Inside Zuckerberg’s AI Playbook | July 15, 2025

0:27you Good morning. Welcome to The Information's TITV. My name is Akash Masritcha. There is a lot happening in tech news right now, but we want to get right to a great conversation that our CEO and Editor-in-Chief Jessica Lesson had with Meta CEO Mark Zuckerberg yesterday. They talked all about Meta's AI strategy. We want to play that for you. Let's take a listen. Thank you, Akash. And Mark, thanks for joining TITV. We're happy to have you. Happy to be here. Already this morning, news that you've spent even more money with your big announcement about your new supercomputers. We'll get to that. But to start, you took a huge stake in scale AI.

1:08You have been on a blitz of AI hiring. Why and why now? Yeah, it's been busy. I think the most exciting thing this year is that we're starting to see early glimpses of self-improvement with the models, which means that developing superintelligence is now in sight. And we just want to make sure that we really strengthen the effort as much as possible to go for it. Our mission with the lab is to deliver personal superintelligence to everyone in the world. So that way we can put that power in every individual's hand. And I'm really excited about it. It's a different thing than what the other labs are doing.

1:51And my view is that this is going to be something that is the most important technology in our lives. It's going to underpin how we develop everything at the company, and it's going to affect society very widely. So we just want to make sure that we get the best folks to work on this from entrepreneurs to researchers to engineers working on the data and infrastructure. And then, of course, we want to back it up with just an absolutely massive amount of compute, which we can support because we have a very strong business model that throws off a lot of capital. Did you feel like Llama was behind?

2:26I think the field is accelerating. You know, we keep on like having goals for where we want to be. And then the field keeps on moving faster than we expect. So, you know, look, I mean, I'm not a person who's ever particularly happy with where we are. I always want to do better. But I think the thing here that I'm very focused on is just the technology that's going to be possible to develop in the over the next period of time is just it's going to be super fundamental. And I just want to make sure that we position ourselves as well as possible. I mean, there's this big debate in the industry, which is like, all right, is super intelligence going to be possible in two to three years and five years and seven years?

3:03And I don't think anyone fully knows the answer to that. But I just think that we should bet and act as if it's going to be ready in the next two to three years. I believe that there's a shot at that. And I think if that's what you believe, then you're going to invest hundreds of billions of dollars in building up the compute and whatever it takes to build out the strongest team in the industry. Now, I want to ask, there have been reports of pay packages at$100 million,$200 million. How do you justify that amount of investment for these researchers? Sure. So look, I mean, a lot of the specifics that have been reported aren't accurate by themselves, but it is a very hot market.

3:41I mean, as you know, and there's a small number of researchers who are just the best, who are in demand by all of the different labs. So I think that, you know, it certainly is quite competitive. You know, what I'd say, I think to put this in perspective, you know, if you're going to be spending hundreds of billions of dollars on compute and building out multiple gigawatts of clusters, then, you know, it really does make sense to compete super hard and do whatever it takes to get the, you know, 50 or 70 or whatever it is, top researchers to build your team. I think that the physics of this is you don't need a massive team to do this.

4:17You actually kind of want the smallest group of people who can fit the whole thing in their head. So there's just an absolute premium for the best and most talented people. So it kind of makes sense that the market would be aggressive when you kind of think about it. And from that perspective, the amount that is being spent to recruit the people is actually still quite small compared to the overall investment and all the compute that we're doing. There are so many companies in this super intelligence race from Google to OpenAI, Anthropic. I'm curious how Meta's vision of super intelligence maps with those or doesn't.

4:54Many of those companies obviously have large enterprise businesses. Yeah, well, look, I mean, I think personal super intelligence is a different vision than what the others are going for. You know, AI, it's kind of like when the internet was first getting to scale. People were like, all right, is the internet going to be for productivity? Is it going to be for entertainment? Or is it going to change how we work at our companies? And the answer is all. And I think the same is going to be true for AI. And there are going to be different companies that focus on different parts of it. So, so far, I think that, you know, you hear a lot from the other labs around, you know, wanting to automate all of the economically productive work in society.

5:32And, you know, certainly I think that that economic automation and economic progress, there's a lot of potential there to do great work that's very helpful to a lot of people. But there's another side of this, which is like, what are the things that people care about in their own lives? And some of that is productivity. But a lot of it is is relationships and culture and creativity and having fun and enjoying life. And my guess is that, you know, there's a lot of the folks at the other labs kind of talk about, okay, we want to develop the central superintelligence and direct it at grand problems.

6:09And I think the reality is that, you know, there's a lot of great stuff to do there. And I'm excited about that. And I think when you put personal superintelligence in people's hands, some of what they're going to want to do is directing it at grand problems. But I actually think a lot of what people care about are just relatively simpler things in their lives. And part of our ethos and values have always just been to try to put that power into people's hands directly. And that's what we're going to try to do with personal superintelligence. So it is a bit of a different focus. I think that at some level, I think this is also going to shape how we build the company and how we write all the software and all the technology and products.

6:46So that'll probably be pretty similar to a lot of the other places. And a lot of the fundamental methods are similar. But I think the products that we're going to build are going to be pretty different. and I think are just going to touch billions of people's lives. I think we're kind of the premier company at developing products and scaling them to billions of people. And we're, of course, leading in the next generation of computing platforms with the glasses that are doing exceptionally well. And I continue to think that that's going to be the best form factor for AI because they can see what you see and hear what you hear and you can talk to them throughout the day.

7:20And once we get a display and the holograms in there, it'll be able to generate a UI for you. So I think that that's going to be, there's just going to be this incredible connection between those things. Picturing my meta super intelligence glasses, companion friend doing things for me. I don't know. Am I on the right track? Yeah. I mean, I think it's going to, I think we're going to reach a point where, you know, you're going through the world and you're having conversations. One thought experiment that I like to do is in almost every conversation that I have, there's like five things I want to follow up on.

7:51But I almost don't ever follow up on all five, right? Just because I go do something else. Maybe I'll follow up on one if I'm lucky. And, you know, I just think like, OK, in the future, we're going to have glasses and they're going to be able to, if you want, you know, observe what's going on in your life and be able to go follow up on things for you and and either, you know, follow up with you later or give you information in real time in the glasses. And it's almost going to be like, look, if you need vision correction and you don't have glasses, just optical glasses, you're sort of at this cognitive disadvantage.

8:23And I think in the future, if you don't have AI glasses, you're basically going to be at a cognitive disadvantage, too. So, yeah, I mean, that's part of what we're going to do. But we're also very focused on the entertainment, the culture, the kind of personal relationships with people. I think that like having an AI that can help you with that is going to be super valuable. So anyway, there's a ton to do. There's no shortage. And in addition to the direct products, super intelligence is also obviously going to really change how we develop things at the company and how all of our software development goes.

8:55And I think our engineers are just going to be way, way, way more productive than they are now. So I think that that's going to be really exciting and is just something that we want to that we really want to focus on, especially the self-improvement loop. We are already seeing glimpses of the models, even based off of Llama 4, improving the Facebook algorithm and other AI systems across the company autonomously. So, I mean, I think that it's pretty exciting to see glimpses of that. And I know you have to go, but I do want to ask you about your news. First of all, at some point, another time, we'll have to talk about how you name these data centers.

9:32Well, they're titans. You know, I love, you know, I'm a fan of Greek and Roman mythology. But I want to ask, is there a limit to how much you'll spend on CapEx? I mean, I think we'll see how the technology trends and we'll see what the results are. I mean, in running the company, I'm sort of always looking for ways that I can convert capital into a higher quality service for people. And one of the benefits of reinforcement learning is it gives you a venue to potentially convert very large amounts of capital into a better and better service and potentially likely a better service than other less well-funded or less bold competitors will be able to do.

10:15So I view that as a competitive advantage if we can get this to work well. And that's why we are basically all in on this. We're building multiple multi-gigawatt data centers. And we can basically do this all funded from the cash flow of the company. We have a very strong business model that supports this. We have a very strong infrastructure team that is doing novel work to build up data centers. I wanted them to not just take four years to build these concrete buildings. So we pioneered this new method where we're basically building up these weatherproof tents and building up the networks and the GPU clusters inside them in order to be able to build them faster.

10:59Yeah, they're hurricane-proof tents. But no, and we have people working around the clock on Prometheus and Hyperion, which are our first two Titan clusters that are going to be greater than a gigawatt. Hyperion is going to scale to up to five gigawatts over the coming years. And I shared this image of it, but it's, you know, the size of the site covers a significant portion of the footprint of Manhattan in terms of space. The thing is just massive. You have a large amount of the energy and kind of people's energy in the area that it's getting built, going towards constructing it. I think it's great for the community.

11:38And these are really exciting projects that, I mean, you can tell I get, I mean, I'm very excited about building them in an innovative way. And I'm very excited that I think we're going to have the largest compute fleet of any company. and focusing on that, on building, on being powered by a small and talent-dense team means that I think we're going to have by far the most compute per researcher to do leading-edge work. And I think that it's just going to be really exciting. Big few years coming up. That was what jumped out at me about your announcement on Threads and Facebook was you saying you have the capital to do this, which made me think it was maybe a slight jab at those competitors who might have to still raise the capital to build that out.

12:20Do you see that as an advantage? Well, I mean, I think it's an advantage. I think it's an advantage. But also, you know, I think a lot has been written recently. I'm obviously spending a lot of time on recruiting. And a lot has been written about money and all that. And like I said before, a lot of the numbers specifically have been inaccurate. But I think it discounts the other key reasons why people are super excited to come work on Meta Super Intelligence Labs. And one of the biggest is that you could just have more leverage as a researcher. You have more compute, right? I mean, the best researchers basically, like historically, when I was recruiting people to different parts of the company, people are like, OK, what's my scope going to be?

12:57And here people say, I want the fewest number of people reporting to me and the most GPUs. And so having basically the most compute per researcher is definitely a strategic advantage, not just for doing the work, but for attracting the best people. And I actually think that that has been underreported as one of the main reasons why people are excited to come do this. The other is just having a clean slate is fun, right? I mean, a lot of times people refer to building up the infrastructure for these things. Someone gave me this analogy that it's like building a house. The first one you build for your enemy.

13:31The second one you build for your friend. The third one you build for yourself. And I think the ability to start from a clean slate where you're not just at another company working on some other infrastructure and you get the opportunity to build a new lab from scratch is going to be, I think it's just really appealing to a lot of people. So anyhow, yeah, I'm out there very focused on this. I know we've got to let you go soon, but I have to ask, it was Sun Valley last week. You have Sam Altman of OpenAI there as well as so many other executives whose team members you have been hiring. Did you get a chance to say hi?

14:04And what do you talk about? Yeah, I mean, I think part of the fun of Sun Valley is you run into all your friends and colleagues. So yeah, I mean, I saw Sam said hi. I also saw Dario and saw Sundar and Satya and everyone who's kind of in the space working on this stuff. And look, I mean, what would I say? I mean, look, we're building up this new team from scratch. And there's like five or six places in the world that have a lot of great researchers. You know, Meta, of course, is one of them. And part of what we're doing is pulling over a lot of the key folks into the new efforts as well. But we're also doing a big recruiting push.

14:44So I don't know. We're not trying to target anyone individually. I just want to make sure that I personally get to know all of the top researchers in the industry. So that way, whether it's now or in the future, if there's a good opportunity to work together, then we can make that happen. Well, it's the hot AI summer, Mark, and you're at the center of it. So thank you for joining TITV. And we'll be watching. Awesome. Good luck. See ya. Back to you, Akash. That was our conversation with Meta's Mark Zuckerberg. Jessica is here with us. And Jessica, I want to take a couple minutes here just to debrief your conversation with Mark.

15:23What were your biggest key takeaways? You know, Akash, it's really interesting. I knew Mark was going to want to talk about all this new compute he and Meta were bringing online. But I think there are two new doctrines, if I want to call that, that emerged out of the interview that I really heard Mark talk about for the first time. One was Meta's vision for personal superintelligence, which I think is a move to differentiate it from what we're seeing from OpenAI, Microsoft, Google, and he went on at length about that. I think the other one was really his approach to CapEx spending, to this infrastructure race that clearly he thinks they're in a wonderful positions.

16:05So we can unpack those a little bit more, but I think those were two themes that we haven't heard him talk about that I think are his new frameworks. Obviously, when I asked him if this hiring blitz was the result of the failures of Lama 4, he had a good answer, I suppose, which is, well, actually the field is moving ahead very quickly. So I thought that was an interesting response to that. He really tried to push back on the notion that he's luring all these researchers with these massive pay packages. Of course, it's been all over the headlines as we've been reporting in the information as well.

16:43Instead, he was saying, look, it's also the ability. And I think he says, the people I'm recruiting want to do these jobs with the fewest number of people and the maximum amount of compute. And so I thought his framework actually on how the landscape has shifted there. It used to be you hired people by promising them the biggest possible scope, I believe was his word, biggest possible set of responsibilities. Now these AI researchers really just want to be with their crew with access to a lot of computers. So certainly a lot to unpack. And I think I'm still kind of making sense of some of the things he was saying and what they mean for the company and the industry.

17:24Yeah, no, I mean, you know, I think those were some of the things I picked up on as well. And I think he really talks about super intelligence in the context of the rest of the company. And, you know, I think he really talks about it, you know, in a way of him trying to productize it and eventually make money off of it, which is, you know, which is something that other CEOs, when you don't have such a big company to work with something they might get be a little bit slower too so I think those are all great I think the money point is a good one in that you know if I had more time I would have pressed him on how do you pay for personal super intelligence right so if super intelligence is a hundreds of billions of dollars arm race and I think it's notable he did acknowledge that the timing is very unclear but he basically says it could be two it could be five it could be seven but you have to act like it's seven or i'm sorry you have to act like it's two but you know meta's framework and mark's framework is we have this amazing core business that throws off the cash to be able to invest and he actually pointed out that they have the capital to do this and you know i pressed him are you implying that others don't and he said yes we have a business that does this but what about the business model of the future and i think that is one question that everyone is really grappling with.

18:45We know right now that businesses are paying for AI. We see some examples around consumer subscriptions, but this, and we can unpack it a bit more, but this personal super intelligence, this intelligence, I think Mark said that thinks of those five things in your head you want to follow up on a conversation, but never do, and does them for you with the help of your meta glasses of course um you know what's the business model around that and so i think this is one of the reasons it is so exciting to be covering these titans now and honestly i've been doing it since 2005 so for 20 years i've been covering um google and meta and so right but i think the question around yes now mark is drawing these lines about how meta is going to focus on these personal use cases, on these things that billions and billions of people use, obviously how those are going to be the next cash cow or engine for the company, or whether they can be, is a big question.

19:50Let me ask you one quick question before we go. You've been covering Meta really since its inception, and you talked about personal superintelligence. One of the questions I wanted to sort of understand from you is, Meta has spent big on other technologies in the past. We've seen it, you know, we've seen it with Reality Labs. I mean, how does this investment in personal super intelligence, how does that compare or contrast to other big investments that the company has made throughout its lifetime? Well, look, if you look on the dollar amount, AI is trumping everything, and that's true of all the big companies across the board.

20:28But I think you can, if you want to understand how Mark thinks, You can look to other big bets, especially reality labs, where they dialed up the spending, obviously acquired Oculus, put the cash cow into it, and have been pruning, I will say, especially through the post-COVID age of efficiency, but still investing. And obviously now, I think the company would say those investments are fueling their position in AI with the next phase of hardware. So I think what you can take from that is there's there's a, shall we say, at least in their mind, a prudence to the wild and crazy investing. You know, when I asked Mark if there's a limit to CapEx, he said, we'll see how the technology progresses.

21:12So, you know, I think I take from that he doesn't want to be a step behind his competitors, but he's willing to adjust course, maybe more so, frankly, than other CEOs, because he can act, you know, so singularly at the company. So, you know, we're absolutely at an unprecedented phase here. But I think you can learn from behaviors of the past and assessing how the company will react, you know, in the years ahead. I mean, but the difference, of course, is you're investing so far ahead. I mean, I guess Marcus may be cutting a little cost on these data centers by putting them inside hurricane proof tents.

21:50But that seems more about speed than anything else. So, but yeah, it'll be wild. One more thing I'll say, Akash, too, and I know it's a small thing, but as an observer of these companies, I think Mark's shirt choice from the interview was very telling. And I know this can seem silly and I have no inside knowledge here, but, you know, we've seen, I just, I think there was a seriousness to him, not super formal. The fact that it had a collar. It had some kind of a collar. There have been many phases of Mark over the past two years. And I think, you know, the one I felt like I was talking to yesterday was someone who was serious, who was dialed in.

22:40You know, if you go way back in Facebook's history, and it was Facebook back then, there was a year he wore a tie to work every day. and it was a and I think he at the time told the team this because it's a serious year and I I just sort of in in his demeanor in you know his outfit choice which was probably random but but I think I I sensed you know he's here to play and um I think when we look at this moment across AI more broadly. It's quite an exciting one. And again, I just keep coming back to his comments, really trying to draw those lines between what he believes Meta is doing in AI and others.

23:25I mean, he talked about the other companies focused on these grand plans. And I think while he thinks AI can do wonderful things for society, he's really focused on what it could mean for individuals and um i thought it was very different as well as saying you know that they're going to be in this infrastructure race they view it as a huge advantage it's an advantage in their recruiting um and it doesn't seem like they're going to stop yeah well look you know we we'll get even more answers in a couple weeks when the company reports its quarterly earnings we've got uh earnings season coming up all the big tech companies coming up um netflix is this week um Thank you, Jessica, for being with us here to debrief.

24:10It was a great conversation. Look forward to us having more with executives like that. Well, that is it for today's show. We'll be back tomorrow with more analysis and breaking down the headlines of the biggest stories in tech. I want to thank AWS, who is our presenting sponsor for this production. Thank you for tuning in today, and we will see you tomorrow on TITV.

24:38Thank you.