Bill Gurley & Vince Hankes at MIT: Careers, Conviction and Investing in the AI Era

What began as a PowerPoint presentation Bill gave to an MBA class at the University of Texas turned into a six-year journey writing this book. At the heart of it is a question that I think resonates with many people: "How do you make career decisions you will not regret?" Bill became especially interested in the idea of career regret, and that work eventually contributed to a study showing that six in 10 people would have chosen a different career path. The book is really about helping people move past fear and giving themselves permission to pursue what they are genuinely passionate about. It is built around seven principles, including chasing curiosity, honing your craft, developing mentors, embracing your peers, going where the action is, and giving back.

We are very lucky to have both Bill and Vince with us today, so please join me in welcoming Bill Gurley and Vince Hanks to MIT.

Thanks for being here, Bill. Thanks for having me. I actually think most people are probably wondering why are the two of us here, because neither of us went to MIT. But it turns out if you back enough founders that went to MIT, they'll invite you back. I got you. I think I was interested in doing this with you, obviously, around your new book. I think so many people come to me asking, "Oh, if you're successful, how do you find that path?" And I think a lot of things in the book resonate. I want to talk about the book a little bit today, but also I'd rather talk about putting it in context of both your career and the industry. Very few people, I think, that are close to the modern AI world have also seen the Internet

and mobile, and so experience is a good thing, and so I think we should try to talk about that a little bit today. You took Amazon public, but now we have some of the biggest modern companies going public, and so what are some of the things that were similar about that and will be similar today but different too? And I also find you to be a lifelong learner. Every time I see you, you're always asking and learning about the new thing. And your book profiled so many people in a way that I thought was very different, where it wasn't really about tech entrepreneurs, but it was a very wide breadth of people. And so really, I want to start with there, which is you probably read so many biographies and met so many people.

What was kind of the most wild or mind-blowing story you came across that didn't make it into the book?

Well, so one thing I would start with. By the way, before I get into the book, just because we're going to transition to other things, this is an interesting opportunity as you guys listen to us talk. Because I spent 25 years in venture capital, in most respects, started retirement, which means I don't do new investments anymore. So I did see two and a half decades of venture. Vince really represents the new guard, drives one of the more disruptive, innovative firms out there in venture capital. And so you're getting two perspectives, but one that's super modern and one that's kind of dated. So I wanted you to have that context. One, I'd say the key point in the book—and by the way, I intentionally, the book's divided into two halves: principles and profiles.

So it's stories of success and tools of success. I intentionally—me and my co-writer read like 100 biographies, but we intentionally only put the ones in here that are in fields your parents would tell you not to go into. That was very intentional. So I left out entrepreneurs or doctors or lawyers or consultants or investment bankers. There's none of those profiles. And because I wanted to push people to believe they could go chase careers way outside their field. And the common trait that exists in all of these people—and we also only included people that started at the very bottom rung, so no one born on third base, also an intentional decision—is this commitment to lifelong learning, which you brought up.

So they have this insatiable curiosity. And what I've come to believe is you can't force that as an action, as like a discipline, an activity. You can't say, 'Oh, I'm going to go be a lifelong learner.' You'll just run out of steam. And Angela Duckworth, who famously wrote Grit, 10 years after it came out... So Grit says that it's half passion, half perseverance. And she said later she should have overemphasized passion because this generation of kids, partially because of this stupid gauntlet you have to go on to get into these amazing universities, have been taught to grind. They've been taught to grind, to persevere. But she now says, 'Perseverance without passion is suffering.' I think the key test of whether you're really—so the mechanism that unlocks lifelong continuous learning is obsessive fascination.

And so if there's a topic that you're so fascinated with that you would study it instead of watching your favorite streaming series on Netflix, that's the kind of thing you're looking for. And almost all the people in the book were able to find that thing that allowed that. For a profile I didn't put in there—there might not be that many sports fans in the room, but it would probably be Mike Leach, this famous coach. He was a lawyer in his... and despite being saddled with debt from pursuing the lawyer thing, he took, like, an assistant coaching job that paid, like, 10 grand a year just to get into football. And then worked—he didn't end up winning, like, national championships, but he passed away, unfortunately, a couple years ago, and the service they did,

in Starkville, several coaches came out and talked about the influence he had on the industry. And his coaching tree is amazing—so the number of people that worked under him. But he was so fascinated with innovating in the game that he just lived on the edge. And there's an AI thing to this as well, which is, I think, people that are high agency and fascinated, they not only study the history of their field and all that, but they're constantly studying the edge of their field, which is a much safer place to be. If you just arm yourself with the rote algorithms that they taught you in school, those are in the models. And so just being able to execute the rote part of the job is a not-safe place.

When you're talking to folks, you obviously have a career of investing and meeting lots of people. How do you—what threads are you looking for as someone's telling you about their story to identify if this person has that potential or has found their kind of fascination early?

Well, I mean, are you getting into, like, betting on a founder kind of thing?

Yeah, kind of, or just in general. Like, we have a lot of—we have young, aspiring investors, entrepreneurs, people in this room, and, like, as they're listening for, from other people who are curious about things or even think about themselves, what were things that you are looking for in people that they should maybe be thinking about in their own right?

One thing about this, this fascination concept, is you kind of know it when you see it. It's really noticeable when someone is captured by something in this way. And because of that, all kind of good things happen to them. So if there was someone in your organization that was just emphatic about robots, and that's all they talked about, when a startup came by that had anything to do with robots, you'd say, 'Oh, go talk to Sally. That's all she talks about.' And so introductions come your way. Information comes your way: 'Oh, you probably want to see this. You love robots.' There's this energy that they put off. And by the way, that shows up in interviews. Imagine if you're going to the career center and you sign up for a job interview, and there's 15 of your peers in that same group.

And one of you loves the industry so much that they know all the history. They've read biographies of all the founders. They pop when they talk about it and they smile. They're going to crush you. They're going to get the job. So I just think, in fact, one thing I might worry about in an AI world is these high-agency, fascinated people might separate even more—even more than they do already because they can learn.

You're saying the amount of information you can ingest with AI means the ceiling can be even higher.

Look how fast you can learn. You could never learn faster than right now. And I think it's exponential, like with AI. Because even before AI, podcasts, YouTube interviews—like, your ability to just go home and self-learn, like, go back 40 years and imagine what self-learning looked like. You get a library card—I mean, like, it was hard. But today, you can learn. And so if you're high-agency and are just ridiculously curious, you can run really fast.

From the investing side, I think about this when we're meeting founders: how do you pull these threads out of them? Do you think it's actually true there, where actually it's very obvious in the first 30 minutes of meeting somebody whether they have an extreme fascination?

I think most founders do. And most founders are obsessed with the new technology and the wave that's happening. And so, you know, I mean, take Bret Taylor, who has been around for 20 years and has done amazing things, but a new thing pops up, and of course he jumps over in front of it and studies it. And now if you listen to a Bret Taylor podcast, like on emerging models to solve real corporate problems, he's, what, top five in the world? And it's kind of my point about high-agency people being able to run fast. But I think most of the people that succeed in Silicon Valley, both from a founder and an investor standpoint, are hyper-curious about the technology itself and where it can go. I would include you in that.

Thank you. How do you know? Or—you've seen a lot of great people, too. There's also, to me, I kind of always wonder this: how do we tell if someone's in the top? There's a big difference between the top 0.1%, top 5%, top 10%. And so many of the businesses that we're in, a lot of it's on the margin. Or, like you're saying, identifying information on the edge of what's happening is where it's most comfortable to live. Because I think about our job, a lot of it is making decisions very much on the margin, not the baseline. Or professional athletes, it's true; restaurants, it's true. How do you distinguish between that—the difference between top 0.1% and top 5%? Or does it matter?

I think it matters a ton in venture capital because of the power law. There's just asymmetric returns in venture capital. So you can lose one times your money, but if you miss out on Google, as I did, that's 10,000x. Larry and Sergey presented at Benchmark. I brought them in, 25 employees. So once you recognize that asymmetric outcome—and you guys live this, but you're just constantly looking to try and put yourself in front of those things. So you're in that game where you're looking for that ultimate. One of the things that I would say about founders that are going to go do that: it requires a little more multitasking for them to be more multi-dimensional than you would, than just being fascinated with one thing, because you're gonna have to learn to—no one ever talks about this, everyone assumes a founder can evolve into being a CEO, but if your company goes from 10 employees to 100 to 1,000 to 10,000, you gotta lead an organization.

And so there's a scenario where you might be fascinated with technology, but are you fascinated with leadership techniques?

What's your advice in that situation? I mean, we see that all the time, where, like, great technologists, but as you become a company of 500, your job is not technology anymore.

I mean, I've watched people like Michael Dell go on that journey. He promoted himself to chairman, kicked himself out, went back in. It's a requirement. This gets into the multidimensionality thing. Bezos and Elon have leadership techniques that are innovative. You know, how did Bezos keep Amazon innovative after 100,000 employees? He's not in every meeting. He's created some heuristic that allows the whole organization to behave.

And his fascination with it, and that was a big thing that people talk about.

And he got interested in, he convinced himself, let's just say that, to be interested in the leadership component. But it's not a fait accompli that a founder's going to solve that problem. It's hard.

One of the other things that you talk about in the book that I find interesting is just age. Age doesn't matter. I think that's a lot in the investing career or just in general when you come out of school and people want to be successful young. I often think about being successful young is either a combination of luck or it's ego-driven. Right. Because who cares if you're 32 and successful or 45 and successful? You'd rather be fulfilled and happy going and doing it. But yet a lot of people say, and at least the world I work in, in venture, it's a young person's business. I've seen it. Yeah. Well, Doug Leone's at Sequoia, you know, and I think he's in his late 60s or something. I still stand by it.

Would you go back to Benchmark?

No. But I do think it's a young person's game.

Why do you think that? Or when I think about it, what advantages does a young person have in the industry?

But before we get there, let me hit at one thing you said just about age. And if you read the book, you'll get this notion. There's a chapter called 'Chase Your Curiosity.' You may not be fascinated with something at the level I'm talking about. And I would just say that's OK. I would just say recognize it and experiment a lot. I think Angela Duckworth now says, 'wander with a purpose.' Just move towards the stuff you find exciting. There's this interesting thing that one of the Acquired podcast guys did, which was at each one of his companies, he created a side hustle. And so he would go to the people he worked with and say, 'Do you mind if I do this on the side?' Which is actually, I think it also differentiated him as someone that took initiative and whatnot.

And so at Microsoft, he created Microsoft Garage, which was this thing that allowed Microsoft to start talking to startups again. And through that, he met venture capitalists. And one of those hired him at Madrona. And when he got to Madrona, he said, 'What if I start a podcast on the side?' They said, 'Sure.' And that got him into podcasting, and now Acquired, I think, from a revenue standpoint, one of the top. And so there are ways to wander and experiment and try different things. And talk to people. Who knows? You might meet someone that does something you would love to do. And here's another thing for young people: don't get overextended financially, because your flexibility to repot is dependent upon being able to jump from one place to the other.

Having low OpEx. In my friend group, we say, 'low OpEx.' Yeah. For young people, this is a question I often get, which is you're trying to compete in an industry.

Well, that's why I'm kind of going there. You can say in the context of venture, because I think some people here might want to go into investing or investing broadly. I think it's true in business, too. Experience versus naivete, or in investing, how do you compete against the people that have lots of references and founder relationships and so forth? People do say venture is a young person's business. And so I would love for you to kind of pull on that thread of the why.

So there's about three or four reasons. One of them is it's insanely hustle-oriented. And I don't care what you say. If you end up with a big house, or maybe two or three houses, and three or four children, and you join a bunch of nonprofit boards, your day just gets taken up. And there's no amount of rocks you can look under and go to sleep at night without anxiety. Once again, because of this asymmetric outcome thing, you live in fear of missing this one or two deals a year. How could you possibly minimize the chance of not finding that? It's hard. And so that makes it a hustle business. And can you be out meeting entrepreneurs four nights a week? It's hard as you get older. Totally.

I have a kid now and it starts to get hard.

Yeah, yeah, yeah. So that's one reason. Two, I think, and this all goes back to the asymmetric outcome thing, as you get older you build more mental models and you become more cynical. And that could make it very easy to miss this next thing. And the next thing may break the rules of the previous thing. You and I have talked about things like, do you invest in hardware or not? An old venture capitalist would say, no, never do. And then you might miss Tesla or you miss SpaceX. And so that, I think, being idealistically naive could be a powerful tool. You know, not stupid. No, no, but optimistic. But kind of optimistically naive could be very helpful. And then you're more likely to be close to the age of the founder you're trying to get close to, which can be beneficial.

I would say we, you know, Thrive itself is a young firm. We're only 15 years old. And I think one of the things that we lean into because often people come and say, oh, well, if you are already getting access to the companies, isn't that good enough? And I'm sure this resonates from the Benchmark's perspective. We're not competing against someone that's nobody. We're competing against the best of the best, and you're living on the margin. And so those competitions to win a deal, you have to sell and lean into things that are very specific. And I think a lot of what we do as young people is, yeah, we'll hustle for you, we'll work harder for you, we'll be an extension of you, which is very different than experience.

Yes, yes. And I would add even to that, if you are deep in a subject matter that matters to the founder, I've seen so many deals won that way. So you could be a 25-year-old venture capitalist that just understands gen AI bots better than you. And if you're in it, you're, you're going to outrun any generalist venture capitalist that's 50 years old, because they can't be up at night putting 50 Claude bots to work, but maybe you can. And when you meet a startup and they know you know this stuff, you can win a deal based on that. I've seen it done. Yeah.

Shifting gears a little bit to AI, many entrepreneurs and investors are sitting trying to figure out what the future of AI looks like. And I don't actually know that it's easy to talk about that. But I do look back at the situation of if we were sitting here 15 years ago, 2011, which I guess a lot of people in this room probably were in middle school or something. iPhone had just crossed 100 million users. You were investing during this period of time. iPhone just crossed 100 million users. We would have said, what are the big companies that will emerge? I'm curious what people would have said based on your experience. But I find it hard to believe they would have said, food delivery from strangers to your home, taxis called over your iPhone, staying in strangers' houses.

And yet those are kind of the big apps that were built. So when you think about the pitches that we might hear as investors or founders who are looking for ideas to build, yeah, what do you think stands out the most to you about AI right now that's not obvious?

Well, I mean, one way you can get there is by talking about it. One thing that I witnessed in the internet, so prior to the mobile, was in the early days of the internet, most venture-backed startups would spend 20% to 40% of their venture dollars on Sun and Oracle. And that's how they set up their web servers. And I think when you're in an early phase of a technology, everyone's interested in differentiation. No one's thinking about unit economics because you just want to get out there and wow people. Five years later, almost all of the websites had switched to Linux and open source and MySQL. It's a completely different stack. And I think it was just you went towards optimization. So I expect we'll go through that window.

I don't know what it will mean. There are certainly inference-focused chip companies that'll tell you that's exactly what's going to happen, and open-source models. And I think companies like Cursor have already been forced into thinking that way. There are other things we've seen before, companies moving up the stack, what Anthropic's done to Cursor. We've seen that movie before. And there's lessons to take from that. I don't, you know, part of why Uber happened is if you look prior to the mobile phone shipping, automating the location of every car would require you to put personalized hardware into those vehicles. And people had done it. There were systems inside of taxi cabs. You've seen them before.

But imagine outfitting every car. And you know who did this? I think some of these people that were trying to do car rental companies based on cars you own. They had these kits. You're crushed. But the minute the mobile phone's there, you don't have to install this hardware. And so it's a massive unlock. So that's the kind of frame of mind you need to be like, what things could happen here that are enabled by this technology?

I mean, there's definitely a thread that's connected, which is if you're living on the edge of something and we meet an entrepreneur who's studying the edge so much that they are stumbling into these insights, do you think that is the thing that you should be looking for in an idea right now? I would just be...

Or is it more like you just want to be on the field and building, and actually a lot of this doesn't matter, you just need to be go-getting the reps, learning to find you? Here's what I would say. So, you and I were talking about this a little bit before, but the obvious stuff is priced in right now. Like, these are the highest valuations, biggest dollars. So, like, I don't think it's a good use of marginal capital to say, "Oh, I'm going to either fund a competitor to Anthropic or give Anthropic more money." I don't think either of those is a good return. I would be insanely curious about anything that looked unusual. That doesn't sound actionable. But I would just have my ear to the ground. An AI psychiatrist.

I'd be wildly open-minded to what might pop up. I met a professor that's going to create an AI version of himself because he gets paid really high dollars to consult. I don't know if it'll work, but I would want to know. I would just be wildly curious about anything that seemed unusual. I guess... Novel.

Novel's a better word. So I guess if you take this over, you've had a 10-year, maybe 15-year period between kind of major platform innovations. And as that happens, you get a lot of maturity in the business cycle. And people, what I find, especially when we're hiring execs, as an example of a company, is people try to take patterns and just apply it and force it over a business as opposed to thinking about what the business is and how should it work. And it feels a little bit like we were at that phase in this start of the cycle where the average company was like a SaaS company, which is what worked, plus AI. And I think a lot of the companies that have worked have defied that logic.

And this may get to the young person's game point, but I've seen, yeah, some of my partners have written letters to their companies saying, 'You have to unlearn everything. Like, applying the SaaS toolkit here is not going to work,' you know.

And so, in some sense, you deprioritize the traditional heuristics of success. In some ways, the market size for anything shouldn't really matter that much today because you're creating all these new markets with AI. Do you agree with that?

Yeah, and someone like yourself who is in some of these model companies, if you had access to the API data and there was a company you hadn't heard of that was moving up, that's the kind of novel thing.

That's like the WhatsApp investment story, where all they did was look at the download sheet and see that WhatsApp was rising in the download sheet to find the investment. The other topic we're talking about in this new era is, unlike software and the internet, where you could marginally distribute for free over the internet, today it's super capital-intensive in two ways. One, to the maybe Sun comment, it costs a lot of money to run inference today. And it costs even more money to run it when you're going through the labs, but most companies are using that. And so there's a high startup cost to serve your customer base. And the second is most companies are raising a lot more capital today. And so, just the business environment is very different.

How do you—yeah, this is a bigger question and we can break it down—but how do you think in general about capital intensity in a business cycle and what it changes about the mindset for a company?

Well, I'd say a couple things. So, one, prior to AI and prior to the current version of venture capital, which I'll talk about in a moment, the proclivity to try-harder-style investments was always cyclical with how well the venture capital world was doing. So people get more speculative and open-minded when they're winning. And the minute the market corrects, all that shit goes out the door. And the reason venture capitalists have stayed away from either long-duration—I used to call them material science plays—or low-margin businesses is they're really expensive. And when capital's highly available, it's easy to ignore. If capital becomes scarce, you can die fast, especially if you took on debt.

And people forget about that, maybe because you transitioned to younger venture capital. But anyway, so I do think it's avoided for a rather good reason. You know, venture capital works best in high-margin industries, and industries where you can predict the timeframe. Because people say—Matt Cohler used to say, 'We don't predict the future. We understand the present very clearly.' If you haven't watched this movie, General Magic, I would highly encourage everyone to go watch it. It's free, and it's fantastic. And it's a bunch of really smart people that tried to build a smartphone too soon. And there were three or four venture-backed companies that tried to do that, and they all failed, and they all raised a ton of money.

And the problem with the material science plays, we were talking about fusion earlier, you just don't know when is when. And if when is 15 years instead of five, and your burn rate's a billion a year, that's a huge problem. So anyway, I want to transition into, and I think Thrive's part of how this happened, so you should comment, but the venture capital world, starting about when I stopped doing new investments, has industrialized. The amount of dollars under management, the way you guys are playing the game, the way you forced, I think, the CO2s and altimeters to react to what you're doing, and the fact that I don't know, 50% of the branded venture firms have now scaled up to meet that as well.

There's just dollars moving around at a size that I don't even recognize. And you may not believe this, but the majority of... companies that get hot, um, are approached for preemptive rounds, so they're not, they're not out raising money. People call them and say, 'I want you to take our money,' and there's an implicit threat that at least that SoftBank Masa was very, very direct about, not implicit, that, 'If I, if you don't take it, I'll give it to your competitor.' And it forces everyone to take the money. And so, I'd like to—you look at either AI programming or AI lawyers. There's seven companies that have raised hundreds of millions of dollars. And that's not the venture capital I practiced. And if you're raising that amount of money, and then you try to put it to work to win competitively, I did live through this with Uber, Lyft.

You just start doing very aggressive things from a go-to-market or sales or distribution standpoint. Everybody's market-share-oriented right now. They're not unit-economic.

So this is something we talk about, which is in a world that is that way, can you be unit-economic-focused as a company when you have to compete against your competitors who are not?

I don't think you can. That's what I'm saying. I think all of those companies, let's leave the big model companies aside and just talk about AI lawyers or AI coders. What's the worst thing that could happen to you? You lose market share, like mindshare. That's all anyone cares about. So you price to gain market share. You don't price to unit economics. You just can't.

If you were on the board of one of those companies today, how would you be advising them to operate?

First of all, I'd have to start meditating more. Because I was on a board. This is another area where being young probably is helpful. Because if you're old, you're just nervous as shit. So you'd be better off being naive. Yeah, I mean, what would I say? I'd say really, really, really understand unit economics. Like it, like, even though it may not, like, the fact that Dara was able—I mean, Uber was burning 2 billion a year, and Dara had six, 10 billion in free cash flow last year. Like, being able to make that shift, it's going to require an innate understanding of your true unit economics. It doesn't mean that you don't say, 'Right now, we're going for market share.' And Bezos lived through this. And if you go back, Bezos wrote annual letters.

So if you wanted to relive through this, you could start with the IPO letter and read through it. Burning $2 billion a year, $1 billion a year, the market fell out from underneath them. Barron's said 'Amazon.bomb,' said they're going out of business. People wrote articles saying they're going out of business. And they had to adjust and lay people off. But if you read those letters, he knew what the free cash flow, he knew the unit economics. It was very apparent from the very beginning.

He was aware. We were also talking about the profile of founder that might do well today is one that is able to be both more aggressive on this end, but also more aggressive on the change of the environment where, if you look at what's made Elon successful through these booms and busts, he's been really reactive to cutting teams, downsizing, battening down the hatches when times get really tough. Bezos did the same. And so, even if you look at the kind of the COVID pull-forward and everyone hired all these teams, everyone got so big. What happened in—I looked at this recently—it's like, what happened in the average downsizing, people, you know, I think there was letters put out there that the median downsize was 12%.

The vast majority of companies, though, have continued to shrink from there. Why? Because the right downsizing wasn't 12%, but everyone did the average as opposed to kind of doing what was necessary in that time.

This isn't a restructuring course, but if you ever work in a company and anyone says they're going to do a 5% to 10% layoff, tell them that's the stupidest thing they could possibly do. Because you're going to suffer all the negative dynamics that you would with a 30% or 40% layoff, and you're not going to achieve anything from a real changing the bottom line. And most of the companies I know that do a 10% do two more after that. So then you're like ripping things with a rusty old saw instead of just a short, you know, quick move. Anyway, not a restructuring. But it does, yeah.

And the capital intensity shifts to other categories. I mean, hardware is one that I've asked you about a bunch. But because it's dynamic where everyone was favoring software, we're now entering a phase where, you know, the physical world is also seeing a lot of change. And hardware is becoming really popular. And the conventional advice is like hardware is hard, and the cycles are long, and it's more intensive to go build inventory. No software company has thought about inventory as a thing. And it'll be a real thing in hardware. I'm even just curious in this room, how many people here, if you just raise your hand, want to work at a hardware company or invest in hardware companies?

Some.

A third of the room or something like that. So it's a lot. Do you think that's part of that cyclicality, or do you think we're kind of entering a more re-industrialized and bring back some of that know-how to America, or Dan Wang's book on China where the US is more IP-driven and China is more physical-engineering-driven? Do you think we're entering a phase or elongated cycle of re-industrialization, or do you think that this is kind of a cycle?

Probably both. Yeah, I think there's an element of both. I mean, if there's innovation around robotics, there's an opportunity. I think even Elon would tell you the majority of the componentry for those things are in China, the actuators. So are you really in hardware? Are you an assembler more than you are? Are you really doing material science work? I doubt it. In that case, you're probably going to buy piece parts. You're not going to do a ton of custom parts, but there are new semiconductor plays. I mean, I don't think the venture industry backed many semiconductor companies for a 10, 15-year period, and now we've had one exit, and hopefully Cerebras will do well. But, like, you know, so the...

But, you know, it's a... It's a free market. Anyone can try anything. And so I do think that we, there's this great book by Carlotta Perez about technology diffusion. And she makes the point that bubbles come with real waves. And if someone says, oh, I think there's a bubble, they say, you don't believe in AI. And her analysis would be, no, it's a bubble because AI is real. And what happens is, with a wave that's real, people get rich quick. When people get rich quick, fools rush in. Like speculators, charlatans, they come in. And that happened in the gold rush. And that's already happening here. I like to point out Rick Perry's SPAC for data center, which already boomed and bust, right? But you will see speculation.

And the fact that we are playing around with these huge chunks of money means there's a lot of randomization and whatnot. And the thing that... you'd have to believe for it to just be OK forever to do hardware and all these things is that capital will never get scarce again. And maybe that could happen. But what's happened multiple times in my career is capital gets scarce. And when that happens, the world's very different. Like, very different. And there's one thing about, and I don't know if you guys have thought about this much, there's a weird construct about private company cap tables. They're not set up well for down rounds. Public companies, it's very easy to raise money at a down.

Say more about that.

So there's no reason to know this, but there's a document that when you create a company that births into your existence, your incorporation, and there's common stock for the employees. The preferred stock that venture investors make has this thing called a liq pref, where they can get their money back if there were a liquidation before the common participate. And as you add in more liq prefs at different price points, it creates this kind of—

What's the word I'm looking for?

It's this brittle thing that's all piled up. And some of those rounds have weighted average anti-dilution. And if you go way down, they can explode. And so you can get to the point where if you had to raise down round, you might have to recap. It's too messy to try and pull off. Anyway, one aspect that hasn't been tested of this stay-private-longer thing is, do any of these companies that are well-known and have 15 years as a private company get into a situation where they'd have to do a down round, and what that would look like?

Yeah. Staying optimistic on hardware and saying there is capital. The other dynamic, and you and I have both kind of, I think, been students of China in this regard, is, you know, can—even if we can, you know, muster up the capital and the talent to go compete—can we compete with China? Has been, you know, something that has been a topic, I think, out there. I'm kind of curious. Yeah, it just—you, you went there. I've been there more recently. Any, any advice or thoughts you want to share from your observations there, here for many of these folks? I certainly have a few.

Yeah, I mean, there are a couple of sectors where I think they have escape velocity where it might be near-impossible for the U.S. to compete. So that would be batteries, solar panels, and probably EVs. And in fact, they're so far ahead. If I were in Washington—and by the way, I don't think Washington's oriented a way to do what I'm about to say—but I would want to try and convince companies like Ford to go do a deal with one of these companies,

where they come and build a joint venture plant here the way we used to do there so that we can better understand— Well, I think this is one thing that I don't think people understand. You—I think you posted the video of the Xiaomi factory. Yeah, and oftentimes people say, "What was most surprising about it?" And my—my most surprising fact is it's so automated. Yeah, I think they automate 98 or 99 percent of functions, how they measure it, but it's something like—3,000 people do 30 or 40,000 cars a month. And so even if you were to bring that technology back to the U.S., you wouldn't create a job. Exactly. And so I think people associate how they are progressing as because of low-cost labor and abundant labor.

And actually, it's quite the opposite of many of these hardware companies.

Yeah. I mean, if you're interested in that topic, read Dan Wang's book. Read Patrick McGee's book, Apple in China. You can get up to speed quick. I did a podcast a couple hours after my trip. The podcast is a couple hours. John Arnold went over there recently and did one with Patrick O'Shaughnessy. Elon talks about it. And there's a lot of China hawks in D.C. that want to dismiss that they really innovate or whatever. They're just wrong. They're just flat-out wrong. One thing I do think, and this is—let me say one more thing. One of the reasons people have a hard time thinking about China as an innovator is because we conflate ideas of capitalism and democracy and all this kind of stuff. They have built, inside of a communist country, they have built a system that's highly competitive.

And it comes from leveraging provincial competition and some other things. And by the way, you can just go to AI and say, "Tell me about the history of the EV industry in China." And you'll see there were 150 EV car companies there. So, hyper-competition. In fact, I could argue the model they've created has more intense competition than we have here.

I think you can see that in, I think, why has enterprise software not been a market in China? I think one of the reasons is because people don't want to pay the margin. And so they will host it themselves or build themselves or use open source themselves. We're seeing it now with OpenLLaMA. Why is OpenLLaMA taking off in China? Because it's a free and easy-to-access piece of software. I think one of my other learnings from China, and this is something that you talk about in the book, is they are very good at clustering talent very quickly. Even when I'm there, just people are willing to flock to Beijing. People fly to where that startup is in Beijing. If you guys go to China, I would encourage you to go to Shenzhen, which is kind of the modern city.

I think I learned this stat when I was there. The population in 1985 was under a million. Today it's— I think it was 50,000.

Today it's 20— 20 billion. Yeah, 20 million, 25 million or something like that. And it's very modern. But the reason why it's become such a center is because everyone, all the suppliers are all densely there. Which is the same thing, I guess, about Silicon Valley.

They're a 15-minute high-speed train ride from Hong Kong.

Yeah, and so it's true with Silicon Valley too. How do you think we get to that kind of talent clustering in the U.S.?

I don't think talent clustering is the issue. So Mike Moritz, the famous Sequoia partner, 10 years ago wrote a piece in the FT that just said they work harder. And so that's a problem. And there's so much voice given to work-life balance and whatnot here in the U.S. that you may not culturally be able to get there. Ironically, the AI startups and startups in San Francisco have all adopted the phrase 996, which was stolen from China, or borrowed— let's say borrowed. So there is a group of young people that have decided that they want to work that hard. But there's some of that. Especially for this hardware stuff, there's a remarkable amount of red tape and bureaucracy in America you'd have to untangle.

So South Korea, not China, South Korea can build a nuclear fission plant for one-fourth the price that we can in the US. And someone needs to be massively alarmed by that. Like, you need a Manhattan Project-type re-evaluation of red tape in America if you really want to be competitive with what's happening in China.

I mean, I think, again, to be— to help find areas that we should be optimistic about. SpaceX exists here. Yes. And China, to my knowledge, has not been able to recreate that. And so when I think about areas where folks should be focused in hardware, I think the question would be, are there any areas that you have stumbled into? Maybe it's in silicon, but they're more advanced or there are pockets of hardware where you think that we will have a native advantage by being in the US?

You'd probably be better at answering your own question.

I mean, we're seeing it right now in areas in the data center supply chain. I think this is one place that if you have a lot of capital pouring in, it's always fascinating to me how people will solve problems because now there's a new bottleneck. And so how could it be true that we have some of the most brilliant people pushing AI forward and the constraint's just gonna be physical data centers? It won't be, you just have tons of people doing it there. So that's an area where I'm optimistic. We actually will break through because we're far more constrained there than I think they feel in China.

Although you're running into the bureaucratic red tape. Yeah. And Dario's not helping you by being a doomer because that then creates local backlash.

The other area is just the intersection of robotics and software where it will be more advanced systems. So someone was telling me about this kind of crazy idea in life sciences with animal studies where they're building a cage that has all kinds of advanced robotics in it to help do these animal studies better. I think those are kinds of things where you see the intersection of intelligence in software that will be embodied in robots, where I think we might have an advantage because it's more of the intersection of IP and robotics.

And I'm sure there's a lot of that in biology because you can run experiments simultaneously in a bunch of different ways.

Two more questions and I think we'll open up to the group for more questions. One is, this world is so ephemeral. If you go on Twitter, the view can change so quickly. As you're trying to think about how do you build conviction in what you believe, but the world is changing so much, what advice do you have for people of how do you ground your conviction in something?

I mean, I don't know who coined this phrase. It wasn't me, but it's used a lot. But 'strong opinions, loosely held' is a requirement to be a good venture capitalist. If you don't have strong opinions, you can't make a bet. But to hold on to an opinion in the face of data that's suggesting the opposite is really dangerous also. And so you should just constantly... One of the reasons why I think a partnership works so good, I think it was a book Annie Duke wrote where she said partners know each other's weaknesses so they know when to ignore the person's advice. And it's an interesting construct. You're good at some things, you're bad at some things, and the group can know what people are good at, bad at, and are able to synthesize as a team together.

And a group can also be really good at helping everyone look over walls where you may have blinders or reassess some beliefs that you have. And I think the best venture firms probably pride themselves on proving themselves wrong because of the cost of missing something big. A way long time ago, maybe you were in high school, the venture capital world had convinced itself selling to small businesses was a bad idea. You had Intuit, but no one else. There was a ton of failure. And Peter Fenton, my partner, brought Tobi and Shopify in twice, and we passed. And he let us hear about that for a very long time. You know, and I think that's the right thing. Like the right thing is, see, and that's another area where a mental model can kill you.

Like, do you have any... When you think about building peer groups like that, do you think that these peer—I mean, it seems like they can obviously serve for testing your belief in something and hopefully doing it empirically, but also I find peer groups can be good motivating forces for people too, because if you're trying to be on the edge of something, you have to find other people who are pushing you to be on the edge. Do you think, as you study good peer groups, do you think they're more motivating, or do you think that it's more about trying to get people to open up opportunities for you in a cluster?

It's all of the above. There's a chapter in here called 'Embrace Your Peers,' and it starts with a story about a gentleman named Jimmy Donaldson. Some of you may know who that is. Jimmy, at 17, was fascinated with YouTube and nothing else. He was a D student. His mom wanted him to go to college. And he told her he was going to college, but he wasn't. He was at home playing around on YouTube. But he found three other people that were equally trying to understand how to gamify and exploit and win at YouTube. And he says they were on a... It was one of the precursors to Zoom.

Which one was it?

WebEx? No, it's the one that Microsoft bought. Yeah, they were on Skype calls 16 hours a day for four years. And they all became millionaires. Jimmy Donaldson is MrBeast. And he said if a fifth person had been on those calls, they would have made a million dollars also, regardless of who the individual was. When I was in school, people said, "You should network." And I thought they were promoting a social activity, like go have drinks and say hi and tell people your name. And what I would say instead, now that I have the hindsight of a lifetime, find four people that are equally excited about the stuff you're excited about. And share podcasts, and share ideas, and root for them, and help them succeed.

And they'll help you succeed. And what happens is your network is now 4x as large as it was. The emotional support is real. If you're having a down day, they'll pick you up. The enthusiasm for the project, the learning is accelerated because everyone can share and bring back to the group. And it's a superpower. And I have numerous examples in the book of people that did it. But if you can find a posse like that at this age, they'll be lifelong friends. And you'll all do better as a result of that.