JOHN COLLISON: It’s time to welcome our next and final guest for today onto the stage. This person has already invented the future more than once. He runs three of the four most downloaded apps in history, and he’s now building the future again with AR and AI. Please welcome Mark Zuckerberg. Welcome. 

MARK ZUCKERBERG: Thanks. Good to be here. Doing the live demo thing. 

JOHN COLLISON: Yeah, I know. Obviously, live demos are so much worse when they’re AI-based because you just get a different thing every time. It’s like it’s over the— 

MARK ZUCKERBERG: It adds an element of fun. 

JOHN COLLISON: Exactly. It just gets you really live. 

MARK ZUCKERBERG: Yeah.

JOHN COLLISON: So the last public speaking appearance you did with Theo Von. He actually took a lot of the questions that I wanted to ask. 

MARK ZUCKERBERG: Same kind of vibe. 

JOHN COLLISON: Exactly, yeah. Here at Stripe Sessions. 

So we obviously just talked a whole bunch about AI. You are, I mean, arguably just the largest AI company, right? You are the largest buyer of GPUs in the US outside of a cloud hyperscaler, from my deep research. How has your thinking on AI just evolved over the past year? It hasn’t been moving slowly. 

MARK ZUCKERBERG: Yeah, I mean, I think the high level is, it’s on track. 

JOHN COLLISON: How so? 

MARK ZUCKERBERG: Well, I mean, I think one of the big questions is—I think it’s like—I think most people at this point view that AI is going to transform pretty much every category of product and every part of the economy and all that. 

The big question is to what extent are we in a bubble versus to what extent is this going to be a big thing soon. And I think one of the—I mean, you were just talking about this, I think, right before I came out—even if it is a bubble and it takes a bit longer to build out, that’s not necessarily the end of the world, although it will be expensive. 

JOHN COLLISON: For certain people. 

MARK ZUCKERBERG: Well, yeah, when you’re spending $65–$70 billion in CapEx a year, you’d like it to come when you think it’s going to come. 

JOHN COLLISON: It looks good on you. [Crosstalk] 

MARK ZUCKERBERG: But I think the main question is, sort of, is this going to be a kind of 5–10-year thing? And in some ways it probably will, to get built into every enterprise workflow and things like that. 

But I don’t know, I mean, the more time that—kind of the more that we work on this, it actually seems like all the things that we think are going to happen, happen sooner. And being even more ambitious has been more predictive over the last few years about where things are likely going to be in the industry. 

So, I mean, like, we keep on building, setting goals for our teams. Meta AI is like right about a billion people monthly active using it across our apps. We keep on setting these goals for, “What do we think is going to be the most used AI, what do we need to do?” And we keep on passing them. And then, like, others in the industry on passing their goals, too. And then we’re like, “Ah, we should have had even higher goals.” 

And how much AI is accelerating all of the business agent stuff that we’re doing and advertising and changing that? It’s all just happening very quickly. 

JOHN COLLISON: Do you want to describe the business agent stuff briefly for people who aren’t familiar? Maybe everyone’s familiar. 

MARK ZUCKERBERG: Yeah, yeah. Well, I mean, there’s two basic versions of this. And I think this is probably good context for folks here to have. One is just the transformation of the ad system. It used to basically be, if you wanted to run ads, you had to come up with your own creative, and you had to come up with who you wanted to reach. You had to kind of have a sense of who your customers were, and do a lot of your own measurement. And over time, we’ve just, like, automated more and more of this stuff. 

So the basic end goal here is any business can come to us, say what their objective is—“I want to get new customers to do this thing; I want to sell these things”—tell us how much they’re willing to pay to achieve those results, connect their bank account, and then we just deliver as many results as we can. And in a way, I mean, it’s kind of like the ultimate business results machine. I think it’ll be one of the most important and valuable AI systems and kind of business systems that gets built. 

Now, I mean, obviously there’s a whole ecosystem around this stuff. No one company is going to do everything in advertising. There are all these companies that specialize in doing creative work, and they’ll keep on doing that. But it will be possible in the future that if you’re running a small business, or maybe even a larger one, you won’t even have to start off with a creative. Maybe you can. Maybe you’ll get better results. But you could just come with a goal, and how much that’s worth to you, and we’ll just kind of deliver results across our platform. 

So I think that that’s cool, and I think, is going to pretty dramatically change what advertising is. Historically, advertising has been about 1% of global GDP, but that includes all the stuff that’s really inefficient. Like you buy a billboard and, you know, like three-quarters of the people who see the billboard are not target customers. So when you actually can change it so it turns into this AI business results machine, I think that—you know, just like the overall internet economy is going to grow—I would guess that what we currently think of advertising will sort of grow as a field and will be a larger portion of global GDP. 

It’s interesting. I mean, a lot of innovative companies are experimenting with different kinds of automation with our platforms. And that’s kind of a fun thing that we can get into. The other part of what we’re doing on business agents is this notion that, you know, just like every business today has an email address and a website and a social media presence, in the future, every business is going to have an AI agent that lives in the different messaging platforms that their customers can do customer support and sales through. 

And there’s sort of this wild precedent that we’ve seen so far. In a couple of countries where—like Thailand and Vietnam, are kind of these interesting cases for us—where there’s relatively low cost of human labor and, because of that, a lot of the economy has sort of developed around messaging. And basically, like a large number of businesses out there can afford to have people who just do manual customer support and sales through messaging.

And it’s at the point where even though Thailand and Vietnam are ranked in the top—I think they’re in the 30s by global GDP—they’re actually Meta’s number 10 and 11 revenue countries. And it’s just because so much commerce in those places goes through the messaging platforms already. But that hasn’t really come to a lot of developed countries yet for a number of reasons, including the kind of cost of labor. But I think when you get to the point where every business just has a business agent that is an AI agent that can do customer support and sales, I think that that’s just going to turbocharge. 

And most of our business today, it’s Facebook and Instagram, and ads in there. WhatsApp really, from a revenue perspective at least, is sort of just getting started. And I think that that’s kind of going to be the next pillar of the business. 

JOHN COLLISON: So sorry, that’s quite interesting. Basically, you’re saying that when cost of labor is not an issue, you see people have much more messaging-based business experience—kind of UIs, at some level, for how you do things. And the reason that we don’t see as much of it in the US is just that it’s too expensive to have someone as your, you know, hair salon, just responding to messages all day. And that’s one of your predictions for how it seems like people actually want to interact with—

MARK ZUCKERBERG: Yeah, totally. I think that there is some path determinism here. And I think that the whole messaging ecosystem developed differently in some of these places where the cost of human labor was lower. But overall, I think people are going to contact more businesses for more things when they get very high-quality, low-latency, personalized responses. 

And now, I mean, of course, in the kind of modern web, you’ll be able to have a business agent that is a chatbot on your website, too. So I mean, how much of that is happening through messaging services versus apps or having this just embedded everywhere, I’m not sure. But overall, what we do see across our platform is just business messaging is growing very quickly. It’s about $10 billion of revenue now. It continues to grow quickly. And yeah, I think that this is probably going to be the next kind of new pillar of our business and the stuff that we’re doing. 

JOHN COLLISON: And then, in the first point you made, it sounds like you’re basically saying this execution friction just gets in the way of a lot of things. Like maybe there’d be more confetti animations in the world if it was just a single line to get them to happen. And similarly, you think that many businesses would actually go out and do more. It’s not that they’re not economically willing to pay $20 to get an acquired user or something like that. It’s just all the steps that go into the data analytics, and the creative and everything like that, mean that they don’t actually get the right outcome for their businesses, and so “less friction” in execution?

MARK ZUCKERBERG: Well, that. And also, I think we’ve just reached a point where the AI is more effective at a lot of things that different businesses do. So, I mean, it’s like, OK, if you’re starting a company, there’s normally, like, some core idea or something that you’re trying to do. And then there’s like all these other things that you need to do to make a company work. And like, OK, if you want to reach customers, you’ve had to, you know, like basically go find ways to do marketing. And then, OK, you want to make sure your marketing works well. So you need to find ways to measure your marketing and all this. 

And I do think a bunch of that stuff, we’re just getting to the point where the AI can do it better than sort of the average people who are doing that, which makes it so that more companies can get started more easily. I mean, this is also a big theme for Stripe, right? It’s like, OK, companies also needed to spend all this time doing all this different stuff with financial services and payments, and now they don’t have to. 

So I think what that allows is just way more people in the world to focus on what their core idea is, and kind of execute that, and have a core talent-dense team of people who are focused on that idea; and more of the stuff that’s not the core part of building your company is going to be able to be offloaded. So in the future, if you were working with a creative agency to make creative, you’ll probably keep doing that. If you aren’t, and you’re just kind of hacking something together and throwing it into Meta’s ad system, well, now we’re going to be able to come up with, like, 4,000 different versions of your creative, and just test them, and figure out which one works best. 

It used to be that, on the targeting side—like, the way that our business kind of got started on the ad side was, because people put all this profile information into Facebook early on, it used to be a great place for a business to say, “OK, I think my product is—” like, “I think women between 18 and 28 in this city are really going to love this thing.” And Facebook, early on, was the best place to go reach them because, OK, we kind of knew who people were and had all this—people had input all this demographic information. 

At this point, we basically encourage advertisers not to restrict what the AI can do. We’re like, all right, look, if you think that women in this age range are going to be more receptive, then yeah, give us that hint as kind of an input to the system. But unless you really have a reason why you don’t want to sell to other people, don’t constrain who we’re going to reach to just that, because the AI is actually probably going to be able to find who is going to be interested in your product better than you can. And that just means that, OK, you just focus on building the best product. More stuff that’s awesome is going to happen in the world. 

JOHN COLLISON: The Hello Kitty effect. So, last question on AI maybe. 

MARK ZUCKERBERG: I’m not familiar with that reference. 

JOHN COLLISON: Oh, never mind. So, just on AI and market structure, obviously you guys are so interesting here because, unlike a lot of the pure-play AI companies, you were off doing stuff, doing a lot of production AI, and then also decided to get into the AI business. And also, you know, Llama—well, one of your big bets is open source. I guess you’re also wearing the—what are your Meta Ray-Bans saying to you about me right now? Like, what are they—what blinking warning is— 

MARK ZUCKERBERG: Nothing. They’re off. I left my phone backstage. Their Bluetooth connection is severed. 

JOHN COLLISON: I see. I’m picturing, like, a Terminator-style UI. 

MARK ZUCKERBERG: Well look, I’m, like, famously awkward. I need some input as we’re doing these, so I can make sure I don’t say anything stupid. 

JOHN COLLISON: Yeah. John, not Patrick. 

MARK ZUCKERBERG: [Laughter]

JOHN COLLISON: No, OK. So anyway, you guys are doing a lot in AI. I’m just curious for your thoughts on market structure, because it’s a super fast-evolving—like, it feels to me like the first time since I moved to the Valley that we’ve been in a tech race. You read about the development of, you know, the computing industry in the 1980s, and there’s a new leader every six months. And we didn’t have a good tech race for a while. And now we do, where there’s constantly evolving leaders, and it’s just not clear even where the value capture happens. And so I’m curious how you see the chessboard. 

MARK ZUCKERBERG: Well, I mean, I think this is going to transform every category. So I think people are like, “Oh, is it going to be primarily a consumer thing, or is it going to be primarily an enterprise thing?” It’s going to be everything. So it’s kind of like the internet—it’s like, “What is the point of the internet? Is it going to be consumer or enterprise?” It’s like, there are going to be a lot of opportunities. Right?

So I think we were kind of in this zone early on where people were competing to build the kind of best general model. And my guess is that that stuff is just going to start specializing a little bit. There’s obviously—I mean, intelligence by its nature is general. So you’re not going to be able to just build something that is focused on one area and only go deep in that and not be good at other things. But I do think it will be the case that the products and ecosystems, and the areas of research that different companies emphasize, will be focused on different areas. 

And, you know, we’re doing some of the business stuff we just talked about. We’re also, on the consumer side, very focused on, like, personal AI, right? More than kind of productivity. So, you know, what is that? It’s going to be probably more conversational, probably index a little bit more towards voice, be more personalized. You know, I think people are going to want a system that gets to know them well and that kind of understands them in the way that their feed algorithms do. And honestly, that they don’t have to like, you know, ramp up from scratch because it’s already connected to some of that stuff.

And as you walk through the world and you have your glasses—it’s like glasses are kind of the perfect form factor for AI because you can let an AI see what you see, and hear what you hear, and talk to you throughout the day. And it should all just kind of get to know you, right? And it shouldn’t be like every time you do a query, it’s like—you know, you’re just kind of starting from scratch and you need to provide all the details. So that’s going to be kind of our game on the consumer side.

And that’s a little bit different from, you know, a company like, you know, Anthropic or Cursor who—you know, Claude can do a lot of stuff, but they’re really focused on coding and coding agents, and like that’s going to be a massive, massive part of the economy, right? Where there’s like all these knowledge workers and engineers, and they’ll all be turbocharged by now having more—you know, every engineer is going to be a tech lead that has like effectively an army that is working for them.

JOHN COLLISON: It is striking to me what an overhang there is of personalization in AI, the ability to improve products. Like, it’s kind of funny to me that Siri cannot understand my voice to save its life. And it’s kind of funny, right? Because it’s a general language recognition model. But you have so much of my voice to go off. Like, your job is not to recognize any voice under the sun. It’s just to recognize my voice. And similarly, it would be trivial to have personalized handwriting recognition models. But we don’t have those yet. And so it’s just interesting to me how little personalization there is compared to how much we’ll obviously have in five years’ time. 

MARK ZUCKERBERG: No, I mean, I think it’s very little right now. I think we’re just starting to scratch the surface with some of these products on, like, OK, they learn from what you’ve talked about in the past and what you’ve told them in that model. But I think it’s pretty clear that you’re going to want to connect more and more things to your AI. So from our world, it’s going to be starting with the other Meta services—messaging, glasses—all of that as well as just being able to infer and kind of remember the key things from all the conversations that you’ve had over time. 

JOHN COLLISON: It’s going to be really good. I’m very excited for the Stripe Sessions in five years’ time and what we’re— 

MARK ZUCKERBERG: It’s going to be—yeah, no, it’s going to be wild. 

JOHN COLLISON: It’s going to be a wild time. Does it feel like a wild time to you? Again, it feels—

MARK ZUCKERBERG: Oh yeah, it’s the craziest. It’s amazing. 

JOHN COLLISON: But like, it’s very different—yeah, it’s very different to three years ago. Things were just quieter three years ago. Maybe not for you. 

MARK ZUCKERBERG: Well, I think it was starting three years ago. It’s ramping, it’s ramping. I don’t know, I think it’s the most exciting time. But it’s like, yeah, no, it’s really good. But I do think that everyone needs to kind of rethink a lot of the different ways that their products work. And over time, I think that there are going to be a lot of opportunities for just building companies in different ways, too. 

I know there’s the version of this where, like, everyone at your company just gets a ton more leverage. And that’s the version that I’m kind of excited about. People are, I think, very fixated on this notion of, “Oh, AI is going to be able to replace people’s jobs.” I kind of think— 

JOHN COLLISON: It’s the wrong mental model, yeah. 

MARK ZUCKERBERG: I mean, I guess if we get into, like, some kind of really deep recession and the order of the day is, like, everything needs to be more efficient, I guess AI can be used to make things more efficient. But across what we do, there are all these places where there’s stuff that just wasn’t cost-effective to do before, like providing voice customer support for everyone who uses Meta. It’s like, all right, there’s billions of people who use it. That would cost so much to do. It would cost us so much. And we don’t charge for our product. Our business model doesn’t quite support that. But OK, maybe now if you could have a voice AI that could answer 90% of people’s questions, maybe now it actually becomes—maybe now it makes sense to go hire the people to go answer the other 10% of questions that the AI can’t. Because you could kind of—again, you’re just getting this 10x leverage, and it’s just super, super powerful. 

JOHN COLLISON: This is where someone whispers “Giffen’s paradox” from offstage. 

MARK ZUCKERBERG: Yeah, yeah, yeah. 

JOHN COLLISON: But, OK, as we live through this very fast-changing landscape, you are positioning Meta for it. And one thing as I was preparing for this that I was observing is, it feels like you are early to a lot of tech trends. So here’s my argument for this. You built your own chat app when you were 11, before AIM took off. This was in 1995. 

MARK ZUCKERBERG: Yeah, I got to connect my dad’s—the different offices in his dental office. I built him ZuckNet. 

JOHN COLLISON: Yeah, so the computers at your dad’s dentist’s office. You know, the Oculus— 

MARK ZUCKERBERG: I had to convince him to get me a computer, you know? It’s like— 

JOHN COLLISON: OK, so it’s a good [crosstalk]— 

MARK ZUCKERBERG: The business model of a kid is, like, do useful things for your parents in order to get computer. 

JOHN COLLISON: See, it’s always been about business messaging. 

MARK ZUCKERBERG: It really has been. 

JOHN COLLISON: You acquired Oculus 11 years ago. Like, AR was not a cool thing 11 years ago. I don’t know, is it fair to say that FAIR, the AI lab, was kind of one of the earliest big things you guys did in AI? But that was 2013, again—

MARK ZUCKERBERG: Yeah.

JOHN COLLISON: —before the entire current trend. 

MARK ZUCKERBERG: Yeah.

JOHN COLLISON: I want to get to Libra later on, the thing people are obviously most excited to hear about. 

MARK ZUCKERBERG: Yeah. That thing’s dead. But the good news is you have reincarnated the idea with Bridge into something that is going to do much better, because now there’s actually a company that is fully focused on this that’s going to absolutely nail the opportunity globally. 

JOHN COLLISON: We are seeing a lot of stablecoin growth, as we showed in the chart. But I guess, is this part of your self-identity, being early to tech trends, and this is how Meta stays ahead, and this is kind of how you have been good at what it is that you do? 

MARK ZUCKERBERG: I mean, well, it’s certainly more fun when you’re early than when you’re late. But I mean, there’s plenty of things that I’m late to and have to kind of claw our way back into the game, which I think we’re pretty good at that too. But I don’t know. What were we late to? I mean, like reels, right? TikTok got to be pretty big before we were like, “OK, we need to go win.” 

JOHN COLLISON: Reels are pretty big. Seems like people are watching a lot of reels. You did fine on reels. [Crosstalk]— 

MARK ZUCKERBERG: We’re doing OK, we’re doing OK. But TikTok is a very big competitor. Yeah. So there’s some things that we’re kind of ahead on, some things that we’re behind on. I think that’s the nature of the world. If you’re behind on everything, that probably doesn’t work. 

For the next computing platforms, and I guess the way I thought about it was, I started the company in 2004. Like the smartphone, or the iPhone at least, came out in 2007. So, you know, we were still in the phase of just existential, trying to survive, and like, kind of being a reasonable company at the time that all the smartphone stuff was getting done. So we obviously weren’t going to play any role in defining that. 

And the smartphone overall, I think, has been great in a bunch of ways. Now everyone has a computer in their pocket that they can bring with them all day long. And great for communicating, great for entertainment, great for all this utility. 

JOHN COLLISON: Phones are pretty good. 

MARK ZUCKERBERG: It’s good. The thing that’s complicated for us is we deliver our services through our competitors. So Apple and Google, probably two of our biggest competitors, it is awkward to basically have the main kind of way that you distribute your service. 

JOHN COLLISON: When you run into Tim and Sundar, is it awkward? 

MARK ZUCKERBERG: God, you know, I’m going to dodge the Apple thing again. You’re really trying to drag me in here. Tim’s had a bad week. I’m not going to pile on. No, but Sundar’s cool. I like Sundar. But no, look, it’s complicated, right? It’s complicated. Delivering your services through competitors who are trying to compete with your thing. You know, so, you know, it’s like if you’re Apple, they really want iMessage to win and they don’t—you know, it’s like, the more things that are cross-platform, the easier it is for people to switch from iPhones to Android phones. Right? So it’s like, all right, they can probably survive Spotify. But like, the more things that are like—iMessage is such a huge lock-in for them. So I think that that’s—it’s a big deal. 

So anyhow, going back to the main question, though, I kind of was—so there are all these things over time that Google, but especially Apple, have just been like—you can’t do these things that I think are good consumer things to do. And at some point we ran this calculation and we’re like, all right, if we were able to do all the things that we thought were good, like between the additional engagement and usage of our apps, and the random taxes and stuff, like, I think we might be twice as profitable as a company or something as we are now. I honestly get more offended by the consumer things that we just think are good that they just block us from doing. 

JOHN COLLISON: You just can’t do the consumer thing you want to do. 

MARK ZUCKERBERG: Yeah, so it’s like, OK, we really should try to play a role in shaping the next computing platform. And now we’re not a company that has built operating systems and computing platforms, so my general assumption is if we show up the same year as Apple, or Google with Android, or even Microsoft with Windows—if they kind of build the same kind of AR glasses the same year—I kind of assume the tie goes to them, because they have a much kind of bigger kind of base of stuff built out. 

So my view is that we have to be better and earlier. And maybe that will mean that we need to invest a little bit—or a lot—in advance of when the market is ready for prime time. But I mean, look, we’re doing it. Our VR products, I think, are better and 10 times cheaper than Apple’s. Our AR stuff, the glasses—I mean, the things that I’m wearing now, it’s like going really well. We’re selling many millions of units, and just going to keep on expanding into more styles, more functionality. We’ll kind of add the holograms in. 

So, yeah, no, I feel good about that. I think that this is going to be like—I think the glasses especially are going to be like the ultimate social platform, because holograms will deliver this like—this feeling of presence, like you’re there with another person. And that’s something that no screen today can deliver. So it’s like, in the future, maybe I’ll just beam in as a hologram to this, and that’ll be pretty cool. 

But between that and glasses being the ultimate device for AI that can see what you see, and hear what you hear, I think this is one that is worth investing in 10 years ahead. 

JOHN COLLISON: OK, I want to ask about leadership. So Microsoft invested in Meta in 2007, that big strategic investment, and our mutual friend Charlie Songhurst was involved in that investment at the time. And he said that a big part of his investment thesis for why it made sense for Microsoft to invest in Meta was because you were able to beat CivAI on deity-level difficulty, which is the top of the eight difficulty levels. What do you think about this as an investment thesis in Meta? 

MARK ZUCKERBERG: Well, I mean, CivAI has always been famously bad. So I think it’s beating all the other people that is actually more impressive than beating the deity mode. But I don’t know. Games, yeah. There’s one theory on hiring or investing in people, which is, like, being able to be excellent at one thing shows that someone has the discipline to be able to become excellent at something. 

Whether you’re, like, an Olympian athlete, or good at any one area—like, I think that that level of achievement, it sort of requires a basic level of dedication, an ability to basically learn new things and master things. And I think in general people who can do that—it’s kind of like the thinking on going to school, right? It’s like, OK, maybe you learn some stuff in your courses, but really in theory, you’re supposed to learn how to think and learn new things. And so I don’t know. I didn’t actually know that he wanted to invest because I was good at— 

JOHN COLLISON: I’m sure they told you [crosstalk]—

MARK ZUCKERBERG: If I were picking the things that I might be good at, I wouldn’t think that Civ is like that meaningful of one of them. But I’m glad to know that that’s what got the deal done. All those thousands of hours playing Civ are actually useful. 

JOHN COLLISON: Exactly. Go back and tell your parents now. It’s like, I just want to correct the records about video games. 

MARK ZUCKERBERG: Yeah.

JOHN COLLISON: So last year we had Jensen Huang here, and he had a conversation with Patrick, and he described his management system—which, as you probably know, is somewhat singular. He has 60 direct reports, doesn’t do any one-on-ones, and it kind of goes from there. Do you have an unusual management system? Like, what would be in the Mark Zuckerberg management book? Where do you differ from the orthodoxy? 

MARK ZUCKERBERG: That’s a good question. So I don’t have 60 direct reports. I don’t even like managing people. 

JOHN COLLISON: I mean, that’s different from the orthodoxy. [Somewhat], you know? 

MARK ZUCKERBERG: No, so actually, the way that we’ve set up the company is we have—there are a bunch of really great leaders who sort of—we have, like, 15 different product groups. Everything from, like, Facebook to Instagram, to our ad system, to VR, to the glasses to, like—all the different stuff. And I kind of think that the people who run those 15 groups, or so, are like the people who are running the real things. But I don’t want to have 15 direct reports. 

And so, we kind of have organized the company thematically, where all the apps report to Chris, and all the kind of cross-cutting services—like the ads, and the infra, and the integrity systems, and all that stuff report to Javi. And a lot of the, like, future tech platform stuff reports to Boz. And that’s great. But I’ve also set it up so I just go to the people who—so I mean, obviously those people are all brilliant, and I work with them super closely. But I also go directly to the people who are running whatever the thing is that I care about.

So we’re very nonhierarchical in that way. I think our management team is not, really, just my direct reports. It’s sort of this broader group of, like, 25–30 people who I try to include in everything. 

JOHN COLLISON: So you don’t have 60 direct reports. That would be crazy. 

MARK ZUCKERBERG: Yes. 

JOHN COLLISON: However, there are 25–30 people to whom you go and direct their work. 

MARK ZUCKERBERG: Probably more, but—

JOHN COLLISON: Sixty direct reports would be crazy. 

MARK ZUCKERBERG: Well, what’s the point of having direct reports if you’re not going to do one-on-ones? Well, I actually don’t do one-on-ones with the people who report to me either. Sorry, sorry. So let me stop that sentence and cut that off. But why would you have 60 of them if you weren’t going to do direct—like, I actually believe management is important, right? And part of the reason why I don’t have all the people— 

JOHN COLLISON: That’s for other people to do. 

MARK ZUCKERBERG: I think if you’re going to report to me, you need to be able to manage yourself. But I believe that for the rest of the company, management is kind of important, and it’s an important way that people grow. But for the team, I like to have sort of the largest group that I think can basically have context on every single thing that’s happening across the company. And so that way, they can go do as much as possible. And so that’s why I have this group that I sort of ironically call “small group,” but it’s like, you know, 25–30 people. And that’s like—

JOHN COLLISON: [Unintelligible] Small group auditorium people. 

MARK ZUCKERBERG:—the people running all these things. And I invest a lot of time in making sure that there’s a group that’s very connected to each other and that has all the context on what’s going on inside the company. And that’s sort of, like, the core army that can go get anything done across the company. But I guess the other thing that maybe is nonorthodoxy—it may be, like, a little pithier to explain this. I basically don’t believe in delegation. 

You know, I think that there’s sort of this theory that a lot of people have, which is like, all right, the job of a leader is you hire people and you delegate things to them. My theory is there’s so much going on across the company that I can’t possibly get involved in all of it, so all these people are going to have a ton of stuff that they’re going to do. But fundamentally, if there is a decision that I want to be involved with, I’m going to be involved in it. I don’t believe in delegating. And I think that that’s generally a good way for founders to operate. 

I think it’s like, if you’re running the company and you’re on the hook for everything, and there’s something that’s important at whatever level of detail in the organization, I don’t get the logic of saying, like, “I’m not going to be involved in that.” Now, I mean, you want to have humility and know that, like, if you’re diving into some decision, you may not have the most context immediately. But I generally think that you want to be able to just have the cultural expectation that things are not going to be so hierarchical and you’re just going to dive into whatever you want. 

JOHN COLLISON: Well, and so what is a week in the life of Mark where—it’s striking to me that you do often have a lot of context on these product things. But at the same time, you’ve been running the company for 21 years. You presumably don’t have to do the things that you don’t want to do. And so I’m curious how you organize your time to actually spend time on the things that you think is valuable for the company that you want to work on, and just so you have the context to be able to make these decisions. 

MARK ZUCKERBERG: Yeah, I really try to minimize standing meetings. So I do two things weekly with this Small Group zone. There’s an open-ended kind of strategy discussion, and then I go through our—I have an operational meeting where I basically go through company priorities. And that is it. So the point of not having one-on-ones—

JOHN COLLISON: Sorry, the only recurring meetings [you ran] are these—OK. 

MARK ZUCKERBERG: Yeah. Well, I mean, I do earnings once a quarter and, like—OK. But, look—yeah, we have a board meeting. 

JOHN COLLISON: Current calendar hold. 

MARK ZUCKERBERG: There’s a, yeah, board meeting, stuff like that. But in general, no, I try to cut down on that. You know, and when I say I don’t have one-on-ones, I don’t have, like, recurring scheduled one-on-ones. I talk to all these people constantly. More than they want to talk to me, I’m sure. But I do it when I have something that I want to talk to them about. Or if they want to talk to me, I try to generally keep a bunch of time open, so that way like—stuff is pretty dynamic, and you wake up in the morning and you’re like, “OK, I need to work on these three things today, and I want to make sure that I have a block of time where I can go do that.” 

I get really frustrated and in a bad mood if, like, my whole day is scheduled, and there’s a thing that I know is really important, and I don’t get time to do it because I’m sitting in other things that are not the most important thing to be doing. It’s like, if you have too many days like that in a row, I just, like, explode. So I think it’s just, like, you want to make sure that you keep a meaningful amount of your time open, so that way you can just go talk to the people that you want to either about whatever you want to get done, or whatever you want to learn about, about what’s going on in the organization. 

JOHN COLLISON: So very few recurring meetings. But you have to turn up to earnings. 

MARK ZUCKERBERG: I mean, yeah, yeah. I haven’t won that argument with Susan yet. 

JOHN COLLISON: OK, well, I’m not going to [weigh] into that. But very few recurring meetings. Sorry, Meta investors. Decided here. 

MARK ZUCKERBERG: It’s like, now look what you’ve done. 

JOHN COLLISON: Exactly. Very few recurring meetings, diving deep with the core—

MARK ZUCKERBERG: I think [going to] earnings is actually pretty good, for the record. For everyone who’s livestreaming this. It’s good. High leverage, good use of time. 

JOHN COLLISON: Excellent. Very few recurring meetings, diving deep with the people that you actually need to spend time on, and tons of free time. It sounds like that’s part of what has allowed you to actually be ready for these new things. 

MARK ZUCKERBERG: Yeah, that and like, really good people. 

JOHN COLLISON: Yeah. 

MARK ZUCKERBERG: Right, I think, like, it’s— 

JOHN COLLISON: That’s what struck me about the Meta senior leadership team. Yeah. Very strong leaders who’ve been at the company a very long time. 

MARK ZUCKERBERG: Yeah, we have a tenured team. I mean, one of the things that’s interesting—so we have—so of these, like, 15 or so different product groups, one of the things that I think is really cool is none of them actually started at the company running a product group. So I mean, it’s just a completely different set of things. Like, some started as a designer, some started as a data analyst. One started as an admin. One—two actually now, on our management team, started as admins. And some started as an engineering director or something, and then eventually kind of got to be a VP running a big group.

But it’s literally like—part of the process, for me, is that it’s not just like, “Is this person good?” It’s like you want to have a dense network of connections, so that way—you know, ultimately getting things done is a people-driven thing, as much as it is kind of an intellectual exercise. And I think it’s just really valuable when you’re running a team, to have a team that is densely connected, that trusts each other, that understands each other’s APIs and how—you know, what they’re good at, what their weaknesses are; and just has, like, a long history and trust of having worked together on a bunch of things. You can’t hire that. You have to build that. 

And so, I think a lot of people think about how they want to run their company in terms of like, what is the org chart, right? It’s like, how many people report to you, whatever. I actually think, like, the people and the culture—you know, what’s the saying, “Culture eats strategy for breakfast,” or whatever? But it’s like, I do think that there’s like—yeah, it doesn’t matter what lines you draw or all that. You need to have really good people, and I think you need to make sure that they work well together. 

And different people have different ways of doing that. One of the ways that I do it is I just think, OK, if you want to get a group to work well together, let’s get a group and, like, we’ll spend a lot of time together going through all the context of everything across the company; and, like, that’s kind of—that’s what’s worked for me. It doesn’t mean that it has to be what is the case for everyone, but that’s at least what’s worked for me. 

JOHN COLLISON: In a way, I think this is like an old-fashioned management philosophy of a high-trust group that spent lots of time together that’s pretty well-bonded. So there’s some ways in which you’re embracing the orthodoxy. 

And sorry, I have a very oddly specific question, but I was just wondering, like, everyone here is running a business where there’s a tension between investing in the current thing that’s working and we could just do an iterative improvement, or, you know, your version of the Meta Ray-Bans, or, you know, Llama or some new thing that could be a huge business in the future, but it’s not going to show any interesting P&L contribution this year or next year. And I think that problem is, interestingly, scale invariant. We had it when we were 10 people at Stripe. We were having it when we were 1,000 people at Stripe. We still have it. You’re always debating between this year payoff, next year payoff, 10 years from now payoff. Do you think about that question quantitatively? Do you just do stuff that you think is cool? You see what I mean by the problem, right, every single business faces with just the competing time horizons. 

MARK ZUCKERBERG: Yeah. I don’t know that there’s one answer to it. Obviously, some of that depends on how profitable is your business, and how many bets can you take long term. Obviously, when you’re getting started, you’re just trying to survive. 

JOHN COLLISON: Step one, be very profitable. 

MARK ZUCKERBERG: Yeah, step one. Yeah. Well, I mean, look. Yeah. I think every business goes through the phase early on where you’re just trying to survive and become a stable and self-sustaining thing. And you obviously want to do that in a way that has an eye towards the future. But I think that there is just sort of a grappling to get to some kind of sustainability part. That’s the beginning of a business. And then after that, you can make different bets. 

I don’t have a specific kind of quantitative or financial framework for this. It might be better if I did. But I kind of just think, like, you want to get really good people to do all of the different things, and then you want to do all of the things that you’re doing well. And so, I think part of this is we can do the long-term stuff because we do really well at the core business. And on a day-to-day basis, I’m not spending really any time doing the improvements to the recommendation algorithm in feed, or Instagram, or the ad system. But the people who are running those teams are amazing, and they get whatever resources they need. Because that’s just fundamentally—I mean, those are the core products that people come to the company to use. And I think that there’s so much more upside in terms of delivering better personalized experiences—the business AI platform that I talked about earlier, all that stuff. 

So I think about stuff maybe more in terms of people than finances, which is like, where are you going to allocate the best people, and who gets to eat first? And then there’s just a question of, OK, the more profitable you are, the more you can try different things on top of that; and the more of a surplus you have in talent, the more you can do good things, good additional things. But, yeah, I’m probably more people-focused on that than [crosstalk]—

JOHN COLLISON: It’s more about managing bandwidth, yeah. 

MARK ZUCKERBERG: Yeah. 

JOHN COLLISON: We’re almost out of time, and I can’t believe I haven’t gotten back to the most important thing to talk about, which is payments and stablecoins. 

MARK ZUCKERBERG: Yeah, let’s do it.

JOHN COLLISON: I guess I’m just curious, what got you really excited about Libra and getting into the payment, the stablecoin space in 2019? Libra obviously didn’t work out. Now we’re seeing huge amounts of stablecoin adoption. But I’m just curious, what was your thesis back then? 

MARK ZUCKERBERG: Well, so I kind of have this view that if you’re starting a social app from scratch today, rather than starting it with a feed—I think that’s a big thing, too. It’s sort of evolved into this, like, algorithmic discovery engine. I think you’d start off as a messaging platform, and basically the atomic unit is a one-to-one encrypted message that’s super private and secure, and then you build really secure ways for people to share all the things that they want on top of that. So encrypted calls, encrypted location sharing, all this stuff. 

JOHN COLLISON: Journalists from The Atlantic and… 

MARK ZUCKERBERG: It seems self-evident to me that—again, not going to take the bait. But, you know, it’s like you’re journalist-level, like, trying to get me in trouble here. 

But it seems sort of self-evident that being able to pay and send money to people should be a core part of this. And so how that works, whether it’s a stablecoin or fiat or whatever, is perhaps less important, except that our messaging systems are inherently global, and as you know, sending money to people across lines, country lines, is surprisingly, crazily difficult. 

So we’re like, OK, we’re not a financial services company, but it just seems like this thing is so broken that someone needs to build a stablecoin, and the thing that needs to exist is basically it needs to have a certain amount of scale in order to become a standard. Our take on that was like, all right, maybe we can do that because Meta is a large enough company to build this in a standard—and like, WhatsApp is a large enough app that maybe if you get people transacting in that, that becomes a standard. 

Honestly, what you’re doing makes much more sense than what we were doing. I think these things are very DNA-driven, and you guys are, I think, the best—

JOHN COLLISON: We think about it a lot, you know, [crosstalk]—

MARK ZUCKERBERG:—the best financial services, payments company. Whereas this was a project for us, this is like—your whole company is making this thing go well, and your talent density—I mean just going back to the question we were talking about before, is super deep in these areas, and you have all these awesome people who are, like, very interconnected to the company who are going to do this. And that’s why I think you guys are just kind of steadily building what I think is going to be one of the most valuable companies in the history of the world, and I really admire the way that you guys run the company and all that. And I think everyone who uses your platforms is along for the ride, and they’re going to all benefit from all the innovation that you bring. 

But no, what you guys did with buying Bridge, building it out as a standard, getting to critical mass on Reach, having it be a global thing across the global network that you’re building out, it honestly—it makes a lot of sense to me. I think it’s going to work. And I think you guys are probably the much better company to do this, I think, than we would have been. And I’m glad it’s happening. 

JOHN COLLISON: Well, hopefully there’s some collabs in the future. We’re very intrigued where I think part of our enthusiasm has been borne out of all our work on this over the past 15 years. And again, this idea of borderless financial services, like you were saying—you know, when you start working on messaging and then you try to put in any kind of payments or financial services, those quickly seem apparent. And so we’re very excited about that, and the adoption shows it. 

And very last question: any advice for this group? You have thousands of entrepreneurs in the room, people at earlier stages in their journey. I know it’s kind of hard to provide generalized advice, but based on everything you’re thinking about right now. 

MARK ZUCKERBERG: Well, I mean, I think we’ve talked about a bunch of the themes, right? So one is the way that AI and these internet platforms are developing, a lot of the things that if you were starting whatever you’re starting 20 years ago, you would have had to have built up all these different competencies inside your company, and now they’re just great platforms to do it. I mean, you can run your whole business using all the different things that Stripe has. And I mean, that’s awesome. Like, we didn’t have that when we were getting started, right? We had to build up all this stuff. You can basically find customers and interact with your customers using all of the AI stuff that I talked about earlier. 

So it’s like, I think that that—just what does that mean? I think it means that you can focus on the core idea of what your company is, better. And I think that this is going to lead to much better-quality stuff that gets created around the world, because now you’re just being able to have these very small talent-dense teams that are passionate about an idea, which maybe gets to the management philosophy thing, which is like, yeah, make sure that, if you actually don’t need to end up duplicating all this different functionality, and you can outsource key parts of what you’re doing to a company like Stripe’s platform—which I think is going to make it so easy to scale in all these places around the world and run your business better—what does that mean? It means that your company can all be people who are really passionate about the core thing that you’re doing, and that you can get started on building that team that hopefully is going to be with you for 10-plus years, and that you’ll just do kind of amazing stuff with. 

And I don’t know, I think that’s a pretty cool vision for the future, is that people are going to be much more empowered to focus on the core idea and be able to work together to accomplish things at larger scale than has ever been possible before at any other point in human history. 

JOHN COLLISON: Please join me in thanking Mark Zuckerberg. Thank you. 

MARK ZUCKERBERG: Thanks.