Uncapped #27 | Vince Hankes from Thrive Capital

[00:00] There's a different way of thinking about concentration, which is when you write a billion dollars into a company, [00:05] you have to have conviction. You can't be like on the fence about is this going to work or not. You have to have almost dogmatic conviction it's going to work. So how do you build that level of conviction? The wind up period to doing these investments is super long. Like years? It could be years. I mean, Stripe, we made our first investment almost 10 years before we made this big $2 billion investment. Or I invested in a company called Isomorphic at the beginning of this year. We spent 18 months getting to know them. I'm very excited to be here today with [00:35] He's worked on most of the major investments at Thrive, including OpenAI, SpaceX, [00:41] Databricks. [00:42] Stripe. [00:43] A lot of ones people have heard about. So very impressive run you've had there, and I appreciate you doing this with me. - Thanks for having me, Jack. - All right, Vince, I wanna start with the evolution of Thrive. So the firm was founded in 2009. [00:56] when you and I are not yet out in the world. And it was a $10 million fund. You fast forward to today, it's a $5 billion fund. Thrive invested in Lattice in 2016. At the time, it was like [01:10] It's been an incredible sort of rise. You joined in 2019. Feels like in the last few years, like a huge amount has happened that's been sort of, you know, [01:18] skyrocketed into new echelons. So can you just kind of like give your overview of like what this [01:24] journey has been, both while you were there and maybe even the broader history back? [01:28] Yeah. I mean, it's been [01:29] a really fun journey to be on. So when I joined, we were, I was the 25th or 26th person

[01:35] We had just raised a billion dollar fund, which was really an early stage fund of $400 million and a growth fund of $600 million. So it's still kind of small in today's dollars. [01:43] And [01:44] I think we were just starting to get into some of the bigger bets we were making. When you joined, did it feel small at the time? Or at the time, were you like, "This is clearly becoming one of the platforms"? It felt small. Because I'd say, I was coming from Tiger. We were investing in almost a $3 billion private fund. We had a $15 billion hedge fund. And so coming from that pool of capital to then a billion dollar fund felt [02:04] And I think the biggest funds at that time were much bigger than we were. And so we weren't kind of in the position of always thinking about-- [02:11] leading or doing the biggest [02:13] checks into a round. We were just trying to really get into the great companies, which was different. But if you go back, [02:19] Part of the evolution that's so fun to think about is Josh was 26 when he started Thrive. [02:24] which is crazy to think about. And so like a 26 year old starting a venture fund [02:28] First of all, it's like, what does that mean? First is he didn't go [02:31] to Silicon Valley and set it up. He did it in New York. He was obviously from the New York area, so that makes sense. But it also kind of made you an outsider to the valley. The people he recruited, it's not like, who wants to go work for a 26-year-old starting a venture fund in New York? So like people who were recruited were his friends and kind of the like misfit-y type people that would go work for a 26-year-old. I mean, you know Miles, like people like that who are coming out of college and young. [02:51] It turned out to be unbelievable picking of people though. There was Will Gabrick, Chris Paik, Jared Weinstein, lots of amazing people. But all those people did something that was unobvious at the time. [03:05] When I think about the evolution, a big part of it is the people that have been around the journey for a long time are people that self-selected into this environment that was different. Do you think that selection itself was part of why the talent hit rate seemed so interesting?

[03:18] Certainly. Today we think about it, now that we do have a brand, people want to work at Thrive. We think about recruiting as a lot of people we want to come join the team. [03:26] are not necessarily the people that come inbound to us. We go seek them out, always be recruiting. - How do you get a contrarian minded person now? [03:34] - Exactly, it's hard because we are more consensus now. So a lot of what we think about and look for are people that aren't looking necessarily to work at Thrive, but at the same time we're trying to set up [03:42] the most important thing about our team is how do we be the place that the most talented young people want to work? And if you don't maintain that, [03:49] then obviously you can't attract those kinds of people. It's an interesting thing about becoming consensus, sort of de facto as you scale like this. When I know basically the whole team there, and I think everybody [04:01] doesn't prefer to be that way, but obviously are all pragmatists and acknowledge that there's sort of this self-reinforcing reality when you become this big. [04:09] So what is the sort of like internal struggle or thinking around this like [04:15] rooted in contrarian, doing our own thing, building something out of nothing. But now you're kind of one of the top groups. Let's go back to that evolution, because I think it's kind of been in the journey. [04:27] People don't know this, but 2012, we invested in Instagram. [04:30] OK, it was important for-- obviously, you got to call it Facebook, because it's a big deal. More importantly, Josh had spent [04:35] I think a couple of years beforehand, just trying to break in and get into this hot company by getting to know Kevin and Mike and the whole team. [04:42] we got in, we got a $20 million allocation. We had a $40 million fund at the time. So in terms of big, bold bets, proportional-- Did you put $20 million? Yeah, we put almost $20 million into the company. We had to raise a little bit on top because we couldn't

[04:55] put a 50% position in the fund. But like big, bold bets were a part of the firm since the beginning. Wow. And obviously it was also important for us because no one knew Thrive Capital, no one really knew Josh. [05:06] But then it's like Instagram, this hot company, Benchmark, Sequoia, and Thrive Capital. And everyone's like, well, who's Thrive Capital? And so it kind of got us into the game. You fast forward a few years. [05:15] And we invested in GitHub. [05:17] Okay, well GitHub obviously was this kind of high-flying developer company. [05:21] All of a sudden, we invested. Literally within months, the CEO stepped down. It kind of went from being consensus to consensus, like in trouble as a company. Nabil, who you know, one of our partners, [05:32] I went in and was interim CFO of GitHub. [05:34] for a period of time. We got to know the business really well, which is part of the whole [05:38] pitch of us as we've deeply partnered with you [05:40] We started to see the data, meet the people, built more conviction such that we actually made-- GitHub became one of the largest investments we ever made in the firm's history. [05:48] at that point in time. And we were able to buy all the shares because it was so non-consensus in the valley. But part of being in New York, part of being an outsider was like we weren't basing our conviction off of-- [05:58] what was happening kind of echo chamber of the valley, we were definitely outsiders to the valley. And so take that even further, like Stripe, more recently, if you think about the evolution, [06:06] Obviously, the quantums have gone up. [06:08] But Stripe to me is kind of the evolution of this exact same confidence in making these big contrarian bets where COVID happened. All these companies were high flying. Obviously, then it all went down and numbers decelerated. Everyone thought, oh, shit, like, is this over? [06:23] Um, [06:24] And

[06:25] if you were really grounded in the numbers, you were scared. Or it looked like an unobvious deal. But if you take a step back and say, [06:33] Okay, well, strike. What is it lever to? It's levered to the growth of payments online and e-commerce. And I said, Jack, [06:39] What do you think e-commerce penetration is going to be in 10 years from now? [06:42] A lot. A lot. A lot. Like no one doubts that. You know, Kareem has this line, "All right, [06:47] It's a lot easier to predict the long term than it is the short term. But everyone in that moment, what are they trying to do? They're trying to predict what will Stripe be in two years from now. [06:54] And so we went in, we spent all of our time with Patrick and John, we've known him for a long time. [06:58] We spent time with product, all the things that you would think about, not just the numbers. And we decided to lean in. [07:03] And then what was unique about that, and I think reinforcing to how we've evolved, is [07:07] we put in almost $2 billion to the company, but they needed to raise five or six. So we had to then go help them raise the money. So a bunch of us went out and pitched to other investors why we had so much confidence to put that much money into Stripe. And this is now in early '23. And so the market environment was not go-go, it's coming off of this post-COVID hangover. [07:24] And everyone to us was basically saying, [07:28] It looks way less profitable than IDN. It's not as good of a business anymore. Patrick and John, they're great, but are they really kind of the founders for the next era of the business? It's interesting because I think you get categorized for that investment as like, [07:40] consensus compounding investment type of mindset. To us, it's so visceral how not consensus it seemed. [07:48] It was a really good formative moment for us in what it feels like to continue to double down at scale with independent conviction. You sort of described also like an investing mindset that I would traditionally ascribe to like early stage investing but applied at...

[08:04] growth plus stage. You're talking about the founders, you're talking about the product, you're talking about the long term and not being too close to the numbers. Do you think that [08:13] when you think about Thrive's great late stage investments, it was [08:17] more of a growth mindset or an early stage mindset that led to those decisions. [08:22] Um... [08:23] I think the... [08:25] Growth mindset, if you want to call it that, which is basically a shorthand way of saying it's very quantitative and numbers driven, is important. But our whole philosophy is [08:34] We start with the qualitative, we develop a hypothesis, and then the hypothesis has to be confirmed by the quantitative. [08:39] And it has to be in that order. Because if you start with the numbers, and you get so excited about the numbers, and then the numbers go down, [08:45] you lose all your confidence. But if you start with the people and what they're doing, and the customers, and the product, and you build confidence, [08:50] and then they miss a quarter or they miss two quarters, you don't react of like, "Oh, shit. My returns are now going away." You try to unpack why and what happened. Is our assumption on the product wrong? [09:00] And that's what should change our confidence almost. And so I think like, what do I think my superpower is? I spent a lot of my career on the financial side. [09:07] But now at Thrive, I spend a lot of it empathizing with how do you build a company that's [09:11] What does it mean to hire an exec? How do you build a team? [09:13] if you kind of intersect those two things, I think that's where our sweet spot is. [09:16] You know, like, is Databricks, Databricks we did a year ago. Why was that a great [09:20] time to invest in the company. [09:22] Well, I'd be like, oh, it's growing really fast. It's got the tail end of AI. But I think the most important point, it became clear that it was going from this single product company [09:29] to a multi-product platform. And the value of the platform is much bigger than the value of the single product. And so inherently there's a mispricing because it's a lot harder to build a multi-product platform than it is a single product.

[09:39] You talked about how the Instagram investment, whatever number it was, was a huge percentage of the fund. You've obviously made a lot of those very bold bets early days. [09:49] Recently, too, you've made bets that, maybe not betting the firm, but are much more aggressive than most venture firms. I think one of the reasons people often talk about founder-led companies having advantages is because people are willing to risk it all many times. Do you feel like you all are still doing that? [10:07] Because it's a founder led firm and you have those dynamics. Is that the sentiment or are you now in a place where you're like, we can't risk [10:15] the firm anymore? Well, I think this is, you know, this is the mentality of why doing what we do is really easy to say but hard to do. [10:23] Like in reality... [10:24] to put that much capital into a single company, it's scary. You've done it with Stripe, Databricks, OpenAI, and obviously you guys have a lot of capital, but on a percentage basis, you're very tied to these companies. We think about an ideal growth fund for us as 10 companies. [10:39] And ideally, it's 10 companies. That's it. It's highly concentrated. And why? Because if you believe the power law is true, [10:44] There's [10:45] In some sense, it's easier to kind of catch a company that's really established going to 100 billion or 200 billion. [10:52] than it is to try to pick the breakout company from a pack of a few thousand. And so what we're really trying to do with these big concentrated bets is find what are those great generational technology companies. [11:02] and then concentrate all of our capital into them. Would you rather try to pick out of $10 billion companies going to $100,000, rather than pick $1 billion companies going to $10,000?

[11:10] I think so. I did this-- we had a thing with RLPs more recently, and so I cut some of the data because it's something that people talk about a lot. And it was basically, what's the number of trillion-dollar companies, hundred-billion-dollar companies, ten-billion-dollar companies over the last decade? And what you'll see is a decade ago, there were no trillion-dollar companies. [11:26] The biggest tech companies were $300 to $500 million large. Fast forward today, there's [11:31] I think that are around over a trillion, the top are three to four trillion. [11:35] OK, in 10 years from now, what will the biggest be? Probably bigger than that. I mean, all of the stuff that's benefiting AI technology means these companies will get bigger. [11:41] But the kind of thing that's missed in this analysis is, [11:44] There's been 75 companies in the last decade [11:46] that have [11:47] reached 100 billion or more in value. [11:49] That's a lot of companies. There was only a few hundred in the $10 billion bucket. So if you think about it, picking the 75 out of a few hundred is much better odds than trying to pick the thousands of unicorns that will get to $10 billion. Yeah, I mean, this actually confirms kind of an intuition. I think a lot of people have been thinking and saying, which is that the winners are getting bigger than ever. Which basically what you're saying is the number of $100 billion companies has grown much faster than the number of $10 billion companies. Yeah, or think about a different lens, a more product-centric lens. [12:18] you know, Stripe. [12:19] obviously they dominate in online payments. [12:22] But now we're talking about stablecoins or AI in payments. Who's going to be the company that wins or takes a lot of market share there? [12:29] Is it going to be Stripe or is it going to be Star? I hope a lot of startups have a shot at gold, because that's the nature of what we believe in. [12:34] But right now, guess who's really excited about it and investing a lot of money at it? It's Stripe. They have amazing founders, a great technology stack, great talent.

[12:41] And so I think what people underappreciate is these really well positioned companies that are getting to scale, they benefit from the scale. Just like Google and Microsoft and Amazon and Facebook all benefit from that same scale. Why is that happening now different than it was happening 10 years ago? Why did these dynamics not exist in 2015 in the same context? [13:02] you know, magnitude. [13:04] I think just people don't appreciate the power of compounding. [13:08] The iPhone came out 15 years ago, about, a little bit more than that. And so mobile internet, I think it's not that old. Everyone wants things to happen in a short period of time, but the reality is most of these companies are 10 or 15 years old. It's actually, if you looked at, obviously I study a lot of these companies at scale, the vast majority of dollars of enterprise value that get created are in the second or third decade of a company. And so if you think about all of the great companies we talk about now, Stripe, Airbnb, DoorDash, Uber, these companies were all built out of the mobile era. [13:38] now getting into the second decade of their lifetime, [13:40] And that's where we know historically, if you look at Shopify or Salesforce or Tesla, [13:44] The second and third decades were far more lucrative in value creation than the first decade. Well, now all these companies are coming of age to the second decade. And I think you will realize that scale is really powerful. And it just took time for them to blossom. [13:55] Do you think that same thing ought to happen with new startups 15, 20 years out? In other words, like, [14:01] Is what you're saying-- does it describe just that all of tech is going to do really well? Or is there some particular reason why right now investing in the big companies is better than investing in these small companies, some subset of which will compound forever too?

[14:15] I don't think this is summarizing Tide lifts all boats. I think it's more of an expression of [14:20] the nature of technology is there are benefits to scale. [14:23] So like, [14:24] if you get to scale. [14:25] You have more distribution. [14:27] you can reach more customers, you can ship new products, you can kind of compound on yourself. [14:32] in the talent flywheel. [14:33] Obviously a lot of these companies started with great talent and they developed it. But now if you're a young person, where do you want to go work? You want to work at OpenAI or SpaceX. Which by the way wasn't true 10 years ago. It didn't used to be that the OpenAI sized companies 10 years ago were appealing in the same way that these are now. Maybe, but the OpenAI sized companies 10 years ago was Google. Yeah, I know, but recruiting against Google as a startup was doable. Recruiting as an OpenAI right now is very hard. In general, I think the market just got a lot more competitive in technology. I don't think this is a rising talent for all folks. [15:03] I think there are a select number of these companies that will benefit from the scale. And when they're very founder-driven, [15:09] they will benefit as they get bigger. And we know with the power law that there are kind of accumulated advantage to those companies. And so all we're trying to do is map our fund strategy to the power law, which is to be concentrated in these [15:20] really phenomenal tech platforms. [15:22] I think on the like [15:23] the whole asset class in general has matured. [15:25] So like [15:26] 15 years ago, how many companies raised $100 million in capital? [15:29] Not that many. [15:30] Now, I think the numbers are something like, over 500 companies a year raise at least $100 million capital. That's crazy. That's a lot. How many companies do you know that you would put $100 million into? I don't know, but 500 a year is a lot. [15:41] And so I just think about it is there's a lot more saturation [15:44] in the game on the field for picking out the billion dollar company. Like that same, you know, because we were pulling for the same LP event.

[15:51] Um, [15:52] The average growth check, so say $100 million plus round, is $150 million round. [15:58] at a billion to $2 billion valuation at 20 to 100 times revenue. [16:02] OK, well, that doesn't-- to me, that's not growth investing. That's kind of like large-check venture investing. And you hope it works. [16:09] But the likelihood is the vast majority will not work. And so in your view, the real growth investing starts at more like... [16:15] the $5 to $10 billion valuation rounds, and things have nine figures of revenue. Or said differently, I think about less quantitatively and more [16:22] The real growth investing happens when you can actually wrap your head around something that's solidified. [16:27] And then you can look at the data to substantiate it. [16:29] And if you can ground yourself in the product and what's solidified and then substantiate it, okay, now we have a baseline that we can work from. But if you're [16:36] Investing in a company that's been around for three years and has $50 million in revenue, that's amazing momentum. [16:42] But as you know, building a company, a lot of things can break from there. [16:48] Given this backdrop, what do you feel like are the winning strategies in venture [16:52] like now. Obviously, there's this [16:55] I don't know how you'd categorize what Thrive does, but like large check, concentrated investing, and like breakout winners. Are there other strategies other than Thrive that you're very bullish on now? [17:05] Thank you. [17:06] I mean, I'm obviously most bullish on the strategy we're executing on. [17:09] I think there will always be a place for [17:13] very focused early stage investing. [17:15] Because at the end of the day, where else can you go put in 10, 20 million dollars into a company and get a billion out? [17:20] And the reality is if you're good at finding people early and you can help them build companies and you can get a large chunk of their company, you'll be able to drive great returns.

[17:28] I personally think it's gotten a lot harder because of how competitive it is, but there will always be a market for that. So what are you skeptical of now? Or what's the hardest part in venture to play? I'm skeptical. I mean, I think we really barbell our strategy. We're either early [17:43] with companies and we [17:44] work alongside and roll up our sleeves, we own [17:46] decent, chocolate company, or we invest when it's becoming a clear [17:50] platform company. We do do stuff in the middle. [17:54] Like we call it breakout companies. Like we invested in cursor. [17:56] It's obviously a phenomenal team. There's so much momentum. It's kind of got the zeitgeist in this AI moment of coding. [18:02] But A's and B's are rare. No, but investing in cursor in some sense should be the exception because that part of the market is so competitive. Yep. [18:10] There's so much capital chasing these $500 to $2 billion companies that really a lot of them don't even have product market fit. But we're capitalizing them like they are a company with product market fit. [18:20] and growth. And so I think that to me is where I'm [18:23] least optimistic, not because I don't think there's going to be winners, there will be [18:27] risk of capital loss matters there. If you're investing $10 million checks or $5 million checks, you can afford to have zeros. [18:33] If you're investing $100 million checks, it's hard to have zeros. If you look at the composition of a lot of these growth funds that are $2, $3, $4 billion [18:41] They have 30, 40, 50 investments in a growth fund. [18:44] which means the average check in that fund is $70 to $100 million. But why do we get to that point in time? Well, because as these funds got bigger in scale, [18:51] They hired more partners. [18:52] But it's like the law of people. If you have more partners, you do more deals. If you do more deals, you're less concentrated. It's that simple of an equation.

[19:01] And if you're not very concentrated and you have a bunch of $100 million checks into [19:05] mid-stage companies, I think it's just hard. [19:07] How many people work at Thrive now? [19:09] The whole firm is something like 75 people. We're about eight investors. This is a good example. [19:15] take our assets and they divide it by investors and they say, "Okay, put you on a benchmark. Where do you rank?" [19:21] And they look at us and say, wow, you guys have a lot of assets per investor. I'm like, that's the wrong way to look at us. [19:26] You should look at, because we don't think about it as, oh, each dollar requires more work. It's basically the investment decision. Each company is a commitment. [19:32] So if you just flip the numerator of that equation from dollars to companies, we make 12 early stage investments a year. [19:39] and we make a handful of growth bets a year. [19:41] That's 18-ish round numbers on eight people. So we're roughly two, maybe three investments a year a person. You just put a lot of zeros when you send the check. Yeah, exactly. That's a good way to do it. But seriously, that-- OK, then just on the same benchmark, who else does two investments a partner a year? [19:57] I mean, there are firms that do that, but yeah. [19:59] But it's very different, I think, in the philosophy, relative to a lot of the other growth funds that we compete with. So actually, when I was asking the number of people that worked there, I was thinking of it this way. Because I was thinking of, let's say, 10 concentrated positions. [20:09] eight investors, whatever. I guess the net of that is most of the time, you're making new investments rarely. [20:19] So what does that look like for you? What's your mentality given the infrequency of decisions? [20:25] I mean, we turn over a lot of rocks. I spend a lot of time looking, evaluating, trying to get to know people. Like, you know, there's a different way of thinking about concentration, which is when you write a billion dollars into a company.

[20:38] you have to have conviction. [20:40] You can't be like... [20:42] on the fence about is this going to work or not. You have to have almost dogmatic conviction it's going to work. [20:47] So how do you build up that level of conviction? [20:49] The wind up period to doing these investments is super long. Like years? It could be years. I mean, Stripe, we made our first investment almost 10 years before we had this big $2 billion investment. I invested in a company called Isomorphic at the beginning of this year. [21:01] We spent 18 months getting to know them. But how did that journey work? [21:04] We were partnering deeply with OpenAI. We were thinking about what are the other domains where it might not be kind of on the path to AGI for the core labs. Life science was one of them. [21:13] I got in touch with Isomorphic almost just for learning, just to meet people in the industry. And that kicked off [21:19] 18 months, almost two years of getting to know the company, which then kind of cracked the door open to investment. [21:24] If you look at my calendar, my calendar looks like meeting a bunch of early stage companies, [21:29] and then a lot of irons in the fire. [21:30] on new initiatives. And then obviously, I spend a lot of time with our portfolio, and internal, and so forth. Is it likely that the next large investment you'll make, you already [21:39] know that company well? Like, are you basically working out of a relatively confined list and you're definitely going to invest off that list? Yes. [21:47] Yeah. I think when people underappreciate about the motion to do this, [21:51] is how much work it takes to go create the opportunity. [21:56] you know we started in new york scrappy having to go get the opportunity i think a lot of the way venture's been built [22:02] is [22:04] You go to Sand Hill Road, you receive the pitch, we need to raise money. Oh, I have money, let me do diligence for a month. And that's how it works. That model to me is not competitive anymore.

[22:12] If you look at Stripe or Databricks, we are going to the company. We are saying, "This is all we know about you. It's a lot, outside in, inside out." [22:21] We were willing to make an investment, but we want to kind of engage with you on these handful of things, and then we go do it. [22:27] That's a very different motion. And by the way, there's no process. We don't wait for a process. We kind of have to wait for our point in time. And sometimes things aren't clear, and so we don't do anything. But in that moment, what am I doing? I'm spending time learning about it. [22:39] learning about what happens with product, customers, people. You get to know people that join the team. You get to know more of the team. So like a lot of times investors only spend time with the CEO. [22:48] CEO is one person in the company. [22:50] Did they spend time with the head of engineering, with the head of product, with the sales leaders, with people lower than that? [22:55] You really learn how clear the flow of information is in a company by meeting a lot of people on the team. [23:01] Are people singing from the same song sheet? Do they understand the decision-making framework? [23:06] The best companies, it's super simple, and it just ripples down, and everybody knows how to act. One of the challenges for a lot of groups, though, is if they can't [23:14] if they're not capable of writing a billion dollar check, these late stage companies aren't going to give them the right to go spend that kind of time. Because there's a small list that can [23:23] The rounds you're talking about, Databricks round, Stripe round, an OpenAI round, you just can't price those rounds. So you don't even get the right to spend that kind of time otherwise. So I guess part of your argument basically here is like the [23:36] scale drives a lot of the competitive advantage. [23:39] Yeah, I want to play in a game on the field that's less competitive. [23:42] I don't know, how many firms do you know that can write a billion dollar check into a company? Three, four, I don't know. It's not that many. So I'm competing with a handful, two handfuls of people.

[23:49] versus if you're trying to write $100 million checks in the growth stage or $10 million checks, you're competing with at least an order of magnitude or two orders of magnitude more [23:57] firms. Maybe that won't last forever and people will get [24:02] you know, capable of doing what we're doing. [24:04] But today it's a more advantaged game in the field, because there's just not that many people that can compete. What's interesting is I think there are a lot of firms in sort of like the [24:13] size rung below, call it one to three billion dollar firms who I've spoken with who I think see [24:19] what you're doing, what Founders Fund's doing, maybe Green Oaks. I don't know what the number of firms would be. [24:25] and want to, but I think it's actually very hard still, even if you're close to it. [24:30] Do you know what the thing is that you think [24:33] What's the dominant thing it takes if you're a leader of a one or two billion dollar firm to get to this? What is the thing preventing [24:41] more people from doing this. [24:45] It's a question we think about. I do think [24:47] while some of this seems... [24:49] easy to kind of articulate as a strategy. As we talked about our journey, [24:54] Part of it has been a long buildup to get to the level of confidence to do that because it's unnatural. [25:00] for a single partner to want to make that big of a bet, because it feels like career risk every time you do it. [25:05] And so just in general, building a culture [25:08] that [25:09] rewards this, I think is not something you can do overnight. Which kind of goes to like, why is it hard for firms to do? [25:14] There's politics. Like, we have a founder-led firm. [25:17] And you can debate, you know, there are pros to that and cons to that. But I think a lot of the pros are what enable us to do this, which is at the end of the day, we have a super small team.

[25:25] We have clear direction. And we can go make these decisive big, bold bets. [25:30] If you're at a firm, [25:31] that's governed by a handful of people. [25:34] and one partner wants to do it, or even one partner in the handle doesn't want to do it, what happens? [25:38] especially if it's viewed as firm, reputational, [25:41] Oh, you write a billion dollars from this company? What happens if it doesn't work? Is it going to tank the firm? Well now all of a sudden you have bureaucracy that creeps into [25:49] decision making. [25:50] And I think it's hard to compete partially just because most firms have that kind of setup. They don't have kind of a singular... [25:57] small team that makes decisions. There's a good segue. I want to talk about Carvana, which was, I thought, one of the most interesting and in the full arc, it was extremely impressive to me. Where basically, you made an investment in Carvana. I think it was public when you initially did it, but you obviously knew the company for a long time. But you make this investment. [26:16] it goes down like a lot. [26:18] and you buy more. [26:19] you hold and it ends up doing great. You distribute shares, you know, and you crushed it. But like, that seemed like a nail biter to me because you're, you know, it was a lot of capital. It was a public company and you know, you're Thrive, you're not public, you know, for the most part, obviously you do it. But I was just like, and you know, and you were, you know, you know, partially proven, but also earlier in your career when you did it and it worked. [26:43] So just talk me through that whole thing. I mean, I think in a lot of ways, this embodies who we are as a firm. [26:49] the journey of Carvana wasn't just random. I think some people say, oh, you guys saw the public markets go down and you went and moved on a company.

[26:59] Just like, it's not how, you know, we're not just looking to go do random things. I first got to know Carbonics when I was at Tiger, one of my closest friends was the guy [27:08] looking at it, spending a lot of time with it. So I got to know it then. When I joined Thrive, it was on a list of things, but it kind of traded to a price in public markets where it didn't make sense. And then the post-COVID kind of tech cycle happened. [27:19] And it was one of the names that went down 50% in six months. And so that kind of-- ears went up. What are the opportunities? Well, the public markets-- [27:27] we're moving faster in the private markets. And so one of the things that's nice about our model is [27:32] I can write a seed stage investment or these big late stage or I can go to the public markets. Well, the public markets move faster than private markets in terms of pricing and so we went and spent time on it. [27:41] But like... [27:42] Going back to the Stripe analogy, [27:45] a lot of people looking at the company [27:47] were kind of grounded in [27:49] "Okay, well the numbers are turning, there's a cycle, [27:52] If you just looked at the product and the team, who have gotten to know for a while, [27:56] It was very clear that companies invested billions of dollars in infrastructure to build out this logistics network that makes it a great company. So it's not just an online listing for cars, it's a logistics company, which is its core advantage. [28:07] The team... [28:08] You know, it's kind of an amazing story. If you look at the LinkedIn profiles of the people of Carvana, [28:12] There are people that have been there their entire career. [28:14] And they've tried to hire Amazon execs and have them come in and it hasn't really worked. And they've kind of found this way to self-develop talent that's super unique. [28:21] And so a lot of things we got excited about were that. And as it gets to scale, it's a business that gets better, it gets bigger. [28:25] More inventory you have, more people that convert, more people that convert, the better economics are, the better economics, the more you invest in that whole flywheel and it spins and spins and spins. And so we made an investment thinking like this is one of these could be generational giant companies. I think we first bought shares. It was like a 10 or 12 billion dollar company.

[28:42] Um, [28:43] And then the used car cycle turned. They actually did an acquisition finance entirely by debt, which was, we don't deal out with leverage in the private markets. And so there was definitely learnings from that. And the combination of that and a cycle and burning money per car sold meant [28:59] the business kind of completely unwound in a short period of time. I mean, the stock went down. [29:03] Like all the way. Like 90%, right? 90 plus percent. So obviously that's not a fun journey to write down, particularly because when you watch it get marked every day. When you do a contrarian investment in the private markets, your friends might give you some crap about it or grief about it. [29:18] or like people you talk to are LPs. [29:20] But when you're investing in a public stock that goes wrong-- Everyone knows every day. Every day you wake up down 5%, down 4%, down 5%, and you have to answer the question, why? Yeah. Which, if that happened in the private markets, [29:30] No way you could-- - No one could do this job. - No, yeah, everyone would sell their shares. - Yeah. - But, you know, I think we really, [29:37] had an advantage mindset is this was the same period of time that we were working with private companies and [29:43] And what were we doing in private companies? [29:45] we were going through the get fit era of private companies. Everyone hired way too much in COVID and now need to restructure their companies or let go of people and right size the P&L. [29:55] And so [29:56] That's what Ernie and Carvana did. [29:58] They basically said, "We can't focus on growth anymore. We have to do a whole 180 shift to focus on profitability." [30:03] So the whole lens you have to look through the company has changed. It's just like if you were operating a company, which you were, Lattice, through that period of time, [30:09] And you said, growth has gone down.

[30:12] Yes, you want the company to keep growing, but during that moment in time, the only thing that matters is getting the trains on track and making sure you're right-sizing the company for what it is. [30:19] So the lens we started evaluating Karvana through was that lens. [30:22] They were making a tremendous amount of progress. [30:24] And so by the end of that year, stock was down a lot. [30:28] But through the dimensions we were evaluating on them, did they have control on the levers of operating the company? [30:32] The answer was unequivocally yes. And they were making progress. Now, it wasn't fully where they thought they could get it. But this little trend line was very good. [30:39] And so we ended up doubling the number of shares we bought at the company. At a very low price. For a fraction of the price. And so there's also this combination that I think is less [30:46] Which by the way, that's a hard move to do. Like, you know, the knife has fallen all the way. Psychologically. That was the moment that I was like, this is crazy. Yeah. Well, I think when you learn, even if something goes down 50%, it can still go down 90%. Yeah. And after it's gone down 95 more is another having. Yeah. Yeah. It's tough. Yeah. [31:01] But I think [31:02] What you realize is [31:04] In private markets, people don't think about the concept of risk-reward that much. [31:08] But in the public markets, everyone talks about it. And the Sharpe ratio is a thing, and you can measure it. And so what we were able to do is double our position size in shares [31:16] for a fraction of the capital. [31:18] And so the risk-adjusted bet we were making on the data points we had were actually very good. And so we were kind of isolating what we were betting on. And in the fullness of time, obviously, it's really worked out. But I think [31:28] It took having [31:29] the ability, one, to operate in a culture where people are used to making big, bold bets. [31:35] And if you're not used to that, I would have got fired, or we definitely would have sold the shares. For sure. [31:39] Then you have to have an environment where people aren't just nitpicking the numbers. Because if all people are doing is nitpicking the numbers, we were going to sell.

[31:46] But we have an environment that's very much grounded in the product. And people believe in the product. I don't know if you ever bought a used car. [31:51] "Use car dealers, [31:52] are literally the canonical example of what you don't want to be in sales. [31:56] And so having a pure, transparent online experience is [31:59] the epitome of what a quality product experience is in this industry. And so people believed in that at Thrive. The number of people that are like, "Oh, I bought a car on Carvana, and we're so excited about it, and sent it to our Slack channel." [32:08] There were a lot. And so I think people believed in the product, and it enabled us to do this. And then in the fullness of time, obviously, [32:15] And it's a good market. I remember at the bottom, I was texting you every day, "There's cars everywhere I go. It's a big town." It's a big market. It's a big market. I just think it's amazing that you were... the buy at the bottom, I think, is what impressed me the most. And the ability... [32:28] Somewhat is you, but mostly I'm going to give credit to the firm to allow you to do that. I think that is hard to have. I think it's a firm thing. I mean, obviously, going through it personally is hard, but I do think the whole firm went through a lot. Our LPs asked about it a lot. Everyone paid a tax. Right. [32:43] for us doing this. And I think in the fullness of time, we've benefited from it. But in the moment, it's really hard. How important is managing conflicts? [32:52] like for Thrive, basically like, you know, you've got to invest in such, you're doing so few with such a large fund. [32:58] that you've got to be investing in big winners with most of your capital. If you pick something earlier, you conflict it out. How do you manage that? Yeah, we take this super seriously. [33:07] Because if you do a series A in a company, you can't go do a series D or E or F in something else. It's part of the calculus. But if you look at the market today,

[33:15] A lot of investors, it's just by the index across the companies that are working. And so they don't care about conflicts. And for us, because we're not doing a lot, we implicitly are making kind of a commitment to the companies that we are all in on your company. Just like a founder's all in on the company, we're going to be all in this investment. No, I'm just thinking it's one thing. If it's open AI or SpaceX, it's one thing. And you can't go invest in Blue Origin, and I feel sad for you for that. Even though it's obviously a good company, but you're in SpaceX. But when you go and invest in the Series A of an ERP or a CRM or whatever, [33:45] Is the category done for you at that point? [33:47] It's not done. [33:48] But [33:49] This is the... People talk about the advantages. I think about the advantages of full stack investing. This is maybe in the... [33:56] disadvantages of it. [33:58] just an early stage investor. [34:00] You look at the seed or the A, maybe the B. If you can't get it there, you never think about it again. "Oh, it's in your anti-portfolio" kind of thing. [34:07] If you're in our shoes, [34:08] It's like, I can invest in the seed, the A, the B, the C, the D, when it's public, who knows when. You can take it private. Exactly. No, I know. But this happened, I mean, we're joking, but this happened in this Thrive Holding strategy that we created. [34:23] Announced. [34:24] What is it? We end up raising a billion dollars. [34:27] in a company. It's not a phone, it's a company. [34:31] Why did we do that? People are like, "Oh, well, you're expanding strategies and you're doing stuff." It was very organic. [34:36] We started looking at accounting AI companies. And one of our partners is very good friends with this guy who's running this accounting role of strategy. And so we started talking to them about AI tools. Why? Because we're just looking at early stage AI companies.

[34:50] We started partnering more with them. We thought their thing was interesting. We wrote a [34:53] growth investment into that company [34:56] the technology started working, our conviction went up, [35:00] And we all of a sudden said, "Okay, well, wait a second. What's the best way to play accounting in AI?" [35:05] Is it to invest in a [35:06] software company early. Or be the accounting firm. [35:10] Because we're taking this kind of scrutinizing lens from all sides of the equation, [35:14] we decided that's the right way to do it. [35:16] Also, the right way to do it is not in our fund structure, because these assets need a lot of time, and you have to do different things than you do in a typical growth fund. And so we raised a dedicated pool of capital. Right. [35:25] towards that strategy. But the genesis of it is very much looking full stack at a company. Which I think is very different. [35:32] There's $800 million of venture investing going into the software tools for accounting. [35:36] Really? Yeah, so we're riding the R&D dollars to the entire industry as the service provider. And accounting's sticky. Are you going to change your accounting provider because they have 10% lower price? No. Probably not. Now, the risk is, going back to the competitive stuff, if these startups are so successful that they radically change the cost curve, that you can take your business. [35:53] and go to ChatGPT. [35:55] and do your accounting [35:56] for 10% of the cost. [35:58] Maybe you'll change. So one risk in our discussion of this is, if the change is so radical, should we be doing it now? [36:03] or should we be waiting? And I think our assessment of accounting is, it was not going to be as radical as it could be, and therefore you actually wanna be the service provider 'cause you'll capture the value. [36:12] Thank you. [36:13] In general, across sort of functions and verticals, do you more fall into the camp? [36:18] at the moment, right now, September 2025, are you more in the camp that like,

[36:24] most jobs are going to be completely overhauled. People are going to have to find new stuff. AI is going to do things end to end in legal, finance, health care, etc., etc. Or are you like, this is just like super sick software and everything's going to get more efficient? [36:38] Do you think about this question at all? - We do. I'm probably more in the latter camp. [36:43] Um... [36:44] I think that's a good thing. [36:45] I just think we're also humans in a lot of places. [36:48] and we want to deal with humans, not with [36:50] There are some places where I actually think we'll prefer to deal with software. Customer support is a good example. [36:57] - You don't ever want to talk to somebody. - But I can call a number and get an instant response from an AI? - And I don't have to wait ever. - That's amazing. And so in some functions like that, I think it'll probably shift towards software. Now, by the way, there's going to be abstractions of that where rather than doing the actual job, you're going to be managing the system. And so those people will get upgraded in what they're doing. [37:16] But in certain things, I think the value a lot of times is with the humans doing it. And so I think making them 10x more efficient is better. Take creative. [37:25] Are we going to automate away all the creative people? Probably not. Or are we going to make them... [37:28] 10 times as productive, so there's 10 times more creative content? I mean, I would guess there's a certain type of creative content that will get automated, and then there's a certain level of it that won't. [37:36] I think it more is like, you lower the floor of entry, so now someone who wasn't talented enough could probably do it. So they can take their culture and lens and now do it. And then you also increase the output of the people that are really good at it. [37:48] I don't think you really-- maybe you eliminate some people on the margin that you don't need, but I think the vast majority is make people a lot more efficient with their time.

[37:54] So you basically so far you've listed, you know, you obviously we all believe in like [37:59] not search, but like, you know, chat. We all believe in support. [38:03] I think Cogen, although we can talk about that a little bit more. Any other areas that you're like quite bullish on that you're like, this is this is working now and is going to go very far? [38:11] I think the thing I'm most bullish on that seems early and broadly the market hasn't acknowledged it yet is life science, like drug development. [38:19] should radically change. On what dimension do you think it will change? Well, the reason we invest-- this company, Isomorphic, started inside of Google. [38:27] Dennis Hospice, who runs Gemini, [38:29] This is his kind of second project. [38:31] or company, and [38:34] He won his Nobel Prize for the work he did on protein structure prediction. [38:38] And so that has now morphed into this company. And the entire objective, well the mission of the company is to cure all disease. It's a pretty big mission. It's one of the only companies in our portfolio that I think about as having the potential size as OpenAI. [38:50] I think it is truly that big. If you have- Yeah, of course, if you could do that, it's huge. Trillions of dollars. [38:56] And then when you go spend time with them and what they're working on, [38:59] The objective is we're going to take an entire wet lab of experimentation and simulate it computationally. [39:05] Well, if you do that, you flip the entire model of how drug development works on its head. [39:09] Rather than having this kind of waterfall like thing-- - You just do it all on a computer. - Yeah, but now the speed of iteration's different, the scale which you can run computer that's different. [39:17] What needs to be true for that to happen? [39:20] Like what needs to be different in the future for that to work? [39:23] a lot of things, which is why it's such an interesting problem to work on as a company. Obviously, there's the AI part of it today, which is how do you take all of this process and embody it in a model with the right data and results and stuff so you can

[39:35] at the end of a computational run, get a good drug. [39:38] That's a very hard problem. But even once you do that, then you have to run it through trials because today it's regulated. So you can't just come and put drugs into people's bodies in the U.S. Which is going to take the same amount of time. Yeah, and so what's another problem we're working on? Regulation. How do you... [39:52] work alongside regulators to change the way the FDA thinks about it, because the top of funnel might go up a lot. [40:00] That's another thing. How do you find people with disease? Yeah. [40:03] How do you look for the right biomarkers? We can use AI for that though. And by the way, there's a huge, [40:08] there's a huge kind of upswing of startups that are working at Daniel X company. Like, let's go scan your body, take your blood, do it in a high class experience. That stuff will become more important as you get... [40:18] better drug development because you're gonna need better biomarkers to target the right people with the right diseases. So this time from [40:24] identification to cure is extremely compressed. That to me is [40:28] Um, [40:30] an area that it feels far away in sci-fi when we talk about it this way. But when you go spend time with these companies, the rate of progress feels very early open AI oriented. How about CodeGen? Maybe one of the things I'm curious about is, can you talk about the stack from [40:49] a cursor subscription down to the middle, like how the 20 bucks moves. [40:55] I mean, CodeGen's so fascinating because [40:58] It's one of these areas where there's been extremely high return on marginal intelligence, to use a fancy word. I like that.

[41:06] And I'm paraphrasing that from one of our brilliant young people that works on the team, Mohit. Yeah. [41:12] But because that's the case, you want the frontier stuff. And so all of it goes through frontier models, or most of it. And so what you have is you have this dynamic where [41:20] the coding companies, which are doing a great job, I think are doing amazing and amazing people. [41:25] But their subscription, they pay a big chunk of it to a model provider. [41:29] and then that model provider [41:30] You know [41:31] Pays a chunk. Pays a chunk to someone running the compute. In Anthropics cases, they're running on Amazon and Google. And then that provider pays money to build the data centers, and ultimately the biggest toll taker there is NVIDIA. And so if you trace this ecosystem. [41:45] And you look at the dollars of profit today, [41:48] I think Stan Druckermiller said this, 120% of the profit in AI is from NVIDIA. Which implicitly means that most companies are losing money and they're the ones making money. [42:00] in this fascinating state where these companies are growing really fast. There's lots of promise and potential. [42:05] But the economic equation is very much [42:08] unknown. And I think they will figure it out. I think there will be big, big winners that come out of it. But today what you have is a lot of dollars that are getting handed around, and there's less discretion. [42:19] on who should get valued appropriately. And really everyone is getting valued at a high multiple on those same dollars. [42:25] To connect back to an earlier part of the discussion where we talked about Stripe and how you were [42:30] Obviously, [42:31] understanding the financials, but that came later and more you were understanding, you know, the product, where the market might shape out, you know, those kinds of things. When you think about let's just take that stack and you're obviously an investor in both Cursor and OpenAI and we talked about a bunch of others. Is it very important for you to think hard about where the money is going to ultimately

[42:51] land or are you more able to just think about, you know, [42:55] Dynamics that have more to do with the product, the adoption, the customer, the founders, etc. I think the answer is both. [43:02] Like I think if you're too qualitative, [43:04] you miss a lot of details that are important. And if you're too quantitative, you miss a lot of the details that are qualitative. So we call this East Coast meets West Coast venture. Where East Coast, I'm surrounded by public investors and people that are very quantitative. West Coast, I think, is highly product and people-oriented. And so we try to sit at that intersection. [43:21] The reason it's important is [43:22] OK, if you're investing something in a billion dollars, [43:25] It's a big price. [43:26] But if it works, we need to believe you can get paid for the risk. So if you have an unknown economic equation, [43:32] and we're investing at those prices. We need to believe it's not a 3 or 4x. We need to believe it's like a 10 or maybe 20x. [43:38] If you're investing in something at 50 billion, [43:41] OK, it's hard to put 20x's on paper or 10x's on paper. [43:44] But that means that the confidence you have to have, or the kind of band of outcomes, you have to have a tighter standard deviation on the variance of the outcome. [43:54] those kind of more certain [43:56] equations versus if you're investing earlier in the curve, you just need to get paid for the upside. The challenge of that philosophy is if a lot of people believe a lot of things can get big, a lot gets funded. And so you live in these periods of time where the economic equation could be distorted because there's a lot of funding in the environment. And Bill Gurley's talked [44:12] you know, at links at this, but [44:14] You can do all you want on paper, but when you're living the strategy in the boardroom and all of your competitors are raising billions of dollars of capital, the economic equation goes out the window because everyone's competing for a theoretically big prize. And so I think right now what we have is a bunch of

[44:29] companies that are operating that environment. And so we have to be grounded in [44:32] where do we think conceptually the value will accrue? And if you look at Cursor, we think distribution matters. And they have amazing distribution with lots of developers, lots of love. [44:41] and people that use it every day. [44:43] That's valuable and amazing team. With OpenAI, they've done a lot of hard work on the model, but it's not just that. They have to scale all the infrastructure to do that. They have to run the inference really efficiently. [44:52] There's a lot of IP in that. There's only, you know, [44:55] Anthropic and OpenAI are basically the two independent scaled model providers. [44:59] and then they're competing against, [45:01] Google and Meta and Tesla, big, big companies. That's hard, and therein lies the value of what they're doing. And so even if you can't know the end state of the financial equation, I think you have to try to telegraph it through [45:15] what the kind of quality of the product and business model is. [45:19] Just to move through a couple of the other areas of AI, how do you feel about, for lack of a better term, vertical-specific AI workspaces? We're both investors in Rogo. There's obviously in Legal, there's Legora and Harvey. There's a bunch for different verticals. It's not the agent doing all the work, but maybe a bit more co-pilot type work. Do you feel bullish on those? [45:45] type of companies? Is it vertical by vertical? [45:48] Do you look at a lot of that? We do spend a decent amount of time looking at it. It's kind of... [45:51] I would say in the fullness of time it's a little ironic that if we're one or two innings into a massive technology wave that we're starting with

[45:59] vertical market software. You think that should come last? Well, if you looked at the last generation of software as a proxy, you start with the big horizontal categories, because they're massive, and then you work your way up. [46:10] kind of product pyramid because [46:12] It's much easier to build the big CRM than it is to build Viva, which is specific for life sciences. Okay, you go to AI. [46:18] and who are some of the early adopters of the technology, [46:21] lawyers and doctors. There are categories that you would not put in the early adopters of technology. Developers? Makes sense. But lawyers and doctors? It's surprising. So I think from that perspective, there's something interesting to study, which is why are these the folks that are adopting early? Do you think it's because they feel like they were a little slow last time around, and so they're a little more front foot this time? I do think that plays into it. Most companies in general don't want to lose out in the internet. You saw that with crypto. How many companies got on the crypto bandwagon? [46:51] They don't want to lose in crypto. Even if it doesn't make sense, they're doing it. The other thing is, a lot of these law firms, just to take one example, they spend all day with startups. They understand the tech. They know it. And law as a... [47:04] modality is highly text-based. And the models are very good at that. And so it makes sense, [47:10] doctors, a lot of it is kind of evidence-driven or publication-driven or diagnosis-driven, which you can ingest that in modalities that are good with LLM. So you could argue a different side of this, which is the reason those are early adopters is... [47:20] product. But I just think it's ironic in the fullness of time that investors are so focused on it so early because it doesn't seem like the biggest markets to go after. But there lies the kind of

[47:30] opportunity, which is we are excited about some of them. We invested in Rogo in financial services. You know, I think the amount of competition in these markets early [47:39] It breaks the heuristics of success historically. Like the old equation of success in a vertical market. [47:44] was you get a lot of market share. Why? Because as people professionalize on you, you take a lot of market share. Like how many competitors are there to Adobe Photoshop? [47:51] or AutoCAD. There aren't. [47:54] How many competitors are there going to be in legal chatbots? [47:57] Today there's a lot. And so now there's a couple of companies that have broken out of the pack and exist. [48:02] But over time, you would hope that it consolidates to a small few. In general, we're in a moment in time where there's probably more competitors than at least I've ever seen in a lot of these categories. I think it's because in that quadrant of looks good and is good, a lot of things look good and might also be good. You have a lot of people flocking there. There's a capital market to back on. To answer your question. [48:26] I think we're selectively excited. Why do we think Rogo was really interested? It starts with a person. You know Gabe, I think he's 12 out of 10. And so we have to get excited about that. The category has properties that I think are interesting. I personally believe... [48:39] that exposing data businesses [48:42] via the chatbot is actually more interesting of a thing to look for than to look for big labor markets. I've been spending time, how do you find data assets that you can now expose that way? Because the models are amazingly good at structuring and looking at and searching across big data. What else would that include? [48:58] I would love to find one in real estate. CoStar is this amazing, almost monopoly-like data asset. Is there someone who can go after that? I think there are other categories, too, that we could talk about. But in finance with Rogo, what's so fascinating is,

[49:14] There's a couple million seats. [49:15] that people can go sell to. If you're competing against ChatGPT, [49:19] Is Chachipiti going to prioritize integrating public market research? I guess the flip is they're willing to pay a lot and they have a lot of money. And Bloomberg got really big. Of course. But the thing is then, if you're Chachipiti, you can't sell them Chachipiti off the shelf. You have to build a custom Chachipiti. Better just go get a billion users. And OpenAni need scale. And so to me, [49:38] that's the advantage that they can do. If they have to compete with Chai Chippity head on at their own game, it's gonna be hard. But if we can go compete on [49:45] and doing spreadsheets really well, and public market research, and exposing all of that data with great sources. [49:50] That'll be an opportunity for us. - Are there any other areas in general, in the application layers or the startup side, that you feel particularly interested in with AI? [50:00] I mean, robotics is the one that [50:02] We didn't talk about it. It's probably next to life sciences, the biggest opportunity. I think the biggest market. It could be the biggest. Today, cars. [50:12] are one of the biggest industries globally. They might be the biggest industry globally. [50:16] which means robots. [50:18] for consumers, should be the biggest market. Because the utility of a robot in your house doing things is way higher than a car. I mean, if everybody in the world had one robot, 10,000 each, it makes a lot of sense. Yeah, exactly. And so to me, that is the biggest market. The question I think right now is, [50:32] you know, are we in 2015 self-driving or are we on the precipice of cracking [50:38] full stack robots. And I think that's what's the hard question to answer right now. And I've heard arguments on both sides. I've heard arguments that,

[50:45] We are not in self-driving. Actually, self-driving is a harder problem [50:49] Because if you make the wrong decision when you're going 60 miles an hour, you kill somebody. [50:53] But if a robot doesn't put the apple in the right spot, or breaks a dish, everyone's going to be okay. And so the problem, while complex, because there's more degrees of freedom, [51:03] There's more complexity in the environment. [51:05] is actually in some sense easier because the risk of failure is lower. I've heard other arguments though that because the components are brittle and there's a lot of hardware that has to get built and also software that has to get built, that's going to take a long time. So I don't know. But that's an area that we are super excited about. We're investing in physical intelligence. I know you know Lockheed and Carol and those guys. I think it's an awesome company. [51:27] Yeah, but it's early and it's an area that we're definitely interested in looking for more. How do you decide putting an incremental $50 or $100 million into one of these companies, versus just, you know what, OpenAI is definitely working, we've got a great relationship, let's just put more there. You could liken it to crypto or something where the right answer looking back might've just been like buy Bitcoin along the way. [51:50] I do think there's some parallels in the buy Bitcoin. The best thing to do is just go long Bitcoin, because at the end of the day, it is kind of the derivative. [52:01] value indication of everything, meaning [52:03] All these AI companies have to use models. OpenAI's the leading model. They're also leading consumer product. Yeah, you could also just say, if you believe in crypto, you're sure Bitcoin's going to... If it doesn't work, Bitcoin's not going to work. If it works, Bitcoin will definitely do the same kind of thing. Being so close to OpenAI, I think this is one of the things that...

[52:17] helps us in a lot of ways and sometimes can hurt us because we kind of know maybe too much but [52:22] Yeah. [52:23] they are well positioned to do a lot of things. The investor community has basically now narrowed OpenAI and said, "Oh, ChatGPT is the next consumer chatbot globally. [52:31] And that's how you should underwrite the company. And that's what it's going to do. I mean, obviously, Sam super well [52:36] I think what he promotes is [52:40] this kind of [52:41] do a lot of small teams betting on new things, and it's a very entrepreneurial culture. [52:46] And so I would be shocked if there are not lots of new products in OpenAI that are massively successful. Well, I mean, as we're talking about a lot of these other categories, [52:56] It's like if they are in fact really big. I mean this was one of the things historically with like, Google. Fang. It was like, it's either not that big of a market and they won't compete, [53:05] or it's a really big market and they'll compete. I think what turned out to be the case in a lot of things in the cloud cycle was they tried to compete and they just didn't do a very good job at it and they lost the market. [53:14] It seems like the companies today are much more, probably including Fang, have woken up. I just feel like the incumbents are much stronger than they used to be. [53:22] I think the incumbents are super strong. And why? We've never had global distribution. [53:27] there wasn't [53:28] every user connected on the internet prior. [53:30] Again, if you think in decades, [53:34] not yours. [53:35] A lot of the capabilities right now are a couple decades old. [53:39] Everyone didn't have access to the internet two decades ago. Not even one decade ago. So now you take today. And so over the next two decades, [53:46] those companies are going to be able to flex some of this distribution and technology they have in a way that

[53:50] will be hard to compete with. And so I think the thing that we absolutely can't have happen [53:55] is that new companies can't break into that. [53:58] I think one of the things that [54:00] you know, is... [54:02] hard is right now everyone's basically ganging up on-- my mental model for you on how the competitive landscape looks like for OpenAI is OpenAI is in a corner, and every big tech [54:13] has a bazooka pointing at them to try to take them down. Because none of those big tech companies want a new big tech company. [54:20] We should all want it to be a competitive, [54:23] fair fight for a new company to break into Mag7. Because that's what our whole ecosystem lives and breathes off of. [54:30] The flip side of this is, [54:31] Yeah. [54:32] going back to this number of $100 billion companies, as the big get bigger, they're just self-reinforcing properties. [54:39] Uber. [54:40] They have a driver network. [54:41] They have millions of consumers in their app in a geography. [54:44] are they able and best off to do food delivery? [54:48] It turns out, yes. If you would have said, five years in Uber's journey, was that their focus? Everyone said no. So now that market has largely coalesced around DoorDash [54:57] So one company did do it without having the ride share, but the rest of the market consolidated around basically an incumbent in a different area. [55:05] I think you'll see the same thing in AI. [55:07] And so [55:08] This is where we have a hard time, which is I think a lot of our bias is [55:11] this isn't a risk-adjusted better return than OpenAI. And therefore, if we're given a decision, we should do it. At the same time, you can't not take risk because of that. And so we are investing in other companies. For sure. I mean, but it's also like, I think one of the things when I think about Founders Fund and why they're so impressive to me is that they made that calculus for a decade into SpaceX and others. And to continually have the discipline to say, it would be really fun to go make a new investment, but if I know these two things

[55:41] more into something that I already own a lot of. I'm like, I find that very impressive. Partially because the patience is brutal. [55:49] I think there's a firm that has the most similar strategy. It's probably Founders Fund. I think they have very similar DNA of making big, bold bets on companies. [55:59] thinking about what matters most for Thrive in the next leg of the journey. And obviously, if I know one thing about Thrive, you're not going to stand still. I've never known Thrive to do the same thing it did last year. [56:10] what matters most sort of either organizationally [56:15] brand-wise, capital-wise, investment-wise, what needs to be true for you guys to get to a place where you look back and in 2025, Thrive feels somehow small? I mean, we talk about this a lot. What worked for the last decade is not going to work for the next decade. And so you have to evolve. [56:30] which I don't think is a given. [56:33] for most firms. [56:35] I think the most important thing for us is we continue to be a place that attracts [56:40] the most talented and ambitious young people. To the end of the day, [56:44] a lot of what we do is [56:46] new, disruptive, and you need the right combination of experience and naivete. And you know, as people get older and older in their careers, I hope I continue to learn, but you learn through young people. [56:56] And if we're not able to attract [56:58] that kind of person to us, [57:01] Will we attract the right companies and founders? Will we attract the right investors? We have people now, you know, Thrive is a product. We are building Thrive as a company with our own products and technology. And so we attract those people that way too.

[57:12] And it's going to be how we attract new strategies to grow the firm. [57:16] And so, [57:17] I do think the single most important thing is we maintain that. [57:21] And then we also have a healthy dose of, you know, we've got to change, evolve or die. [57:25] in this industry. [57:27] Love it. Vince, this was super fun. Thanks for doing it. Thank you so much, Jack.