s, I don’t need to wade through all of Taylor Swift’s boyfriends and songs to be able to get through that.” 

So distillations of models to be able to serve at the exact moment of time that’s necessary, whatever the language or the function is. I think that you’re going to see that distillation will help us with this. The most recent Blackwell chips from Nvidia were about four times more efficient. Lighter weight models are super important to us. So we’re going to solve a lot of the inference problems and the cost associated with inference. I’m seeing it drop dramatically for what we do, and so it’ll be very manageable over time. I think that the real big cost for a lot of these companies is the training of some of this stuff, and that is going to come in line as well. We’ll get to a point of diminishing returns, like do you really need to go to 10 trillion-parameter models, or do you need something that’s just lighter weight to be able to do chemistry, biology, security, or coloration, as an example? I think, right now, we’re in the era of big models, and I don’t think that’ll last.

Do you think we need to spend $500 billion on data centers all around the world?

We need to lay down a new infrastructure of silicon. The silicon that’s around the world right now is highly optimized for general compute, and this is a new mathematical model that has to be supported with silicon. Otherwise, we’d actually consume more power if we didn’t have specialized chips that run this math equation. So all that’s happening right now is that, yes, we need to run our current compute and now we have a new algorithm we have to run. We’re going to need optimized silicon to be able to run that extra algorithm. So the short answer is that we actually do have to duplicate the silicon around the world.

Is that something that you can drive at Vimeo? We talked about needing to serve 8K and the price of storage and compute for those falling on a pretty predictable basis. We need to invent new silicon to support AI workloads. That’s like a whole industry effort, right? And the pressure is all on maybe a handful of companies in one foundry to pull that off. How do you make those bets?

I would tell you that we’re doing a lot in evaluating quality across a whole spectrum. One of the things that, as I said to you at the very start, we’re obsessing about is how our creators really want to use AI. They don’t want to be replaced and all this. And so what we’re doing is we’re picking each one of these areas and actually establishing quality frameworks inside of Vimeo — I almost said Google — inside of Vimeo where we’re saying, “Hey, this is high-quality translation,” or, “This is what we need to do to be able to support understanding what changed frame to frame.” And so a lot of what we’re doing is that we’re saying, “Okay, what’s the best model for the job that our creators are going to need to get done?” And then under the surface, we’re stitching all that together so the creator doesn’t even know there might be multiple models that are supporting them.

We’ve established quality and then also handoffs for that creator because, as I said to you, we’re creating AI that’s going to be unique to that creator. And so we’re going to remember whether it’s over here in the translation world, over in asking a question, indexing, or otherwise. We stitch it all together for the creator so they don’t even know that we might be using multiple AI, but it’s establishing quality bars for each one of those things. And then also economics — making sure we get the best economics and best performance, like queries per second from the model providers for that area so they can serve our massive minutes of video and number of creators. We’re managing performance, cost, and quality on behalf of the creator across multiple models.

Well, Philip, as you can probably tell, I could talk to you forever, but we are out of time. What’s next for Vimeo? Give people a preview of what’s coming up next and we’ll let you get out of here.

I’m probably most excited right now, as I mentioned to you, about the massive formats that are coming at the creator. I’m super excited about what can be done with immersive formats. I’m also starting to see a lot of people who want to go back into these sphere-like experiences. I do think that that’s going to be exciting and you’ll see us continue to push the edge there. You’re going to see us invest more in the filmmaking community. 

Literally on Monday, I’m headed over to the Berlin Film Festival after hopefully, my Philadelphia Eagles do well in the Super Bowl. So you’re going to see us do even more around staff picks and celebrating filmmakers in every geography we serve around the world. I mean, it’s been fairly US-centric, and you’re going to see us get a lot more global in supporting filmmakers.

Also, I would tell you as we look over at our enterprise customers, we think we can support customers in their customer journey. This mass proliferation of video across every part of the organization in the service of customers. We’re going to do really well at just-in-time video serving just the right video to just the right person at just the right moment of the customer interaction. So you’ll see us really come out with some exciting things about that between the format and the AI, things we can do to transform storytelling.

Amazing. Well, we’ll have to have you back when we do just-in-time immersive video, AKA to both eyes at the same time. Phil, thanks.

That’d be amazing. Thank you.

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}

I’ve been the editor-in-chief of The Verge for over 10 years now, and a tech journalist for about 15 — and until recently, it felt pretty safe to assume that there was a zero percent chance that the US government would punish me for doing my job. There’s that whole First Amendment in this country, after all — we don’t usually tolerate government interference with speech, and we have a high bar for how public figures like politicians and celebrities can use defamation law to shut down reporting they don’t like.

So it’s been fairly disconcerting these first few weeks of the second Trump administration to realize, suddenly, there’s a nonzero chance the government will punish our work. All that talk about the media being the enemy of the people is turning into concrete legal action against publishers, broadcasters, and platforms that don’t do what the Trump White House wants. 

There are new examples popping up basically every day. This week, for example, The Associated Press was barred from a White House event because the AP Stylebook says reporters should refer to the Gulf of Mexico by its original name, and not the “Gulf of America,” like President Donald Trump wants. It sounds silly, but as AP executive editor Julie Pace said, the government punishing a company for speech it doesn’t like “plainly violates the First Amendment.”

And there are bigger, deeper attacks on free speech coming from Brendan Carr, the new Trump-appointed chair of the Federal Communications Commission. The FCC is pretty much the only government agency with some authority to directly regulate speech in America because it controls the spectrum used to broadcast radio and television. Carr has started using that authority to punish broadcasters for speech Trump doesn’t like or even for having internal business practices that don’t align with the administration. He’s opened investigations into ABC, CBS, and NBC in a way that no previous FCC would ever do, and he’s even launched an investigation into NBC’s parent company, Comcast, over the existence of DEI policies.

I have to disclose here that Comcast and NBCUniversal are investors in our parent company, Vox Media, but it doesn’t even really matter in this context, as Carr is hell-bent on punishing every media company unless they fall in line. In the course of my work, I’ve talked to virtually every FCC chair going back to the Bush administration, and Carr’s interpretation of what the FCC is for and what authority it has over speech is completely out of line with any of his predecessors and is as autocratic as it gets. He wants to be America’s chief censor, and so far, he’s getting his way.

This is really serious stuff, and it could have far-reaching consequences on how journalists try to report on the Trump administration — primarily by creating a real and sustained culture of fear, one that’s already ripping through the corporate parent companies of media organizations. They’ve begun rolling over and settling various lawsuits instead of fighting for our First Amendment rights.

It’s dire out there, so to help understand the scope of the problem, I brought on Matt Wood, general counsel and VP of policy at Free Press (not the Bari Weiss publication, but the much older First Amendment advocacy group). Matt has been a fixture of the open internet and press freedom movements for decades now, and his expert analysis of the history of speech in America can help us see the context of what’s happening today and why the Trump administration’s unprecedented attack on free speech is such a big deal. 

If you’d like to read more about what we discuss in this episode, check out the links below:

• Carr’s emerging agenda and its dangerous effects | Tech Policy Press
• Trump’s MAGA enforcer is having ‘the time of his life’ | The Daily Beast
• FCC to investigate Comcast for having DEI programs | The Verge
• Trump amends CBS ‘60 Minutes’ lawsuit and demands $20 billion | Los Angeles Times
• No apology over Trump lawsuit, ‘60 Minutes’ top producer says | The New York Times
• The FCC is investigating NPR and PBS | The Verge
• ABC News to pay $15 million to settle Trump defamation suit | The Wall Street Journal
• Top Trump donor wants SCOTUS to reverse press protection | The New Republic

Today, I’m talking with Sen. Ron Wyden, a Democrat and the senior senator from Oregon. He’s been in the Senate for almost 30 years, which makes him one of the longest-serving members of the institution. He’s also deeply involved in tech policy — in fact, he co-authored Section 230 in 1996. That’s the law that says tech platforms like Facebook, TikTok, and YouTube aren’t generally liable for what their users post, making it the law that basically allows the modern social media internet we have today to even exist.

We scheduled this interview with Sen. Wyden a while ago — he’s got a new book out called It Takes Chutzpah: How to Fight Fearlessly for Progressive Change, and it’s a combination memoir and call to action for young people to take part in politics. Wyden was going to come on and talk about that book.

But recent events made it vastly more important to talk about the state of our federal government and, specifically, what Elon Musk and his Department of Government Efficiency (DOGE) are doing as they seize power in various federal agencies. DOGE has some level of access in the Treasury Department and now has its sights set on the Labor and Education departments, too. Numerous agencies have had their funding frozen, with workers locked out and told to resign or put on leave, all while lawsuits are being filed and courts are telling the administration to hold on. It is utter chaos, and it’s not slowing down.

So I asked Sen. Wyden as bluntly as I could: What the fuck is even going on? And can Congress and the courts even keep up with it? Wyden did not mince words here; he called it a coup. 

You’ll hear us get into the mechanisms Musk has used to seize power in the government and why it feels like any attempts to slow this down feel so flat-footed. We also discuss whether we can trust anything these people say — Wyden basically calls bullshit on virtually every promise and declaration coming out of DOGE, especially around how much access its staffers have to sensitive data and systems. 

It’s the story of the moment and we dove into it. But if not for DOGE and its associated chaos, the biggest tech policy story of the moment would be TikTok, which was legally banned by an act of Congress and then upheld by the Supreme Court. But enforcement of that ban was paused for 75 days by President Donald Trump so the company could figure out how to sell itself. 

Right now, it’s just in a legal gray area, and that clock is ticking. So I asked Sen. Wyden, who sits on the Senate Intelligence Committee, what’s going on with that ban and if Trump can really just keep delaying enforcement of the law, which is also not a great legal situation. Wyden seems to think a deal can get done, and I did my best to get any hints about who might buy it.

There’s a lot more in this one, including some discussion of our networks themselves and Trump’s newly censorious Federal Communications Commission, but it’s probably best to just get into it. If you pay attention, you’ll notice that Sen. Wyden manages to seamlessly plug his new book throughout this episode. 

Okay, Sen. Ron Wyden. Here we go.

This transcript has been lightly edited for length and clarity.

Sen. Ron Wyden, welcome to Decoder.

Hey, thanks for having me again.

Thank you for being on. There’s an awful lot to talk about, senator. You have a new book. There are an escalating series of catastrophes in the government. There’s TikTok to talk about. It’s still banned, but unbanned in a liminal state. I want to ask you about that.

I was listening to our audience, and I was talking to our staff. Can I ask you the first question in just the bluntest way that I can think of? Can I have your permission for that?

Permission granted, bluntness.

All right. Senator, what the fuck is going on?

That is the question of the hour. It’s almost impossible to divine an answer because it moves minute to minute. Donald Trump essentially governs by whim. He gets up and out of nowhere, he will put our whole economy in play with a really colossally stupid approach to tariffs — Tariffs that he has lied nonstop to the American people about for well over a year. He always tells them that foreigners will pay the tariff. That’s factually wrong. All of the people who are listening to your program, the fans of the good work you all do, they pay the tariffs. So, this weekend was just perpetual bedlam, and foreign leaders were basically laughing at us. They were making fake promises. They knew that Trump just wanted to be able to get some kind of headline.

I’m sure we’ll talk about other aspects of the bizarre nature of these times, but you probably know that last Friday, I blew the whistle — as a result of getting input from whistleblowers — that the Secretary of Treasury had signed off on giving Elon Musk access to the federal payments program, which has enormous implications for hacking, abuse, and violations of privacy. The list could go on and on. So, that’s just a little bit of the last few days.

Let’s start there. Again, just hearing from our audience, our listeners, and readers of The Verge, there is a sense that what is happening right now is Donald Trump is sitting in the White House, signing executive orders that may or may not be legal, may or may not have impact, may or may not mean anything at all, talking about building hotels in Gaza. And that is all a distraction from Elon Musk and DOGE taking over the government. We’re all the way at “is this a coup?” I really want to start there. Is that where you’re at? Do you think that there’s a coup going on?

I’m very troubled by the types of policies they are pursuing, that when you add them up, look and feel like a coup. The reality is we are working very hard to derail them at various steps. Let me give you a couple of examples. Last week, Donald Trump basically undertook a kind of freeze of the budget, and he wanted to do it unilaterally. There was a lot of back and forth about what authority he might possibly have. We knew there wasn’t any, but we heard him out. Then we went to court, we won, and he had to drop it. 

Let me give you a small snapshot of the kind of work I do. My background is working with senior citizens. I was the director of the Gray Panthers for about seven years before I got elected to Congress. So, during the freeze, I found out that the Medicaid portals, which they have in all states, were essentially off. They weren’t available for the senior citizens and the disabled to get information about their medicine, nursing homes, and the like. 

So, I went online immediately, and said that the whistleblowers, which were truthful, had told me about the tremendous problems being caused for the vulnerable, seniors, and the like, and that it was important for citizens to get in touch with their elected officials, community groups, and organizations and talk about why this was so unfair. Nobody voted to cause harm and discomfort to seniors and the disabled. And there was this outpouring of calls and letters. Within about three hours, the Trump people were backpedaling. By late at night, they said that it would all be fixed the next day. 

This is just a small snapshot from one day, but people mobilized. The whistleblowers that I work with, particularly on senior citizen issues, came through. This was something where Trump and Musk were stopped cold as they tried to abuse the federal budget. It’s a small snapshot, but it’s the kind of thing that I tried to help with. If I were to embrace it in a sentence, apropos of what the hell is going on here: Trump and Musk are ignoring the checks and balances that are the foundation of our republic. And for all practical purposes, I’d call that a coup.

When you think about the mechanism that you just laid out, you heard about something that was happening that was bad because of a policy that was hastily rolled out, you mobilized some attention, there was a swarm of negative feedback for the policy, and eventually, the administration backed down. Maybe it’s fixed, maybe it’s not. This is an open question. Programs are still being unfunded, right?

I will tell you on this point, we check every day to make sure that seniors can get their Medicaid, their medicine, and access to nursing home information. But your general point is correct. There’s still a flood of abuses in a whole variety of programs across the government.

Can that mechanism of “I need to blow the whistle, then maybe a flood of outrage that’ll fix it, or maybe we’ll file a lawsuit and our courts will fix it” keep up with the pace and scale of the chaos happening right now?

We’re going to use every possible tool: court actions, legislative actions, whistleblowers, the bully pulpit. We’re going to use it all. We are still trying to find our way through the array of remedies that we have. When my wife agreed to marry me. She said, “You aren’t a real lawyer. You worked with the Gray Panthers. So, I’ll marry you because you’re not a real lawyer.” We’re now looking through all of the legal tools to make sure that we can protect our rights. But it’s not going to be a walk in the park. In some of these areas, like with the federal payments, nobody ever even thought of doing what Trump is doing, which is why it’s so important that we stop him.

Let’s talk about that for a second. Do you know how much access Elon Musk has to the overall federal government right now? Has that been quantified for you or is it just ever increasing because it’s chaos?

We believe that it’s ever increasing because to a great extent, Mr. Musk tells us that it is increasing. There was a big story that there wasn’t much being done, just general information. I got asked by the press about it late last night, and I said, “You got to be kidding me. Just look at all the stuff that he claims he is doing.” The first thing he went after was a charity. It didn’t have anything to do with waste, fraud, and abuse.

I believe that we will move quickly in the days ahead to build more tools for a legal strike approach. We’re talking about that as members of a Democratic caucus. I’ve been tasked particularly with promoting privacy for those concerned that their data would be peddled hither and yon in the payments system. I can tell you that from a privacy standpoint, giving out people’s home addresses and the like, this will be a horror show.

There’s a big question right now over whether Musk and his team’s access to that system is read-only or if they have the ability to write code. They have said it’s read-only. There’s some reporting that says they have the ability to write code. Do you know if it’s read-only?

I believe that this is data that’s being abused now. Part of this is also semantics. I’m not taking anything off the table in terms of what they’re doing. They can write code. There’s no question about that.

So, you think this argument over whether or not it’s read-only is a semantic distraction? That they do have full access to the system?

Correct. There is no question that this description is designed to lull people into thinking that Elon Musk is not doing what he tells people he’s doing. I mean, he tells people that he is doing all these things, and then they put out these cover-up stories about how he’s not really doing it.

We have a very technical audience. They’re very worried that we have increased the surface area for cyberattacks–

No question about it! Let me save you some time. No question about it!

Have you heard this from our intelligence agencies? Is this something Congress is worried about?

I am on the [Senate] Intelligence Committee, so I can’t say anything about that.

Is the Trump administration broadly being responsive to these concerns? Because we are at risk of cyberattacks. It’s a pretty bipartisan concern.

No, the Trump people aren’t telling us anything about anything.

Do you think they know what Elon is doing fully?

You asked me if Trump just sits in the office and issues executive orders. I think that’s largely what’s going on. I’m sure they have some conversations, but I don’t get the feeling that there are any checks and balances on Elon. That’s for sure.

There’s a weird gap in every Trump administration. I think it’s even weirder here in Trump 2.0 where Republican officials will not criticize the president in public. Are they understanding or aware of the scale of the problem in private when you talk to them?

They’re saying very little in the consideration of the Robert F. Kennedy Jr. nomination. For example, yesterday I brought up that our committee, the Senate Finance Committee, had passed a bipartisan measure to rein in these pharmaceutical benefit managers (PBMs) that are ripping off seniors, taxpayers, and people of all ages purchasing medicine. I pointed out that Medicare and Medicaid are under the jurisdiction of our committee and are subject to the kind of abuses that could happen with Musk and the federal payments.

And they looked up. It was clear that what I said was registering them. I said, “This is important. This is something all of you have voted for.” But they’re not speaking out. In fact, yesterday, there was apparently some kind of Republican show trial where they basically all lined up and said, “Nothing to see here.”

Do you think that holds?

Let me put it this way. There’s no way this will be something they can just continue to stonewall. Let me give you an example. I announced five days ago that the Secretary of Treasury had given the keys to the kingdom, so to speak. Yesterday there were thousands and thousands of Americans outside the Treasury Dept. speaking out against these practices. Something like 30 members of Congress addressed this, making it clear that we were going to keep at it until we secured the protections that our citizens deserve. That happened all within five days. 

I give great credit to the whistleblowers. We heard a rumor that there was some activity going on with the Treasury payments, and the whistleblowers were the ones who specifically told me that Scott Bessent had signed off on giving the keys to Musk.

That set of whistleblowers is largely the federal workforce, which is appalled. They’re filing their own actions from their unions as private citizens — just worker protection lawsuits. They’ve also all been told to quit their jobs, right? They’ve received these letters that say, “You should resign.” Musk famously used “fork in the road,” which is from a letter he sent when he took over Twitter. That may or may not be legal. There’s some big questions around that. Have you been hearing from that constituency?

Yeah, they’d like some answers about it. You can hear the headline, “Oh, we’ll buy you out. We’ll give you all these opportunities to make a transition.” But most of the experience that the federal workforce has had with Trump — and it goes to the first administration — has not exactly been confidence building. They’d like more details about how they and their families will be affected. This is not just their current and future income, but making sure that the benefits they’ve accrued get treated fairly.

What advice would you give to those workers that have received this offer?

Proceed with great caution.

Do you think that’s a farce as well? Do you think this is just a cover for mass layoffs? Is that something you’re worried about?

Yeah! I don’t trust any of this stuff. I don’t want to sound like the old-fashioned version of a broken record, but I don’t trust any of this stuff. Often you go to these meetings and they act like they’re on the level, and then they go and do everything they want. That was the first administration, so I don’t trust any of this. When it sounds appealing and you can’t get specific answers, that ought to be a wake-up call to make sure you check. The law limits his ability to make mass layoffs, so that’s a big part of the debate too.

We’re talking on a Wednesday. We usually publish Decoder interviews on Mondays. Usually, that’s fine, right? There’s two business days between us talking and publishing the conversation. Usually, not a lot happens in two days.

You can be sure that a lot’s going to happen tomorrow, Friday, Saturday, and Sunday. Tomorrow, we have a hearing on international trade. I described some of the bizarre practices that they’ve pursued in this space, like how they keep saying that the foreigners are going to pay the tariffs but it’s going to be your listeners instead. So you bet, Thursday, Friday, Saturday, Sunday, there’s no days off here with the Trump whims, as I call them.

The specific question I had there involves Elon tweeting, “I work on the weekends when the other team is not in the field.” The other team that he’s describing is the federal workforce, so that is already problematic. But Congress goes into recess. People leave town on the weekends. Are you and the Democrats ready to be here all weekend to see what he’s up to?

They’re making a big deal out of saying they want to work around the clock. Look, I told you that I have this book, It Takes Chutzpah. I’ve been using some of the lessons in It Takes Chutzpah all week to try and tackle the Trump administration. I was flying across the country and taking red eyes to make sure I didn’t miss a minute in the effort to blow the whistle. That was what we were doing late on Friday night, in terms of Musk being given the keys to the Treasury payments. I know the Republicans are leaving town this weekend, so they managed to find ways to set an agenda that’s favorable to them.

Has the DOGE team been responsive to you or other Democrats at all throughout this period?

Absolutely not. I know no one who’s being briefed on the Democratic side consistently. At all, really. 

Is there a way for you to get that insight as a member of Congress?

We’re continuing, for example, to follow up with individuals who used to work at the Treasury and tap them for their knowledge. That’s been very helpful. Obviously, I can’t get into that.

Are you planning to find a way to hold hearings on any of the committees you’re on or call Elon as a witness on any of the committees you’re on?

We’re going to use every tool to excavate the information that the American people deserve. This is one of the most important issues I’ve tackled in my time in public service. It’s actually one of the first rules of Chutzpah that I spell out: “if you want to make a difference, you have to make some noise.” We’re making plenty of noise right now. We just have to keep following through until we get the job done.

One of the weirder aspects with Elon Musk in all of this is that he owns a communications platform, the preferred communication platform of many members of government. Is that a problem, that he can command attention, that he can run that algorithm any way he wishes? Do you think you have to go somewhere else?

There are certainly questions with respect to how he runs his operations, and there’s certainly some questions involving China as well. We’ll have some more to say about that before long.

You are famously one of the authors of Section 230, which protects service providers from liability for the content their users post. Elon is the service provider. He’s also the user that is posting the content. Do you have any newer thoughts about Section 230 in all this mix? Do you think it’s doing the job the way it’s intended for it to be working?

I’ve said this on your show, but I’ve always felt that the big guys — of which Elon is about the biggest — always have the biggest checking account to be able to buy what they want. Elon Musk just bought a presidential campaign. He basically was the big giver. Section 230, as you and I have talked about, is all about making sure there’s a voice for the little guy: the #MeToo [movement], Black Lives Matter, the people who probably couldn’t get a message out without what Section 230 involves.

I think that Section 230 continues to be an immensely valuable tool for the person who doesn’t have a political action committee, doesn’t have power, doesn’t have a deep checkbook. We’re going to have a problem reining in the Elon Musks no matter what because they can buy the kind of government they want… at least until we get rid of things like Citizens United.

The other rich guy in the mix this time is Mark Zuckerberg, who also owns several large communications platforms. He’s notably changing his moderation standards to favor the Trump administration’s desires. He’s allowing more hate speech, particularly against trans people, on his platform. That’s something that Section 230 protects, right? That platforms can moderate however they want. Ideally, you have some competition for moderation standards to push and pull and give people places to go, but we don’t really see that today, right? We see sort of a herd mentality.

There’s some competition. Anybody can sue Elon right now. He posts lies himself. Section 230 protects Elon’s competitors against Trump’s FCC. There are protections, but the bottom line is that when you have somebody as rich as Musk, who literally bought a presidential campaign, you’ve got to use every tool you got, and a lot of them are not Section 230.

There’s another platform in the mix that seemed very relevant during the presidential campaign, TikTok. It was banned, the law passed, the Supreme Court upheld it, and Trump paused enforcement of it. Do you think that Trump’s enforcement pause on TikTok is legal? Is that just him pushing his authority some more?

I don’t think it’s legal, but here’s what I think this debate is all about. The reality is that from day one I had the same position, unlike Donald Trump. Donald Trump was against it, and now he’s for it. I always said it would be a good thing to have an American company run TikTok and have them tightly vetted. That would, for example, have limited [former Treasury Secretary] Steve Mnuchin and others who talked about getting access to it. I’m still in favor of that as a policy issue. The question now is what the courts are going to say given that there’s been an act of Congress, but I have not changed my preference once since the very beginning.

What were the problems with TikTok under ByteDance that required that original ban-or-sell law?

I’m on the Intelligence Committee, and the issue is the closeness to the Chinese government. That was a national security issue.

There’s a lot of young people today who would say, “Sure, but then, Elon Musk is going to buy it and we see what he did to X. That is as much of a national security problem today as the Chinese government owning the algorithm.”

Musk could not pass the Ron Wyden test because you would have to have a reputable American buyer who’d been thoroughly vetted. I don’t think Musk, with all his China connections, could pass the test that I’ve described.

Have you heard about any buyers for TikTok so far?

Rumors, but nothing that I’m going to splay out in front of a national radio audience. The reality is Trump, as he usually does, just governs by whim. He starts off by saying, “TikTok is bad.” Then he learns that there are a bunch of young people who like it, and suddenly, he’s trying all of these exotic ways to deal with it. 

I said from the very beginning that my first choice was to have an American buyer who could pass rigorous vetting to show that they were independent and would meet public interest standards. I don’t think Trump’s interested much in that. In fact, you’d think that at some point, he’s just going to figure out a way to hotwire a deal where he sells TikTok, in some way, to Musk directly. And if he does, that’d be about the most corrupt example I could find.

That vetting process is not contemplated in the bill, right? It’s an interagency process. 

Yeah, this would take extra work to lock it in place, but that’s been my position from the beginning.

The reason I’m focused on this is that I see a lot of danger of algorithmic ownership. You see how Musk has managed to distort the conversation on X because he owns that algorithm. I don’t know how changing the content moderation standards on Facebook and Instagram will play out, but my expectation is a lot of minority and queer creators are going to face a lot more harassment on those platforms because the rules have been changed. The algorithm will allow more things to happen. If the danger of TikTok is, the algorithm is owned by somebody–

Hey, can I just make one point I think I may have misstepped on? I am not for the ban. I am for finding a way to get this right with an American buyer.

Right. I guess my question is if divestiture was required due to Chinese ownership of the algorithm, which would run counter to US interests, do you think that having some amount of algorithmic transparency or privacy protections would solve that problem with American buyers as well? From our audience, what I hear is that all you really need is algorithmic transparency, and then, China should be able to do whatever it wants.

Not only do you need algorithm transparency, you got a guy who’s authored a bill that moves in that direction called the Algorithmic Accountability Act, my legislation.

Under a Trump administration with this Congress, do you think that will go anywhere?

We’ll see. A lot of people now understand that algorithms are pretty complicated, but when you talk about protecting people’s education, healthcare, and jobs, they get interested.

Sure. Trump gave 75 days, which is just a number he apparently made up, and if we hit the 75 days and he says, “I’m going to extend it again,” do you think there will be litigation, at that point, about whether the law is actually in effect?

Hard to say.

Would you support that? We can’t just keep indefinitely postponed enforcement of a law?

I think I’ve described the path to get this done, and I’m going to keep working for that.

What do you say to TikTok creators right now who don’t know if their businesses will last another 50 days?

I’d tell them I’m busting my chops to get them something resembling an American buyer who’s been carefully vetted and protects their interests and all American interests.

Do you think Microsoft is a good buyer? For the audio listener, Sen. Wyden looked away first before he began answering that question.

Put me down as pretty dubious, but we’ll see.

All right. What about Oracle?

No.

All right. What about Frank McCourt’s Project Liberty?

No.

No, across the board.

Right. I think we ought to bag it on these kinds of folks.

I’m just curious. Okay. Amazon?

I’m just not going to get into a lineup evaluation batter by batter.

In the tech world, this is everyone’s parlor game beyond “does Elon run the government.” Those are the two things everyone talks about at the parties I go to. So I’m just curious. It has to be done. The clock is running out.

I think we got pretty far down the road in terms of Elon running the government and what I’m going to try to do to stop him. We probably have a ways to go in terms of TikTok, but I think I’ve spelled out what I think ought to be done. At least it’s consistent from day one.

Last question on TikTok, and I promise I’ll move on to the FCC because I’m very curious for your opinion there as well. The clock is running, right? There’s only 50, 60 more days. Do you think the deal can be made in time?

Yeah, but you’re really going to have to hustle to get the right kind of parties and go through the process. But yes, it’s doable.

Let’s talk about the FCC again. You are a First Amendment advocate. You have said to me on this show, to people on other shows, that Section 230, which you co-authored, is really about the First Amendment, and most people’s problems with 230 are really about First Amendment-protected speech. The First Amendment, from my point of view, is under the greatest full-frontal attack I’ve ever encountered as a journalist.

Could not agree more.

It is bizarre to me. It’s particularly bizarre because it’s wrapped up in the language of free speech. In particular, that language comes from  new FCC Chairman Brendan Carr, who is attacking CBS, NBC, ABC, and radio stations about how they’re using their spectrum and threatening to take that spectrum away. Is Congress paying attention to this? It feels like a sideshow, but to me, this is the most overt government speech regulation I’ve ever seen.

It’s not getting a lot of attention in Congress, but I’m doing everything I can to light the place up around it because I share your view that a big chunk of what the Trump crowd and the far-right want to do is act in the name of free speech. They make all these proclamations about how they’re about free speech, but they’re trying to do everything possible to deny the voices they don’t agree with.

Have you spoken to the chairman about the FCC? He put out a blog post today saying that he’s worried about TV commercials that are too loud, which sounds like it’s straight out of the 80s. At the same time, he’s opening inquiries into CBS News. Have you spoken to him? Have you pushed anything?

I have not since his appointment, but it’s very predictable. They find something to try and make sound relevant, but it is completely irrelevant to the big question. That makes it less likely the public’s going to pick up on the big question. I’ve had 1,103 town hall meetings, and you can read about them in It Takes Chutzpah. They’re open to all, and I don’t give any speeches. I just show up and say, “For the next 90 minutes you can say your peace. Ask me any kind of question.” It is, to me, my vision of the First Amendment.

My dad was a journalist and an author. Right next to me is one of his books on the Bay of Pigs, with a picture of Fidel Castro next to my dad in the book. The caption is “Peter Wyden knows more about it than we do,” coming from Castro. So, I’m a journalist kid, and I think the Founding Fathers, who said that a free press was more important than government, got it right. And I believe that.

Normally you would see these big companies look at an FCC (particularly a 2025 FCC) complaining about news distortion over their broadcast spectrum, and they would just bat it away, right? They would send the lawyers, the lawyers would fight, and they would bat it away. Meta was sued by Trump himself for banning Trump while he was president, saying that was somehow a state action. Very confusing. Meta settled this case and agreed to pay Trump $25 million for his library. It seems like CBS is going to end up settling the 60 Minutes case.

Hey, let me give you the real–

But I’m just asking, the big companies are caving on the First Amendment. They’re the ones that should be fighting.

They are caving like crazy, and nobody has caved more than Mark Zuckerberg. So Zuckerberg goes to Mar-a-Lago, and he tells Trump what he wants to hear. There aren’t going to be any fact-checkers. “Okay, Donald, just what you wanted. No fact-checkers.” He scores some points with Trump. My guess is that seeing how this was successful for Zuckerberg, a lot of the other big players are going to go down to Mar-a-Lago and look for favors also. 

In the process, what’s going to happen is that we’re going to re-consolidate big communications and have even fewer small voices than we had before the internet came into being. We ought to be protecting those small voices. What Zuckerberg is now a part of is doing more to help the big voices, and in effect, push the little guys out of communications.

That visit by Zuckerberg to Mar-a-Lago and what he did with the fact-checkers was a god awful thing to do for the future of communications.

But if Zuckerberg caves, Disney caves, and Shari Redstone caves because she wants to sell Paramount to Skydance so she rolls with the CBS lawsuit. No one shows up in court and says, “Actually, these threats violate the First Amendment.” Don’t we just lose the First Amendment? Doesn’t the pressure just work? How do you push back?

We’ll have to find ways to crank up the pressure, and I’m going to do that. You have people like the Electronic Frontier Foundation (EFF) that might step in, or other kinds of groups.

ACLU.

The ACLU, EFF, Ron Wyden, and those are your three. Sign me up.

The argument I’ve heard from the other side — I’m not saying I agree with it, but I’ve heard it and I understand where it comes from — is that these are mercenary moves. Meta spending its time and energy fighting a First Amendment lawsuit instead of just paying the money runs the risk of them losing, right? It’s worse for Meta to roll into courts — especially all the way up to the Supreme Court, which is 6-3 conservative — and lose a First Amendment lawsuit than to just pay the money and leave the law effectively where it is, even if the norms change. Do you buy that argument?

No. I think that both of them are harmful, and I think Zuckerberg is looking at buying whatever influence he possibly can because he sees it works. It’s not that complicated. Write big checks and you get influence with Trump.

Carr is now targeting NPR and PBS. He’s saying that they’re technically running ads by asking for viewer support even though they’re publicly funded, and that he should look into their content. Is there a way for us to try to protect NPR or PBS?

The best way to do it is by statute, and we’ll see if any of these rural Republicans whose farmers really like those services in the morning are going to stand up for them. What we’re describing is continued good business for the big guys, and they’re doing all the things possible to run out their competition. That includes virtually everything you can see. I’m trying to find new ways to get voices for people who don’t have power and clout join with a Republican trying to get more streaming services. So we’re busy looking for all the options, like getting C-SPAN out and streaming.

You’re talking about rural farmers. You spoke earlier in the show about the outcry from people and Republicans backing down. Are you seeing the policy hit on some of these Republican constituents actually driving change? Are you able to reach those populations? I think a lot of people are worried that they’re fully captured by whatever TikTok right-wing influencer they see.

It’s very early. I gave you some examples where we had some effect, but let me give you the bottom line here. I think very rarely does political change start in Washington, D.C. and then trickle down. I believe it’s almost always in the exact opposite direction. When people at home get concerned about access to communications they care about and whether they’re getting clobbered by tariffs, they go to their elected officials. They’ll see them in a grocery store, they’ll see them at the gas station, they’ll see them someplace and say, “What the hell are you doing? I’m getting clobbered on that.” And you end up making change. It’s a grassroots, juggernaut setting in that really changes things.

I probably use that as much as any legislator. The way we were the architect of the PIPA-SOPA movement when all the power, all the money, was out trying to redesign the internet for the big guys. I organized a lot of grassroots groups, and [Sen.] Harry Reid set a date for a vote, and suddenly, 15 million texts and emails came in, and we beat all the money.

You’re saying, “Do you see all of this happening yet?” When I was with the activists yesterday at the Treasury talking about trying to rein in Musk, I said, “This is really stunning.” This is less than five full days after I blew the whistle on the problem. And look, we got thousands of people and 30 legislators in the streets. So put me down as continuing to believe that citizens who mobilize can make a difference.

That’s what I describe in the book. The book, for example, gives 12 rules, “Ron’s 12 Rules of Chutzpah,” that describe not just how people are looking at past examples, but how you can use the rules outlined in the book to make a difference. That was the first rule that we used yesterday: fight bad things, make a difference and speak out for good things, and above all, make noise.

The book is really about pushing for progressive political change. You have a history of doing that. I think the question on everyone’s mind right now is when the other side doesn’t seem to care, and you can’t reach their constituents because they own the media platforms and their algorithms are driving all kinds of views that honestly just dismiss any concerns or turn the Democrats or any opposition at all into cartoon characters, how do you drive that kind of grassroots change? You have a lot of people who voted against their own interests here. They voted for cheaper egg prices and they’re getting Trump saying they’ll understand when terrorists drive prices up.

Yeah, but they’re also catching up in a couple of weeks. Musk’s popularity numbers are really low. They’re finding out that Donald Trump tells them, “I am going to lower the prices of your groceries” — he said that point-blank — and then, he puts up a white flag of surrender a couple of weeks later and says, “Nope. Can’t do anything about it.”

We’re going to build every single day between now and 2026 as many wins on the issues and matters people care about. It’s not sitting around and waiting. When we did that during the first term, we gained a few seats at the midterm. Two years later, we beat them all together. This is not going to be an easy walk in the park, but I’ve been part of efforts where I was able to beat incredible odds, which is what chutzpah is about. Chutzpah is about being bold, taking on big odds, and reaching out and finding people with all points of view. That’s what we’re going to be doing.

Can you do that when you don’t own the communications networks? Just to boil the question down. Elon Musk owns Starlink in a world where–

I’m talking to you, and you guys get the message.

That’s true.

You guys get the message out, and this guy’s working to make these changes that we’re talking about. He’s got a book that tells you how you can do more of it.

There you go. I hate to hammer on it, but you can see the bigger picture. You have a Brendan Carr at the FCC putting pressure on legacy broadcast media. You have a Mark Zuckerberg and an Elon Musk telling their algorithms what to do, what kind of media to disseminate, and what’s acceptable speech and not. You have a TikTok sale in the offing that could go to a Trump loyalist. Is there an alternative media? Do you look at Bluesky and say, “Okay, decentralize social media. Provide a necessary check.”

I’m on Bluesky. I love it. But if you want somebody who’s just going to say, “All I’m going to do all day is hang crepe because it’s hopeless,” I’m the wrong guy. I’m a 29-year-old Jewish kid who wanted to play in the NBA. That was a ridiculous idea. I got a scholarship, but I couldn’t get that dream, and I ended up being one of the youngest people elected to Congress by beating an incumbent. 

Guess who else did that not very long ago? [Alexandria Ocasio-Cortez], again, beating incredible odds. So, you’re talking about ways where you can really make a difference and beat the odds. Frankly, it takes a lot of chutzpah to do it. People say it’s hard to beat the powerful and the money, but I just gave you some examples of doing it.

There’s a big split right now in the tech community. You have a lot of regular workers at the big tech companies who are horrified at what’s happening, and then you have their bosses who all showed up, sat at the inauguration, and are trying to get various things out of the Trump administration. As you said, I think they realize they can just write checks and get what they want. 

But their interests are not aligned. I don’t think Tim Cook’s interests and Mark Zuckerberg’s interests are aligned. I don’t even think they like each other. I don’t know if Jeff Bezos and Elon Musk like each other. I think they’re pretty ferociously competitive about their rockets and general cowboy demeanors. Is that a split that you see that can be exploited for political gain here? That the interests of tech, which has now taken over, are not actually aligned?

If you look at the chutzpah rule, you say you’ll work with anybody regardless of political philosophy to get the public interest. I think that if Trump equals chaos, we’ve got more to work with than people think.

Have you heard from that constituency at all? From the big tech companies? What I have heard from their executives, very directly, is that they thought the Biden administration hated them, that they could not work with Biden at all. They couldn’t even get the meetings they needed or meetings with the right people. What they generally got was, “We’re going to break you up,” and then there were a lot of lawsuits aimed at breaking them up. What you might still get out of the Trump administration is, “We’ll break you up, but you might be able to buy your way out of it.”

Let me tell you what I think a bunch of these big tech companies are saying. Yes, they didn’t like the policies of the Biden administration, but they’re seeing more and more particularly young people — you see it in schools, on TikTok, and the like — who want to have their own businesses. They want to be entrepreneurs. They want to be the future. I think that big businesses who don’t pay attention to the small guys that have ideas and creative approaches are making a mistake because that’s who I’m throwing in with. I’m throwing in with the little guys because I think they can make a big, big difference. 

Biden was a big opponent of mine in technology for 30 years. I don’t think that the Democratic Party is about being for Big Tech. I think being for small business is going to be the future of the Democratic Party. In fact, small businesses and users. The two people that I fought for in the tech space are going to be my vision for the Democratic Party going forward. You can count on that.

Is there any lesson you took away from DeepSeek? The big tech companies are spending or want to spend enormous amounts of money on data centers, on energy–

The real message is how important open source is. Open source communications have got to be in the ballgame. I’ve continued to push them through all my time in public life, and it’s even more important today than it was.

The other lesson there is that restricting chips to China didn’t work. The Chinese companies were able to take somewhat underpowered Nvidia chips, over-optimize them, and then outcompete.

Because I think we’re wrapping up pretty soon, let me just mention that the Chinese are engaged in all kinds of abusive practices. Salt Typhoon, for example. The phone companies were supposed to have secured the phones 20 years ago, and I’ve been screaming my lungs out about it. I’ll be introducing another bill on this because when this latest set of breaches took place, a bunch of conservatives in the Congress said, “This is the biggest hack in American history.” Well, I’m going to make sure that we get another debate on this whole set of abuses. My bill will be going in soon.

Salt Typhoon didn’t collect a lot of attention, but this is a massive breach of the American phone system.

Stunning breach!

The Biden administration was telling people to use secure and encrypted communications. Why do you think that went so unnoticed?

I think that a bunch of the big companies you mentioned, like Microsoft, all said that they were taking care of things and tried to lull people through it. It wasn’t until the Chinese actually got away with it that anybody woke up and said, “Oh my goodness Mildred, we’ve got a problem.”

On the same note, the notion that we would ban them from having our chips and that would keep them from competing in AI? I don’t know if that worked. I am looking at DeepSeek and saying, “Well, constraint breeds creativity. They got really creative, and they built a cheaper model that performs as well as our most expensive models.” Do you think that it’s time to rethink that approach to technology limitation?

I think that’s a fair judgment. We understand that we have national security interests, and we’re certainly trying to compete with these people all over the world. And rethinking this is a sensible idea. We can’t stop China from getting chips, but we can certainly notice if it’s more expensive. We can make it more expensive.

To bring that all the way back around to the top. Elon Musk, who may or may not be running part of our government, has significant ties to China. It’s a huge market for Tesla.

You’ll hear a bit more from me on this very soon.

Can you give us a preview?

Nope. I’ve got to go through all the security rules and all the rest, but you’ll hear more about it soon.

Do you think whatever you have to reveal will provide the necessary momentum to stop the chaos run of Elon Musk through the government?

I think people are going to pay attention to it. It’ll be an important part of the debate.

Sen. Wyden, I think we’re out of time. Thank you so much for being on Decoder. This was great. Thank you for letting me ask that question so bluntly at the top as well.

No, I appreciate it. You guys, really care about policy issues, and thank goodness.

I’m going to ask you for your advice, and I mean this sincerely. Last question. The Verge has always only cared about policy issues. This has been a thing that we say. We write about policy, not politics, because policy is important to people. It’s real. But it seems, here in 2025, politics and policy are the same thing. Do you feel that?

It certainly makes me reflect on the policy that I’ve always said. It comes up at town hall meetings all the time. I say, “The best politics is good policy because you get in a position to be able to win people over for the right reasons. The reality is that these are really serious questions that don’t lend themselves to quick thought. We have to be ready to babble on both fronts. Let’s get the best policy ideas that we can and then work like hell to have the grassroots support in order to get them passed.

But here’s my pushback. It seems like right now, the politics are winning. The politics are about collecting as much attention, being as loud on social media, and being as “authentic” as you can be. That lets you do whatever harebrained policy idea you want, including announcing a takeover of Gaza from the White House. That feels like a meaningful split. How do you bring that back into alignment?

I’m going to continue to say that it’s possible to make this temporary and we can move on. But it’s only going to come about with good ideas and political organizing that will have to mobilize people all across the country. I mean, you can’t fake the price of eggs. Donald Trump’s got a problem, and if we pound that issue, a specific policy matter… I am going to talk until I run out of larynx space about him saying he was going to lower food prices and then putting up the white flag of surrender two weeks later.

Now, the homeowner is worried about the price of eggs and you can’t fake that stuff. We’re going to do the best policy and work like hell to make sure people know about it. This last campaign was particularly hard because we made a big candidate change very late in the game, and it’s pretty hard to get your issues out that way, and we can do better.

Sen. Wyden, I’ll let you get back to work. It sounds like you have a lot to do. We’ll talk to you soon.

Thanks.

On today’s episode of Decoder, we’re discussing a very big problem with extremely far-reaching consequences: how dysfunctional has the federal government in the United States become, and how much of it has been handed to Elon Musk? 

If you’ve even glanced at the news over the past week, you’ve likely come across some alarming headlines about Musk’s Department of Government Efficiency, or DOGE, and how it’s begun commandeering control of some pretty important government agencies. Ostensibly, Musk’s goal is to slash spending and reduce waste across the government. But what’s actually happening is a lot more destructive than that. Musk and the DOGE team have gained access to the entire Treasury Department payment system, they’ve frozen lifesaving foreign aid programs, and they’re systematically locking entire chunks of the federal workforce out. And I do mean locking out — they can’t access the buildings. 

If you followed the Twitter takeover, it’s a familiar playbook — Musk moves in, breaks things faster than anyone can keep track, and generally causes chaos and dares the market and the legal system to stop him. At government-scale, it aligns perfectly with the Trump 2.0 strategy of shock and awe, moving so fast and so relentlessly that it’s impossible to figure out what to focus on, what’s actually changed, what’s been broken forever, and what’s merely noise.

In many cases, DOGE operates literally when everyone else has left the building — at night and on the weekends. On Monday morning, at 1:54AM, Musk tweeted what feels like an already infamous line about the nature of his work in the government, writing, “We spent the weekend feeding USAID into the wood chipper. Could [have] gone to some great parties. Did that instead.”

Obviously, a big issue here is transparency — DOGE is operating without any meaningful oversight, and it’s actively resisting attempts to reveal information as simple as the names and work histories of who’s on staff. When Wired reported the names of six DOGE employees, the new right in Silicon Valley called it doxxing, even though Musk himself has directed massive amounts of harassment at government workers over the past few months. On top of that, efforts to impose transparency on DOGE have caught the attention of at least one hardcore MAGA federal prosecutor who now claims he’ll go after anyone that attempts to impede Musk’s work. 

The whole point of transparency is to have accountability, which is the other big problem here. It seems like a lot of what Musk and DOGE are doing is simply illegal. There have already been dozens of lawsuits filed challenging this Trump administration, including DOGE’s approaches to headcount reduction, funding freezes, and access to sensitive agency data and controls. But these lawsuits will take time to work their way through the courts, and there’s no telling how much damage DOGE will do before that happens. As it stands, Congress does not seem equipped in any way to check the actions of the world’s richest man. 

That leaves us with some big — frankly scary — questions about where the power is now concentrating in Washington and whether any of the institutions we thought of as steady and resilient to chaos are truly as safe as we thought. 

For this episode of Decoder, I didn’t simply want to go through everything that’s happened. I wanted to know how all these changes are affecting the people who’ve so far been the most newly supportive of Trump because they have the most to gain and to lose: the money, the billionaires. So I invited New York Times reporter Teddy Schleifer on the show. 

Teddy has been covering Musk, DOGE, and the new Trump administration as part of a group of reporters at the Times pretty much every hour since the inauguration. They published a blockbuster story on Musk’s first wave of DOGE actions earlier this week, and Teddy himself specializes in the intersection of money and politics and how the influence that buys you has evolved pretty dramatically in recent years. 

So I wanted to ask Teddy what Musk and DOGE are up to, how the government is responding to that, and more importantly, is this what Jeff Bezos and Mark Zuckerberg and all the other billionaires wanted? Elon Musk in charge of the federal government? What are they saying about it, and how much further is it going to go? The answer might surprise you, but not in a good way. I think, over time, that part of the story might become the most complicated of all.

If you’d like to read more about what we talked about in this episode, check out the links below:

• Inside Musk’s aggressive incursion into the federal government | NYT
• ‘The biggest heist in American history’: DC is just waking up to Musk’s takeover | The Verge
• ‘Scared and betrayed’ — workers are reeling from chaos at federal agencies | The Verge
• Treasury Department sued over DOGE takeover | The Verge
• Can anyone stop President Musk? | The Verge
• Elon Musk’s team now has access to Treasury’s payments system | NYT
• Elon Musk’s bureaucratic coup | The Atlantic
• Trump: Elon Musk won’t do anything ‘without our approval’ | NBC News
• The young, inexperienced engineers aiding Musk’s government takeover | Wired
• USDS head Mina Hsiang wants Big Tech to help fix government (2023) | Decoder