• As a private chef, I shop for groceries via Amazon delivery at least four times a week.
• Organic carrots and multicolored bell peppers are great for crudités.
• I like the organic chicken thighs from Whole Foods when I can't make it to a butcher.

I'm a private chef and a mom, so convenience is key for me when it comes to errands and essential tasks.

I often don't have the time to run to the grocery store when I have three different clients to cook for in one day. When I'm in a pinch, I turn to Amazon's grocery delivery services.

The online retailer has a few options, like Amazon Fresh and GrubHub, but I typically choose from its Whole Foods items.

Now that I've checked out dozens, if not hundreds, of shopping carts via the Amazon app, I know exactly which items I can confidently order online.

Parmesan is pretty much always in my cart.

The one item I always add to my cart is grated Mitica Parmigiano Reggiano. The Whole Foods product costs $26 a pound, but it's worth it.

It's an imported dream of beautiful, crystalized, salty cheese that adds creaminess and complexity to my pastas, soups, and roasted vegetables.

I like the grated version because it's ready to use whenever I need it.

Bell peppers fill out a number of dishes.

I love 365's three packs of peppers for stir-fries, crudités, and frittatas.

I can order an assorted pack of three for $5 (less than $2 a pepper), which is great when I'm on a budget.

I use cilantro a lot in my cooking — at home and with clients.

I use a lot of cilantro, but unfortunately, herbs can be expensive.

Luckily, the fresh cilantro bunches from Whole Foods are generously sized and super flavorful. A regular bunch is $1, and an organic one is $2.

I can usually make one bunch last a few days between cooking for my clients and my family.

Chicken is one of my go-to protein sources.

I prefer buying meat from a butcher, but sometimes, I just don't have the time.

I'm actually very happy with the quality of Whole Foods' organic chicken. It's flavorful and cooks up tender and juicy.

I love getting skin-on thighs for grilled tandoori chicken or any one-pot dishes that call for crispy skin.

At $6 a pound, the price also feels reasonable to me.

Carrots are great for dinner parties.

Organic carrots with their tops on are my go-to for dinner-party crudités. They also go great on a pan of roasted veggies.

The $3 bunch of carrots cooks up beautifully, and the array of purple, orange, and yellow hues impresses kids and adults alike.

Tacos are a hit with my clients, so tortillas are a must.

Many of my clients request different types of tacos throughout the week, so I always make sure to order good tortillas to elevate the flavors.

Vita Hermosa tortillas aren't available everywhere, but when I order via Amazon, they're usually in stock at Whole Foods.

I get an eight-pack for about $6.

The 365 fire-roasted corn is one of a kind.

This might seem random, but not many stores carry pre-roasted frozen corn.

I love adding 365's $3.70 pack of fire-roasted corn to my cart when I'm shopping online. I'll toss the kernels in salads, pile them on top of my tacos, or even add them to soups.

• Meta CEO Mark Zuckerberg told employees there will be no changes to the hybrid work policy.
• In a meeting held Thursday, the Meta chief said "the status quo is fine."
• Most staff are expected to work in an office at least three days a week.

The future of in-office work was top of mind for Meta employees on Thursday during a wide-ranging Q&A with CEO Mark Zuckerberg. The bottom line? No changes for now.

Most Meta employees are expected to head into the office at least three days a week, but some staff have worried the company would change the policy and demand a full in-office return.

Following a long opening monologue that touched on AI assistants, AR glasses, and the "intense" year ahead, Zuckerberg turned to questions submitted by staff. The top-voted question was about remote work.

"Basic summary: no change," Zuckerberg told staff. "There's a lot of things going on in the world. I just have not been focused on this at all. I think the status quo is fine. Carry on."

Zuckerberg's answer was met with applause by employees watching the all-hands live. The Meta chief remarked that two of the top-five voted questions submitted concerned in-office work.

Meta's three-day policy is in line with some other big tech companies, including Google. However, some companies want workers back full time. Dell sent a memo to employees on Friday calling them back into the office five days a week.

In Thursday's all-hands, Zuckerberg told employees to "buckle up" for an "intense" year ahead. He defended several recent changes around content moderation and fact-checking, and said the company had an opportunity to have a "productive partnership with the United States government."

Zuckerberg also discussed how DeepSeek could benefit Meta's open-source AI strategy, and why Meta was slow to respond to TikTok's rise.

Do you work at Meta? Contact BI reporters from a nonwork email and device at [email protected] or [email protected]. You can also reach them via Signal at hughlangley.01 and jyotimann.11

• Chinese startup DeepSeek shook the tech world and markets when it released R1, its new AI model.
• The West is now trying to reproduce R1 on its own terms and cut out Chinese servers.
• Recreating R1 from scratch can help researchers build better models and validate DeepSeek's claims.

Silicon Valley doesn't want to get caught out again. It's scrambling to replicate DeepSeek's AI model, the cheaper Chinese tech that shook Wall Street and is freely available for anyone to adopt.

Companies like Microsoft and Amazon have already made versions of DeepSeek's R1 models available on their cloud platforms. This allows people to use the models, which appear to match the capabilities of rivals like OpenAI, while keeping data from being sent to servers in China.

But there are also attempts to replicate DeepSeek's cost-efficient AI from the ground up — and see if all of the Chinese AI lab's market-moving claims hold up.

One major effort is being led by Hugging Face, a platform for researchers in AI's open-source community to collaborate and share their trade research notes and ideas for free.

Leandro von Werra, head of research at Hugging Face, told Business Insider that the company expected to complete its replication efforts within "weeks." He described the mood at Hugging Face as "kind of like Avengers assemble" as they dissect the inner workings of R1.

DeepSeek obtained open-source licensing for its model from MIT, which means a lot of the vital components of the recipe needed to build R1 have been laid out in the company's publicly available technical paper.

However, there are some elements of R1 that remain unclear.

In a December paper on V3, DeepSeek's earlier model, the Chinese company said training cost $5.6 million in total. The cost was calculated based on its use of H800 GPUs, a less powerful version of Nvidia's top chips, at a rental price of $2 per GPU hour.

Right now, no one can be quite sure what the actual development cost of R1 was, von Werra told BI.

DeepSeek's research paper also did not share what was required to bake reasoning capabilities into V3 to then produce R1.

That said, von Werra thinks it won't remain a mystery for long. "I don't know about the compute number, we can only guess at this time," he said. "I think one thing that's exciting about our reproduction is we're going to find out pretty quickly if the numbers hold up."

Learning from DeepSeek

Some quarters of Silicon Valley responded swiftly to the launch of DeepSeek. This week, Meta set up "war rooms" for its researchers to analyze DeepSeek, The Information reported. Sam Altman, the CEO of OpenAI, said Tuesday his company would accelerate the release of "better models."

DeepSeek's decision to publish its findings and make its R1 model open gives researchers worldwide an insight into its novel approach.

The main technique used to make R1 so capable was "pure reinforcement learning," DeepSeek's paper said. This, Hugging Face researchers said in a blog on Tuesday, can "teach a base language model how to reason without any human supervision."

The researchers also know more specific technical details about why R1 caused such a stir in Silicon Valley and wiped $1 trillion from US stocks on Monday. For instance, the reasoning model is what's known as a "mixture of experts" model — industry-speak for a model that can be "pre-trained with far less compute." It also involves subtle changes to its architecture by introducing techniques like "multi-token prediction," first introduced by Meta, that make models more efficient.

Hugging Face's von Werra notes that details like this from DeepSeek have helped the industry better understand how a reasoning model like OpenAI's closed-source o1 was built. "Everybody thought this is the secret that is going to take a while to crack," he said.

DeepSeek comes to platforms

Days after its launch, DeepSeek flew to the top of Apple's Top Free Apps chart. While everyone wanted to try the latest AI tool, DeepSeek's policy of storing user data in China has prompted security concerns.

This spurred US companies to make R1 available on their own platforms so customers could use the Chinese AI model while cutting out China's servers.

Lin Qiao, CEO of Fireworks AI and former head of the PyTorch team at Meta, told BI that one clear reason for doing so was to ensure AI developers and users continue to get access to top model innovations.

"The approach we have been taking is always to enable state-of-the-art models for developers the fastest," she said. "DeepSeek is one example."

Her company, founded in 2022, made R1 available on its platform after congratulating DeepSeek for "pushing the boundaries of what's possible in open models." It has been made available via its serverless service, as well as through on-demand and for enterprise customers.

Others have followed suit. On Wednesday, Microsoft announced that it was making R1 available in its model catalog on its AI development platform Azure AI Foundry to make it "accessible on a trusted, scalable, and enterprise-ready platform."

Asha Sharma, corporate vice president at the tech giant, wrote in a blog that R1 "offers a powerful, cost-efficient model," but one that it had done "rigorous red teaming and safety evaluations" on before introducing as a new model to its library.

Amazon Web Services is making a similar move. Swami Sivasubramanian, vice president of AI and data at AWS, said this week that the company's "commitment to AI accessibility" meant R1 was being made available on its platforms like SageMaker and Bedrock.

Perplexity CEO Aravind Srinivas, whose AI startup made R1 available to users of its search platform this week, said in a Tuesday X post that downloading the model onto its servers could also help control the way it responds to user queries as well as ensure those queries don't go to servers in China.

DeepSeek's AI appears to censor sensitive information about China, such as refusing to answer questions about the 1989 Tiananmen Square protests. Srinivas said that Perplexity's version of R1 had no censorship and shared its accurate response to what happened in Tiananmen Square.

David Sacks, the White House's AI czar, offered one reason why Perplexity's R1 integration was an important way to reproduce R1 in the West. "This is one of several ways that you can try DeepSeek R1 without downloading the app or sharing any data with a Chinese company," he said on X.

Many see DeepSeek as an example of China challenging American AI hegemony using a tried-and-tested playbook. OpenAI, which has a lot to lose from DeepSeek making its technology freely available, said on Wednesday it's investigating whether the Chinese firm "inappropriately" replicated its models for training.

For von Werra, it's a full-circle moment. The whole field started as open-source, so seeing efforts to make a leading reasoning model available for free is welcome, he said.

"I think in the end, everybody's going to get better models and do cooler things," he said. "I feel like it's a win-win situation."

• OpenAI and the White House have accused DeepSeek of using ChatGPT to cheaply train its new chatbot.
• Experts in tech law say OpenAI has little recourse under intellectual property and contract law.
• OpenAI's terms of use may apply, but are largely unenforcible, experts say.

This week, OpenAI and the White House accused DeepSeek of something akin to theft.

In a flurry of press statements, they said the China-based upstart had bombarded OpenAI's chatbots with queries, hoovering up the resulting data trove to quickly and cheaply train a model that's now almost as good.

The Trump administration's top AI "czar" said this training process, called "distilling," amounts to intellectual property theft. OpenAI, meanwhile, told Business Insider and other outlets that it is investigating whether "DeepSeek may have inappropriately distilled our models."

OpenAI is not saying if the company plans to pursue legal action, instead promising what a spokesperson termed "aggressive, proactive countermeasures to protect our technology."

But could they? Could they sue DeepSeek on "you stole our content" grounds, much like the grounds OpenAI was itself sued on in an ongoing 2023 copyright claim filed by The New York Times and other news outlets?

Business Insider posed this question to experts in technology law, who said challenging DeepSeek in the courts would be an uphill battle for OpenAI, now that the content-appropriation shoe is on the other foot.

OpenAI would have a hard time proving an intellectual property or copyright claim, these lawyers said.

"The question is whether ChatGPT outputs" — meaning the answers it generates in response to queries — "are copyrightable at all," said Mason Kortz of Harvard Law School.

That's because it's unclear that the answers ChatGPT spits out qualify as "creativity," he said.

"There's a doctrine that says creative expression is copyrightable, but facts and ideas are not," explained Kortz, who teaches at Harvard's Cyberlaw Clinic.

"There's a huge question in intellectual property law right now about whether the outputs of a generative AI can ever constitute creative expression or if they are necessarily unprotected facts."

Could OpenAI roll those dice anyway, and claim that its outputs actually are protected?

That would be unlikely, the lawyers said.

OpenAI is already on the record in the New York Times copyright case arguing that training AI is an allowable "fair use" exception to copyright protection.

If they do a 180 and tell DeepSeek that training is not a fair use, "that might come back to kind of bite them," said Kortz. "DeepSeek could say, 'Hey, weren't you just saying that training is fair use?'"

There's arguably a distinction between the Times and DeepSeek cases, Kortz adds.

"Maybe it's more transformative to turn news articles into a model" — as the Times accuses OpenAI of doing — "than it is to turn outputs of a model into another model" as DeepSeek may have done, Kortz said.

"But this still puts OpenAI in a pretty tricky situation with regard to the line it's been towing regarding fair use."

A breach of contract lawsuit is more likely

A breach-of-contract lawsuit is much likelier than an IP-based lawsuit, though it comes with its own set of problems, said Anupam Chander, who teaches technology law at Georgetown University.

The terms of service for Big Tech chatbots like those developed by OpenAI and Anthopic forbid using their content as training fodder for a competing AI model.

"So perhaps that's the lawsuit you might possibly bring — a contract-based claim, not an IP-based claim," Chander said.

"Not 'you copied something from me,' but that you benefited from my model to do something that you were not allowed to do under our contract."

There's a possible hitch, Chander and Kortz say. OpenAI's terms of service require that most claims be resolved through arbitration, not lawsuits. There's an exception for lawsuits "to stop unauthorized use or abuse of the Service or intellectual property infringement or misappropriation.

There's a larger hitch, though, experts say.

"You should know that the brilliant scholar Mark Lemley and a coauthor argue that AI terms of use are likely unenforceable," Chander said. He was referring to a January 10 paper, The Mirage of Artificial Intelligence Terms of Use Restrictions, by Stanford Law's Mark A. Lemley and Peter Henderson of Princeton University's Center for Information Technology.

To date, "no model creator has actually tried to enforce these terms with monetary penalties or injunctive relief," the paper says.

"This is likely for good reason: we think that the legal enforceability of these licenses is questionable," it says. That's in part because model outputs "are largely not copyrightable" and because laws like the Digital Millennium Copyright Act and the Computer Fraud and Abuse Act "offer limited recourse," it argues.

"I think they are likely unenforceable," Lemley told BI of OpenAI's terms of service, "because DeepSeek didn't take anything copyrighted by OpenAI, and because courts generally won't enforce agreements not to compete in the absence of an IP right that would prevent that competition."

Lawsuits between parties in different nations, each with its own legal and enforcement systems, are always tricky, Kortz said.

Even if OpenAI cleared all the above hurdles and won a judgment from a US court or arbitrator, "in order to get DeepSeek to turn over money or stop doing what it's doing, the enforcement would come down to the Chinese legal system," he said.

Here, OpenAI would be at the mercy of another extremely complicated area of law — the enforcement of foreign judgments and the balancing of individual and corporate rights and national sovereignty — that stretches back to before the founding of the United States.

"So this is, a long, complicated, fraught process," Kortz added.

Could OpenAI have protected itself better from a distilling incursion?

"They could have used technical measures to block repeated access to their site," Lemley said. "But doing so would also interfere with normal customers."

He added, "I don't think they could, or should, have a valid legal claim against the searching of uncopyrightable information from a public site."

Representatives for DeepSeek did not immediately respond to a request for comment.

"We know that groups in the PRC are actively working to use methods, including what's known as distillation, to try to replicate advanced U.S. AI models," OpenAI spokesperson Rhianna Donaldson told BI in an emailed statement.

"We are aware of and reviewing indications that DeepSeek may have inappropriately distilled our models, and will share information as we know more," the statement said. "We take aggressive, proactive countermeasures to protect our technology and will continue working closely with the US government to protect the most capable models being built here."

• When we travel, my partner prefers heading outdoors but I like exploring big cities.
• We plan trips we both enjoy by finding experiences and stays that feature our shared interests.
• We stay open-minded when trying activities one of us is way more excited about than the other.

In many ways, my partner and I are proof that opposites attract.

For starters, he adores mountains and hikes, while I prefer a big city and spas. Although we both love to travel, our preferred itineraries, activities, and destinations can be pretty different.

Over time, though, we've figured out how to plan couples trips that we can both enjoy together.

We try to book experiences and stays that combine our interests

Whenever possible, we try to find activities and excursions that feature something we're both passionate about, like cycling.

I love a spin class, and he adores mountain biking. So, when we travel, we look for bike tours. One of our favorite trips featured a bike tour through the French countryside that included a picnic.

We also combined our interests on a day trip to Tolantongo, just a few hours north of Mexico City.

Since we both enjoy being in and around water, we relaxed in hillside hot-spring pools, ziplined, explored caves hidden behind waterfalls, and lounged by a thermal river to end the day.

The experience combined the relaxing vibes of a spa day (my favorite) with a stunning outdoor landscape (his favorite) that blew us both away.

Plus, we try to combine our priorities and preferences when booking accommodations.

Last summer, he wanted to go horseback riding and camping in Montana's Big Sky. So, I found an Airbnb close enough to the wilderness (for him) with a massive bathtub (for me) to relax in after a day in the great outdoors.

We're patient and open-minded when doing activities the other is really excited about

Sometimes, the best gift you can give your partner is letting them introduce you to an activity that makes them light up inside.

This helps us lean into our individual interests while still making memories together. For example, I love spending a day vintage shopping for the perfect pair of boots or a jacket that I definitely don't need.

My partner is wonderful and patient as I rummage through racks of secondhand goodies — and I'm the same way when he takes me on a three-hour hike.

Plus, while trying activities my partner is excited about, I've learned more about myself and what I enjoy.

I'll never forget the rush of endorphins I felt while whitewater rafting in Colorado or how special it was to see a moose and her baby while hiking in Wyoming.

When in doubt, we take turns planning

If we're struggling to find common ground, we'll take turns choosing destinations and activities.

Usually, if I pick one trip, he'll choose the next. If neither of us really has a preference, we'll draw a destination out of a hat.

Doing so is easy because we prioritize each other's happiness and know each other's limits — he knows I'm open to camping, but only if there's electricity nearby.

• Sam Mewis launched the women's soccer podcast "The Women's Game" about a year ago.
• The former US soccer star has grown the show into one of the most listened-to in women's sports.
• Mewis explains her conversational approach and expansion plans with the media company Men In Blazers.

Former US soccer star Sam Mewis is part of a new wave of media personalities reshaping women's sports coverage.

The 32-year-old former US Women's National Team player was considered one of the best in her position until a knee injury forced her to retire from soccer in 2023. Mewis could no longer play the game she had dedicated her life to, but that didn't mean she couldn't stay close to it.

Mewis joined the soccer-media company Men In Blazers to launch about a year ago its first women's sports vertical with a podcast, called "The Women's Game."

Until recently, women's sports were underserved by mainstream media. Because TV coverage was not widely available, women's sports have expanded in other areas like podcasting.

Mewis has built "The Women's Game" into one of the most listened-to podcasts in women's soccer. According to podcast database Podscan, the show had averaged 11,700 listeners per episode since its launch and was the 22nd-most listened-to soccer podcast in the US on Apple Podcasts as of January 30. The other podcasts on the list largely focused on the men's sport.

"In many ways, what we were able to accomplish last year exceeded my expectations for going into the first year of a brand new vertical," Mewis told Business Insider.

'The Women's Game' feels like talking to a good friend about soccer

"The Women's Game," which has over 30,000 subscribers on YouTube, has stood out for its conversational approach.

Guests make regular appearances, some of whom are friends of Mewis from her days playing professional soccer, while others play on international teams. To date, its 45 guests have included USWNT captain Lindsey Horan, US teammates Rose Lavelle and Kelley O'Hara, and international players like Daniëlle van de Donk.

In addition to the flagship podcast, Mewis has shows under "The Women's Game" banner, including "Friendlies, " where she interviews athletes one-on-one, and "Good Vibes FC," which she cohosts with her former teammates Becky Sauerbrunn and Lynn Williams.

The podcasts aim to create an environment that makes listeners feel like they're talking with friends about soccer rather than getting hard analysis. In a recent Christmas special with 39,000 YouTube views, Mewis brought on players including Lavelle, O'Hara, and former US teammates Ali Kreiger and Megan Rapinoe, who debated topics from egg nog versus mulled wine to their favorite holiday songs.

Mewis attributed the tone of the show to her strong relationships with her cohosts, including Williams, with whom she previously hosted the Just Women's Sports podcast "Snacks."

"I feel safe when I'm with them to be myself and to explore topics that maybe I don't know every single thing about," Mewis said.

She said the approach also helped her take ownership of her personality. As a professional athlete, she had to be careful about what she said to the press, but she could control the narrative with her own show.

"It is an opportunity for players, people in the media, and in the public eye to show a side of themselves that we don't always get to see," said Mewis.

Podcasts offer regular content to hungry women's soccer fans. When the USWNT isn't playing, or the National Women's Soccer League isn't in season, coverage of the sport has been hard to find. Research firm Nielsen found through fan surveys that the biggest barriers to seeing women's sports have been a lack of information and access.

With "The Women's Game," Mewis aims to cover all sides of the sport, said Roger Bennett, one of the cofounders of Men In Blazers, who followed Mewis during her playing days and approached her for the podcast.

"She said, 'I want to cover this women's game, not just American, I want to cover it all globally, have the biggest names in world football come and talk about themselves to the audience and to build a regular narrative around the biggest games in women's football,'" Bennett said. "That's exactly what we've set out to do."

The popularity of women's soccer has risen worldwide. In a 2023 Nielsen survey, 41% of the global respondents said they were excited about the Women's World Cup, rising from 34% before the 2019 tournament.

Fresh funding could help expand 'The Women's Game'

Looking ahead, Mewis is preparing for two major soccer events: the 2027 Women's World Cup and the 2028 Olympics. She also wants to bring international tournaments like the Euros or the Women's Africa Cup of Nations to North American audiences.

She wants to expand the vertical with more weekly programming, as well. It currently publishes two to three times a week across its three podcasts.

"The Women's Game" is set to receive fresh funding from a $15 million Series A round that Men In Blazers announced earlier this month.

"As women's soccer explodes and just continues to grow, there's more and more space for people like me and companies to invest in content production around women's soccer," Mewis said.

• I grew up as a latchkey kid and loved roaming the city by myself.
• But I struggled to give my tween the same independence I loved so much.
• At 13, I caved and let her walk alone to a café.

For months, my 12-year-old daughter begged me to let her walk alone to the strip of stores and cafés a half-mile away from our house in Los Angeles. Craving independence, she wanted to get ice cream, pick out a friend's birthday gift, and do her homework at the coffee shop (anything) so long as she could do it by herself.

Each earnest request was met with a resounding "No." Resentment radiated from my daughter's big eyes. I was the evil witch to her Rapunzel. Instead of a tower, she was trapped in a California bungalow.

Attempting to compromise, I offered to let her go solo on one condition. I'd trail out of sight behind her.

"You won't even see me," I argued.
"Not the same, and you know it!" she shot back.

I was terrified of giving her the freedom I grew up with.

I was a latchkey kid

I understood her desire to explore on her own. I was a latchkey kid growing up in Seoul, where my father was stationed. Once school was over, I'd scarf down some Spam with kimchi while watching my favorite soap opera and head out the door of our apartment. There was nothing I loved more than roaming the bustling city.

I'd walk to the corner store and buy a grape-flavored Popsicle in the shape of a shark that was raspberry red in the middle when I bit into it. I'd ride my bike along the Han River, ferry boats and gleaming skyscrapers streaking by in my periphery. Sometimes, I'd peruse my favorite stationary store, running my fingers along the crisp paper of Keroppi the Frog notebooks. Embedded in these memories was the gratification of doing it all without adult supervision.

I was terrified of giving her the same freedom

But as a 42-year-old mother, I was terrified of allowing my daughter the same freedom. Watching cars blazing through our neighborhood, barely coming to a rolling pause before blasting past a stop sign, my chest seized at the idea of her navigating the streets without me to protect her.

My fears didn't stop at speeding vehicles. Push notifications from apps like My Citizen and Nextdoor were frequent reminders of nefarious activities taking place. In one harrowing post, a girl was accosted and groped by a man as she walked to her middle school in a nearby suburb. Reading the scary details, I thrust my phone into the air. "See!" I declared. My daughter sighed.

Thirty years earlier, when I was 12, I escaped an attempted abduction shortly after my family moved from Korea to the US. In Washington State, a man followed me and tried to get me into his car as I made my way home alone.

The terrifying incident imprinted me with a deep paranoia. It was leaching into my child's adolescence, causing strife. I wanted to keep her safe, but tethering her tightly to my side only created distance between us. She was frustrated and angry. More than once, she called me "Smother."

I needed her to know I trusted her

How could I expect her to be confident and self-sufficient if I didn't show I trusted her? I knew I needed to let go, to let her experience and navigate the world without me holding her hand. To help quell my paranoia, I deleted My Citizen and Nextdoor. Those apps were gasoline to my fiery anxieties.

One Saturday morning, I swallowed the bile gathering in the back of my throat, gripped the kitchen island for support, and told my daughter she could walk by herself to the coffee shop. Her face shone as she promised to abide by my safety precautions. Watching her skip down the front walkway without me next to her, my knees buckled.

Pacing the kitchen, phone clutched in my sweaty hands, I started at her shared location. For a half-mile, the blue dot moved from block to block. Then it stopped. My phone chimed. "I'm here, Mama!"

An hour later, my daughter returned, beaming and exuberant, sipping an iced matcha latte. Meanwhile, I looked like I'd just crawled out from the trenches.

Now 13, my daughter has ventured out by herself on many occasions. Though my nerves still rattle, they've improved with time. When she walks back through our front door, pride and confidence fill her eyes — and mine, too.

• SoftBank is in talks to invest up to $25 billion in OpenAI, which could value it around $300 billion.
• The deal could make OpenAI the joint-second most valuable tech company after SpaceX.
• The funding round could see OpenAI pump $15 billion into the Stargate AI infrastructure project.

SoftBank is preparing to lead fresh investment into OpenAI at a $300 billion valuation, Business Insider understands — a move that would value the ChatGPT maker the same as TikTok owner ByteDance.

The Japanese investment giant is in talks to invest up to $25 billion in OpenAI, the Financial Times first reported.

While SpaceX is the most valuable private company in the world, the funding round could push OpenAI up to second spot alongside Chinese tech giant ByteDance. OpenAI currently ranks as the third-most valuable private tech firm.

Talks between OpenAI and SoftBank are ongoing, which means details around the the round size and valuation are subject to change. If the AI juggernaut hits its target valuation of $300 billion, it would nearly double its current valuation of $157 billion.

SpaceX's valuation is around $350 billion after the Elon Musk-owned company agreed to buy back $1.25 billion worth of stock at $185 a share in December. ByteDance, the parent company of TikTok, similarly catapulted its valuation to $300 billion following a buyback offer, offering investors a price of about $180 per share, the Wall Street Journal reported in November.

OpenAI could become one of three companies globally with a centibillion valuation; Stripe and Shein follow suit with valuations of $70 billion and $66 billion, respectively.

Prior to SoftBank's latest investment talks, OpenAI raised nearly $20 billion from investment heavyweights, including Thrive Capital, Khosla Ventures, and Nvidia.

If SoftBank invests $15 to $20 billion into OpenAI, it would overtake Microsoft — which has poured around $13 billion into the company so far — as OpenAI's lead investor.

Such a deal would mark the most significant bet yet on the generative AI boom from SoftBank and form a key part of the wider ambitions of its billionaire founder, Masayoshi Son, to usher in an era of "artificial super intelligence."

Son, who made his original fortune from a timely bet on Alibaba in the dot-com era, has previously spoken about his nonstop use of OpenAI's ChatGPT, and predicted AI that is 10,000 times smarter than humans will arrive by the midpoint of the next decade.

A fresh investment from SoftBank would also further deepen its relationship with OpenAI after the companies announced plans last week to form Stargate, a joint venture that aims to spend up to $500 billion on AI infrastructure projects in the US over the next four years.

SoftBank's Son will serve as chairman of the project, which was unveiled at the White House by President Donald Trump. Initial equity funders include Oracle and the UAE's MGX alongside SoftBank and OpenAI. The project will begin deploying $100 billion immediately, according to OpenAI.

It is not yet clear where the companies will source the capital from, with OpenAI currently lossmaking. It has been suggested that SoftBank's proposed equity investment could allow OpenAI to invest around $15 billion in Stargate.

SoftBank declined to comment. OpenAI did not immediately respond to a Business Insider request for comment.

• There have been 66 album of the year winners since the first Grammys ceremony in 1959.
• Taylor Swift made history in 2024 as the first person to win album of the year four times.
• Beyoncé, Chappell Roan, Sabrina Carpenter, and Billie Eilish are among those nominated this year.

The 67th Annual Grammy Awards will be held on Sunday in Los Angeles. 

Every nominee hopes to clinch the top prize of the night: album of the year. The coveted prize has previously gone to Hollywood icons like Judy Garland, Barbra Streisand, and The Beatles, and is always the last award of the night to be announced.

In 2024, Taylor Swift took the prize and made history as the first person to win album of the year four times. She previously tied with Stevie Wonder, Frank Sinatra, and Paul Simon, who all hold three album of the year wins. 

Swift could break her own record again this year for most album of the year wins if "The Tortured Poets Department" takes home the award.

However, she's got strong competition in the category. Grammy darlings Billie Eilish and Beyoncé, as well as pop superstars Sabrina Carpenter and Charli XCX, whose hit songs "Espresso" and "360" dominated pop culture this summer, are all nominated. Another nominee, Chappell Roan, has had a standout year, with six Grammy nominations to show for it. Rounding out the category are the multi-talented André 3000 and Jacob Collier, whom BI's Callie Ahlgrim called "dark horses in a pop-heavy album of the year contest."

Here's every winner of album of the year throughout history.

1959: Henry Mancini — "The Music from Peter Gunn"

Mancini, pictured left, was the inaugural winner of the award.

1960: Frank Sinatra — "Come Fly With Me!"

Sinatra won his first of three album of the year awards in 1960.

1961: Bob Newhart — "The Button-Down Mind of Bob Newhart"

Newhart starred in "The Big Bang Theory" as Professor Proton.

1962: Judy Garland — "Judy Garland at Carnegie Hall"

Garland was recently played by Renee Zellweger in "Judy," for which she won an Oscar — one award that eluded Garland.

1963: Vaughn Meader — "The First Family"

The album was a musical spoof based on the Kennedys.

1964: Barbra Streisand — "The Barbra Streisand Album"

Released on February 25, 1963, it was the debut album by Barbra Streisand. She is a rare EGOT winner: She's won an Emmy, Grammy, Oscar, and Tony.

1965: Stan Getz & João Gilberto — "Getz/Gilberto"

This year marked the first time two people won the award.

1966: Frank Sinatra — "September of My Years"

Sinatra won consecutive album of the year awards in 1966 and 1967. He is one of only two artists to do so, the other being Stevie Wonder.

1967: Frank Sinatra — "A Man and His Music"

Until 2024 — when Taylor Swift won for the fourth time — Sinatra held the joint record for the most wins for this award.

1968: The Beatles — "Sgt. Pepper's Lonely Hearts Club Band"

The Beatles became the first band to win album of the year.

1969: Glen Campbell — "By the Time I Get to Phoenix"

Campbell beat The Beatles and Simon & Garfunkel to the award this year.

1970: Blood, Sweat & Tears — "Blood, Sweat & Tears"

Johnny Cash and The Beatles lost the award to Blood, Sweat & Tears.

1971: Simon & Garfunkel — "Bridge Over Troubled Water"

Paul Simon also won twice as a solo artist. Therefore, he has technically won this award three times.

1972: Carole King — "Tapestry"

Carole King has won a total of five competitive categories, plus three honorary awards.

1973: George Harrison & Friends (Ravi Shankar, Bob Dylan, Leon Russell, Ringo Starr, Billy Preston, Eric Clapton, and Klaus Voormann) — "The Concert for Bangladesh"

Harrison also won the award as a member of The Beatles in 1968.

1974: Stevie Wonder — "Innervisions"

Stevie Wonder won his first of three album of the year awards in 1974.

1975: Stevie Wonder — "Fulfillingness' First Finale"

Wonder won consecutive awards in 1975 and 1976, the first person to do so since Frank Sinatra in 1966 and 1967.

1976: Paul Simon — "Still Crazy After All These Years"

This was the first of Simon's wins as a solo artist.

1977: Stevie Wonder — "Songs in the Key of Life"

Wonder won his third album of the year this year, making it three wins in four years.

1978: Fleetwood Mac — "Rumours"

Fleetwood Mac beat John Williams and his "Star Wars" score to the award this year.

1979: Various Artists — "Saturday Night Fever" soundtrack

This year marked the first time the winner was listed as "various artists," as well as the first time a film's soundtrack or score won the award.

1980: Billy Joel — "52nd Street"

Billy Joel beat Donna Summer and Kenny Rogers to become the first winner of the 1980s.

1981: Christopher Cross — "Christopher Cross"

Christopher Cross beat three-time winner Frank Sinatra to claim this award.

1982: John Lennon and Yoko Ono — "Double Fantasy"

John Lennon won his second award with his wife, Yoko, following his first win with The Beatles in 1968.

1983: Toto — "Toto IV"

Paul McCartney lost his first nomination for this award as a solo artist to the band.

1984: Michael Jackson — "Thriller"

Michael Jackson won eight awards this year.

1985: Lionel Richie — "Can't Slow Down"

Lionel Richie beat legend Tina Turner to the award this year.

1986: Phil Collins — "No Jacket Required"

This album contained two US No. 1 hits: "One More Night" and "Sussudio."

1987: Paul Simon — "Graceland"

Simon won his second award as a solo artist this year.

1988: U2 — "The Joshua Tree"

This was the first of U2's two album of the year wins.

1989: George Michael — "Faith"

"Faith" contained hits such as "Faith" and "One More Try."

1990: Bonnie Raitt — "Nick of Time"

Bonnie Raitt beat Tom Petty's "Full Moon Fever" to this award in 1990.

1991: Quincy Jones & Various Artists — "Back on the Block"

Surprisingly, this was Quincy Jones's first win in this category, despite serving as a producer on several album of the year-winning albums.

1992: Natalie Cole — "Unforgettable... with Love"

Natalie Cole won the award this year, preventing nominee Paul Simon from winning his third award for album of the year.

1993: Eric Clapton — "Unplugged"

As well as winning this prestigious award, Clapton is the only three-time inductee to the Rock and Roll Hall of Fame.

1994: Whitney Houston — "The Bodyguard"

This was Houston's only win in this category and only the second time a movie's soundtrack won this award.

1995: Tony Bennett — "MTV Unplugged"

This album was created as a result of Bennett's appearance on the MTV show "MTV Unplugged."

1996: Alanis Morissette — "Jagged Little Pill"

Alanis Morissette beat Michael Jackson and Mariah Carey to win this award.

1997: Celine Dion — "Falling Into You"

Two years after her win here, Dion also won four Grammys for her "Titanic" song, "My Heart Will Go On."

1998: Bob Dylan — "Time Out of Mind"

In 2001, Dylan added to his awards collection with a best original song Oscar for "Things Have Changed" from the film "Wonder Boys."

1999: Lauryn Hill — "The Miseducation of Lauryn Hill"

Lauryn Hill beat Shania Twain and Madonna to win this award.

2000: Santana — "Supernatural"

Santana won this award with their 18th studio album.

2001: Steely Dan — "Two Against Nature"

Steely Dan beat Radiohead, Paul Simon, Eminem, and Beck to win this award.

2002: Various Artists — "O Brother, Where Art Thou?" soundtrack

It was the third film to win.

2003: Norah Jones — "Come Away with Me"

Norah Jones won this award with her debut studio album.

2004: OutKast — "Speakerboxxx/The Love Below"

OutKast is an American hip-hop duo consisting of Andre 3000 and Big Boi.

2005: Ray Charles & Various Artists — "Genius Loves Company"

This same year, Jamie Foxx won the best actor Oscar for playing Ray Charles in the biopic "Ray."

2006: U2 — "How to Dismantle an Atomic Bomb"

This was U2's second win in this category.

2007: The Chicks — "Taking the Long Way"

The Chicks have won a total of 12 Grammys, including five in 2007 when they won this award.

2008: Herbie Hancock — "River: The Joni Letters"

This album is only the second jazz album to win this award and is a tribute album of cover songs written by Joni Mitchell.

2009: Robert Plant and Alison Krauss — "Raising Sand"

Robert Plant was previously the lead singer of Led Zeppelin, who were never nominated for album of the year.

2010: Taylor Swift — "Fearless"

Taylor Swift became the youngest artist to ever win album of the year. This record has since been broken by Billie Eilish.

2011: Arcade Fire — "The Suburbs"

Arcade Fire beat Eminem, Katy Perry, Lady Gaga, and Lady A to win this award.

2012: Adele — "21"

This was Adele's first Grammy win for album of the year.

2013: Mumford & Sons — "Babel"

Mumford & Sons beat Frank Ocean's "Channel Orange" to win this award.

2014: Daft Punk — "Random Access Memories"

Daft Punk won five Grammy awards this year.

2015: Beck — "Morning Phase"

Beck beat Beyonce, Sam Smith, Pharrell Williams, and Ed Sheeran.

2016: Taylor Swift — "1989"

Taylor Swift won her second album of the year award this year. She was the youngest person ever to win two.

2017: Adele — "25"

Adele matched Taylor Swift's two wins (at this point) in this category with her own second win, following her first in 2012.

2018: Bruno Mars — "24K Magic"

Bruno Mars has won 15 Grammys, six of which came in 2018.

2019: Kacey Musgraves — "Golden Hour"

This was the first year that the Grammys expanded this category to eight nominees. Musgraves beat Post Malone, Brandi Carlile, Janelle Monáe, H.E.R, Cardi B, Drake, and the "Black Panther" soundtrack.

2020: Billie Eilish — "When We Fall Asleep, Where Do We Go?"

Billie Eilish became the youngest-ever winner of this award at age 18, beating Swift's previous record of 20 years old.

2021: Taylor Swift — "Folklore"

At the time, she became the only woman to hold three album of the year wins, and only the fourth person to ever hold the distinction. Other artists who have won album of the year three times are Stevie Wonder, Frank Sinatra, and Paul Simon.

Swift broke that record in 2024 when she won album of the year for her 10th studio album "Midnights."

2022: Jon Batiste — "We Are"

In 2022, Jon Batiste became the first Black artist since 2008 to win album of the year. Only 10 other Black artists have won album of the year since the award show's inception.

Batiste beat artists like Justin Bieber, Olivia Rodrigo, Taylor Swift, and Doja Cat. He was the most-nominated artist of the night with 11 nods. 

2023: Harry Styles — "Harry's House"

In 2023, Harry Styles won the award for his critically acclaimed album "Harry's House," beating out artists like Beyoncé, Adele, Bad Bunny, Kendrick Lamar, and Coldplay for album of the year.

"There's no such thing as best in music," Styles said while accepting the award. "I don't think any of us sit in studios thinking about what is going to get us one of these. This is so, so kind."

2024: Taylor Swift — "Midnights"

Taylor Swift made history when she took home the top award for album of the year at the Grammys in 2024 for her 10th studio album, "Midnights."

Swift, who was previously tied with Stevie Wonder, Frank Sinatra, and Paul Simon for most album of the year wins, is now the first and only person to have won the award four times.