Alexandr Wang: Building a Frontier Lab From Scratch
Meta's Alexandr Wang at Startup School 2026 on rebuilding a frontier lab, swarms of agents, and finding the steepest curve in the world.
Alexandr Wang’s advice to his 18-year-old self: develop your own internal compass for how the future will unfold, and hold conviction in it against the noise.
At Startup School 2026, the Scale AI (YC S16) founder — now leading Meta’s superintelligence lab — talks with Garry Tan about rebuilding a frontier lab from scratch, why talent density compounds, and how to spot the exponential worth betting your twenties on.
Timestamps
00:07 — How Alexandr Wang Started Scale AI
03:25 — Pivoting to the Right Idea
06:23 — Conviction Before Consensus
09:06 — Why This Is the Best Time to Start a Company
11:27 — What Personal Superintelligence Looks Like
13:10 — Building a Frontier AI Lab
16:36 — Why AI Models Need to Be Cheap
20:01 — Vision Will Matter More Than Intelligence
24:06 — Systems Thinking in the AI Era
26:51 — The Biggest Opportunity in AI Today
29:25 — Advice to My 18-Year-Old Self
Transcript
Garry: All right. Full rockstar treatment for Alexandr Wang, everyone. All right. So why don’t we start out—backstage we were saying one of the cool ways to think about this event is this room is actually full of people who are just like us, but when we were 18 or 20. There are some 16-year-olds in this audience. Let’s jump to your story. You came up always really smart, Math Olympian. Jump us to the Alex of that time. What were you feeling? What were you thinking, and what drove you down this road?
Alex: Yeah. Well, I grew up in New Mexico, Los Alamos, New Mexico, which is now Oppenheimer-famous, but it really was the middle of nowhere. I did all these math competitions, all these computer science competitions, but I knew I wanted to do really big things, and it was not exactly clear how or what the exact paths to do that would be. I had a friend who was really into programming and, after high school, got an internship in the valley. I think his first internship was at Palantir. He was an influence for me. After I finished high school, I ended up working at Quora here in Silicon Valley. I worked there for a year. I took a gap year to work there, and then I went to MIT. I was 19 when I worked at Quora.
I was 18 when I went to MIT, and I was 19 when I started Scale. I remember this period from 17 to 19.
I felt like I was constantly changing. Exactly what I wanted to do was constantly changing. I was learning so much just from the people around me. I felt like I was drinking from the fire hose pretty constantly during that time. I would definitely recommend the two things that were really important. One is, I think working at a company was really valuable because from the outside, you have no idea how companies work. You have no idea what it looks like to actually build something. You have no idea what it looks like to iterate on something. You have no idea what it looks like for groups of people to make decisions. I thought that was really important. And then going to school at MIT was actually really important because it gave me a lot of opportunity to explore what was interesting.
It was at MIT that I started training my first models and played around with TensorFlow, which had just come out that year at MIT, and where I ultimately came up with the idea of Scale. After one year at MIT, I applied to YC. It felt like a miracle to get in at that time. And YC was really critical to my entrepreneurial journey.
YC is this amazing blend of being very supportive—they obviously want you to succeed—but they also give it to you very real and tell you when you’re being a dumb ass, which I think is what we all need in life. So yeah, that was the story till then. It was 19, start Scale, and the rest is history.
Garry: I guess you worked with Jared Friedman at the time. And you came in with actually a very different idea than what ended up becoming Scale.
Alex: Yeah. So we wanted to build an AI agent, funnily enough, to help people get medical care. And it was a great example of an idea that I think will ultimately exist. I think we’re even seeing it now. AI agents to help people get medical care are very real, but it was the wrong timing. We worked on it for about a month or two before Jared pulled us aside and said, “Guys, I don’t know if this is going to go anywhere.” And that’s exactly what we needed to hear. At that time, I had studied AI at MIT, I had trained models, and we went back to the drawing board, thought deeply about where the opportunity was, and came up with Scale.
Garry: I guess selling data at the time, large language models had not really come to the fore yet, but self-driving cars were coming up, and computer vision suddenly became important. So that was the first market. Is that right?
Alex: Yeah. So the story here is that when I was at MIT, I did a bunch of projects, trained models of various forms. And these were, by comparison today, little toy models. I remember to train a model, I needed three things. I needed a GCP account, an account on some cloud service to get compute. I needed the code to actually train the model. And I needed data, a dataset. For two out of these three things, you could just press a button online and get them. But for the last one, data, there was no effective way to get data for training these models. So it felt incredibly obvious that this was going to be the future, that there was going to be a way to press a button, so to speak, and get data.
It was very funny because in the years that followed, in the first many years of Scale, data was very unsexy still. Every time we would go out to fundraise, even though our numbers were great and we had great revenue, VCs and investors would always be very skeptical. They’d say, “Oh, I don’t know if this is a good business. Does this have longevity? Is this durable?” It was really weird to me, but none of the investors had ever trained a model, so I guess they didn’t really get it. Fast forward to today, we managed to raise money, we managed to keep going, managed to keep growing the business. But the very same investors who passed on us and were very dour on the potential of AI are writing think pieces today about how data is so critical and is one of the biggest business opportunities in AI.
So it’s very funny to see that whole thing come full circle.
Garry: It seems like that’s actually a really good case study in first principles thinking. You can’t start a company by opening the pages of the Wall Street Journal and saying, “Well, data’s hot. We’re going to go work on that.” You literally couldn’t have started Scale that way. You had to start from simple statements that are about the world that you know to be true and then build something for that.
Alex: Yeah. I think the key thing is you need to develop conviction in a set of beliefs that nobody else agrees with. If you look at all of the most successful companies in the world, they were started at a time long before the core idea was popular. They work on that. They toil in obscurity for years before the idea or the space or the concept of the business becomes consensus. The only way you’re going to be successful is if you’re able to identify these truths about the world early, long before everyone else. One of the most surprising things at Scale is we’ve been working on AI for a decade. You can’t base your business decisions on what everyone else is saying around you. If you go too much with the herd, you will get immensely confused and end up nowhere.
And so you have to develop your own compass of what you think the future’s going to look like because everyone else will just confuse you.
Garry: It seems like one of the things you got incredibly great at was you start with this kernel of, we believe X and nobody else believes it. But then the mechanics of building the business are talking to investors and convincing them and not letting them demoralize you. Talking to customers who should just get it. And then especially convincing people to come work for you.
Alex: Yeah. I think that these early mechanics of building a company, you might have some predisposition to be good at, but nobody is good at starting a company when they start a company. I remember talking to a lot of the investors who I met very early on, and a lot of them would say, “Oh, you just grew so quickly and you changed so quickly. And I didn’t see it at the time.” I think that’s probably true for literally everyone who starts a company. Nobody is good at something they’ve never done before. For all entrepreneurs, you start out pretty bad at everything. The whole game is: how do you develop yourself to continuously improve, to get better and learn quickly?
Garry: Backstage, we were talking about how this is actually a really lucky time to start a company because obviously you can come do YC. The people in this room have each other, which is kind of wild. But not only that, now you have ideally a personal AI that’s going to tell you, “Hey, these are some ways to do it.” Do you think that would have helped you accelerate even faster? What do you think it’s like to start a company today, with AI in the age of AI?
Alex: Yeah. I really think we’re at this amazing moment in the world where the bottleneck is not the progress of the AI models. The bottleneck is diffusing that through the rest of the world and helping the world adapt to this amazing technology that already exists. If the models didn’t improve at all from today, there would still be decades and decades of total upheaval and change in the economy and how the world operates and everything around us. As a result, it’s one of the most incredible—it’s probably a once-in-a-civilization opportunity to be a dreamer and to have a vision and to have ambition and to impose a view of how the future world should look by building something amazing.
One of the things that we were chatting about backstage is when I started Scale or 10 years ago, if you start a company, you had to be, you know, it was David versus Goliath. You had to be clever and find an angle into the market and figure out a way to compete even though you had much fewer resources. Now, with the power of agents, and AI broadly speaking, it’s much closer to Goliath versus Goliath. But maybe the startup is like a Mecha-Goliath that is vastly enhanced by the power of agents and AI, and the large companies are the more traditional Goliath, so to speak. I think that startups now, if you properly embrace AI agents and figure out the way to leverage their strengths in the most ambitious ways, you can easily outcompete incumbents.
Garry: So let’s talk about superintelligence because that’s clearly, that’s even in the name of your lab. What does superintelligence mean operationally inside Meta right now?
Alex: Yeah. I think that we, a year ago, Mark wrote this memo about personal superintelligence, which I think actually is very similar to your concept of personal AGI. We believe that everybody in the world, all the billions of people in the world, are going to have a superintelligence that is adapted and tailored to them, that enables them to accomplish their goals, knows their context, and ultimately is an expander of their own agency. The thing that we think a lot about is agency expansion. How do we help people accomplish things that they couldn’t have ever dreamed of before? What would everyone in the world do if everything was just easy? We think about this in an ecosystem way as well. I think Patrick mentioned it, but we don’t believe in this totalizing, totalitarian view of AIs that control the world.
We believe that these are going to enhance this very broad ecosystem. We believe in billions of people all around the world all having their own personal superintelligence. We also believe in an explosion of entrepreneurship. There are 200 million businesses that are on Meta’s platforms today. We think that number should go to billions with this explosion of creativity and using AI tools. Ultimately, we think that it’s going to be this dynamic ecosystem of business agents working with personal agents and developing this complex ecosystem that is fully AI-supercharged.
Garry: So I was really psyched to see Muse Spark 1.1. My OpenClaw absolutely loved it. How has running a frontier lab been? The Muse Spark level is sort of the Opus level. What’s coming down the pipe? And also, I think that you’re increasingly looking at open source, which I think this audience really loves.
Alex: Yeah. So I’ve been at Meta for about a year now and it’s been quite a year. Getting in, and Meta, we’ve talked about it publicly—Llama 4 wasn’t on the trajectory that was needed for Meta. So I got in there and we did a zero-based build of how do you build an entire frontier lab, in some ways from scratch, obviously using a lot of what we had, and move as quickly as possible. Within nine months of that moment, we launched Muse Spark 1, and then two months later we launched Muse Image and Muse Spark 1.1. There are a few things that have really struck me about this. The first is talent density was incredibly important. That was the core thing to bet on. Talent density is something that compounds naturally.
The more talented people you have, the more of the most talented people want to join you.
And I think it’s amazing to see on the inside, but frontier AI work is research. It is scientific work. We are exploring what you can do with these models, how you can push these models, what can be accomplished with these models, which requires a totally different mindset and operating model than existed for internet companies or internet products. There’s a lot more about experimentation, about science, about scaling. And everything ultimately is about how you develop a lab, an operating model, a system that will be able to compound with all of the exponential growth that will happen in the ecosystem: the exponential growth in capabilities, the exponential growth in compute, the exponential growth in adoption and usage. We are on this very, very steep exponent across maybe every dimension of the ecosystem. And it’s important to develop like an organism.
That’s how you think about the lab that is able to grow with that. It’s been very exciting and we’re going to be shipping a lot more. I think we just launched Muse Spark 1.1, which was a great model. We’re going to continue to have updates on the Muse Spark line. We’ll also have bigger models on the way that I think will be much more competitive with even the very best models that are out there today. We’re going to be launching a harness soon and have been working on a harness to help empower all the developers and agentic developers out there. And then, as you mentioned, we’re working on open source models. As I described before, we believe in a decentralized world of AI capability and progress and development. We want to empower the broader ecosystem and everyone in the world to be able to build and develop using this technology.
So we have a lot of exciting things on the way. We want to empower the ecosystem and developers as much as humanly possible.
Garry: It sounds like one of the ways—certainly when I was using Muse Spark with my OpenClaw, it became clear that it was as good as Opus, especially for that agentic flow with skill files, but it was 8X cheaper actually.
Alex: Yes. Well, I think this goes to it. We don’t believe in a world where these models are so expensive that they get rationed only for the most wealthy of developers and companies. It’s important for everyone to be able to use the technology and to build whatever they want to build with it. We take a view that the best AI products haven’t even been developed yet. If you look at the AI ecosystem and everything that’s happened, every wave is 10 times bigger than the past one. When I started Scale, the first wave was maybe self-driving cars. Self-driving cars are really awesome. They’re really, really cool. But that pales in comparison to large language models and chatbots. Chatbots became this thing that was probably 10 times bigger even than self-driving cars. And then there were coding agents, which came a few years later.
Coding agents are probably 10 times bigger than chatbots. I think we’re just on this deep curve. We’re going to keep seeing these new modalities and form factors and developments of the AI paradigm that will each be dramatically bigger than the last. So our point of view is, let’s unleash the ecosystem. Let’s explore and let’s see. Let’s build the future of the world together.
Garry: So what’s the best way to actually take advantage of the coding model for Muse Spark? It’s OpenCode, right?
Alex: Yeah. Today, the easiest way is to use OpenCode, we have onboarding on the website. And then soon we’ll have a harness of our own. Ultimately, we want great models that plug into all of the available harnesses and empower as much combinatorial innovation in the ecosystem as possible.
Garry: Yeah. I know the harness is under wraps still, but can you tease us with it? I still use OpenClaw. I still use Hermes Agent. These things are, I call them Ferraris that break down on the side of the road all the time. Is this a Ferrari that won’t break down? Tease us a little bit.
Alex: Yeah. Hopefully it doesn’t break down. We’re really focused on speed. For anyone that uses these tools, speed is probably one of the most critical things. Also reliability, like you mentioned, we want to be extremely reliable. We want it to be very extensible and to scale to as complex and interesting of a multi-agent setup as you want to have. There’s so much innovation that will occur even above the harness, frankly, in terms of how to orchestrate and set up loops and develop very complex ecosystems of these agents working together. We want to be really extensible. Ultimately, we want to empower people to harness this technology—harness, actually, pun not intended—but I truly believe these models are already just incredibly powerful. They should be so powerful to fuel many, many points of expansion of GDP growth.
And I think it’s up to smart people with vision and ambition to make all that happen.
Garry: Let’s see. So one question. When you look back on the decade, what do you think they’ll say was obvious in hindsight about AI that people are just missing in real time right now?
Alex: So much of the debate that happens these days is around, oh, how good are the models actually getting? And can the models actually bridge this issue? And when are we going to get superintelligence? Is that in two years or five years? Are we going to hit a wall? So much of that debate is, in some ways, a little bit of a waste of time because I think it’s inevitable that we’re going to have very powerful models. I think we’ll look back and say, all this arguing around when exactly it was going to happen was shortsighted because the reality is we are, as an entire human civilization, on this incredible exponential. You cannot look at the progress of AI over the past decade and not be totally awestruck by how far it’s come.
A decade ago, the best AI models could recognize cats in YouTube videos. And now we’re talking to a digital god that can, you know, I think we’ve all seen some of the hacks and things these systems are capable of. You just can’t help but be awestruck.
I think this trend will continue. These models are going to become more and more powerful. Looking back in a decade, it’ll be obvious that intelligence became abundant and that agency became abundant. The current trends we’re on are just going to keep continuing, and this will be very strange. For the history of humanity, groups of smart people getting together towards a shared goal was the bottleneck of progress. The United States of America, in some sense, was an example of this. The United States of America is formed from a group of very smart people getting together and having a vision for the future that they wanted to enact. That’s the story of nearly every company in America and the story of every YC company.
That’s going to change. All of a sudden, the scarce resource isn’t going to be intelligence or agency. I really think it’s going to be vision and ambition. Do you have a clear view of what you want the world to look like in the future? What is the one way in which you want to put your finger on the scale for how the future of the world will develop and how the world will look in five to ten years that it does not look like today? Do you have the ambition and drive to go through all the crap to make that happen? AI will make that easier. Agents and AI make that maybe ten times or a hundred times easier than it was a decade ago. But the flip side of that is then all of a sudden you can dream bigger.
And the world is, there are so many things that need to evolve for us to be able to fully embrace this technology. The world is really just barely even ready for this technology today. And I think as a builder, we have a responsibility to prepare the world. We have to help enterprises and governments adapt to this new technology. We have to help figure out how we secure the world from a biosecurity perspective or a cybersecurity perspective. We have to figure out how we’re going to manage all these risks that we see with this new technology. But on the flip side, it’s also the time of unprecedented opportunity for humans. We can develop new sciences. We can solve problems in health and biology that have been forever unsolved. We can build new businesses that you couldn’t have even imagined before. There are new creative opportunities that couldn’t have existed before.
So it’s just this incredible cradle of opportunity and risks that I think makes it no better time to be someone who’s a builder and has a strong view of how the world should change.
Garry: Do you think the path has changed? One of the things I saw, I think at Stanford, is that the number of computer science majors actually dropped by some double-digit percentage as people worried, which is sort of insane to me. You still need those skills to even create agents that are that good. Maybe that won’t be true. I’m not really sure. What would you say to people in this audience right now? This is a real question that people are facing. Should they become more wordcel and less shape rotator? What’s the move? And has that changed the kind of people you’re looking to hire and how you manage your teams right now at Meta?
Alex: I think systematic and rigorous thinking are still incredibly important because the abstraction layer—I didn’t use to believe that this is how this was going to play out, but it really has. The abstraction layer just keeps changing. When I started a company back in my day, we wrote code. Now I’m sure nobody here writes code anymore. That’s ridiculous. But now it’s about how you orchestrate the agents together. Then it’s how you develop these organizations of agents. How do you get a million agents to work together well? And then it’ll be, how do you get a trillion agents to work together well? I think there’s going to be this continued need to figure out how you structure workflows at the abstraction layer that we’re going to be operating at.
And that form of rigorous systematic thinking, you now, traditionally, the way this would work in my era of certain companies is you would start by writing code, and then you would have organizations of humans, and you’d figure out how to organize those humans. And that required systems thinking. Now maybe it’s much closer to first you orchestrate the agent, then you figure out how to orchestrate these armies of agents. But I think systems thinking is never going to go out of style. So I think it’s definitely a mistake to go all in on wordcel. I think you need to shape rotate. But then I think what’s necessary going into the future is having a deeper compass and philosophical view on how the world should develop. Because I think there are many lessons from human history around how we think civilization should go through this period.
And humanity will change more in the next decade than it has in the past hundred years, probably. So I think the imperative for us to have positive visions for that and have coherent articulations of how that should develop are more important than ever.
Garry: Let’s get a little more concrete. One of the things I’m curious about is, are there applications of AI that you’re seeing among your friends or internal to Meta that you can talk about that are obvious near term, maybe people haven’t figured out yet? Give us some alpha.
Alex: I think there’s still just astronomical opportunity in agentic looping and figuring out how you develop systems that enable you to spend 1,000x more or 1,000,000x more on tokens to drive an outcome in a continuous feedback loop. If you think about most companies, companies are just these large-scale feedback loops where humans are operating each of the edges. Companies get customers, and they figure out how to make those customers happier. If customers are happier, then they spend more. If they spend more, then you can hire more people who can then go figure out how to get more customers and make those customers happier. That, in some sense, is the feedback loop of every startup or every business. And within that, there are micro feedback loops that exist. I think developing agentic systems that can operate and optimize these feedback loops is—there’s just huge amounts of alpha there.
I think we’ve seen internally at Meta cases where if you can develop the right agentic loop and you have the right eval or the right metric for the agents to optimize, you can have a swarm of agents accomplish more than a team of a hundred engineers very easily, actually. And so I think figuring out what the world looks like with lots of these agentic coordination problems, I think that is one of the most interesting problems today.
Garry: So mechanically speaking, markdown files, cron jobs. And then basically pointing the agent at enough data so that it can figure something out that maybe isn’t in distribution. Yeah.
Alex: Mechanically figuring out what the metric is. And then, yeah, it just comes down to skills, markdown files, cron jobs.
Garry: Slash goal.
Alex: Yeah, slash goal. I think it’s always funny how mundane everything is once you really dig into it.
Garry: So it’s not magic. Some people put a lot of magic. There are some LinkedIn threads out there about some magic stuff.
Alex: Top advice. Ignore LinkedIn.
Garry: Hey.
Alex: LinkedIn is where you get customers.
Garry: So I’d like to end on this, which is you get a telegram to send to the 18-year-old version of yourself. What do you say to that person right now, given all? Thank you for coming back and sharing your wisdom with this audience by the way. What would you send in a message in a bottle to the 18-year-old version of yourself right now?
Alex: Yeah. I think it really boils down to developing your own internal compass for how you think the future will develop, and have strong conviction in it because you will get inundated with noise and people telling you things, and you’ll be very confused, and it’ll be very hard. Especially when you’re young and you don’t have experiences, it can feel very difficult to have true conviction in what you believe and what you want to do. But I think that’s the most important thing. As we talked about, it took a deep, deep conviction in what we were building to be able to weather the storms of many years of chaos in the market, in the industry, and the people around us. And then the other piece of advice I would have is try to identify what is the exponential in the world that has both the steepest curve and will go the longest.
Many decades ago, this curve was Moore’s law, and that was, at the time, clearly the right thing to invest in. I think right now it’s AI progress, but there will be more of these very steep curves in the future. And it’s fine if these curves start, the starting point is very boring, or it doesn’t even seem that interesting.
When I started working on Scale, we had cat detectors and YouTube videos, and that felt—it’s hard to explain the story that that’s the most important technology of our time, but it was on just this unbelievable exponential. And I think I have one last thing I got to say.
Garry: Yes.
Alex: Yes, which is Meta is proud to offer everyone in this room $1,000 of free credits for the Muse Spark API. *applause*
Fantastic. And we’re going to keep making the models better. And right now Muse Spark is, I think, 8X cheaper than Opus. So if you convert that to Opus dollars, it’s a lot more. But no, everyone here, we’ll work to get everyone the details on how to get these credits, and we’re really excited to see what everyone builds.
Garry: Alexandr Wang, everyone.