Apple is considering enlisting the help of OpenAI or Anthropic to power its AI-upgraded Siri, according to a report from Bloomberg’s Mark Gurman. As Apple continues to struggle with the development of an upgraded “LLM Siri,” it reportedly asked OpenAI and Anthropic to create versions of their large-language models to test on the company’s private cloud infrastructure.

For months, Apple has been working to get its AI-enhanced Siri back on track after delaying the overhauled assistant’s launch in March. Apple later appointed Vision Pro head Mike Rockwell as the leader of AI and Siri after CEO Tim Cook “lost confidence” in the team’s former chief, John Giannandrea.

As reported by Bloomberg, Rockwell asked his team to test whether Anthropic’s Claude, OpenAI’s ChatGPT, or Google’s Gemini performs better at handling basic requests compared to its own models, with Anthropic’s apparently seen as the most promising. While Google has Gemini AI features for Android and its Pixel lineup, Samsung licenses Google’s AI model for its phones. It is also reportedly close to cutting a deal with Perplexity, which already has a tie-up with Motorola. Earlier this month, Bloomberg reported Apple executives had considered acquiring Perplexity to help boost its AI ambitions.

LLM Siri was largely absent from Apple’s Worldwide Developers Conference earlier this month, where SVP of worldwide marketing Greg Joswiak admitted that the technology “didn’t hit our quality standard.” Bloomberg notes that Apple’s plans to incorporate a third-party AI model into Siri are still at an “early stage” and that it’s still considering using in-house models.

Finding a Nintendo Switch 2 hasn’t been easy since its launch on June 5th, with sporadic availability over the past few weeks at various retailers. Luckily, if you’ve missed chances to pick one up, Best Buy will have more stock at its retail stores across the US starting tomorrow, July 1st. The Verge has reached out to Best Buy as to whether locations will primarily stock the standalone system ($449.99), the console bundle $499.99) that comes with Mario Kart World, or both. We’ll update this post if we hear back. To see a list of participating locations, head on over to Best Buy’s Switch 2 landing page.

While you’ll have to physically go to a Best Buy for a Switch 2, the good news is you won’t have to deal with an online queue or a potentially finicky checkout process. Despite stock being hard to come by in its first month of availability, the Switch 2 has already become the fastest-selling console of all time, surpassing 3.5 million units sold in its first four days. The handheld’s 7.9-inch 1080p LCD display (up from the Switch’s 6.2-inch 720p screen), magnetic Joy-Con controllers, and improved performance make it a solid upgrade over its predecessor. Or, as The Verge’s Andrew Webster referred to it in his Switch 2 review, “exactly good enough.”

A glass screen protector is one of a few "set it and forget it" purchases that every Switch 2 owner should make, along with buying a microSD Express card and a protective case for the console. In fact, it should be a priority to stick one onto the console's screen as soon as possible to avoid accidental scratches.

I've been installing and removing Switch 2 screen protectors to test others on my console at an alarming rate, more than any sane person should, and here are the most important takeaways:

• Do not buy a flimsy plastic screen protector. They won't protect your Switch 2 enough.
• It's impossible to tell different brands' protectors apart once they're on your screen; product packaging and the installation process are the biggest differences (as well as the key factors that increase cost).
• Some protectors are glossy and some are anti-reflective, each with its own tradeoffs. For my money, glossy is the way to go. The matte finish looks good, but it reduces the screen's vivid quality and its viewing angles.

Of the options below, it's easiest to recommend amFilm's three-pack of glossy screen protectors. Not only is the pack incredibly affordable ($7.19 from Amazon at th …

Read the full story at The Verge.

The US Department of Justice notched an initial win in its antitrust case against Apple today, with a federal judge rejecting Apple’s attempt to dismiss the lawsuit outright. The government’s allegations are “sufficient to demonstrate Apple’s specific intent to monopolize the smartphone and performance smartphone market,“ Judge Julien Neals wrote in an opinion on Monday.

Apple filed to dismiss the government’s lawsuit in August 2024, arguing that the Justice Department failed to show that Apple monopolized the smartphone market or acted in an anticompetitive manner. The lawsuit was based on the “outlandish” premise that Apple’s success comes from “intentional degradation of iPhone to block purported competitive threats,” the company wrote.

The case’s progress is still early, and the judge isn’t ruling on any of the government’s claims just yet. But he is saying that the allegations are “sufficient” to support the claims that Apple acted in an anticompetitive manner.

“We believe this lawsuit is wrong on the facts and the law, and we will continue to vigorously fight it in court,” Apple spokesperson Marni Goldberg wrote in a statement sent to The Verge.

The government’s win here wasn’t necessarily a given. When the Federal Trade Commission sued Meta (then Facebook) for anticompetitive practices in the social media space, the judge initially dismissed its claims and forced it to refile before the case was allowed to proceed.

The DOJ lawsuit against Apple claims that the company monopolizes the smartphone market by limiting the functionality of would-be competitive products, from apps to accessories. Jonathan Kanter, then DOJ Antitrust Division chief, said at the time that Apple used “contractual rules and restrictions” to raise prices on consumers and developers and “throttle competitive alternatives from rival technologies.”

Since it launched in November 2023, you could catch everything from rocket launches to documentaries about space on the NASA Plus streaming service. Starting this Summer, NASA Plus’ live programming will also be available on Netflix, and continue to be ad-free.

NASA Plus’ live programming also includes “astronaut spacewalks, mission coverage, and breathtaking live views of Earth from the International Space Station.” Although you do need a subscription to use Netflix, the NASA Plus content will be a free addition and won’t add to the cost of that subscription.

According to the agency, the goal of the partnership is to make NASA’s work in science and exploration “even more accessible, allowing the agency to increase engagement with and inspire a global audience in a modern media landscape, where Netflix reaches a global audience of more than 700 million people.”

The new Netflix partnership won’t be an exclusive one. NASA Plus will still be viewable for free on the agency’s website and through the NASA app which is available for iOS, Android, Apple TV, Roku, and Amazon Fire TV devices.

In a memo to Meta staff, CEO Mark Zuckerberg announced the company’s new “Meta Superintelligence Labs” group that will head up its AI work, Bloomberg reports.

Former Scale AI CEO Alexandr Wang, who joined Meta as part of a multibillion dollar deal earlier this month, will head up the group as the company’s chief AI officer. Former GitHub CEO Nat Friedman will “partner with” Wang to help lead the division, according to Bloomberg. Meta has also made 11 new AI-focused hires, including former Anthropic, Google DeepMind, and OpenAI employees. 

Zuckerberg has been making huge offers to potential hires – The Verge’s Alex Heath has reported that they’re “well into the eight-figure range” – to bolster its AI efforts, which have fallen behind other big AI companies. Meta has also held talks to buy companies like Mira Murati’s Thinking Machines Lab, Perplexity, and Ilya Sutskever’s Safe Superintelligence, though “none of these talks progressed to the formal offer stage for various reasons,” Heath reported.

Meta is also planning to “start research on our next generation of models to get to the frontier in the next year or so,” Zuckerberg said in the memo, which was published in full by CNBC.

A summer camping trip is the perfect excuse to stop staring at screens, but it doesn’t mean you need to go completely off the grid. Jackery’s Explorer 1000 Power Station V2 lets you charge your gadgets and have more fun while you’re camping. For instance, you can plug in a projector, speakers, or a space heater. In other words, it’s great for entertainment and for maintaining a certain quality of life you may not be willing to give up while camping. This charging station, which also includes a SolarSaga 200W Portable Solar Panel, costs $698.99 ($601 off), its lowest price ever at Amazon. 

The Explorer 1000 v2 will typically output 1,500 watts, which is enough juice to run a mini fridge, coffee maker, and other small appliances, or recharge a smartphone dozens of times. It can handle 3,000 watt power surges if a gadget draws a lot more energy briefly. The Explorer 1000 has three AC Outlets, two USB-C ports, one USB-A port, and a DC power port, and can power or charge multiple gadgets at the same time. A small screen shows its battery percentage and how much power is being drawn. You can also view that information on your smartphone using the Jackery app. 

The Explorer 1000 recharges in under two hours when plugged into an outlet, but you can reduce that time to one hour by enabling its Emergency Charging setting using the Jackery app. Fast charging will sacrifice battery health, which is something to consider. Jackery says the Explorer 1000 can retain up to 70% of its original capacity after 4,000 charges. 

The SolarSaga Portable Solar Panel that’s included can refill the generator’s battery in about seven and a half hours with direct sunlight, and you can cut that time in half by connecting a second one. The panel is waterproof, and can be folded up to protect the panels from damage. The panel uses a sensor in the upper-right corner to automatically adjust its positioning to ensure the sun is hitting it at the optimal angle. 

Four more deals worth a look

• Thousands of games are heavily discounted during the Steam Summer Sale, which is running through July 10th. Indie hit Stardew Valley is currently $7.49 ($7.49 off), which matches its lowest price ever, and play through the remake Resident Evil 4 for its all-time low price of $19.99 ($19.99 off) to whet your appetite for Resident Evil: Requiem’s release early next year. 
  
• If you need help working through your Steam backlog, the Lenovo Legion Go handheld is on sale for $499.99 ($200 off) at Amazon. The Windows handheld supports Steam, Epic Games Store, the Xbox PC app, and many other platforms to allow you to play your entire library on the go.
  
• A three-pack of TP-Link’s Deco X75 Wi-Fi 6E routers is down to $249.99 ($100 off) its lowest price ever at Amazon. The tri-band Wi-Fi 6E routers can create a network covering 7,200 square feet, transfer data at up to 2.4Gbps, and connect to 200 devices without throttling. One router must be hardwired to a cable modem, but the others can be installed throughout your home. Devices will automatically connect to the closest router and on the fastest band they support to reduce network congestion. Plus you can select your most important gadgets to get priority bandwidth. 
  
• If you’re planning on going snorkeling, biking, or adventuring this summer, GoPro’s HERO13 Black Ultra Wide Edition is $349.99 ($130 off) at Amazon. This is the lowest price we’ve seen yet for this model. The camera can record video in 5.3K and up to 240 frames per second in 1080p resolution. It can also capture 27-megapixel still photos and video frame grabs. You can take it underwater to depths of up to 33 feet, and attach it to a grip or helmet using the included mount. The Ultra Wide Edition has a 177 degree field of view, which is 36 percent wider and 48 percent taller than the regular HERO13. Video and photos you take are stored on a microSD Card, and can be transferred to a device wirelessly over Wi-Fi or using a USB-C cable. 

Microsoft will soon no longer let you use its Authenticator app to store or autofill passwords. Starting in July, you won’t be able to autofill saved passwords using Authenticator, and you’ll have to use Microsoft Edge or another password management solution instead.

Microsoft also plans on deleting your saved payment information in Authenticator this July before erasing passwords in August. Last month, Microsoft Authenticator stopped accepting new passwords as part of plans to consolidate its password autofilling feature within Edge.

Microsoft will automatically sync saved passwords to your account, allowing you to access them in Edge. You can set Edge as your device’s default autofill provider by finding the option in your device’s settings and selecting Edge instead of Authenticator. If you don’t want to use Edge, make sure to export your passwords to another service by August.

Microsoft Authenticator launched as a multifactor authentication solution in 2016, and it added support for password storage in 2020. Though Microsoft Authenticator is ending support for passwords, it will continue to support passkeys, the solution that lets you use your device’s authentication method to sign into accounts, such as a PIN, fingerprint, or face scan.

You can find more information about how to export your passwords or make Edge your default autofilling provider from Microsoft’s website.

Date Everything! isn't the dating sim it appears to be. Though the game's cutesy art style and lighthearted premise indicate an unserious game that's making yet another joke at the expense of the genre, Date Everything actually contains novel explorations of human relationships and sharp-as-shit political critique.

Date Everything (the title technically includes an exclamation point) is a game where you, with the help of special glasses called "dateviators," can turn everyday household objects into potential romantic partners. Every in-game day you're given five opportunities to find new and talk to already discovered characters called dateables. As you interact with them, the things you say and do influence their feelings for you. Winning a dateable's love, hate, or friendship will also increase the level of one of your personality traits - smarts, poise, empathy, charm, and sass - which unlocks more advanced dialogue responses. Say the right thing, and Dorian (your door, voiced by Ben Starr) will be your best friend boosting your poise. Say the wrong thing and your toilet Jean-Loo Pissoir (Max Mittelman) will swear eternal enmity, boosting your sassiness.

Date Everything is y …

Read the full story at The Verge.

Kallie Plagge, The Verge’s senior copy editor, has been working in the media for over 10 years now, mainly at gaming websites. She was primarily a video game critic for a lot of that time and was the reviews editor at GameSpot for about three years, before eventually becoming the senior copy editor at Polygon. “I was there for three years before Vox Media sold the brand, and I was fortunate enough to be offered the same role at The Verge,“ she explains.

We asked Kallie to tell us about her home office setup.

That looks like a great space. Where in your home is it?

I currently live with my parents, and they were generous enough to give me a separate office space. It’s next to my bedroom, and then my dad’s office is just down the hall. When I lived here during the covid lockdown, I worked from the dining room table, so it’s a big upgrade!

Could you tell us about your L-shaped desk? It looks like you have separate working and gaming spaces.

I was intent on getting an L-shaped desk when I bought this a few years ago from Target, and I can’t go back. At my old desk, I had a KVM switch so I could use just one monitor and keyboard with both my work laptop and my gaming PC, and that worked fine. But once I’d been working from home for a while, I realized I preferred to have “separate” work and gaming spaces, even if it only involved swiveling my chair 90 degrees. That bit of mental separation helped me really feel like I was transitioning out of work mode. 

Even with all the space of an L-shaped desk, I still wanted a big drawer unit for all my spare cables and stationery. This one is the Ikea Alex drawer unit on casters, and it doubles as even more surface real estate for things like my PC tower and personal laptop. I just really need to be able to spread out to focus. 

Tell us about your chair.

My dad actually bought this chair for me during lockdown. Because I’d been working in the dining room, I was using a dining room chair, and he could tell it wasn’t very comfortable. This is a Respawn gaming chair, and I added foam pads from Amazon to the armrests for extra comfort.

Here’s the long one: tell us about the various tech devices you’re using.

Hoo boy. On the work side, I have a Keychron K2 wireless mechanical keyboard with red switches. I chose it mainly because it has a Mac mode with Mac-compatible keycaps, which means I can easily use shortcuts that require the Command key. My mouse is just one I had on hand — a Razer Mamba wireless gaming mouse, though I use it wired because I prefer wired mice. Please ignore my dirty mouse pad. 

On the gaming side, I have a Logitech G713 wired mechanical keyboard with GX red linear switches. It came with the cloud wrist rest, which was admittedly the main reason I bought it over other keyboards. (I got a rainy day-themed mouse pad from a small stationery shop called Unicorn Eclipse to complete the look.) I also liked the tenkeyless layout; my previous keyboard had a number pad and two rows of programmable keys, and even though setting up macros was nice, it took up way too much space on my desk. Plus, I only ever used the number pad to input the PC shortcut for em dashes.

I also got the matching Logitech G705 wireless gaming mouse, which, again, I use wired. I was worried it would be too small — it’s specifically marketed as having a “smaller fit,” and my hands are a pretty average size — but I actually love it! It’s comfortable and made me realize the Razer mouse is slightly too big for me.

The Dell monitor is a hand-me-down, and I’m not really particular about monitors. I only wish it had a second HDMI port, because right now, I just manually switch the HDMI cable from my Xbox Series X to my gaming PC and back, depending on what I’m playing. It’s not a huge deal, but an extra port would make it easier to set up my Nintendo Switch 2 dock at my desk.

Because I have two stations at this desk, I also have two pairs of headphones. For everyday use, I have the Bose QuietComfort 35 II noise-canceling headphones that I wear constantly. (I don’t even listen to anything. I just really like dampening outside noise!) And for gaming, I have a SteelSeries Arctis 7X wireless headset that’s compatible with both PC and Xbox. It also has a dial on the right ear cup that lets me fine-tune the balance between game and chat audio, which is necessary for me when I’m playing multiplayer games with friends.

My gaming PC is a whole other thing, and I don’t have to get into the specs, which are pretty midrange. But since you can see the chassis: it’s a Cooler Master MasterBox NR200 Mini-ITX case. It was a gift, and I love its form factor.

That’s an adorable Pikachu — love the suit!

Thank you! I have a massive collection of Pokémon merch, including a lot of Pikachus in different outfits and with different themes, and the rest of them live in my bedroom. But I like having businessman Pikachu on my desk! I imported him from Japan in early 2015; his official name in the US is Intern Pikachu, but I think that’s underselling his qualifications.

And a Pokémon tabletop fan, as well! 

I am obsessed with this Poliwhirl fan. I also imported this from Japan — it was part of a recent onsen-themed collection, and a lot of these collections never come to the US for a variety of reasons, so I pulled the trigger right away. It has three speed settings, there are felt pads on the feet so it doesn’t slip or wobble while it’s on, and you can angle it upward. I love getting useful items that also have a clever and cute design, and this is the epitome of that. 

You have a nice, comfortable-looking couch. 

I love this couch! It’s from Article. It’s comfortable, but it’s not so squishy that you sink into it, and that was the goal. 

Is that a desk bike? Do you use it much?

It is! It’s a Flexispot V9 Pro desk bike. I’m using it more now that it’s getting hotter outside, since taking a walk is less appealing in 100-degree heat. I do find that I have a hard time focusing on editorial work while riding the bike, so it’s really only good for when I’m in a meeting with my camera off. But I do like it — I have very bad ankles, and I can’t really run, so it’s a nice way to get some more intense cardio.

Is there anything we didn’t cover that you’d like to add?

Business Pikachu says thank you for reading.

Photography by Kallie Plagge / The Verge.

The Insta360 Link 2 is more than just a webcam — it’s a smart camera that actively follows your movements whether you’re presenting, streaming, or pacing during a call. It’s also packed with an impressive array of productivity features. And right now, it’s available at a new all-time low of $149.99 ($50 off) at Amazon.

The webcam is a modest upgrade over the original Link, which was our favorite 4K webcam when it launched in 2022, boasting all the same impressive features with slight refinements. Like the original, it offers sharp, detailed 4K/30fps video even in low-light settings. But what really sets the Link 2 apart is its motorized three-axis gimbal, which keeps you perfectly centered in frame as you move around the room. The Link 2 also comes with a magnetic mount, letting you attach it to a wider range of surfaces without the need for additional mounting gear.

In addition to its exceptional tracking abilities, the Link 2 features an AI-powered framing tool that automatically zooms and adjusts the shot to focus on your head, upper body, or full body, depending on your distance from the camera. It now supports customizable “no-follow zones,” allowing you to prevent the camera from tracking movement in specific areas of the room for greater framing control. The Link 2’s also far better suited for group presentations than its predecessor, thanks to an upgraded tracking system that can now follow multiple presenters simultaneously.

At the same time, the webcam retains a suite of software features that help cut down on the extra gadgets typically needed for presentations. DeskView angles the camera slightly downward to showcase your workspace, while Overhead Mode captures top-down shots perfect for demos or unboxings. Whiteboard Mode utilizes included stickers to lock onto a presentation surface, keeping it in focus throughout meetings or tutorials.

After retreating from watchOS several years ago, Google now only offers a handful of apps for the Apple Watch. That list is expanding to four in total with the launch of a new Google Calendar app rolling out alongside the existing iPhone version, as spotted by 9to5Google.

Apple’s native calendar app on watchOS  includes customization options that let you see alternate views, like a week summary and a full month. But the Google Calendar app only shows a week’s worth of events on your Apple Watch with their time and location, as well as Google Tasks in a simple list view showing the day of the week in the upper corner as you scroll.

You can click on an event to see additional details, but it’s very limited. You’re prompted to instead “Open Calendar on your phone to view more” if you want more specifics. According to 9to5Google, the app’s functionality is very similar to what’s offered on Wear OS devices.

There are also two home screen complications offered. The circular Today’s date shows the abbreviated day of the week and the date, while What’s next, which can be a circle, rectangle, or Smart Stack widget, shows your next appointment. Both complications can be tapped to open the Google Calendar watchOS app.

Although it may be a welcome option for those who want as much Google on their Apple Watch as possible (Google Keep, Maps, and YouTube Music are the only other watchOS apps the company offers) the standard calendar app can already sync with your Google Calendar and lets you see more details for a given event.

Automattic’s plan to move Tumblr’s backend over to WordPress is now “on hold,” Automattic founder and CEO Matt Mullenweg says in a Decoder episode published today.

The company announced the plan to move over the more than half a billion blogs on Tumblr last year, saying that the change would “make it easier to share our work across platforms.” But Mullenweg says on Decoder that, “what we decided is that we want to focus as much on the things that are going to be noticeable to users and that users are asking for.”

It seems as if the move might be on the table in the future, though. “I still want to do it,” Mullenweg says. “It’s just cleaner. But right now, we’re not working on it.”

The decision to halt the change also appears to mean that Tumblr posts won’t be available in the fediverse in the near future. WordPress.com currently offers an ActivityPub plug-in, so Tumblr moving onto WordPress would theoretically let people bring Tumblr posts to the fediverse. “That would’ve been a free way to get it,” Mullenweg says. “And so that was one of the arguments for migrating everything to WordPress.”

In the meantime, however, “I think if there was a big push to implement fediverse, we would just do it on the Tumblr code base,” according to Mullenweg.

Automattic purchased Tumblr from Verizon in 2019.

Apple Music is celebrating the 10-year anniversary of its streaming service by giving users a throwback to their favorite hits over the last decade. “Replay All Time” — a special version of Apple Music’s annual Replay experience — is being launched for subscribers that allows them to view and listen to their most played songs since joining the platform.

The personalized Replay All Time playlist will be available from the Apple Music Home tab. If you joined the service back when it first launched on June 30, 2015, then you’ll be able to relive how your music tastes have evolved (or not) since the very beginning, all from a single playlist. That’s easier and more comprehensive than manually revisiting your yearly Replay playlists, given Apple Music only launched the feature in 2019.

It’s unclear if this playlist will be permanent or only available for a limited time. If it’s the former, then Replay All Time will presumably be continually updated to reflect the listening history of Apple Music subscribers going forward. We have asked Apple to clarify.

Apple Music is one of the oldest offerings under Apple’s services division, which includes Apple Arcade, Fitness Plus, TV Plus, Podcasts, and Books. These services accounted for $26.6 billion of the company’s $95.4 billion revenue last quarter and are a continued focus for Apple as it shifts towards making customers pay for its platforms, and not just its hardware.

For a broader throwback, Apple Music Radio will reveal the top 500 most-streamed songs from the last 10 years starting on July 1st. The service will count down 100 songs per day, culminating with the top 100 on July 5th, after which the full “10 Years of Apple Music: Top Songs” playlist will be made available to stream. Other celebrations include a week-long block of specials and live programming that’s kicking off today, starting with “Don’t Be Boring: The Birth of Apple Music Radio” between 9AM and 11AM ET.

Apple Music is also opening a new studio space for artists in Los Angeles this summer. Apple says the three-story facility measures more than 15,000 square feet, and includes a 4,000-square-foot soundstage, rooms for photography, editing, and social media content, and two radio studios and a mixing room equipped for spatial audio.

Canada has scrapped its tax targeting Big Tech companies in the US after President Donald Trump threatened to cut off trade talks in response to the fee. On Sunday, the Canadian government announced that it will no longer implement the Digital Services Tax as it works to reach a trade deal with the US by July 21st.

The Digital Services Tax would’ve targeted US-based tech companies like Meta, Amazon, Google, and Uber with a three percent fee on revenue earned from online marketplaces, advertising, social media, and the sale of Canadian user data. The tax officially went into effect last year, but Canada planned to begin collecting payments, which would’ve applied retroactively to 2022, on Monday. The first round of payments would’ve amounted to nearly $3 billion in fees, according to a trade group backing the Big Tech companies.

The Trump administration has similarly put pressure on the UK to drop its digital services tax, which also goes after tech companies in the US. The tax still remains in effect, though White House trade adviser Peter Navarro said last month that the US is still in “negotiations” with the UK over the fee.

Last week, President Trump called Canada’s digital services tax a “blatant attack on our country” and said the US is “terminating ALL discussions on trade with Canada” as a result of the country’s decision to move forward with the fee. Canada hit back with a 50 percent surcharge on steel imports that exceed a newly-imposed quota.

“Rescinding the digital services tax will allow the negotiations of a new economic and security relationship with the United States to make vital progress and reinforce our work to create jobs and build prosperity for all Canadians,” François-Philippe Champagne, Canada’s Minister of  Finance and National Revenue, said in a statement. Canada still faces a 25 percent tax on goods imported from the country, along with a 50 percent tariff on its steel and aluminum.

It's been more than a decade since Apple unveiled the original version of CarPlay. In that time, the system has become indispensable for countless drivers all over the world. But while new iterations of iOS have introduced user interface tweaks and some additional customization features, its basic functionality hasn't advanced much since that announcement back in March 2014. This changes with the introduction of CarPlay Ultra.

First shown at WWDC 2022 and now available on any iPhone 12 or newer that's running iOS 18.5 or later, CarPlay Ultra is designed to deliver greater integration with the vehicle itself, offering not only the ability to have CarPlay on multiple displays within the vehicle but also the option to adjust things like climate controls and drive mode settings from within the CarPlay interface. Much like the original version of CarPlay, CarPlay Ultra has had a protracted gestation period and is finally making its debut in Aston Martin's high-performance SUV, the DBX707.

The fact that Aston Martin is the first out of the gate with support for CarPlay Ultra illustrates just how much the company's approach to technology has evolved in recent years. Like many low-vol …

Read the full story at The Verge.

Today, I’m talking with Matt Mullenweg, the founder and CEO of Automattic, the parent company of WordPress.com, Tumblr, and a whole host of other products like the new cross-platform messaging service Beeper. 

This is Matt’s third time on Decoder; back in 2022, we had him on twice, first to talk about Automattic and WordPress broadly and then to talk about Tumblr and the future of social networking. He’s back now because Automattic just turned 20, and I really wanted to talk about how the next 20 years of running one of the most dominant platforms on the web might look as changes to search and AI threaten to change everything else, and various lawsuits threaten to change the nature of WordPress itself.

Make no mistake, WordPress is one of the most dominant platforms on the web, if not the most dominant. Something like 43 percent of websites run on WordPress, in one of its many flavors. That includes The Verge — the backend of our website is hosted by WordPress VIP. So this might be the first reverse disclosure on the show. Technically, we’re Matt’s customer, and like any good customer, I made feature requests. 

A big reason for that is WordPress is open-source, and like so many open-source projects, WordPress has a very complex structure. There’s the nonprofit WordPress Foundation that owns the WordPress trademark. There’s WordPress.org, from which the open-source project is managed by Matt himself. Then there’s Automattic, which is the for-profit company that offers its own site hosting and enterprise services on top of the core WordPress technology, and which contributes an enormous amount of code back to the open-source WordPress project.

Understanding that structure is really important because there has been a lot of drama in the world of WordPress recently. Last year, Matt essentially went to war, publicly and in the courts, against a hosting company called WP Engine that competes with Automattic. Matt felt WP Engine wasn’t operating in the spirit of open-source by contributing very little back to the WordPress code base.

So Matt filed a lawsuit against the company while revoking its access to core WordPress technologies. Many people felt this was incredibly out of bounds for Matt and a violation of his position as a central steward of the WordPress project, and there has been significant fallout at Automattic and the broader WordPress community.  

It’s been a long, drawn-out saga. WP Engine countersued, and Automattic was forced to reverse some of its retaliatory efforts against the company. But the lawsuits are ongoing, and they’re far from resolved. That said, Matt was willing to come on the show and talk through some of this thinking here, why he made some of the decisions he did, and also what he regrets about how some of this went down. 

Matt and I talked about the future of the web, too, and how he’s thinking about the changes we’re seeing to search and website sustainability as the generative AI boom continues to upend how people use the internet. Matt is notably a lot more optimistic about this than many of the website owners we hear from regularly here at The Verge, and he’s not convinced AI is going to wreck the web. 

We also talked about Beeper, the cross-platform messaging service that Automattic acquired last year. Beeper got into some hot water with Apple when it tried and ultimately failed to bring iMessage to Android. But Matt is really excited about Beeper’s core product. Automattic has acquired a couple other startups and effectively combined them all to try and supercharge Beeper’s growth in the coming months and years.

There’s a lot in this conversation, and Matt is as candid and sincere as ever. I think you’re going to like it. 

Okay: Automattic CEO Matt Mullenweg. Here we go.

This interview has been lightly edited for length and clarity. 

Matt Mullenweg, you’re the co-founder and CEO of Automattic, the parent company of WordPress, and many other things. Welcome back to Decoder.

Thank you so much. The world has changed. So much has been going on since I was last on. It’s great to catch up.

I feel like we had you on twice in one year, and that was three years ago, then many things happened, and I’ve been dying to talk to you about updates to the stuff we talked about the last time you were on. And then a bunch of new things, including some very dramatic new things, have occurred since last time.

We have some cool acquisitions and launches coming up. So yeah, a lot to cover.

Let’s start there. People are obviously familiar with WordPress. I imagine people are familiar with Tumblr, which is another thing that you were technically the CEO of, I believe. But since then, you’ve acquired other companies. You’ve acquired companies like Beeper and some others. What do you think of as Automattic today? What’s the thesis of the company?

I know you love talking about org charts and organizational structures, so it’s interesting. The WordPress side of the business we call “Ecosystem,” right? It’s like gardening. There’s a vast number of players, and that’s really kind of what we’re best known for. I’ve been doing WordPress now for 22 years. I started when I was 19, and that’s WordPress.com, it’s WordPress VIP, which we’re very proud to have Vox as a customer.

Oh, wait, I need to disclose this. This is the first-ever reverse disclosure, I think, on The Verge or Decoder. Usually, I disclose when there’s a business relationship, but in this case, we are your client. 

[Laughs] Yep, so thank you. We’re honored. 

The Verge runs on WordPress, so the conflict is the other way. You’ve got to keep me as a customer.

Well, I’ll do my best on this. And no, that’s actually the thrill of it too, is seeing publications you follow and things like that use the software, it’s very rewarding. Now, the other side of Automattic we call Cosmos, and that’s the apps, and that’s been a very exciting place to work in the past few years.

Now, you mentioned Beeper, which is actually doing its big public launch this July in New York, so I hope to get you there. We did two acquisitions, Texts and Beeper. We combined them, and we’re very excited. I was actually just with that team a little bit earlier today. The other one [we added] just last week was Clay. So, if you aren’t familiar, Clay is a personal CRM. One of the top requests we heard from Beeper users is like, “Okay, I’ve got all my messaging apps…”

Actually, I should probably say what Beeper is. You have probably more than one messaging platform you use regularly, right?

Yeah.

So the cool thing about Beeper is that it can bring them all in one. WhatsApp, Signal, Telegram, all of that. The other cool thing it does is that it can use multiple accounts of those. So you can have multiple Signal accounts and multiple WhatsApp accounts, all in one device, and you can offer it on your desktop, mobile, everything, and it does it all securely. That’s the new stuff. We figured out how to make this all run locally on your device, and so it’s just as secure as using one of these native apps.

But what we heard is people say, “Well, now I’ve got too much stuff going on. Who’s important? What should I do?” That’s where Clay’s CRM comes in. CRM stands for customer relationship management, so usually apps like Salesforce. But this is a personal one that’s pretty slick.

When you say “personal CRM,” what’s interesting is that the throughline for a lot of what you are doing is that small businesses, businesses of all kinds, use WordPress. They’re on the web, and a lot of what the web is used for lately is a small business storefront, or a business storefront, or some commercial enterprise. Things like a CRM, a text messaging platform, all of that stuff — it connects very deeply to just e-commerce in general.

But you’re talking about a personal CRM. What’s the split for you between WordPress as an enterprise e-commerce company and WordPress as a purveyor of consumer products?

I say we’re very lucky to be part of the generation that was… they called it “the consumerization of IT.” It was led by folks like Slack who came in and said, “Hey, we’re just going to make a great user experience and it’s going to be quirky, it’s going to be fun, it’s not going to be boring, it’s going to be colorful and we’re going to build the business features. So it scales and it does that sort of stuff, but we’re going to start from that great user experience.”

There’s some enterprise software historically that didn’t start there. If you’re on an old SAP install or something like that, no one’s waking up being like, “Yeah, I got to get in there.”

Slack, notably now owned by Salesforce, which is fascinating. That’s the two sides of that equation coming together.

Well, I think these companies, including Salesforce, have been doing a lot to reinvest in their interfaces and everything else. So with Clay — and the URL’s clay.earth; unfortunately, there’s another big company called Clay that does more enterprise stuff — so if you go to clay.earth, it’ll be the personal one. 

So what makes it personal versus enterprise? We start from things on your computer, like your address book. We hook in social networks, so you can put in all your social networks. We’ll kind of de-dupe it. We emphasize things like birthdays. It brings in your calendar, iMessage, and WhatsApp messages right now. That’s kind of what Clay was doing before, and what we’re going to be building now is that it’ll just plug into Beeper, and then whatever you have plugged into Beeper, however many networks, whatever you’re doing there, it’ll bring all those messages in. 

Then it can start to, for example, do fun stuff like sorting your contacts by how close your relationship is, which can do clever things. Like, not just look at how often you contact them, but who’s sending more messages, which way, and is it weekdays or weekends? There’s a lot you can kind of infer once you have communication, and we’re very excited to see where that goes.

I would understand if you were saying this was an enterprise product. I would completely understand why you’ve invested in Beeper and Texts, and why you’ve invested in something like Clay, right? You’re starting a website, you’re going to do some e-commerce, you need an outbound messaging platform, and you need a customer relationship management tool. You want to know who your best customers are. This is all just value-adds to the commerce platform.

What’s the revenue model for these as consumer products? Are you going to charge for them?

Yeah, so we’ll definitely cross-promote this stuff. But what you just said — who are your best customers, that sort of reporting — that’s just all built into WooCommerce and WordPress, so you don’t need to go between apps.

As for the revenue models we’re imagining… Clay already has a Clay for Teams, so your entire team can share contacts, share updates, and you can use it in a team-like fashion. And they’ve got a pricing model there, and they were making revenue. So it was nice to acquire a company that already had a revenue model.

With Beeper, we’re still figuring it out. But what I suspect is there’ll be two things. It’ll be free up to a certain number of accounts, and if you’re a super power user and you want to connect more, there could be a monthly charge. And then the other thing is that there might be certain connections that are always paid. So let’s say you want to connect to a Bloomberg terminal chat, we’re probably going to charge you for that because you’re more of a business user.

Also, a lot of the messaging platforms support business features now, like WhatsApp for business. There’s even iMessage for business. I don’t know if you’ve ever seen this? 

Yeah.

But famously, iMessage, which previously was like, “No APIs, we’re going to crush Beeper,” in a previous iteration, now has a whole integration system for it. So as we support more and more of these bridges to different networks, I think there are a lot of opportunities there.

Is the Beeper architecture still running an instance of viewer messaging apps in the cloud on your behalf and relaying them to you, or are you more integrated now?

There are still two cloud bridges, but part of what we’re relaunching is that we’ve re-architected the whole thing with the Texts technology, so it all goes local now.

So iMessage is coming locally to your phone through Beeper?

iMessage is the one that we do not support on mobile. It is supported on desktop, just like TextIt, in this sort of way that Apple said or indicated it was okay with. But yeah, we’re not going to fight that fight again. We’re way, way, way away from it. Not touching that with a 100-foot pole.

Is the goal here that these apps should be revenue-generating consumer products? Or are they good additions to the enterprise stack that also might have consumer elements?

I’m thinking of it 100 percent as creating something consumer-first, actually, that then has paths if you’re a power user to do more team-like collaborative or business-like things on it. But first and foremost, I want to get Beeper to 100 million users. I feel that’s actually sort of the first product from Automattic that has the potential to actually be really, really large because its usage is kind of a superset of every messaging network, and the power users are the most frequent users on each of those.

Right now, it’s relatively small. No one feels that threatened. If you get to 100 million users and the primary interface for WhatsApp, Signal, and iMessage is actually Beeper, those apps will feel threatened. There’s nothing keeping those companies from saying, “You’ve taken our customer away, we’re shutting you down.” Have you had those conversations yet?

Well, I think the difference is that we’re not trying to take the customer away. We’re trying to give them an interface where they can use the network even more, connect with more people, and use the business features. If they have any way to monetize, we’re just going to link to it. We’re not trying to avoid that. So we want to support each of these networks, which by the way, we’re investing huge amounts to run everything, so we’re trying to be complementary.

Because they want to be the app on your phone through which you do many, many other things. If the app on the phone is Beeper…

The model that we see people use Beeper is that they don’t get rid of the native app. They keep it. Because it’s always going to be something there, like a functionality or something, that you’re going to want the app for. And Beeper is more for managing lots of messages, getting the local LLMs, giving you intelligence across networks, things that are power user features. But I think there are a lot of power users.

So, a strange thing that is happening with Beeper now is that France is the number one country, and it started to go viral in certain ways, and some of that is people wanting to be able to check their messages without getting too distracted. They call it “Friends without Feeds,” but they still go back to the feeds, don’t worry. They’re doing just fine. It’s just sometimes, if you’re in a meeting or something, you just want to take a quick look, that sort of distraction-free mode is really nice.

That’s really interesting. One of the reasons that I’m pushing on this is that as we’re speaking, I think you’re days away from Automattic turning 20 years old as a company. Automattic, WordPress, and these other apps are all part of the fabric of the web. Broadly speaking, about 43 percent of the web runs on WordPress.

Some of these fights that Beepers has had in the past are like how do we build open architectures out of services that are fundamentally closed, use web technologies to pass that stuff to people, and what kind of fights will we have there. There’s the open source fight and WordPress that I want to talk about in depth.

But it feels like here at 20 years old, the web is changing in meaningful ways. And the web as an enterprise platform might be headed toward ever higher heights, and as a consumer media platform might be headed toward ever lower lows. I’m wondering, as you think about these investments, these tools, and the apps versus the ecosystem, whether you feel that tension playing out?

No. So I’m actually going to ask you to expand a little on what you see as these lower lows, because it feels like as we get more compute at the edge, as the devices become more powerful, broadband becomes ever more ubiquitous — you can’t escape it anywhere now with Starlink — that there’s a bit of a sort of swinging back towards these apps and user-centric things. Even the regulatory environment is very friendly.

And that’s what I would pull it apart. As an application platform, the web is, I think, at its peak and maybe with higher highs to come. Web apps are the most interesting they’ve ever been. Every powerful AI application is mostly expressed as a web app, especially on desktop. Google is literally demonstrating Veo 3 as a web app. Google CEO Sundar Pichai is like, “I’m drawing people to the desktop web to use these applications in Chrome on the web.”

Figma exists, right? It’s one of the most powerful design tools, and it’s a web app. As an application platform, the web is at its highest it’s ever been. Some of the regulatory changes, the Fortnite lawsuit, as a transaction platform, the web is going to hit higher highs because we’re going to see more transactions pushed to the web. That’s something else I want to talk to you about.

I’m saying, as a media platform, as a document viewer, the question I’ve been asking every CEO is, “Why would I start a website today?” I would start a TikTok channel. I would start a YouTube channel. That’s how you reach consumers with media. The web as a media platform seems to be at its most perilous moment, and that’s really the split I’m talking about. Applications, transactions, and the web are very clearly at a peak with potentially higher highs to come. As a media platform, media companies are going out of business on the web basically every day.

There’s a lot in there, right? Because you kind of hit the creator thing. We hit social networks. I’ll start with your question: why, if I’m popular on TikTok, would I start a website? Well, one, so you’re not a one-hit wonder, and I think we’ve seen even some of the biggest creators on a certain platform will often have trouble getting as popular on another one.

So you need to develop a direct relationship with your audience because as long as your audience is fully mediated by this thing you don’t control — YouTube, TikTok, Reels, whatever it is — you’ll have a run. But these things change generationally sometimes. People move from one to the other. The business models change, and what they emphasize changes. If you’re a creator who was just all-in on Facebook 10 years ago because you thought nothing would ever replace Facebook, you might be facing some of the same things that these media companies are facing that haven’t adapted and really embraced their users.

I think that the media thing is also kind of complex because we had a real degradation of the user experience and sort of the speed of sites, the way advertising would work, and slow down your browser, and everything. Present company excluded, but for some other media sites, you’d load them, and it’s almost like they’re having trouble. It’s hard to read the article because it keeps moving around as the ads load, and so I felt like that was a death spiral for some of these sites that might’ve over-monetized.

I’m going to go all the way now to the local stuff-

Actually, can I just push on over-monetized for one second?

Yeah.

Over-monetized. You can read that in several ways. I think what you mean is they put too much shit on the page and then the user experience was degraded, and nobody ever wants to go to a local news site again.

The other way to read it is that they had no distribution except for Google or maybe Twitter, and every page view was so scarce that they needed to eke out every single penny they possibly could because that visitor was never coming back. And that’s the distribution that’s going away. That’s why I’m saying that as a media platform, the web is at a low because all of the audience is on somebody else’s distribution, which are by and large closed platforms.

I think there’s also an aspect that you’re competing to be the best in the world, as audiences become more discerning, as there’s sort of this global competition to raise the discourse and have the best analysis. I mean, in some places we see single-person newsletters killing it, as well as media organizations and everything like that, but many others have struggled. 

That’s why I was going to say the local media, because it’s a great example of where we historically had thousands of local newspapers in the U.S., as a geographic monopoly type of thing, and many of them have gone. But I’m going to go back to the ecosystem side of Automattic. 

One of our most exciting mini-companies inside is called Newspack. It’s led by Kinsey Wilson, who used to be chief of digital at The New York Times. He’s sort of taken everything he learned there and is bringing that to these small newspapers with this product called Newspack, which you can think of as distribution for WordPress. So, it’s WordPress plus hosting, plus a bundle of plugins that enable all the things that these small-town papers need, like classifieds and all that sort of stuff. The fun thing about it is they’re learning from all of them and sharing the business best practices. So, porous paywalls, or I mentioned classifieds already, but people really love local news. It’s just that it couldn’t support some of the old business models.

Now, the sad thing is some publications that switch to Newspack actually save hundreds of thousands of dollars. There were some of these legacy software companies that were just charging way too much. It has a whole print component and everything like that, so you can still print it out and distribute it at the local coffee shop and everything. But I’m far more excited about actually growing their revenue through new things that are allowed, like paid newsletters or sports scores. All those sorts of things that, when you go ultra local, you can support three, five, or 10 journalists to cover a small area, which I think is important for democracy.

How are you thinking about the distribution of that? If I were to compare you to, I don’t know, Ghost or Beehiiv or Substack? Fundamentally, what Substack is selling to a large number of its top newsletter authors is growing its audience, right? If you’re Heather Cox Richardson, you’re the most popular Substacker out there. It is crazy that she is paying 10 percent of her revenue to send emails, right? Mathematically, she could get a better deal to send emails than what Substack is offering her based on that cut.

But I think what Substack would say, and what I’ve heard the company say to others, is, “We will generate new subscribers. Our network will provide distribution. That will get you new customers. This is a cheaper way to get new customers than if you move to some other standard email service, and you have to do your own marketing, your own customer acquisition.” That’s the distribution puzzle that they’re solving that appears to be worth it for some people. 

As Google goes away and other platforms stop linking, how are you thinking about solving that problem for the Newspack customer? Are you going to move people around an ecosystem? Are you going to build other forms of distribution?

You said it like it was a set thing that Google’s going away and no one’s going to link to websites anymore. I think what we’re seeing aggregate across everything is that there is a lot more traffic being driven by the OpenAIs and the Perplexitys of the world. This also feels like the early days of that.

I don’t know. Maybe if LLMs never hallucinate again, people will stop visiting links, but for now, I actually find myself sometimes clicking on three or four things, even from the Google summaries that they put at the top of the search results. So I’m probably clicking on more things than I used to. When I just had the 10 blue links of the old Google, I would pick one of them and then spend time on that webpage. Now, here’s a summary, and it links to three different things, so I find myself exploring a little bit more. What we’re seeing aggregate in traffic is that what happened from 2020 to 2022 was actually worse than what’s happening now. So things are actually starting to come back a bit. I don’t know. Have you seen that with your traffic?

I think, broadly, what we see is the same thing as everyone else, which is that the shape of Google traffic is changing, and some Google services are sending more traffic, and some are sending less. They only just started saying what traffic comes from AI Mode and Search Console, so I can’t actually tell you. This is like, as we’re speaking, I believe this happened yesterday, that they started breaking out AI Mode and Search Console. 

So, it’s too early to say, but we have reported on website after website that has just disappeared. The Daily Dot went from millions of Google referrals to thousands, and then the business was over, and that’s the end of The Daily Dot, and it doesn’t exist anymore.

Remember back in the day, Jason Calacanis had Mahalo, right?

Yep.

Google has always been mercurial, especially if you optimized your business around that. So, imagine that one of these creators was only on one network. You want to have many paths to the ocean. That’s what I’d recommend for any business, really, right? You don’t want to stake everything on just one partner or one business model.

I think the question I’m asking you is, there are vanishingly few paths left to the web. There are lots of paths to the web as applications. I think there’s going to be an increasing number of paths to the web from iOS apps looking to escape transaction fees by doing commerce on the web. The paths to the web as media are obviously changing. And it sounds like you’re saying you are actually seeing more traffic from the AI search engines than people expect.

It goes to different people. So, in aggregate, I am optimistic. In the short term, there’ll be a changing of the guard, perhaps, or maybe it’ll reward different sites. Again, one of the things that I think hurt some of these media sites that we talked about before, the ones with too many ads, Google started taking in site performance as part of its ranking. So, if you had a pop-up there or something like that, they would start to de-rank you a little bit. When you think of the incentive of these engines, they want the user to have the best possible experience.

Yeah. I mean, WordPress is 43% of the websites out there. I’m assuming that you can see-

We see a lot, yes.

You can see a lot. Are you seeing more or less traffic from Google than a year ago?

I don’t know off the top of my head, but I think it was flat-ish to some and then up for others.

But you haven’t seen these dramatic declines that are wiping out some publishers?

Not in the past year, no. We saw some of that four years ago.

Interesting. So, you think it’s the same amount of traffic, which is expressed differently across your network?

That’s right, yeah. And more of it is starting to be driven by LLMs.

That’s really interesting because Google will happily tell you the same thing, and then we get website owners in our inbox saying, “They took all our traffic away.”

I believe them, actually.

You hear that across the board. Business Insider just had layoffs because their Google traffic went away. That seems like the dynamic where maybe there are going to be a bunch of new media websites that have a bunch of traffic driven to them by engines.

But to your point, what’s going to be the backstop against that? There’s only one referral source left of huge value, and it’s Google. Maybe these new LLM companies and these search engines will drive some traffic over time, but there isn’t another user behavior that drives a lot of traffic to the web in that way. 

There are search and search-like things, and chatbot-like search. But there used to be Twitter, which would drive a lot of traffic to some websites. Facebook used to drive a lot of traffic to some websites. Those other things have faded away. Do you see something else coming up that might balance out the incredible search dependency?

What people sleep on is the Google articles. I don’t know what they call it. It’s not Google News, but it’s that thing that if you scroll to the left on an Android or you open the Google app. 

Google Discover. I know entire media companies whose business is Google Discover, just programming Google Discover, which to me feels the most brittle of all. 

I actually dream of a day when Twitter doesn’t de-emphasize links again. So who knows? There are a lot more niche social networks. For example, if you’re an engineer, Hacker News from Y Combinator, but there’s one called Lobsters, which can have cool little spikes. 

Honestly, I think the video stuff actually can still drive really great web traffic. People allow links in them now. They talk, or they just say something, and people click on it. It’s like this thing we saw for Beeper. The French usage surpassed English just from this one viral Reel. So, that was people going to a website.

Fair enough. I want to get to the Decoder questions because I want to talk about the other side of the web, which is the open web and the open-source nature of a lot of the stuff you work on at Automattic and WordPress. You’ve described your chart as Ecosystem and Cosmos. Cosmos is the app side. Ecosystem has the core technologies you’re building. You’ve had a bunch of buyouts and layoffs this year. How big is WordPress today?

We’re about 1,500 people.

Is that substantially smaller than it was at the top of the year?

Yeah, like most tech companies, we hired a lot in the 2020 to 2021 range. And also like most tech companies, we found we can be more efficient and move a little bit faster with smaller teams. I don’t love that, but it is a business reality.

Did anything about your leadership structure change? Did anything in your org chart change? Or is it still the two main groups?

Yeah, actually. Internally, we just did a big switch, where we have traditionally been sort of independent product silos with their own engineering, marketing, everything, and we just did a huge centralization effort. So, product engineering and design are now all centralized, and we have some new leaders there as well. We have a new colleague named Pedraum Pardehpoosh, who was at Apple for 15 years at the App Store and Airbnb. He’s helped us really reimagine how we think about product, which I’d say historically we’ve really… because my proclivity is very much on the engineering side. So that’s been really nice and really exciting.

It is actually a very different organizational structure. I think I said this last time, but I feel like all org structures are just a series of trade-offs, and sometimes you just need to make the other trade-off for a while. So, if you’ve been in one place for too long or doing things one way, you need to do the opposite to break out of whatever rut you’ve found yourself in. It could just be how you’re thinking.

There’s a real pendulum between centralization and decentralization at most companies, especially 20-year-old companies. Was the decision really just, “We’ve been doing it this way for a while. We’re just going to swing the pendulum the other way and see what happens.”?

[Laughs] Yeah. I mean, that wasn’t just it. It’s not like, “Oh, let’s just swing the other way.” It was really like, “Hey, what are some of our issues here? Ah, we’re not having some global quality. We’re getting some local maximums in certain areas. Performance management across all of this can be inconsistent. Let’s try one roadmap for the whole company and see what that looks like. Should we try this ‘every six months’ thing that a lot of companies are doing?”

We examined all of that and then looked at how that fit with what we’re hearing from the customers, what’s happening with the business and the environment, what we’re really excited about, and this is what we ended up with. We’re only two months into it. So, there are a lot of changes in the first three to four months of the year, and it’s kind of been baking the last two months. To be honest, I won’t be able to tell you if it worked until probably towards the end of the year. 

What are you hoping to get out of it?

All the things the business wants. So I want happier colleagues. I want better business results. I want better retention and acquisition. I want Beeper to get to 100 million users. We have all our ambitions. Ultimately, I want to solve open source for publishing, commerce, and messaging. This is my life goal. 

Everything in between is a means to an end. I try not to be overly attached to almost anything to see what happens. So much so, I’ll tell you something a little wild, which is that we were famously like the most remote distributed company ever. There are many others, but we were pretty early, and I advocated for it quite heavily. I’ve actually been exploring whether we should bring a team entirely to New York for six months to do a sprint. So essentially, that co-locates, to do the opposite of remote, to see what would happen.

Actually, I love the idea of co-working weeks, maybe not co-working for six months, but co-working weeks. We are pretty remote, and I always think, “Well, we can get together for an offsite,” but then the offsite is its own process. But actually, I just need everyone to work in the same room for a while and be in The Office together. Like the show The Office, not the office office, to like to goof around and tell jokes and just watch each other doing the actual job, not the job of making decisions at an offsite. Let me know how that goes. I’m very curious. We haven’t done it yet, but I’ve been thinking about it a lot. 

A lot of what’s happening in the WordPress community is kind of a reflection of the fact that being virtual all the time breaks some trust, right? It feels like there are some trust issues with WordPress and the larger WordPress community that you might have to work through. Do you think any of this restructuring or any working together will help you through that?

It was tough during COVID because the secret ingredient for Automattic has always been our meetups. The secret ingredient for WordPress was these WordCamps and also meetups with Meetup.com. When that went to zero during the pandemic, we all know about it, but I think that there are still echoes of that period of time, which stay with us today.

I’m so excited it’s back. Part of the reason I’m on this side of the world is that I was at WordCamp Europe. It’s a couple of thousand folks. We had a lot of new folks coming. It really energizes everything. Our contributor day had like 600 to 700 people at it. So, those things are really the gateway drug, if you will, to open source participation.

The last time you were on the show, I asked you about your decision-making frameworks, which I ask everybody, and you said your goal was to push as many decisions down as fast as you could, and that you documented everything across WordPress. There was an internal blog system that you guys used. Is that still the case? Is that still your style of decision-making?

Yeah. I would say the main thing that’s changed is that we started doing some of these product reviews. So, on a periodic basis, I’ll go to basically every product in the company and we do presentations, and get feedback. Again, not something new. Something many other companies have done for a long time.

But the thing that actually made me realize we really needed to change is that last year, I took a sabbatical for three months, which is a benefit Automattic offers. Every five years, you get three months off. I had never done it, so I was setting a bad example. So I’m like, “I’m finally going to do this.” Also, that was interesting, just to get a little space. But then, when I came back, I was like, “Okay. What can I do differently?”

So I did 100 days of support with all of our different products, shadowing people, and talking to customers. I came out of it with a real sense of where we had accumulated technical debt, where we’d accumulated cultural debts, and where we had golden handcuffs in our business model, which I think is one of the most dangerous things. Because it’s not the business going down, but you can kind of see how, “Oh, if we just stay in this forever, it’s not a good look five years from now or 10 years from now.” So that was part of why I started recruiting some of the new executives we brought in and thinking about more drastic changes to how we worked.

What are some of the golden handcuffs you had in the business?

Well, just to give an example that you mentioned, our enterprise business is incredibly strong with media, and in some ways, we’re almost reaching a saturation point. There are not that many great publications left that we could bring on, and some of those publications are feeling the squeeze. 

So, just industrially, from an industry point of view, there are currently — although I think this will come back — secular headwinds to that business. If we were all media, that would be trouble. But of course, we have other things, finance, a lot of startups, Facebook uses it, all these other things. But is the product as well suited? So, that’s something we have to think about.

All right. Let me ask you about the big decision. You decided that WP Engine, which was a rival WordPress host, was free-riding on open source. You decided that you were going to cut them off. Many, many things happened. I just want to start at the very beginning of this. Your decision-making process involves all team decisions. Everything is pushed down. Everything is documented. Was the decision to take the fight to WP Engine a team decision or a Matt decision?

That was a team one with a lot of community feedback.

Walk me through that specific decision, that you’re going to put a lot of pressure on this rival hosting company.

I don’t know if I can right now. I think there’s a period in the future when we can dive a lot more deeply into this. Well, you’re a lawyer, right?

My job is to get you to talk. I was a horrible lawyer.

I think something I’ve learned in this process is not to talk as much while it’s going on. We’re very much deep, where both sides are spending millions of dollars a month on lawyers. I think that there will be things that play out, but the legal system moves a lot slower than I would like. So, we’re a little bit in the middle of it right now.

But you did put a lot of pressure on this company. You cut off some of their access, and you changed some code in different ways to further cut off their access. A lot of people saw you making unilateral decisions in a way that didn’t feel compatible with open source. Were those just Matt decisions, or were those also team decisions?

I take full responsibility for it all. So, ultimately, yes. If people are unhappy with it, they should hold me to account.

One of the ways that people held you accountable inside of Automattic was that you said, “If you don’t like it, you can leave,” and a bunch of people took buyouts and left.

We did a very generous six-month and a nine-month buyout offer. We called it an alignment offer. So, at that point, we were at our very peak, around 2,100 people. Actually, we’d already started coming down. But it was clear that there were some folks who just weren’t in line with where the business was going, or we had some folks who already had another job. It was this mix of everything in there.

So, the way I read it from the outside was that Automattic is a very idealistic company. You are a very idealistic person after all. All the times we talked, that’s my impression of you, is that you are a very idealistic company — WordPress and Automattic, very idealistic projects, particularly the Ecosystem side, the open source side of WordPress. 

And here is Matt, the benevolent dictator for life of WordPress, saying, “I’m going to squeeze this player out of the ecosystem in a way that reads against the ideals of open source itself.” Some people at your company were so incensed about this that you said, “You can walk, I’ll pay you to leave.”

That’s not how I would fully characterize it. So, first and foremost, I will say that it is true that my bias is towards optimism and radical openness. That’s my whole career. The downside of that, and this is not the first time this has happened in WordPress’s history, is that that can be taken advantage of. This is probably the fourth big time that there’s been a controversy like this in WordPress. It’s the first time in this media landscape, or when we’re this big. But there have been similar things in the past where, as a community, we had to say, by the way, not all the community, but a good portion of it, “Hey, there’s something that’s not okay here. And if we don’t stand up to it, it could threaten the future of us existing at all.”

I believe this was one of those decisions, or one of those situations, rather. I’m sure with the benefit of hindsight, there could be things that were done differently, probably on all sides. There’s a lot that we’re doing to make sure this particular thing doesn’t happen again. So, we’ll only make new mistakes. I will say that nothing like this has quite happened in the past, and I hope it never happens again in the future.

Wait, can you be specific about what you think happened that you need to prevent from ever happening again?

Let me think about how to put this. I think where WordPress has had the greatest success is when we can get everyone around the table — the commercial folks, the agencies, the developers, the geeks, and the anarchists. We get us all around the table and say, “Well, what are our shared goals and how will we achieve them?” Even the most rampant capitalist knows that you can over-squeeze that lemon. 

You don’t have to look far to find other open-source projects where the commercial interests sucked the life out of the project. That could be talent, it could even be as simple as… I’ll give one example of something people have criticized us for, and we get a lot of pushback on. It’s like, “How come the WordPress.org directory — the app store, if you will — doesn’t allow you to buy stuff?” We allow commercial things, but you have to buy them from the developer directly. We don’t have the one-click type thing.

It’s actually a very complex sort of incentive answer. As we saw other open-source projects do that, and some of them did it kind of early, what shifted in that community was a collaborative nature, both for the core software and for add-ons, to every person for themselves. So, imagine you and I were both plugin developers, and I created a better widget and you had another widget. Now, often, what happens in WordPress is I say, “Hey, what you have is cool. Let’s combine that.” And we’ll just combine the plugin. Or say, “Hey, this should be in Core. Let’s build it. Let’s submit it. Let’s get this so everyone has it.”

Now, if you and I are both selling that, we might each be making $20,000 a year or something off this. And so, there’s a local incentive for us not to open source it to the Core or to work with each other. If you play that out over the years, what happens is probably what happened with Joomla, where users felt nickel-and-dimed with every single feature. The core software can atrophy, because all the best development went into these extensions, and it hollowed out a bit. It’s my view from the outside.

You said “we” when you talked about WordPress.org. I think one of the things that a lot of people realize throughout this entire saga with WP Engine is a quote from you, that you control WordPress.org. This is a quote that you gave to us.

[Laughs] I regret that.

“WordPress.org just belongs to me personally,” said Matt Mullenweg in an interview with The Verge.

That’s been taken out of context so many times. I regret ever saying that. It’s the worst thing ever.

So, you’re saying that’s not true?

Because you were asking about specific ownership. Like, okay, what entity? I think you even said, “What entity is it that owns this?” which is very different from how the site runs, how decisions are made, how the code works, and everything like that.

So, why doesn’t Automattic own WordPress.org?

Why doesn’t Automattic own WordPress.org? Because I wanted to keep it separate, especially in the early days of Automattic. It was controlled by investors. So, I wanted this balance of power. 

WordPress predates Automattic, by the way, that there was this sort of a thing, which was not just the C Corp working for the fiduciary responsibility of shareholders. But that’s something that, yes, I kept separate and technically just I support, but I’m one of many, many people who support it. 

If you look at how WordPress.org is run, just how everything works, it is the result and the fruits of clearly tens of thousands of people and hundreds who work on it daily. So, I was very sad that it got mischaracterized as just me making all the decisions and just being the thing that belongs to me. I do regret that a lot, I will say that.

I appreciate that you regret it. I’m going to push back on the fact that we’re mischaracterizing, you saying the following.

I didn’t say you mischaracterized it.

Sure.

I said a lot of other people now take it out of context and are like, “We can’t have this thing rely on someone’s personal website.”

But that is an outgrowth of this quote and this fight. I’ve seen others say there’s a supply chain weakness if you have a WordPress dependency, where Matt Mullenweg owns WordPress.org, which is the plugin repository and the update repository. And if he gets mad at you, he can cut you off, which is what happened to WP Engine, right?

And it was reversed.

Well, sure.

So, my authority to do it in the long term is apparently not that big.

But you had a commercial dispute, and because you were just in control of this, you were able to cut off their access. And I think a lot of people-

We had a moral dispute, actually. It’s a moral dispute, as well as a commercial dispute. It’s an ethics dispute.

If they had paid the money, I don’t think you would’ve had a moral dispute. If they had said, “We commit this many engineers to the open source project,” would you have still had a moral dispute?

I don’t know if I can comment there.

Well, I’m asking because, to me, the dispute reads as if this is the classic open-source free-rider problem. There’s a big, vibrant open source ecosystem. Someone’s just going to take it, run the software, and sell the hosting. Maybe they provide a better level of customer support. They’re not going to pay into the project. Now they’re free riding on what should be the margin that funds the project as a whole.

That, in the abstract, has happened all the time throughout the open source community. Usually, the answer is, well, that’s just the price you pay for open source. This is a thing that happens. This is a thing the licenses enable. But the point of open source is that, eventually, the free rider will get far enough away from the thing that they will feel required to pay back into it to do what their customers want. This is the self-correcting mechanism. It sounds like you just didn’t see that happening, or you felt that was never going to happen.

By the way, there are probably $10 billion of hosting companies that I don’t feel are a threat or harm to the future of WordPress. So, I think that you can say I’m crazy or off my rocker, or whatever, but I would like to point out a long career with very few things like this happening, and that not everything is public. If I really felt like it was that much of a threat to the community, yes, I’m going to stand up. But we try to run things in a way that that happens very rarely, or not at all.

Do you think that a good solution here is for you to cede more control of WordPress.org back to some sort of central body? Because that feels like one of the main criticisms of this entire situation, is that you personally have control of what feels like a very centralized dependency for the WordPress ecosystem.

Yeah, and I can see why people see that because I’m really the spokesperson for a lot of these things.

Well, again, I know you feel this quote is mischaracterized. I’m just saying that I’m looking at a quote from you to us that says, “In my role as owning WordPress.org, I don’t want to promote a company which is legally threatening me in using the WordPress trademark. That’s why we cut off access from the servers.” To me, we quoted it. I don’t feel like we’re mischaracterizing it. But that’s you saying, “I have an enormous amount of power here and I will use it.”

I feel like I have an enormous responsibility to the tens of thousands of people who contribute to WordPress.org and WordPress on a regular basis. I feel like I have a responsibility to be the focal point for the arrows and the hate. So, developers who are actually writing all the code for WordPress and people who might not feel comfortable being the face of some of these objections don’t have to be. I’m happy to do that for them.

You’re 20 years into it. Open-source projects rely on characters like you a lot. Linus Torvalds, I think, is the most famous example of what we literally call “benevolent dictators for life.” At the end of the day, you can have all the hippie ideals you want, and then you need one person to take the arrows. That’s what you’re describing, one person to make some decisions and hold the value. Again, this is a pattern that repeats in many, many, many open-source projects.

Sometimes, that might be representing a minority. In many of these fights, it might not be the most popular thing to do. You really have to believe it’s the right thing to do, though. And history will show whether you’re on the right side of it.

How long do you think until this dispute is resolved? Is it just going to happen in the courts? Are you going to settle?

That’s a good question. I wish I could answer that. I really, really want to get back to the most collaborative version of WordPress possible. I think the complement, WP Engine, has an incredible set of talent. They have a lot of customers who are happy. I think they have a lot they can give back, to do, and to grow WordPress and their use of it. So, yeah, I’m optimistic. I’m generally pretty optimistic. There have been times I wasn’t, though. And I will say that the legal system and the whole law thing are incredibly challenging. If you’re an entrepreneur listening to this, I would say avoid it as much as possible.

We’re lucky to work with incredible, credible lawyers, but I think my disposition is much more on the product and engineering side. So, I remember that Steve Jobs quote where he said, “Look at how you’re spending your day and how you feel at the end of the day and everything.” And definitely, I want to get back to where that’s where I’m spending the vast majority of my time.

I just want to go through some of the moves here, and then I promise I want to talk about Tumblr to wrap this up. Mostly, I just want to talk about Tumblr with you, but I just want to go through some of the moves here. You cut off their access to WordPress.org, and you rewrote some plugins in a way that I think people felt was way over the line. You called them a “cancer.” Automattic stopped contributing to WordPress Core. You ramped it down. You’ve ramped it back up. Do you think that you were too aggressive? Do you think you went over the line? A lot of those moves are ones you’ve walked back.

I don’t know yet. We’ll see.

Why’d you walk them back? For example, not contributing to WordPress anymore? You walked that back. Why’d you walk that back?

So, remember I said you sometimes have to try the opposite. [Laughs] I mean, after 20 years of Automattic, there’s basically not a day, including weekends, that we haven’t done some work on core WordPress or other open-source projects. And again, not just me. When you’ve devoted so much of your life to giving back to something and it’s a very, very small group, but there’s some of that that’s just attacking you, everything you do, and you sometimes need a little break. So, think of that almost like an Automattic sabbatical, just like a person might need a sabbatical. 

What I was really hearing internally is that people want to regroup. It’s also a good opportunity because we’re making some of these other big changes to bring some, for example, core WordPress engineers over to WordPress.com and say, “Hey, what can we do? What could make this a better WordPress experience?” It was also an opportunity to reboot some of our development efforts that I’m glad we did, actually. But we missed it.

You missed doing it.

[Laughs] I can’t quit it. Honestly, I think I’m going to have to do open source for the rest of my life. It was torture for me, because we were like, “Okay, now I guess we need to not do any more relief. We don’t have enough stuff for a release, and so we’re going to have to push that.” That broke my heart, and then I just couldn’t stop thinking about it because we kept developing technology. And one thing we do is develop it on the Automattic side, and then we bring it to the open-source community. So, we had all this stuff developing, and there’s real-time co-editing and really exciting stuff. I’m like, “Oh, I got to get this out there.”

I’m going to tell you right now, the entire Verge team just heard you say real-time co-editing. When are you shipping simultaneous editing? As you know, when Vox Media moved to WordPress, this is the only thing I asked you for, and you told me it was coming. When is it coming?

I’m excited about it.

Is it this year?

It’s a “this year” thing.

It’s this year?

Murphy willing, yeah.

All right. And we’re going to hold you to that. By the way, you mentioned an Automattic sabbatical. I connected your sabbatical to this fight. There was a lot of hostility, a lot of back and forth, a lot of just pent-up anger. And then you took a sabbatical. Did this push you to do that?

No, that was like eight months away [from the lawsuit] or something. It wasn’t that close.

Okay. All right, let’s talk about Tumblr. You bought Tumblr ages ago from Verizon for some reason.

I actually just ran into [Verizon CEO] Hans Vestberg at an event. We talked.

I don’t mean you bought it for some reason. What I meant is Verizon owned it for some reason, and the idea that you bought it from Verizon is still very funny to me, because why did they ever own Tumblr? You migrated the entire backend of Tumblr to WordPress. 

No, no, we have not. We put that project on hold.

I swear I read a press release.

We announced that we were starting work on it.

Is that still not going to happen? Or why’d you stop it?

What we decided is that we want to focus as much on the things that are going to be noticeable to users and that users are asking for. This was more like an infrastructure thing, kind of like any big re-architecture. I still want to do it. It’s just cleaner. But right now, we’re not working on it.

One of the reasons that you wanted to do that back then, or at least when we first talked about the very idea of doing this, was that you saw some opportunity for ActivityPub. And there’s an ActivityPub plugin in WordPress. You bought the plugin from the developer and hired the developer. I’ve met him, he’s lovely. You can see, “Oh, that would really work for Tumblr. You move Tumblr to this backend, you have this plugin, and now you’ve got Fediverse Tumblr.” Do you still see that as an opportunity?

Yeah, so that would’ve been a free way to get it. And so that was one of the arguments for migrating everything to WordPress. And still going to do it someday, but in the meantime, I think if there was a big push to implement Fediverse, we would just do it on the Tumblr code base.

Is Tumblr growing? What are the user numbers there right now?

I would say Tumblr has a passionate, “never going to give you up” cohort of users, and it still acquires users at the young end. I’ll also say that the elements of how it works, I think, are very confusing if you’re new to it. So, one thing we’re thinking about is how we can make it a bit more accessible. And also, the constant thing you deal with at any social network is how to keep it friendly. So we want it to be someplace where you can go on the internet and leave refreshed, interact with art and artists, and your friends. On these networks, you have to do a lot of work to keep out the spam and the bullies.

You’ve personally waded into some Tumblr content moderation controversies. 

Oh my goodness, yeah.

Particularly around trans people. It’s a very bad time for trans people in America right now. Is that something you still want to be the face of? There’s a lot of pressure on that community, and that community is very, very focused on Tumblr.

Yeah, and so what I learned from that is that… I jumped in because it’s a community I care about deeply and want to show support for, but everything I did to try to show that this wasn’t targeting, then got twisted, or turned around, or a lot of misinformation. So, the learning is, and by the way, many other people would’ve told you this, is just don’t comment on content moderation decisions. Just point all that to the terms of service, the team, and everything like that. More explanation does not make it better. 

Again, probably one other lesson from 2024 is that, particularly in open source, often we talk through it, we post through it. And we have these vigorous debates, and we come out the other side, often with a consensus, or at least a shared view of reality. And in the media landscape of today, that doesn’t happen, sometimes ever. So, it’s just better not to try to engage with everything. And, as an early internet person, this drives me crazy sometimes, but I think I’ve learned that lesson.

One of the reasons I’m asking about this, again, is that this is just a bad time for a lot of marginalized communities in America. I think of Tumblr, I think of young people, I think of queer people, I think of people of color. A lot of the culture begins on Tumblr, from those communities. It is not clear to me that the big platforms are committed at all to making those communities feel safe any longer. They might’ve been.

Why do you say that?

There was a time when, performatively, these platforms wanted these communities to feel safe. I think Instagram is deeply aware that Instagram is not the same without gay people and Black people on it. They know, they just know, but they are playing a different game with the Trump administration.

Tumblr is sort of out of the limelight. Are you going to do anything to try to make those communities feel safer there? It feels like an opportunity to take people from the big platforms and put them in a place where creativity and free expression are more valued in that specific way.

I know there’s a narrative that these bigger platforms have changed their mind there. I would challenge that. They might be doing things performatively there, but I bet if you went to the on-the-ground person who does the very, very tough job of looking at the worst stuff on the internet to protect the rest of us, that’s someone like a firefighter or a police officer. I think we will appreciate that job so much more in the future because they’re essentially sacrificing some of their mental health to protect the rest of us. I would say they are — with very, very few outliers who get weeded out and fired immediately — there for all the communities you spoke about, and driven by a sense of trying to protect and help safety.

On Tumblr, it’s the work we do every day, and I want us to be, again, judged by the results and the actions. It doesn’t mean that there will never be a mistake. In fact, we’ve actually had to let someone go before because they weren’t operating in line with the values that we have as a company, or in line with our terms of service and everything like that. But yeah, that is the exception that proves the rule, I believe. It’s hard to see that from the outside, though, because you don’t see the hundreds of thousands or millions of things that are moderated every day. You see the one time it messes up.

Yeah, I understand, broadly and across the industry. I certainly understand that content moderators, no matter where they work, have horrible jobs.

I’m just saying that I look at Meta, which is saying very publicly that it’s shutting some of this moderation down and going to use community notes, and then Mark Zuckerberg is going to be in the White House. I think, “Well, there’s a whole community there that feels under attack by just the gestalt of that. Just the way that feels is bad. And Tumblr exists, and Tumblr has always been the place for those communities to go. It is what makes Tumblr, Tumblr. 

I’m just wondering if you see it as an opportunity to claw back some share from the billionaire-owned big platform companies that feel like they’re playing a vastly more political game.

Yes, and we do see waves of people come over when these incidents happen in various places. I think what we have to do now is retain those folks because the app needs to be super fast and performant. The ads need to be good. By the way, some of the criticism I made of media things, you can make about some of the advertising we have in the Tumblr app today. Not all the ads are to the standard that I think we should be held to. The app needs to just be more intuitive than I think it currently is.

I think what happened is that Tumblr invented so much stuff. It was like the pioneer that had images on posts, things like that, before any other social network did. We got kind of attached, as a team, to some of the different ways we did things. Now, in the meantime, across every social media platform, there’s almost like a set of primitives that are universal, like the Platonic ideal of certain interaction modes or how things should work. For instance, replying, commenting, liking, and liking something versus viewing the likes. 

There are affordances in the interface with so many of these things. And when you drift too far from that, it just creates a lot of mental friction when you use it. So what we have to navigate is maintaining the character of Tumblr, and everything that people love about it, while also not frustrating you when you go between different apps or different experiences.

Everything is turning into TikTok. Do you think Tumblr needs to turn into TikTok?

No, I don’t think so. And there are definitely ways I want to evolve the business model. We’ve invested a lot in subscriptions and things like that. Self-serve, first-party advertising, which we’ve gamified. You can advertise someone else’s thing. So I think those models, at scale, could be a non-TikTok incentive. Like, it will be smaller than TikTok forever, but it could create different incentives in the business model that I feel pretty good about.

Is Tumblr sustainable today?

It is still not profitable. So we’re still supporting it and subsidizing it with our other products at Automattic.

How much runway do you want to give it?

Everything. Obviously, we’ve invested a ton in Tumblr. I’m a believer in its future. And so that’s part of why I want to make it sustainable, because that means it doesn’t have to go on the benevolence of myself or anyone else. It can stand on its own.

There was a report last year from 404 Media that part of the revenue model would be to sell Tumblr data to Midjourney and OpenAI for training purposes. Is that true? Is that going to be part of the revenue mix?

Gosh, where are content licensing things? So, Automattic has done content licensing deals in the past. Often, it’s things that people already had, because it was on the web, so it was already part of the indexes and other stuff like that. I don’t think that’s going to be a big part of the revenue model going forward, and the whole AI content thing, fair use, everything, feels like the question that will work itself out in the legal system over the next five to 10 years.

So wait, do you have a licensing deal for OpenAI to train on Tumblr data?

I don’t think I can comment on which deals we have or which we don’t, but I will speak more generally about how the AI content licensing thing feels very much up in the air.

Do you want to have those deals?

I want creators to get paid for their creative work. Absolutely.

So if you had that deal, you would pass the revenue through to Tumblr creators?

Well, passing revenue through, you have to have a certain threshold of revenue. [Laughs] You have to know your customer, and you have to get tax information. So there are thresholds at which I think we need to get to, where this can become part of what can contribute to creatives. I don’t see a path for that right now, but it’s definitely something that I know at least some people at the AI companies are thinking about. And I’m hopeful for it, something there. But I don’t know what it’ll be. I don’t think it’s going to be micropayments. I don’t think it’s going to be crypto. We’ll see.

I want to end by just talking about the future of the web, and really just about the open protocols: ActivityPub, AT Protocol, and Bluesky. The last time we spoke, ActivityPub felt like it was going to be the winner, just by default. Threads had adopted it, everyone was talking about it, and there were a lot of services being built on it. You were really high on it.

Then Bluesky showed up, and all the people on its network are sort of on that protocol, right? Threads is, obviously, a big player, and they’ve launched a bunch of Fediverse features with ActivityPub, but it’s not driving adoption, maybe, in the way that Bluesky’s community might drive a bunch of adoption of AT Protocol. There’s stuff that bridges them together, but we’re having a protocol fight. Here we are having an open interoperable protocol fight. This is some Matt Mullenweg stuff. How do you think that’s going to play out?

Oh, that’s a good question. I think the failure mode of internet idealists is protocol-first thinking. So what I would love to do — and actually, maybe I should host this — is a summit where we get together and don’t argue about how the servers are talking to each other, but what our current and potential audience, the customers, the users, wants the most, and how are they not being served by the competition. Because that’s the other thing we do, we say, “Oh, we just make this network, the people will come,” and miss all the network effects and the lock-in effects of these social networks with billions of people on them. So it’s tricky. 

Even email, sort of the most famously open thing that’s still running, is effectively closed down for most people. If you run your own email server, most of your email is going to go to spam, to whoever you send it to, on Gmail, and stuff like that. So we’ve had a re-centralization of a decentralized protocol. And this is the kind of thing that can happen if you don’t have the right incentive structures in place for something that is very, very open. And this is kind of like a version of the problem that I’m trying to avoid in the WordPress community and the open-source projects that we support.

So, I wish I had an easy answer for you. I think the best way that we’ll get there is with a relentless focus on the design, the user interface, and iterating as fast as possible, and not getting locked in these local maximums of what your existing user base might be. Sometimes you have to do something incredibly unpopular with your existing user base to get to that next level, to unlock the order of magnitude of growth that’s the billions of people who don’t use it yet.

Another mistake we make in open source is that we talk about how the community wants this, or the users want this, but we have to account for the voices of the 7.9 billion people who don’t use WordPress yet. And that’s tricky.

I was really high on interoperable social networking, and interoperable sites using ActivityPub or AT protocol. It doesn’t matter to me, actually. At the end of the day, I just want to see more interoperability of these networks. Do you still think that’s going to happen?

Yeah, if I had a bet on one, actually, you know what I’d bet on is the Matrix protocol, the dark horse in the corner.

A third one out of nowhere. Just briefly explain what that is, and why.

Matrix is taking more of a messaging-first approach. So think of it kind of like an open-source Slack. It’s what Beeper actually used as a backend in its cloud version. There are also cool extensions to it that can do things like pure peer-to-peer. That feels like a disruptive technology, that if I had to just put five, a bet for five bucks on it, a long-term call option or something, I would maybe put it on that one.

All right, Matt, this has been great. You’re going to have to come back again sooner than three years. I feel like I have a million more questions for you, but this is so great. Thank you for all the time.

Well, maybe we’ll celebrate the co-editing.

The day you launch it, I’ll give you a full day, a 24-hour live stream on Decoder, the day you launch co-editing.

Awesome. Thank you. It’s been a pleasure talking. And thank you so much for what you do, and also for advocating for the open web and open protocols. I really appreciate it.

We run the last website on earth, Matt, you know how it is. Thank you.

_{Questions or comments about this episode? Hit us up at [email protected]. We really do read every email!}

Google announced a variety of new features for its Classroom software suite, including free Gemini AI tools for educators and NotebookLM for users under 18 — the first time the tool has been available to minors.

Teachers with a Google Workspace account will have a new dedicated Gemini tab in their Google Classroom, offering tools that can help brainstorm lesson plans, make math problems, or draft emails, among others. For those with Education Plus Edition accounts, Gemini audio lessons will also be offered as add-ons. Another new tab for Analytics will allow teachers to track student performance, highlighting any who have been showing recent improvement or who are missing assignments and may need more support.

Google will also be opening up its NotebookLM research and note-taking tool to users under 18 “in the coming months.” Students will be able to use NotebookLM to access things like interactive study guides and podcast-style audio overviews based on source materials their teachers upload. And educators can create their own custom Gems, mini custom Gemini agents, based on their curriculum to share with students. Students will be able to interact with a Gem to get extra help or learn more about a topic it’s based on. This feature goes one step further, as well, as administrators will also be able to share Gems with the teachers they oversee in their district, to help with lesson planning or whatever else administrators might like to micromanage.

A few other new tools coming to educators include Google Vids for generative AI video creation, a new “help me create a form” feature in Gemini, and the launch of Google’s Class Tools. The Class Tools feature was announced earlier this year, and now that it’s live, it will allow teachers to manage their students’ Chromebooks — like sharing content to their screens and minimizing distractions — during a lesson. By connecting directly to their students’ accounts, teachers can push a workbook, video, presentation, or even a quiz directly to their screens.

Speaking of Chromebooks, Google is also highlighting that it has more than 12 new Chromebook, Chromebook Plus, and Chromebook OPS (Intel’s Open Pluggable Specification) devices for front-of-classroom displays coming — at some point. There’s no specific timeframe given, though the list includes the recently announced Lenovo Chromebook Plus 14 that I got a preview of, which is launching today.

With all these new Gemini tools for teachers and students, Google is also announcing that Gemini Education is now the standard tier for its base-level education accounts. It offers Gemini 2.5 Pro AI models, which educators can access with higher limits than base-level consumer accounts. Existing Gemini Education and Gemini Education Premium add-ons will be unified as Google AI Pro for Education.

Apple seems to have big plans to expand its lineup of extended-reality (XR) devices by launching new Vision series headsets and smart glasses, according to a report from supply chain analyst Ming-Chi Kuo. In his report, Kuo shares Apple’s potential roadmap for seven new head-mounted wearables, which could include the release of Meta Ray-Ban-like smart glasses, followed by XR glasses with a display.

As noted by Kuo, Apple could launch an M5-powered Vision Pro headset this year before showing off a “substantially lighter” Vision Air in 2027. This device will reportedly cost less than the $3,499 Vision Pro and use Apple’s flagship iPhone processor.

Kuo has also detailed Apple’s rumored plans to launch smart glasses in 2027, which could support voice control and gesture recognition, along with the ability to record videos with a built-in camera and an “AI environmental sensing” feature. Last month, sources told Bloomberg’s Mark Gurman that Apple’s upcoming smart glasses will be similar to the Meta Ray-Bans smart glasses “but better made.”

Rumors about Apple’s upcoming smart glasses have been floating around for over a year now and have only ramped up following the success of Meta’s Ray-Ban smart glasses. Meta recently announced a new pair of glasses with Oakley that capture 3K video, while Google is working on Android-powered XR glasses of its own.

Further down the road, Apple could launch a 2nd-gen Vision Pro that features a new, “significantly lighter” design and a cheaper price in 2028, according to Kuo. The company Apple is also reportedly working on display-equipped XR glasses with voice control and gesture recognition for 2028 as well.

There are a couple of other devices with murkier timelines, including a second-gen pair of XR glasses and a device that users can tether to an iPhone or Mac. Kuo reports that development on the tethered device has been paused due to “insufficient competitive advantage,” as Apple’s device would weigh more than rival products.

Google has agreed to purchase electricity from a forthcoming nuclear fusion power plant, the so-called holy grail of clean energy that scientists have been chasing for more than half a century. 

While the fusion industry reached a significant milestone a few years ago, the technology has yet to prove whether it can be a technically feasible or commercially viable option. Nevertheless, the deal Google announced today shows confidence in the possibility of harnessing nuclear fusion to power its data centers.

The news follows the release of Google’s latest sustainability report on Friday, which shows its greenhouse gas emissions continuing to climb despite its clean energy commitments. Even in a best-case scenario, fusion reactors wouldn’t be online in time to help Google meet its goal of slashing emissions by 2030.

“It’s a world-changing technology in our view.”

“It’s a world-changing technology in our view,” Michael Terrell, head of advanced energy at Google, said in a Friday call with reporters. “Yes, there are some serious physics and engineering challenges that we still have to work through to make it commercially viable and scalable. But that’s something that we want to be investing in now to realize that future.”

Specifically, Google has agreed to purchase 200 megawatts of “future carbon-free power” from Commonwealth Fusion Systems (CFS), a private company that is building the fusion plant in question and in which Google is also an investor. Offtake agreements like this are common for other sources of electricity as a way to fund new projects. What’s different here is that the timeline for nuclear fusion is far more uncertain.

Nuclear fusion researchers are attempting to recreate the way stars generate their own light and heat. In our sun, hydrogen nuclei fuse together, creating helium and a tremendous amount of energy. If someone can figure out how to do that in a controlled way on Earth, they would unlock a potentially limitless source of carbon pollution-free energy. 

But doing so takes extreme heat — more than 100 million degrees Celsius. With such high temperature and pressure requirements, scientists weren’t even able to achieve a net energy gain from a fusion reaction until 2022. And so far, only the Lawrence Livermore National Laboratory has been able to do this. (Today’s nuclear power plants generate electricity through fission — releasing energy by splitting atoms apart rather than fusing them together; a process that leaves behind radioactive waste.)

CFS says the technology is finally advancing fast enough to connect its first fusion power plant to the electricity grid in Virginia by the early 2030s. Virginia is also home to “data center alley,” where tech companies have built or expanded facilities to develop new AI tools. Other energy experts The Verge has spoken to over the years, however, think it could take decades longer for fusion to become commercially available. CFS is currently building its pilot plant in Massachusetts.

Google and CFS aren’t alone in their ambitions. Microsoft inked a deal in 2023 to purchase electricity from a nuclear fusion generator being developed by Helion Energy, which is supposed to be ready by 2028. Around $8 billion from primarily private investors has flowed into fusion startups in recent years, the Washington Post reported last week.

Google first announced an initial investment in CFS to support R&D back in 2021. It’s now making a second capital investment, although the companies aren’t disclosing any concrete numbers. Google has also been an investor in another fusion company, TAE Technologies, since 2015 — although the recent deal with CFS marks Google’s first offtake agreement for fusion. 

In 2021, Google pledged to reduce its planet-heating pollution by 50 percent by the end of the decade compared to a 2019 baseline. The company’s latest sustainability report, however, shows that the its carbon emissions have actually ballooned by more than 50 percent since 2019 as it doubles down on AI. 

The 200MW deal with CFS represents a fraction of Google’s carbon-free energy purchases. It says it has signed more than 170 agreements since 2010 to purchase 22,000MW of clean energy — much of that in wind and solar projects that it sees as more feasible near-term solutions.