Garry Tan Decides Who Gets to Build the Future
- YC pairs a rapid selection process with $500,000 in funding intended to broaden founder access. Tan says 80,000–90,000 applications compete for about 800 spots, with a 10-minute final interview testing the user, problem and solution: “We just want to know what it is and why it's awesome.” Isaacson argues that lacking a financial safety net can keep people out of startups; Tan calls meritocracy an aspiration YC will never fully achieve.
- Tan sees antitrust and interoperability as conditions for startup growth. He credits the DOJ’s action on Microsoft with helping Google emerge, then argues Google now closes markets through self-preferencing. Isaacson points to successful new AI entrants and Google’s Gemini; Tan agrees and worries AI companies are becoming “the new big tech” and might gatekeep frontier models.
- Tan estimates spending about $100,000–$200,000 annually on tokens for his personal AI experience. His agent prepared for Isaacson in roughly 15 minutes by downloading, transcribing and diarizing Isaacson’s top 10 podcasts. Tan says it “seems inevitable” that everyone uses such assistants in two to three years—a “Jiminy Cricket on your shoulder”—while acknowledging that current costs make his experience a privilege.
- AI-driven job creation remains Tan’s hope, with room for near-term displacement. His open-source project, gstack, aims to teach founders the tools he uses for “400 times” the output; people whose roles become redundant could combine domain expertise with software skills to start companies. He warns that CEOs choosing only cost cuts could deepen wealth concentration.
- Control of personal agents could determine who captures the next wave of commercial value. Tan expects another “browser war” next year and cites AI-lab insiders’ “five kings” scenario, in which potentially all commercial activity collapses into five kings. His preferred alternative is a Linux for personal AI, inexpensive and simple enough for everyone to choose.
- Tan says investment returns must coexist with responsibility for the businesses YC helps scale. He recalls an applicant focused on ADHD medication for children whose commercial prospects did not establish a societal need. Isaacson’s closing reflection is that AI should address cancer and climate change, with YC accountable for the future its selections help produce.
1. YC tests for outliers while trying to widen access
Tan says YC receives 80,000–90,000 applications annually for about 800 spots; 8–10%, sometimes fewer, get interviews. The final interview lasts 10 minutes because Paul Graham thought five could be enough, but five minutes felt rude. The test is to explain “a real person who has a problem” and how the technology solves it. Tan says 10–15% of self-made billionaires age 50 and under are YC alumni.
Tan describes YC’s institutional test as whether it could say yes in 10 minutes to founders such as Musk, Gates, Jobs, Drew Houston, Brian Chesky, Brian Armstrong or Tony Hsu; if not, “I'd have to rewrite everything.” Walter’s Musk hypothetical specifies seeing him before his success was known.
Isaacson argues that people who need to pay for their next meal may not be able to gamble on startups. Tan answers with YC’s $500,000 funding, versus $18,000 when he did YC, intended to let founders participate without rich parents. He contrasts that with seed funding of $30 million at a $300 million post-money valuation for what may be just an idea and one person.
Tan calls meritocracy an aspiration: “it would be a mistake to say that we were actually a meritocracy,” but “it is a virtue to attempt to achieve it.” On immigration, he says about 50% of startups that “really make it” have an immigrant co-founder. His mother’s family came from Burma as refugees; his father came from Singapore.
Tan says his experience founding Posterous—“dead-simple blogs by email”—is important to his work as an investor. Its demo made immediate sense to YC, and the company got funded. At Palantir, Alex Karp taught him how buyers infer product quality. Young engineers wanted Motel 6 rather than Washington, D.C.’s Mandarin Oriental; Karp explained that a general or Fortune 500 executive might ask where they were staying. “If you say Motel 6, they're going to think that they're buying Motel 6 software.” Palantir, Karp insisted, was “Mandarin Oriental software.”
2. Startup opportunity depends on dominant platforms opening markets
Tan argues that the DOJ’s consent decree on Microsoft helped Google emerge by checking Microsoft’s use of Windows to dominate browsers. He now sees Google self-preferencing, pointing to the shrinking share of Google search results occupied by non-Google properties. Europe’s DMA is a model for opening platforms through data transparency and privacy, portability and reverse engineering. His railroad analogy is that a monopoly over transportation can let its owner set prices across the commodities carried on it.
Isaacson points to OpenAI and Anthropic’s success, Apple’s inability to embed its own system in its phones, and Google’s effort to push Gemini in Search. Tan agrees, then calls AI companies “the new big tech.” He shares critics’ concern that, as the AI competition progresses, frontier companies might withhold access to their best models. Openness, he argues, would support startups and consumer choice.
Tan says YC’s policy work responds to groups funded by astroturf money and hired actors who claim to represent small software companies while advocating positions that mainly help big tech. He says YC spends about one-hundredth of what even a single big-tech firm spends on policy, with one full-time hire in D.C., Luther Lowe. Its leverage includes more than 2 million YouTube subscribers: finding founders comes first, and ensuring they can access markets comes right under that.
3. Personal agents offer broader agency at an exceptional price
Tan rejects treating AI as merely chat. YC’s internal agent, Straylight, has a “company brain”: context on every founder YC has backed, people who have emailed YC and all its applications. Tan contrasts a human’s ability to track “7, plus or minus 3” things with an AI agent’s capacity to work across a stack of 1,000 pages—roughly three Harry Potter books—and says that capacity could grow to five, 10 or 20 books in another year. Isaacson challenges the example: retrieving facts is useful, but “that's just data retrieval.”
For a more concrete example, Tan describes an agent downloading and transcribing Isaacson’s top 10 podcasts, diarizing them, then using an LLM to extract themes line by line and timestamp by timestamp, about 15 minutes before their interview. Citing Peter Steinberger of OpenClaw, he says Claude Code and Codex currently trade off some of the appealing, humanlike parts of agents in favor of repeatability and safety.
Tan says his agent has read his email, Dropbox, Apple Notes and therapy notes going back 20 years. The ambition is anticipatory help: a “Jiminy Cricket on your shoulder” that knows its user and can help them see around corners. Tan says widespread use in two to three years “seems inevitable,” but estimates spending $100,000–$200,000 annually on tokens himself: “that's a crazy privilege.”
4. Employment gains require retraining and new businesses
Tan says he does not know for certain whether AI will create more jobs, though that is his hope; in the short term, it may not, and some positions will become redundant. He argues technology is over-vilified and that the response should target abuses of technology, not stop it. He also warns CEOs against treating cost cuts as the endpoint, which could produce a “dystopian moment” of wealth accumulating without being shared.
Tan’s open-source project, gstack, works backward from teaching YC founders to use Claude Code for “400 times” their previous output. His retraining example pairs someone with experience in a field such as medical billing with a 22-year-old who can build software—or has the experienced person learn to do so. The result might reproduce their former work or be “two clicks away”; either way, Tan says, it requires more agency and applying knowledge from lived experience.
Tan recalls an applicant focused on ADHD medication for children as an example of an ethical concern: the founders were impressive and the business was likely to make money, but the need for greater distribution of the medication was unclear. He says the company was later sanctioned by the FDA and shut down, and that there may have been criminal charges. Walter says he has not checked recently. Tan says YC must not only make money, but fund “the things and the futures that we want to live in.”
5. Personal AI needs an open alternative to concentrated control
Isaacson proposes social media that strengthens society. Tan imagines Facebook and Instagram accessed as APIs through a user-controlled agent, with a personal “cognitive filter.” Isaacson objects that personalization already exists while business models reward outrage. Tan’s answer is that the algorithm’s owner has the control today; personal AI could put control over what enters someone’s mind back in the user’s hands. He gives his own example: wanting to focus on himself and see local, but not national, politics.
Tan predicts next year will bring another “browser war,” centered on whose agent people use, with probably a Mac-like and a PC-like option. Isaacson asks whether consolidation to two or three providers would be troubling. Tan says the market is already concentrated, citing AI-lab insiders’ “five kings” scenario, in which potentially all commercial activity collapses into five kings. His counterweight is “Linux for personal AI,” easy and cheap enough for people to choose.
Tan argues regulation should open access, citing the Open Markets Act and the possibility of apps operating on iMessages. He rejects Apple’s claim that only Apple can be safe enough, calling it a smokescreen for market power that can prevent the next Apple from emerging. He also says there are “zero open-source AI companies in the United States” and argues regulation could allow distillation, an advantage he attributes to Chinese companies over American and European companies.
Isaacson asks whether left-wing populism or concentrated technology worries Tan more; Tan says both outcomes are terrible and argues for a middle way. He rejects Bernie Sanders’s proposal to seize assets, and argues for more Googles, choice, prosperity and jobs rather than either big-tech oligarchy or mass asset seizure. Isaacson’s closing reflection is that AI should build “what humanity actually needs,” including curing cancer and fighting climate change. Choosing who builds the future carries responsibility for the future they build.
Full transcript
The Builders is a production of Kaleidoscope and iHeart, presented by Ramp. I'm Walter Isaacson. Welcome to The Builders, my series of conversations with the most consequential minds shaping our future. Today, I’m speaking with Garry Tan. Y Combinator has become one of the most powerful and important brands in Silicon Valley. The companies that emerged from YC include Airbnb, Stripe, DoorDash, and Dropbox. Together, they have transformed how we travel, spend money, work, and communicate.
Now, Garry Tan is deciding which builders come next. Garry first entered Y Combinator as a founder. He became an early investor in Coinbase and Instacart, and today he leads YC itself. In some ways, he’s helping shape those who shape the future. Here's my interview with Garry Tan. Garry Tan, welcome to The Builders.
Thanks for having me.
As a leader of Y Combinator, you have to pick founders who are going to make it. Do you look for one particular thing, or are there 20 different models of founders that can work?
The wild thing about Y Combinator is that we get 80, maybe 90,000 applications a year now for about 800 spots. So it’s less than a 1% acceptance rate. Maybe 8% to 10% of those people would get an interview, sometimes less now. The wild thing is we actually spend only 10 minutes with them for that final stage.
Paul Graham, the founder of YC, figured out that you didn’t need half an hour. You don’t need an hour. You could probably do it in about 5 minutes, but 5 minutes feels rude. So that’s how YC ended up at 10 minutes.
What did Paul look for, and what do you look for in those 10 minutes?
It sounds almost too simple, but 10 minutes is a wonderful time constraint, like how 140 characters for Twitter way back in the day was a wonderful constraint. If you can say something very simply that hits us—“Here’s a user, here’s a real person who has a problem, and here’s how the technology that we build actually solves that”—we just want to know what it is and why it’s awesome.
That founder is the person to go and build that thing. You can do all of that in 10 minutes, and it’s an up-down decision. Something like 10% to 15% of all self-made billionaires as of the present time who are age 50 and under are YC alums.
Let me ask you a real quick question, then. Suppose Elon Musk walked in and you had 10 minutes with him. What do you think the percentage chance is that you would actually fund him?
Oh, 100%. I mean, in current form, 100%. No, no, no.
I’m saying you saw him. You didn’t know he was successful. He was a guy who’d just been, say, ousted from PayPal.
I had a managing partner meeting the other day, and that’s actually one of the example questions we ask ourselves to figure out whether we’re on the right track. By definition, if YC is operating properly, if an Elon or a Bill Gates or a Steve Jobs or a Drew Houston from Dropbox, a Brian Chesky from Airbnb, Brian Armstrong from Coinbase, or Tony Hsu from DoorDash walks into our office, can we say yes to them in 10 minutes? If that answer is no, I’d have to rewrite everything. I’d have to retool everything.
That is almost the only question that really matters when it comes to whether YC is doing well or not.
If you had walked in years ago, would you have accepted yourself?
I think so. My startup, Posterous, was dead-simple blogs by email. I had a pretty banger demo. Paul and Jessica and Trevor Blackwell, the founders of YC, really got it immediately. They didn’t grill us at all. We came on with a really good demo, and they said, “Well, that makes sense.” Then we got funded. Sometimes it’s that simple, though.
Having created and run that company for a while, how fundamental was that to your ability to be at Y Combinator and make venture capital decisions?
I think it’s quite important. There’s a Jay-Z line that I love. In one of his songs, he goes, “Everybody wants to tell you how to do it, and they never did it.”
The entire act of being an investor in technology is bewildering and strange if you really think about it too long, because what we’re doing is selling money. Tech has now become such a fundamental force in the financial world of society that there’s way more money that wants to go into tech than good opportunities. That’s really what YC is.
If you look at most venture capital, it’s crazy that it’s grown so much. The S&P 500 is almost entirely driven by tech. The wild thing Paul Graham and Jessica Livingston figured out when they started in 2005 was that the type of people who would get funding were your buddy at Stanford Business School, your golf buddy, maybe someone whose son went to school with your son at some prep school like Harvard-Westlake. It was access, right?
As it was for a few millennia.
Time immemorial, right?
But do you all think you broke the back of that, or is being a startup junkie or a startup entrepreneur something that you still have to be pretty privileged and pretty connected to do?
It would be a mistake to say that we were actually a meritocracy, but I will say and affirm loudly that we aspire to be. We will never be. I think the woke age of 5 or 10 years ago, we sort of lost sight of that: meritocracy could never be achieved, but it is a virtue to attempt to achieve it.
The reason I ask about that, in terms of a startup culture, which is what you very much nurture, is that it used to be somebody coming out of college—take Steve Case, for example—could go work at a company like Procter & Gamble and learn a business. Nowadays, it’s very hard to get a steady job, and the economy for young people tends to depend more and more on startups.
Yet it seems to me that unless you’ve got a safety net, unless you’re somebody who doesn’t have to make sure your next meal can be paid for, you can’t be in that game, and it leaves a lot of people out.
That’s why YC gives $500,000, for instance. When I did YC, it was only $18,000. Today, you hear about seed funds and Stanford kids raising $30 million on a $300 million post-money valuation. It’s literally just an idea and one person, and the company is somehow worth $300 million.
When we were starting many years ago, it was a very small amount of money. You had to show that you had something that was great. That’s why YC didn’t stay at that amount. We easily could have just stayed at $12,000 or $15,000.
I think there is a recognition that if we want to be a thing that allows people to participate, you don’t have to have rich parents. With the half a million dollars you get getting into YC, you could just have an idea and come in and meet us for 10 minutes.
You were one of the earliest employees at Palantir. You got to see how Peter Thiel and Alex Karp in particular really worked. What did you learn from their style?
I realize now, if I hadn’t rage-quit, I would have been a peer to some of the real legends who really built Palantir. I would go on walks with Karp, and he was a philosopher. He wasn’t a technologist, I would say, but he was incredibly knowledgeable about the world.
One time, I was sent to DC to do some demos for customers. I remember being 25 years old, in my ripped jeans and band T-shirt, and getting sent to stay at the Mandarin Oriental in DC. We would wonder, “Alex, why are we staying at the Mandarin Oriental? Look at us. We’re just kids. We don’t have to do that. Just set me up in the Motel 6. We’re a startup.”
Alex looked at me and my other early engineers for the Palantir finance product and said, “Guys, you’re going to go and talk to such-and-such general or such-and-such VP of some Fortune 500, and you’re going to show this incredible software that you’ve built. Then, as you go out, you’re going to make small talk, and they’re going to ask, ‘Well, how long are you around, and where are you staying?’”
“If you say Motel 6, they’re going to think that they’re buying Motel 6 software, and Palantir is Mandarin Oriental software.” I was like, “Okay, now I get it.” There are a lot of different things embedded in that story, and I learned a lot from Alex from that.
I know you’ve been fighting this antitrust battle. How important is that in order to keep innovation and startup culture?
True free-market enterprise has never been tried. In tech in particular, we have very short memories. I would argue that Google itself, one of the big tech companies, was able to exist and thrive and become a trillion-dollar company because of the actions of the DOJ with the consent decree on Microsoft.
Microsoft was ready to take over the web, and it used its platform advantages in Windows to default-install and take over the browser market.
Do you think Google is now doing the same thing?
I would argue yes.
The governance structure itself insists that you’re supposed to do your duty, and your duty is to maximize shareholder value. People are following the letter of the law and the spirit of their Delaware C Corp to close platforms. It’s a smaller and smaller percentage of the screen real estate in a Google search that’s non-Google properties, right?
And so that’s self-preferencing.
So a lot of your tech bro friends are very much against antitrust enforcement. How do you have a startup culture if you’re going to allow big companies to self-preference?
My view is that we shouldn’t allow self-preferencing. My super pro-startup stance is that we should stop self-preferencing. We should require open markets. The DMA in Europe, for instance, is a great example of forcing big tech to open its own platforms—not just data transparency and privacy, but also portability and reverse engineering. These are the sorts of things that allow for open markets and open access to startups.
If you ask me, “Is big tech too powerful?” I say, “Absolutely.” The answer and the antidote are little tech. I think it’s not unlike the robber baron times, when if you owned the railroads and there was a monopoly on the railroad, you could set the price on literally everything else. One thing would lead to another, and you’d have multiple monopolies across both commodities and the railroads the commodities traveled over.
Well, wait. Let’s look at AI. I’m going to push back a bit. We get a whole lot of new companies—Anthropic, OpenAI—they’re all popping up. Somehow, Apple wasn’t able to self-preference and embed its own system in its own phones. Even Google is trying to do that, where it almost forces you to use Gemini when you’re using Google Search. And yet these other companies are doing well.
Yeah, that’s right.
What’s funny about the AI companies is that they’re almost certainly—I think you can fairly say that they are—the new big tech. So I have concerns there as well. Critics of some of the AI companies would say that, as the AI wars progress, there might not be open access to the latest and best frontier models, and those companies may gatekeep.
You think they should be open, that all the AI models should be open?
That would certainly be the thing that supports startups and the ability for new people to create new companies and, frankly, new wealth, and to give consumers choice.
So what is the role that Y Combinator wants to play in AI?
What’s funny is I think of YC almost as—we’re not a union, but we have union-like characteristics. There’s not really a group out there that can speak full-throatedly about policies and, frankly, actions that both governments and companies should take to make sure that the golden goose can still exist. Whether that’s the billionaire tax in California or antitrust laws in D.C.
Or in Europe.
We’re going to start talking about it and just having a seat at the table. One of the things I discovered as I started examining politics is that most tech people are not interested in politics, but politics sure is interested in us.
And so I decided—
Three or four years ago, I thought, “Oh, well, I should probably spend some time and reciprocate. Let me understand what’s going on.” And what I discovered was pretty wild. Millions of dollars in astroturf money were going to groups that would take around basically hired actors. They would claim to speak for small software companies or little tech.
And they were just paid shills, actually.
They would go to the halls of power, whether it was Sacramento or D.C., and say, “No, no, this thing that actually mainly helps big tech—we, the little guy, really want that.” I was like, “Oh, wow, this is crazy.” No real startups actually go out there and talk to real people in D.C. I saw a system that was broken, and I wanted to fix it.
We do very little things. We have a tiny budget. It’s probably one-hundredth of what even any single big tech firm spends on policy. We have 1 lone full-time hire in D.C., a man named Luther Lowe, and we run it on a shoestring. But we’re people-powered. I see a repeating pattern: There are broken things in society. There are systems out there that have money. Just trace the money: Where’s it coming from, and who benefits?
And then the heartening thing—
A lot of people talk about social media in a very negative way. They think Facebook ruined democracy. I don’t really think I agree. I think it’s the little guy not realizing he has a voice. He and she do have a voice.
When I look at YC, our job one is to find the next great founders. But right under that, it’s: How do we make sure those people can have access to markets? How do we help people who would not have a voice, who would not be able to go to D.C.? How do we give them that megaphone? YC itself now has a megaphone. We have more than 2 million subscribers on YouTube. We can sort of speak in aggregates, and I’m really excited about that because I think that’s us trying to speak truth to power.
I don’t want to deny that YC is also a power in and of itself, but the world is full of systems. These are all systems that were created and constructed by, as Steve said, people no smarter or better than you or I. We can mold them and change them. Once you realize that, you’ll never be the same.
1. Immigration and Family History
You were born in Canada, with a Chinese background. Tell me how important you think immigration is to the United States.
I shout to everyone who will hear that about 50% of all the startups that really make it have at least 1 immigrant co-founder. What we realize is that’s one of the things that makes America incredibly unique. People from around the world want to come here. In tech, they want to come and build the future. That’s certainly my story. My mom’s family came over from Burma as refugees, and my dad came here from Singapore.
Well, you talk about your dad, and we biographers tend to believe it’s sometimes about dad, especially when somebody’s successful. You had a pretty rough time, though. I mean, your father was at times unemployed, at times an alcoholic. Tell me how you had to deal with all that.
It really made me who I am today. I can put myself back in, let’s say, 1992. I’m 11 years old. I don’t know when my dad would come home, maybe around 5 or 6, and I wouldn’t know how many beers he would have already had. I realized there’s this scanner that happens: If you don’t know what father is coming home that day, then you get very attuned to other people and their inner states, and how to react to that.
Now I realize that’s a scanner I use all the time. I can sit down with someone and, not every time, but relatively easily, feel what’s going on for other people. That’s useful for someone who spends time with hundreds of startup founders every single year.
That’s actually a pretty powerful superpower.
2. The AI Agent Revolution
Let me talk about what may be the next step in AI, which is AI agents. You’ve said that you don’t want to fund companies that look at AI agents as a feature. You want them to look at it as the core operating system. Explain how you think that changes what we think about when we talk about AI.
Yeah, absolutely.
We’re going to get a little sci-fi here, but this is something we’re starting to see ourselves as we rebuild the way we operate at YC. I think the number one thing that people get wrong is they just think that AI is just chat.
And it is so much more than that.
We now have an internal AI agent, for instance, named Straylight, that kind of has a personality, and then it has a company brain. It actually has the total context of every founder we’ve ever backed, anyone who’s ever emailed us, all the applications.
The way I think about it from an interesting tech perspective is that a human can only keep track in their heads of 7, plus or minus 3, things. An AI agent is already super intelligent. I was explaining this to my 11-year-old: What an AI can do is keep a stack of 1,000 pages—think 3 Harry Potter books. It can needle-in-a-haystack any page, any fact, anything that happened in those 3 Harry Potter books. It can just know, synthesize, and work with it. The working memory is absolutely gigantic.
In another year, it won’t be 3 Harry Potter books in the brain of an AI agent. It’ll be 5, or 10, or 20. Then the set of things it can operate on is even bigger.
Well, give me a better example and make this more concrete for me, because data retrieval, even from 3 Harry Potter books, is kind of cool, but that’s not—I mean, that’s just data retrieval. What makes something an agent?
I would point to what Peter Steinberger of OpenClaw figured out. If you use Claude Code or Codex, I would say right now they’re so worried about repeatability and safety that they actually trade off all the beautiful parts of working with an AI agent, which basically can be so similar to a human that it sort of takes on human qualities.
So tell me what you ask your agent to do.
Oh, gosh.
One of the things I did—I, you know, it’s read all of my email and all of my Dropbox.
It’s read all of my Apple Notes, even my therapy notes going back 20 years. For instance, I used it to prepare to talk to you—not to get too meta about it, but about 15 minutes before this talk, it was able to download and transcribe your top 10 podcasts you’ve ever done. Then it diarized them. I had an LLM go line by line, timestamp by timestamp, to pull out the major themes.
And that was really helpful for me.
To be able to do that for almost anything you do in your day-to-day life, and for it to do it somewhat more autonomously, seems inevitable. In 2 to 3 years, everyone’s going to do it, and it’s going to feel awesome. Think of it as a Jiminy Cricket on your shoulder at all times. Almost anything you want to do, it knows you, and it’s there to help. It’ll actually run ahead and say, “What about this?” or help you see around corners. The reason it’s too expensive right now is that it’s literally too expensive.
In order to have this experience.
I spend maybe $100,000 to $200,000 a year on tokens. I realized that’s a crazy privilege—not everyone can do that. A lot of CEOs I know are doing it, and they have wings.
Do you think AI, especially the agents you’ve been describing, will end up destroying jobs, or will it enormously increase the number of jobs in society?
My hope is—and I don’t know for certain, and I’m not clairvoyant—but I do believe that technology is over-vilified. It does have problems, and AI safety is something we need to be concerned about. But the way to fix that is not to stop the technology from happening. It’s to go after abuses by people when they use or misuse the tech.
But do you think it’ll create jobs? Will there be more jobs because of AI?
That’s certainly my hope. In the short term, it may not, and that’s quite frustrating. One reason I’ve been working on open source and trying to teach people how to use AI is that my open-source project, gstack, works backward from the question: If I were trying to teach YC founders how to use Claude Code and get 400 times the output they had before, what would that look like? It’s a project they can drop into Claude Code to get exactly the stack I use. Things like that are necessary to teach people how to use the tools properly. Some people will lose their jobs because AI will make their positions redundant. I think there’s a big retraining issue here, but gstack and lots of other efforts could help solve that.
If you have that—
—context, if you were doing job X—for example, if you’re a medical biller—I want someone like that to be able to pair up with a 22-year-old. They have domain expertise, and the other person really knows how to make 400 times more software than before. Maybe that person who knows the space inside out, paired with someone younger—or even by training themselves—can start a company in that space. Maybe it’s exactly what they did before; maybe it’s something two clicks away. Either way, it requires an increase in agency.
What we want people to do is look into their own history and their own lives: What do you uniquely know? What life did you live? What obscure hobbies do you have? What have you seen that nobody else has seen? Then pair that with this incredible ability not just to code, but to think and reason. I think that’s raising the floor. On the flip side, the ceiling is sky-high. I think the responsibility of CEOs and people who have a lot of privilege is not to cut costs—that’s the easy way. That can result in a dystopian moment where capital accumulates more and more wealth, and we’re not sharing it.
You say you push back against people who vilify AI, but what about those who are very pessimistic about it? When you’re looking at companies to back, fund, or encourage, do you consider an ethical component—whether what they’re going to do is driven by being good for society? Absolutely.
We have so many examples at YC where a team would come in, and it was pretty clear they were going to make a lot of money. The most extreme version I remember wasn’t when I was at YC, but I have to call out my partner, Gustav Alströmer. He was head of growth at Airbnb for many years, and then he became a partner here at YC. A company focused on ADHD meds for kids came through and applied, and it had a remarkable set of founders. But the idea was tantamount to increasing the distribution of ADHD meds in a society where we couldn’t tell if that was actually a need. It was definitely going to make money. In the end, that company was sanctioned by the FDA and shut down. I believe there were potentially criminal charges, too.
I haven’t checked recently, but that was a bullet dodged, right?
I’m very proud that YC is an institution where that kind of thing can happen. We not only need to make money; we have a responsibility to fund the things and the futures we want to live in.
Give me a couple of examples of things that would fund the future we want to live in.
Let’s see.
The hard part is that it’s almost anything we actually fund.
I’ll give you one: social media that, instead of dividing us and enraging us, helps us be stronger as a society?
I think that’s entirely possible. This idea of personal agents is actually, I think, a very direct way to do it. Going back to Meta, this is a huge company that controls the algorithm, and at some level, we all wish it didn’t. Now it might not have to. You could imagine a personal AI that you control and program. You don’t download the Facebook app or the Instagram app; it’s just an API. You have a social network, but it’s mediated through your own cognitive filter. You figure out what you want to see, and then you put that in there. Maybe it could even be programmed not just by you, but by the people closest to you.
I think that’s one example of putting control over what goes into your brain back into your own hands. Before, that wasn’t really possible. Could a normal person sit there and program their own algorithm to benefit them? They don’t know how to program, and what would that even look like? It’s so complicated. But with AI agents that can code, it’ll know you and your goals. Right now, I want to focus on myself this year. I don’t want to see anything about national politics; I might only care about local city politics.
We’ve had things like that in the past. We’ve had personalized social media, and yet the algorithms have a business model that works best when they enrage or divide us.
The business model is hosted by someone else, by a big tech company. That’s their control and their power. What I hope happens with personal AI is something more like the personal computer revolution, kind of what Steve Jobs did in 1984. If you remember the original Mac commercial, “1984 doesn’t have to be like 1984.” Big Blue, IBM, doesn’t have to win the whole thing. I think next year will be one of the defining years, when we have the next browser war. It’ll be about whose AI agent you use. I think it’ll be similar to operating systems: there will probably be a Mac, and there will probably be a PC.
3. Who Controls the AI Future
So you’re saying Anthropic, OpenAI, Gemini—it’ll consolidate to just 2 or 3? Wouldn’t that be pretty horrible?
I think we’re already there.
Last year, our friends inside some of the AI labs talked about the “five kings,” which is a pretty astonishing vision: a light cone of all future value, where literally all commercial activity—anything a business would do—could potentially collapse into five kings.
So what can you do to disrupt that?
We’re trying every day, right?
Whether it’s policy, being the funding vehicle to find the next generation of talent and help them thrive, making sure the markets are open for them, or making sure there’s open source—what I want is Linux for personal AI.
Basically an open-source operating system?
And it just has to be easy enough and cheap enough to use that everyone chooses it.
I think that’s possible.
Do you think we should have regulation that pushes in that direction, or is that an impossible dream?
I wouldn’t say anything is impossible. That might actually be preferable to some of the other things some of the politicians are putting out there. Bernie Sanders wants to seize assets.
Speaking as an immigrant and a child of immigrants who came to America to flee asset seizure, we want prosperity. We want the ability for people to start new businesses and continue to do so. That’s one of the defining traits of America.
But what would smart regulation look like?
I think it basically just looks like the Open Markets Act: being able to create an app that actually operates on your iMessages, for instance.
AI can’t touch your iMessages.
What Apple says is, “We’re the only ones who could possibly be safe enough for you.” I don’t think that’s actually true. We shouldn’t allow them to say that nobody else could possibly do this. That’s a smokescreen for their own market power and making money through their own moats, but it also just kills markets. The next Apple can’t emerge.
If Google, Apple, and all the trillion-dollar companies don’t open a lane and create markets, we’ll live in a world where there aren’t startups. That’s a bit of a nightmare scenario for me. I think this is the counterweight against tyranny and, frankly, mismanagement in big tech. You need to have alternatives.
To the extent that the government can create regulation that makes sure these markets are open, I’d even say, looking at how open-source models are progressing now, we’re just handing Chinese model companies a massive advantage. There are zero open-source AI companies in the United States, and it doesn’t have to be that way. There are probably regulations that could allow distillation, which is the exact advantage Chinese companies have over American and European companies. We can have a vibrant ecosystem. I think it’s real that some of these things are too big.
And that’s correct.
If you go all the way back to Brandeis and trust-busting, we live in a better world when there isn’t concentration in markets or self-preferencing. I think that’s more true than ever when it comes to tech and startups.
In some ways, when you talk about Louis Brandeis and the railway companies, you’re really saying that we’ve been through this before: the 1890s, an explosion of technology even greater than what we have today. Then you create these great robber barons who control it, and then you have a period when people like Teddy Roosevelt or Louis Brandeis take on Standard Oil or the railways and say, “You’ve got to open up. You’ve got to allow everybody else to ride your rails.” Do you think we could just say, “Hey, let’s learn from history and apply it now”?
I think so.
I think we have to learn from history in 2 different ways. When you look at the Sanders and AOC end of things, or in San Francisco, Saikat Chakrabarti, or even what Tom Steyer was saying in his failed bid for governor in California, I think you’re starting to see something that is fundamentally un-American to me.
But wait, let me ask you: what are you more worried about? That populists on the left are going to have some socialist system, or that we’re going to have a technology system that gets too controlled by too few people, that isn’t democratizing and doesn’t incentivize startups?
Honestly, both scenarios are terrible, and I would stand against both of those scenarios. I think there has to be a middle way. We should follow American history. How do we create markets? How do we not make it illegal to create prosperity and wealth?
I don’t begrudge Google.
There are people on the hard left, for instance, who would say Google should not exist. We want more Googles. We actually want more. OpenAI was a spinout of Y Combinator. Sam Altman sat in my seat and got to see the front row of all the builders, all the young people with crazy ideas. He, Ilya Sutskever, and Greg Brockman sat in a room and created YC Research, which became OpenAI with Elon.
Now we live on their timeline, an accelerated timeline where the age of intelligence is much closer. I don’t think that’s a bad thing. AOC and a great many Americans today unconscionably believe that all billionaires are ill-begotten. When people say that billionaires shouldn’t exist, they’re also saying that society should ossify, that LLMs should not exist. They’re also saying that things like Google, which organizes the world’s information, shouldn’t exist.
To that, I disagree. We should have more billionaires. We want more choice, more prosperity, and more jobs. There is a middle way; it doesn’t have to be big tech oligarchy. But it also doesn’t have to be mass asset seizure. I think there’s a middle American way, and that’s what we very much will fight for and hope for.
Garry Tan, thank you so much for being part of The Builders.
Thank you so much for having me.
Silicon Valley is very good at finding people who think differently. It rewards ambition, competitiveness, and ideas that can scale. But the interview made me think about the fact that so much of Y Combinator’s energy is focused on AI. The biggest opportunity isn’t simply to build what the market will reward. It’s to build what humanity actually needs.
If AI could focus more on helping humanity cure cancer and fight climate change, rather than just making chatbots and threatening our jobs, there would be a lot less backlash against what the industry is doing. I’m curious to watch how Garry and Y Combinator will place their bets. When you get to choose who builds the future, you have some responsibility for the future they build. That’s it for today’s episode. I’m Walter Isaacson. We’ll see you next time on The Builders.