Inside Meta’s Big AI Pivot

0:05Last week, Meta said it was laying off 10 % of its roughly 80 ,000 employees.

0:11Meghan Bobrowsky:Meta says it'll lay off about 8 ,000 employees starting next month. The company's also canceling plans to fill 6 ,000 open roles as it plans to invest more on developing artificial intelligence. The layoffs are part of a larger transformation that's happening within Meta right now. As the company tries to reinvent itself as an AI powerhouse. Meta said it would use the savings from the layoffs to balance out its huge investments in AI. This year, the company is planning to spend up to$135 billion on the technology. And for the employees who are left, they are being asked to incorporate AI into their jobs.

0:47Teams are being flattened. And in performance reviews, workers are assessed by how much they use AI. And Meta isn't just using AI to make its employees more efficient. The company is also using its workforce and the way they work, like at their desks, to train the company's most advanced AI models.

1:08Meghan Bobrowsky:A memo went out on Tuesday from a researcher who works on building the models. And they said, hey guys, our models need to get better at learning how to use computers. That's our colleague Megan Bobrowski. And so therefore, we are now going to be monitoring your keystrokes, your mouse movements, and your click locations, feed that data to our AI models to help them understand basically how to use a computer. Hmm. That sounds kind of dystopian. A lot of employees were not happy about this. The top-ranked comment on this post was, this makes me super uncomfortable. How can I opt out? Spoiler, there is no way to opt out.

1:55So it seems like Meta is going all in, like in every conceivable way, from the products that it's making to what it expects of its own employees.

2:06Meghan Bobrowsky:Yeah, it's AI all the time. Mark Zuckerberg himself is working on building a CEO agent to help him do his job. They've also just announced initiatives across the board, trying to get their employees to adopt these things. and then it's almost for not to be punny, but... We love puns here. Well, it's very meta, right? They're building these AI products that they want billions of people to use. And the way they're doing that is by trying to get their workforce to adopt AI to build the AI. So AI to build the AI. Where does meta stack up in the AI world right now? And how hard would you say it's fighting to catch up?

2:48Meghan Bobrowsky:Yeah, so those are two different things. Where it is, it's not the best. How hard is trying might be trying the hardest of any of the companies.

3:02Welcome to The Journal, our show about money, business, and power. I'm Ryan Knudsen. It's Tuesday, April 28th.

3:16coming up on the show how meta is going all in on ai

3:29this episode is presented by sap your company's ambitions can't be held back by a long implementation, surprise costs, or empty AI promises. SAP Grow AI Cloud ERP gets you live fast, keeps pricing predictable, and delivers built-in AI that gets results the first day, not someday. All on a single platform that's easy to manage, industry-ready, and designed to scale with your business. Bring it with SAP Grow. AI Cloud ERP for any size business. sap.com slash grow. Putting off replacing your window treatments because you think it's complicated? At Blinds.com, we've spent 30 years proving it doesn't have to be.

4:05And today is your last chance to save big on Spring Black Friday deals. Whether you want to DIY it or have a pro to handle everything from measure to install, we've got you. Free samples, real design experts, and zero pressure. Just help when you need it. Shop up to 45 % off with minimum purchase. Plus, get a free professional measure during the Blinds.com Spring Black Friday Last Chance Sale. Rules and restrictions apply.

4:34Meta started an AI research lab in 2013. But in the last few years, Meta's AI tools have trailed behind companies like OpenAI, Anthropic, and Google, which have been making big strides developing AI chatbots that can do research, write cover letters, and even code. Last year, to try and up its game, Meta started poaching top AI talent with huge offers.

4:59Meghan Bobrowsky:Meta now notching another name in its expensive hunt for AI talent, poaching a top Apple executive. Coming from the buzziest and earliest AI native firms, OpenAI, Anthropic, DeepMind. Meta finds itself sort of racing to catch up, right? And so they were giving out$100 million offers to researchers trying to basically like rebuild this team and become competitive. Multi-year deals worth$300 million. These are for scientists and engineers. This is wild. It's like NBA superstar money. With some of them receiving$100 million straight up in year one,$100 million. Wow. Yeah, yeah, exactly. And so they build this new team.

5:42Meghan Bobrowsky:They hire Alexander Wang from Scale AI, and they basically just redo their whole AI efforts from scratch. Wang and his team have developed the most powerful AI model Meta has produced so far. It's called MuseSpark and it was launched earlier this month. The new model is competitive with OpenAI and Google and Anthropik. It's not the best model, but it's good enough to keep them in the race. And they've sort of proven that like, hey, we're still here, we're still fighting, we're still in it, don't count us out. And now they're trying to race to get ahead to the frontier.

6:23Meta AI, as its chatbot is known, is designed to compete with chatbots like ChatGPT, Cloud, and Gemini. And it's baked into the company's existing apps, like Facebook, WhatsApp, and Instagram. And compared to some of its rivals, Meta has a key advantage. It already has a huge potential user base.

6:41Meghan Bobrowsky:Meta has 3.5 billion daily users around the world. That would be a lot of people that could use the chatbot. It's a lot of people that could use the chatbot, exactly. And so, you know, investors and analysts argue that Meta has the distribution. That's not the problem, right? Like if you're Claude or you're ChatGPT, the company's behind those chatbots, you're trying to grow. You're trying to get people to use these things. But like you're starting from scratch. Like you don't already have a user base. Meta already has a huge, huge user base. And so what they're trying to do is more about getting people to adopt this chatbot.

7:15Meghan Bobrowsky:That's what their kind of unique problem is that they're trying to solve. Meta hopes to use its AI chatbot as a way to supercharge its already highly profitable ad business. In December, Meta started using your conversations with their chatbot to target and show you ads on Instagram. For instance, Megan is going on a trip to Japan soon. And if she uses Meta AI to research and plan her trip, the company can use that information for targeted advertising. So I talked to the chaperone. I'm like, hey, what are some good temples to see in Kyoto? And so if I, you know, use those conversations that I had, Meta can take that and start showing me ads for things in Japan on Instagram now that I might want to click on.

8:03Meghan Bobrowsky:Maybe it's tours, maybe it's restaurants to go eat at, whatever it is. And so they sell ads. They're very good at this. So this is also another way for them to make that ad business even better and more lucrative for that. I mean, Google searches is a massive business. And this is almost a way that Meta can steal some of that business, essentially, by giving people a place where they can ask questions. And then Meta can mine that for data it can use to deliver more ads to you. Exactly. It's sort of twofold. Yes. The ultimate futuristic AI product that Meta is working towards is AI agents. These agents would be able to act as personal assistants, or more.

8:45Meghan Bobrowsky:So what Mark Zuckerberg has said is he wants everyone to have their own personal superintelligence. We think people, as I reported last year, he thinks people have the capacity to have more friends than they do. And that AI can solve some of these problems. The average American, I think, has, I think it's fewer than three friends. Three people that they'd consider friends. And the average person has demand for meaningfully more. I think it's like 15 friends or something. And also, you can get an AI to go out and do things for you and just help you in your life. I think as the personalization loop kicks in, and the AI just starts to get to know you better and better, I think that will just be really compelling.

9:36Employees at Meta already have access to one such agent called MyClaw.

9:42Meghan Bobrowsky:MyClaw has access to a bunch of your different things, right? It might have access to all your GChats, all your work projects, your whole drive, all your emails. and so it doesn't need the context that you would have to tell a chatbot to do, right? It can just go and like, it has access to everything so it can actually be more proactive. It can say, hey, Ryan, I see you have like a dinner with your wife in the calendar for 6 p.m. Like, do you want me to order her some flowers ahead of that? Yeah, great idea. Or like, you know what I mean? Stuff like that where it can actually be more proactive.

10:20You have an interview in 30 minutes. Do you want me to research some bad puns?

10:23Meghan Bobrowsky:Exactly. So with the agents, I mean, you can still tell it to do things for you. But I think the idea with those in some situations is it's like it's almost better than you at remembering everything because it has access to everything you've written down. And this is kind of like the Silicon Valley vision for like the movie Her with Joaquin Phoenix, where he's just got, you know, this super smart companion. that's like, hey, you've got an important email that you might want to know about. And it just talks to him and he talks to her. That's all. You mind if I look through your hard drive? Um, okay.

11:00Meghan Bobrowsky:Okay, let's start with your emails. You have several thousand emails regarding LA Weekly, but it looks like you haven't worked there in many years. Oh, yeah. I think I was just saving those. Yeah, like a chatbot, it's like you have to go to it and you have to converse with it. It can reach out to you. It can text you first, but like it's different. Like an agent, I think, takes things like just one step further. Do you know how Mark Zuckerberg is using that CEO agent that he's been working on? It's still early. I think the agent is helping him retrieve information faster. So before, he might have to go through multiple layers of people to find information.

11:37Meghan Bobrowsky:Now he can just ask his agent to go find wherever it is and, you know, the emails or the drives or whatever and get the answer versus having to, you know, do that. like telephone game of like, hey, can you go ask this person this? Yeah, let me go ask this person this, et cetera, et cetera. That takes time. So it's still early days, but that's one of the initial things that we know that he's using his agent for. Coming up, how Meta's AI transformation is affecting the people who work there.

12:20Study and play. Come together on a Windows 11 PC. And for a limited time, college students get the best of both worlds. Get the Unreal College Deal. Everything you need to study and play with select Windows 11 PCs. Eligible students get a year of Microsoft 365 Premium and a year of Xbox Game Pass Ultimate with a custom color Xbox wireless controller. Learn more at windows.com slash student offer. While supplies last. Ends June 30th. Terms at aka.ms slash collegepc. Today we helped a... Latte for Sam. ...coffee shop get an insurance quote simply and easily. And made sure... A floral delivery van was able to make someone's day.

13:01We're the Hartford, with decades of experience insuring millions of unique small businesses. When it comes to your small business insurance... Thank you. ...one size absolutely does not fit all. Get a quote or find an agent today at thehartford.com slash smallbusiness.

13:26So can you walk me through how this giant pivot to AI at Meta is affecting the company's employees?

13:34Meghan Bobrowsky:Yeah, I'm going to back up a little bit more. So Meta, just to get a sense of how this company has gotten to where it is. During the pandemic, it roughly doubled its workforce. It got up to like 87 ,000 people. And then Mark Zuckerberg declared year of efficiency and they did a bunch of layoffs. That got the headcount down to about 67 ,000. And since then, it's sort of continued to climb back up. It's now at around, prior to all these cuts, was around 78 ,000. And then last year, it comes out that employees are going to start to be graded on how much they use AI. And Meta's leadership expects employees to use AI a lot.

14:21Meghan Bobrowsky:What we've seen, there's a lot of internal memos that have kind of come out over the last few weeks. And one of them, the CTO, Andrew Bosworth, says in the future, AI agents are actually primarily going to do the work. and that the human's jobs will be to supervise them, direct them, and help them improve. So almost like everyone's going to be a manager and you have your own little agents going around doing things for you, but you're not doing the work anymore that me and you used to do.

14:56During an earnings call in January, Zuckerberg said AI was making it possible for a small number of employees to do a lot of work. We're starting to see projects that used to require big teams now be accomplished by a single very talented person.

15:11Meghan Bobrowsky:We sort of got a preview of how he was thinking about things. And then over the last few months, you've started to see it all go into action. Meta also believes it can rethink the company's org chart. In March, Megan reported on an internal memo that laid out how Meta was creating a new team focused on AI development. they would have a very flat organizational structure. It would be 50 employees reporting to one manager, for instance. So like really getting rid of the middle layers of management. Basically empowering employees. Exactly. And so, you know, at this point, Meta is a huge, huge company.

15:49Meghan Bobrowsky:I mean, it was 20 years ago, it was a small startup. It's now grown into this like juggernaut. And, you know, part of what like Mark Zuckerberg and Andrew Bossworth, the CTO, talk about is they've sort of come to the conclusion that to get them to work more efficiently, you just need fewer people. Inside the company, employees say the mandate to use more AI is creating chaos. People I talk to would say, like, there's just people creating duplicate tools. There's people posting all the time. It's like the Wild West again, right? It's like the move fast and break things era again. People are being given, like, free reign to do this, right?

16:23Meghan Bobrowsky:Like, the company is saying, like, go out and try and just, like, innovate. See what you can figure out about AI and how to use it. So how did Meta get the idea, though, in its head that its employees could be used to train its AI models? They've actually explained a lot of their thinking in memos that they sent to staff. So let me pull up. I just want to read from you because it's actually pretty interesting. They said, we're on a really strong trajectory with our models. And one of the ways we can accelerate our path is by tapping into our own work day to day. While AI models excel at research and technical skills like coding, they still lack some of the basic ways that humans use computers, like choosing from dropdowns and keyboard shortcuts.

17:09Meghan Bobrowsky:For agents to understand how people actually complete everyday tasks using computers, we need to train our models on real examples. This is where all the Meta employees can help our models get better simply by doing their daily work. I'm a little bit surprised, actually, that Meta thinks that AI agents are going to be, like, choosing from drop-down menus and, like, clicking buttons, kind of. Because when I imagine how an AI is going to operate, I feel like, why does it need to interact with a computer the way a human does? It's already a computer. Well, the internet is optimized for humans. The internet is not optimized for other chatbots to use, right?

17:48Meghan Bobrowsky:And so at a certain point, this is what people talk about. at a certain point, the internet might not be for us anymore, right? The internet might be a place where agents go to talk to other AI agents. Right now, the internet is a place for humans, and so it's sort of built for us. And so there's a lot of cases where the AI has to act like a human to get to what it needs because it's not built for the AI.

18:15Some employees don't seem to like all these changes. Megan looked at data from a website called Blind, where people who work at tech companies can post anonymously about their employers.

18:24Meghan Bobrowsky:In 2024, roughly 20 % of the posts about Meta were negative. This year, more than 80 % of the posts about Meta are negative. Hmm. Because they're being asked to train their digital replacements, essentially. That is exactly how someone put it to me, was, am I automating away my own job? A Meta spokesperson said, quote, if we're building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them. The spokesperson also said that there are safeguards in place to protect sensitive content and that the data is not used for any other purpose.

19:08What's at stake for Meta through all this?

19:11Meghan Bobrowsky:Meta is an ads business at the end of the day. It's a very successful ads business. What this is all about is who's going to be the leader in this space in five or ten years. Like, we're playing the long game here. If Meta gets their way, they're going to revolutionize the internet again. Think what they did the first time around with Facebook, what they tried to do the second time around with the metaverse. It didn't work. This is sort of like their attempt to do the metaverse thing again, but in a way that they think is going to be more successful now. And if they're right, then you're going to have everyone using their chatbots and they're going to be making a lot more money.