# Which Businesses Will Survive AI? | Michael Dempsey Empire · 2026-10-19 · 71 min · https://www.youtube.com/watch?v=lR8a8bLI8ak ## Transcript Speaker 1 Nothing said on Empire is a recommendation to buy or sell any investment or product. Okay, everyone. I’m very happy about this. Welcome back to Empire. We have Michael Dempsey, managing partner of Compound. Welcome to the program. Michael Dempsey Thanks for the invitation. Speaker 1 I don’t have notifications for anyone, but I do have a private Twitter list called Brains, where I don’t allow myself to have more than 10 people at any given time, and you’re on the list. Michael Dempsey Wow. That’s pretty big. Speaker 1 We’ll see if you stay on the list after this podcast, but I wanted to invite you. We can talk about cryptocurrencies, but I feel like you—can you maybe explain what you do? I don’t even follow what you’re doing, but I like your tweets. Michael Dempsey I have a close friend who describes investing as sitting and thinking, and sometimes I just sit. I run, or help run, an investment firm. We provide seed-stage venture capital in highly technical areas. We also invest in the public market, both in public shares and crypto tokens, on a long-term and short-term basis. Our company is called Compound. We do this in a thesis-driven way. To put it simply, we spend a lot of time on research. We think about what will happen at the very beginning, and how it will affect markets and technology. Then we try to figure out which futures we believe in and invest in them. ### How Do You Invest Without FOMO? Speaker 1 This is a very rigorous approach. Does everything start with research? Michael Dempsey Yes. Speaker 1 So how do you invest if there’s something like, damn, you know, I think these markets are driven by the fear of missing out, right? So how do you invest in a market like this, where there’s a hot opportunity and all this research has been done in 10 other areas, but none of it has been done in this new area? Do you rush to do your research, or just skip it? Michael Dempsey Usually, you just skip it. We have a pretty narrow field for high-tech companies: machine learning, robotics, healthcare, biotech, and cryptography. Adjacent areas include things like materials science, energy, and space, but unless it carries a lot of scientific or engineering risk, it’s usually not for us. I think what we’ve learned over the last 10 years is that we’re pretty good at analyzing companies as an n-of-1 and saying, “This could be very interesting and very big.” We don’t think we’re good at analyzing five early-stage companies and saying, “This is the best one. This is the best forecast market out of 45 markets.” Speaker 1 Yeah, it’s like AI for lawyers, AI for accounting, all these different types of businesses. We just think, “That’s not our style, and that’s not where we’re getting the highest growth. We just shouldn’t be doing that.” ### Ads (Fidelity Crypto) I think investing is about figuring out the specific areas where you have alpha, if someone wants to use that word, and then continually developing those. Are all the founders of your portfolio companies scientists and engineers? Michael Dempsey Many of them, yes, but not all. Some are more creative or imaginative technologists. Some are doctors, who I think are considered scientists. Sometimes they’re just operators who have novel insights. Speaker 1 Have you been doing all these deep-tech things since 2016? In 2021, I would largely have called it SaaS plus cryptocurrencies. Was it a bubble? Michael Dempsey Yeah, we started in 2016, so we’ve been doing this for a while. My life has been talking to LPs and managing funds. In 2016, we were doing a lot of things related to artificial intelligence, and then we started doing cryptocurrencies as well. Then we changed our LPA and said, “Hey, we want to buy crypto tokens,” and our LPs were like, “Why are you buying things that people were buying drugs with online?” In 2018, our cryptocurrency portfolio decreased by about 95%, and they were asking, “Why, why, why, why did you do cryptocurrency? We told you not to.” In 2020, our cryptocurrency portfolio grew significantly, and they asked, “Why aren’t you doing more crypto?” In 2021, they asked, “Why aren’t you making more crypto? What are these AI robotics companies? Why don’t you have any fintechs, consumer unicorns, or SaaS? What are you doing?” We raised a new fund in 2021. We hadn’t invested for about 16 months, and they were like, “Why aren’t you investing? Everyone is investing now. Now is the time. Don’t have good deal flow? What’s going on?” In 2023, they were saying, “Thank God you didn’t invest in 2021.” Our lives are like sitting in front of the 2 biggest tech waves—or bubbles, if you want to call them that—of the last decade and trying to figure out how to consistently play our game despite all this excessive hype around us. Speaker 1 Do you feel like you’ve been rewarded for sitting idle? You raised money in 2021 and then didn’t invest for another 16 months, which is very difficult, I would say. Do you think you’ve been rewarded? I’m sure in ’23 people said, “Oh my God, thank you. You didn’t cost us money.” But were you rewarded based on the amount of pain you had to endure? Michael Dempsey That’s a good way to put it. I don’t think it was that much pain. The pain was more like an irritation, not real pain. Real pain would be an LP saying, “Hey, buddy, come on.” Speaker 1 Yes. Yes. Michael Dempsey But real pain is losing money. That sucks. I think we didn’t handle it as well as we could have. I will say we made a mistake. We started really investing in March of ’23, after the fund closed in 2021, but we were still in the ramp-up phase. This venture capital fund will have a five-year rollout. I think we weren’t aggressive enough coming out of it. If you look at the 2023 class of companies—and we’ll see, who knows how much of this AI stuff will matter—there are some really interesting things that we didn’t get into as quickly because we weren’t able to get any signal from downstream capital that they were able to make meaningful investments. That’s the worst part of venture capital: you rely on other people to make money to invest after you. Speaker 1 So, in 2021, there’s this thing from Charles Hudson. I think Turner Novak posted it on Twitter, and I even have the quote on my computer. That’s my lesson from 2021: it’s not like people who were sober and restrained in 2021 received a lot of credit from their long-term investors for not participating in the circus. ### Can AI Transform Biotech? I think to myself, is this really what modern venture capital looks like? If there’s a speculative frenzy, the rational solution, if your goal is to remain a venture capitalist, may be to simply participate, even if you’re deeply skeptical about the ultimate outcome. The reason is that you have to stay relevant, stay in the industry, and be able to raise capital. To do that, your company has to be relevant and feel like it gets the message. If that’s true, it takes you in a completely different direction from thinking, “Hey, people who go crazy and lose a lot of money on bad things are going to be punished, and the punishment is that you won’t raise another fund.” Michael Dempsey Unfortunately, I really don’t think it works that way. I think this is partly true. In venture capital, there are institutions that are too big to fail, or institutions where you reach a point of no return. The reason you get to that point of no return is because your LP investors don’t want to look foolish either. When someone from the fund office who runs the books comes in and says, “Listen, this new fund is the best,” they often come in and say, “We’ll invest $20 million in three funds over the next three vintages, and it’ll take, let’s say, a year.” To do that for a year and then say, “Actually, I messed up a bit, and this person is no good,” is also detrimental to their career. Instead, they might say, “Hey, everyone else did it, so they’re just like everyone else who participated in the bubble.” I just don’t operate under the principle that if everyone else is doing it, we should too. It doesn’t make sense, and I think it works both ways. We raise funds every 4 to 5 years. Most venture capital funds effectively operate on 24-month cycles, sometimes faster. In 2021, some funds invested their entire funds within 9 months. You could say all these companies are great, people will fund the next fund, and you can collect fees and make a ton of money. But we can’t say that. We have 5 years, so companies that are like shooting stars, or companies that are great in 2021, will eventually fail in the third or fourth year. And that’s when we start raising funds. So we can’t play a pyramid scheme where we say the next company is the best. We must be able to maintain the portfolio we build. ### Is The AI Boom Built To Last? Speaker 1 Because time is the true arbiter of truth. What’s the Soros quote? It’s like, “I see a bubble forming, and I jump in and add fuel to the fire,” or something. What are you doing? I’m curious about 2021 and how you’ve been doing during this period. Maybe we should talk about 2026. I’m curious what you think about this market—is it crazy, is it a bubble? This should be the best time ever for you because your whole thesis from the last decade is coming true. You’re like, “Robotics, I told you. Biotechnology, I told you.” I’m very curious how you’re doing now. Michael Dempsey Yeah, back to this point, we think a lot about sustainability. A lot of people in the seed sector like to say now, “Well, that’s nice, because as these funds get bigger, they want to own more of these companies and buy more of them.” Speaker 1 In these crazy times, you can sell secondary assets and rake in the cash as a seed investor, right? And that’s part of the whole strategy. Michael Dempsey The truth is, if you make a good seed investment—and we have a few of these—you have hundreds of millions of dollars of capital in these things. So you won’t actually sell out of a position unless you own a small number of assets. Today, those assets would essentially be OpenAI, Anthropic, and Roll. Maybe that’s it. Speaker 1 Hmm. Michael Dempsey Hmm, there might be a few other second-tier assets that could be liquidated, perhaps worth tens of millions of dollars. We think a lot about the idea that we’re going to have to hold on to these things forever, and therefore we have to believe they’re sustainable through different periods of innovation cycles. That means not only massive inference, debt-fueled spending, and everyone trying everything in the AI wave that’s happening today and has been happening over the last few years, but also the next order of things: maybe post-AGI frameworks for how people start thinking about margins for these models, forking their models, and using open-source models. Again, our company was built on some very large seed investments that are now big companies in AI. In particular, there’s a company called Wave in the autonomous-driving space, and another company called Runway, which is built around— Speaker 1 I just watched a video about them and Mercedes, right? Michael Dempsey Yes, yes, yes. We were their first investors, and now they’re multi-million-dollar companies. In 2021, and in this latest fund, we have 3 investments in AI. Speaker 2 Well, your 3 AI investments were quite strange, as was an LP who said, “This is your moment. Why didn’t you earn more?” Of the approximately 23 companies in the venture capital fund, 16 are biotech companies. What were the 3 AI companies? Michael Dempsey One is Prime Intellect. It started out as a distributed-computing engine, and now it does a lot of things for open-source post-training models. Another is the General Intelligence Company. It’s a platform for agents that lets you build solo businesses or businesses with native agents. The third is Orbital Materials, which is focused on building novel materials through a foundation model. Its main product is something called a modular data center. It’s a spinout of DeepMind, and they have a physical lab. In 2022, everyone was asking, “Why do we need a physical lab? We can do everything in silico.” We said, “We need physical data to train it.” I think it works now. The most important thing is that, as early-stage investors who think in terms of a 10-year horizon, we know it’s sometimes not the best time to make those investments. So we try to understand what we think will survive. I think everyone in the AI world feels today that it’s unclear whether any of this stuff will last or be sustainable. As you know, we’ve talked about this: there’s a huge dichotomy. Either everything wins in many different ways, or the labs will own everything and nothing else will win. There are a lot of nuances to that, but I actually think it’s more true than not that the labs have owned—and actually destroyed—a lot of companies in a lot of areas. We had difficulty justifying why some companies would continue to exist, even though we were wrong about the revenue growth of some of them in the short term. ### Will AI Labs Eat Every Startup? Speaker 2 Yes. Can we elaborate on this further? What do you think about it? I told you before we started recording: I was at a founders’ dinner last week, let’s say, and there were a lot of founders there, let’s call them Series B and Series C. The big question in that room is, “Are the labs going to eat us all?” So I’m curious what you think about it. Michael Dempsey All companies, especially high-tech or applied-research organizations, have short-term, medium-term, and long-term moats. I think the problem is that no one can clearly define these moats in the medium and long term. I think the biggest problem with technology today is that people throw up their hands and say, “Everything is changing so quickly. How can you imagine where the world is heading?” We often say internally that the only free thing in building a company is thinking. What you should focus on is understanding how to continually iterate your way toward the ultimate future you believe in. For us, I would say that I’m a moderate believer in the Bitter Lesson: these labs are going to continue scaling compute, data, and performance, and a lot of things are going to break down in the intermediate phase. Then you have to think about which layers will start accumulating value. This could be a product layer. A group of people will go to specific places to get a set of products, which will evolve as models improve. But I think there are a lot of companies that are basically canonical GPT wrappers or something like that, and they don’t deserve to exist. There are also many areas of the market where so much is hidden. A simple example is that everyone is fascinated by inference companies these days. You can look at a lot of these companies, whether it’s Fireworks, Together, Modal, or Foul, and there are a lot of them. They’re all growing very quickly, and they all have slightly different approaches and slightly different parts of the stack. But today, inference is growing, say, 10-fold year over year. If you look at these companies, you’ll think, “Wow, they’re growing so fast. These are amazing companies.” But if the market is growing 10× and one of those companies is growing 6× year over year, you might start to wonder, “Is this company any good if it’s not even above replacement level?” If a company isn’t good above replacement level, eventually, when things calm down, the companies that are great above replacement level will start to win. That means companies that looked amazing on the way up will start to look really bad as the momentum stabilizes. I think you’ll see underperforming assets that don’t look underperforming at all, because the numbers and the growth are so large and so heavily subsidized. Secondly, people are starting to think about where they want to own parts of the value chain. All these companies revolve around the idea that there’s unlimited demand for intelligence. Everyone agrees on that. Everyone says it, and everyone has these general numbers. They say Anthropic will spend a trillion dollars on compute, OpenAI will spend a trillion and a half dollars, and NVIDIA— There’s a famous scene in The Big Short, or rather a quote from the book, where he did something no one else wanted to do: he went and read all the mortgages. I think there’s a perception that if you look at all these moving parts, you start to say, “Okay, obviously consolidation has to happen. Obviously people are going to start understanding more and more of the stack they want to own, because owning more different parts of the stack has its advantages.” You’re going to have software optimization, you’re going to have compute, and what you’re seeing is that all these companies have gone from telling the story that they’re all playing in very specific slots in the ecosystem to, for example, some inference vendors starting to buy GPUs, and some neoclouds starting to work on software that optimizes their inference stack. That’s when you start to enter a completely different game. As an investor, you’re playing a game where you have assets valued at about $30 billion that could be displaced by competition in 9 months if someone decides to acquire a specific part of the stack they own. Or there could be drastic margin erosion, or you could just get squeezed out because you’re working with a limited number of GPUs. We think a lot about all these steps and the long term. I’ll stop there, but the next order of business is: “Okay, in terms of understanding these steps, what are you doing?” You have to understand the entire value chain. You have to understand your very early-stage companies and which ones you can invest in. Then I have to understand what the biggest companies in the world are doing, because the technology market has a very interesting structure where the most innovative companies are actually the biggest. With pharmaceutical companies, it’s the opposite: the biggest companies don’t spend anything on R&D. That’s why we look at the public markets: they allow us to learn a little more about how these companies work, what they do, and which of these 12 neoclouds are the most interesting to own and which ones to short or not hold. ### Which Software Companies Survive AI? Speaker 2 Yes. OpenAI and Anthropic, long or short? Michael Dempsey If we assume each of them is worth about $2 trillion or something like that, I kind of see it as SpaceX. Maybe on a 10-year time horizon it’s interesting; on a 24-month time horizon, I’m like, no, I’m neither. Michael Dempsey I don’t hold it long. Speaker 1 Yes, I ignore it. Yes. Yes, yes. Okay. Salesforce or HubSpot? Michael Dempsey Salesforce probably has a lot more inertia than people think. So, again, I’m not long, but I’m not short. HubSpot, probably a little more short. Speaker 1 Overall, short toward SaaS companies? Michael Dempsey Not all of them. I think there’s a small subset that will actually be very interesting. I published something yesterday. I’ve been doing a lot of work looking at open-source repositories and seeing how many people are starting to create clones of public SaaS companies. I have a list of about 100 public SaaS companies, and then I go through a lot of GitHub repositories, filter them by stars, and make sure they’re real. You start to see a very interesting dynamic where the graph basically goes vertical about 6 months ago, as people just create alternatives and clones. And now that Claude Opus has shown itself to be so good at decompiling software, I think it’ll only get better. Again, enterprises aren’t going to buy random open-source forks of enterprise software, but for many of these companies, there are some pretty negative structural factors. Speaker 1 Yes, interesting. So, which of these stocks do you have long positions in? I saw your figure. Do you have a long position in Figma? Michael Dempsey Yes. Speaker 1 Yes. Let’s talk about Figma. I very rarely tweet about it publicly. I really don’t like this. I went through it before and prepared for it. I looked at my portfolio, and I’m looking at this thinking, “What companies are these?” You invest in a company that, I think, generated the first human eggs derived from stem cells, and you do all these things, and then you’re in Figma for a long time. Michael Dempsey Well, I think what’s funny is that Figma is one of those things I wrote about publicly, because I wanted to see what people thought. I talk to a lot of people in the tech world, and Dylan is a very central figure in technology. Dylan Field is the CEO of Figma. I’d say it’s very much a family business, with San Francisco networks and all that. I’m very curious how the company sees itself in the future. Speaker 1 I heard the same thing. I know the bull and bear cases well. The bull case is really interesting if you believe this idea of designing creativity actually needs to be grounded in some way, and if you believe AI can optimize it. You also have to believe that spending will do what it’s done for all the other AI-based business software: 10 times more usage than anyone would have expected as we move to a usage-based pricing model. Michael Dempsey Figma has a group of products that, in my opinion, took 18 months too long to ship, and now they’re starting to become available. The stock has fallen significantly since its IPO. It’s still not cheap, but Figma isn’t even in the top four. It’s just something I wrote, and now a lot of founders message me every day. Friends say about Figma, “I don’t use it at all anymore. I’m just on Claude Design. It sucks.” And I simply answer, “I know, for sure.” When I finally replied to my tweet because I got so many messages, I thought, “I know, I know. It’s just a facade.” Like, “It’s obvious that agents win, AI is overwhelming, design doesn’t matter. We’re all just slop cannons. This business isn’t cheap, considering all the factors.” Speaker 1 There’s been some management turnover over the last few quarters. Anthropic, of course, had someone on Figma’s board for a long time, and then he left. They launched Claude Design, and that became a huge thing. Michael Dempsey I completely understand that, but I think it’s a spontaneous reaction. Going back to what you said, it’s capitulation by founders in the tech industry who just say, “If this even remotely fits, AI models win. End of story.” And I just don’t like that. I just don’t believe it. Speaker 1 Do you know what? I think the people who are sending you this stuff are 3 months behind because they haven’t used Figma MCP. It’s amazing. Michael Dempsey Well, here we are. We’ve moved everything. We’ve completely moved from Figma. We were 100% Figma 6 months ago. I would say Figma died. We designed everything in Figma, and now we’re completely on Claude Design. You can just feed something into Claude Design, and it just feels like Claude Design. It seems a bit sloppy because of the AI. And now that Figma MCP is out, I don’t know why it took them so long to release it and make it good. Now we’re designing things where my slopometer is off. Speaker 1 Yes. That’s how it is. I think I’m telling you that it is. Here’s your thesis. Look, I think there’s a certain level of romanticism in the way Dylan talks about business, and it’s probably not the right tactic for a SaaS company when an AI lab CEO is on the warpath. But there’s a broader thing, a distant digression I’ll take 30 seconds for. I was telling people that I think we’re at rock bottom when it comes to reading books. I think society is actually going to shift toward reading more and more books over the next 10 to 20 years. Everyone says the same thing: “Everyone uses TikTok, Instagram,” and so on. The biggest question we talk about internally is whether society is on a pendulum or a runaway freight train. Everything is the result of that question. Michael Dempsey I have more of a swing bias. We wrote a letter to LPs last month, and one of the main themes was misery. Everyone in the financial markets is miserable. If you read the quarterly hedge fund letters from last quarter, they all basically start the same way: “Here’s a war story from the 19th century. Here’s how we underperformed the Nasdaq. Here are some charts showing why this is all a bubble. This is why passive investing sucks,” and so on. We have this dichotomy as a company, sitting in these two very strange areas of effectively 12-year investments, while our public fund has a 5-year horizon. We call it “seeing the world as it should be” and “seeing the world as it is.” We think that sitting in those two mental frames keeps us reasonably sharp, hopefully. Speaker 1 Interesting. You believe we’re more of a pendulum? Michael Dempsey I think so. Speaker 1 Interesting. Is technology socially a pendulum, or does it feel like a speeding freight train? ### Why Is Tech Losing Public Trust? Michael Dempsey Socially, I think society pushes technology, and technology pushes society. I think it oscillates. Society swings on a pendulum within the context of how technology pushes it. Speaker 1 Yes, yes. I agree with that. Artificial intelligence from a social perspective—there are lots of conversation topics I’d like to explore. One of them is data centers. Another is a tweet you had, where you said, “I think a few things happened in technology from 2022 to 2026 that created the perfect storm, alongside the breakthrough of the scaling laws and AI efficiency, that led to a pathetic approach and result in the way AI has made its way into the public consciousness.” Michael Dempsey Yes. Disgusting. Let’s see. What was I thinking at that moment? That’s the cool part about going on podcasts and people bringing up tweets. I think the main emphasis was on a more hardcore culture. They moved toward a very experimental, not very serious approach: founder-led communication, the spread of financial nihilism, and the complete abandonment of climate. I think that moment in technology was a rejection. It was that thing with the pendulum, right? Everyone was like, “Okay, we’ve become a little too sensitive about everything. Let’s take a step back.” That’s where it all started. I think Brian Armstrong from Coinbase is a perfect example. The guy started by saying, “We’re not a political organization. This is not what we do.” Around 2021, he tried to get rid of all the woke warriors in his organization. He was the first to do it. Speaker 1 Yes, and everyone said, “That’s amazing.” Michael Dempsey Technology was like us, and I think there came a point where we said we were basically affected by everyone else’s rejection of us. Technology entered this framework of, “Fuck everyone but us.” As a result, we became more closed off and more convinced that we know each other better than anyone else in the world. At that point we also had this idea of founder-led communications, which is fun. Even Lulu backed out of it, and she was kind of the person who promoted it. Lulu has withdrawn from founder-led communications. Speaker 1 I don’t see people opting out of founder-led communications. Unfortunately, they still do it, but I think she realized it’s probably not the best. You know what conversation you had that I thought was amazing? Michael Dempsey What? Speaker 1 Was it Travis? You know Travis? Travis Kalanick did this PR tour. With all this consolidation, like Atoms and CloudKitchens, he actually wrote to my co-founder. Michael Dempsey I said this is peak founder mode. Speaker 1 Yes, yes, yes. In a negative way, not like, “Oh, he’s tough.” It was like, “This is too much.” Michael Dempsey Yes, this is a perfectly optimized PR campaign. Speaker 1 Yes, yes. Michael Dempsey With Lulu behind it, and he’s making hits with David Sacks. I was like, “Oh, this is…” Because if something can’t get better, it has to start spinning. Michael Dempsey I think that’s when the pendulum swings. It was a perfect PR campaign, so that’s when it has to swing back. Speaker 1 It was so heartless. Michael Dempsey Yes, yes, yes. Speaker 1 It was so wrong, so disingenuous. The only thing that was honest was his anger, which is the least interesting thing to be honest about. Michael Dempsey Yes. Anyway, I think all of these things happened, and at the same time we thought, “Besides, maybe we created God,” which was an absurd first-order approach. I think we got into this world that led to the construction of this whole negative view of technology, and this closed view of how people see the world in the context of technology. Then you compound it with what we build and what we spend all our money on, which could destroy your entire livelihood. We will eventually achieve prosperity, but in the meantime, let us tell you that we could put all of you out of work within the next 12 months. If you think about how AI labs might have developed if regulations on scaling all of these things had come into effect a decade earlier, people would have a completely different approach to technology. It’s very utopian, very abundance-oriented, obviously left-wing. There would still be tension between the biggest tech companies and the government, but I think there wouldn’t be as much financial burden and economic reckoning behind it. It was just a perfect storm for everyone but us. Speaker 1 Yes. And how do you think it will end? Michael Dempsey Like a social pendulum. VCs are like, “People don’t like tech people.” Speaker 1 I was just in Cleveland, Ohio, this weekend. I was talking to people downtown while we were watching a baseball game, and the amount of hatred and anger toward AI, and specifically data centers, is a lot more intense than I think you might realize from spending time on Twitter. So what do you think is going on here? Michael Dempsey I think in the short term, the government will play a more radical role than people think, because opposing AI and technology will become very popular among politicians. I don’t know if it will matter in the long run, but I think there will be some volatility. The funniest thing for me is in Wedding Crashers. Sometimes I watch that movie when they’re at dinner. They ask the 2 guys who crash a wedding, “What do you do?” and they say, “We’re venture capitalists from New Hampshire.” When the movie came out, you were wondering, “Who is this venture capitalist from New Hampshire?” People were like, “Yes.” Then they say, “We’re investing in a company that makes sweaters or whatever,” and those shirts and pants, right? When Wedding Crashers came out, nobody knew what a venture capitalist was. Now, if I’m at a friend’s party in Los Angeles, or somewhere with mostly entertainment people, I just say I work in technology. I don’t say anything. Speaker 1 Yes, yes, yes. You can’t be a VC at a party. Michael Dempsey No, absolutely not. Speaker 1 Interesting. Do you think this could change if the companies getting funded became more interesting? VCs are hated because you funded overvalued SaaS, and then AI told society it would use 3× as much electricity and maybe end the world. But if you’re funding something that cures cancer… Michael Dempsey I don’t think so. As someone who funds a lot of things that will hopefully cure cancer, I don’t think people care. I actually think this is a symptom of the wealth gap, and of things accelerating, leaving people behind and leaving them afraid of change. People are incredibly good at dealing with change. In fact, we’re invincible at coping with change, but very weak at processing it emotionally in real time. That’s why, as venture capitalists, we at Compound don’t call ourselves a venture capital firm anywhere, on purpose. We think of ourselves as an investment firm. The idea of venture capital today is that people think of technology as first-class citizens and people as second-class citizens, as collateral damage, or as sharing in the collateral benefits of the technological advances people are promoting. Unless there’s some significant change—something I deeply believe in—and people in the tech industry start to signal in a significant way and actually show some dedication, there’s not going to be positivity. Speaker 1 Where has the wealth gap gone? Michael Dempsey If I’d known this would happen, I would have tried to do more. It’s hard to imagine this improving in the short term. Speaker 1 Right. So, do we have any major social unrest? Michael Dempsey Yes. People would tell you that people are so pacified now that they don’t act. But my hunch, again, if we believe in the pendulum, is that this whole billionaire bunker thing is really stupid. If you read any story, eventually, layers of people just want to kill you, and the guards of your bunker kill you. They think they’re going to have humanoid robots or something. I think it’s very likely there will be more attempts at violence against major tech leaders in the next 5 to 7 years. Speaker 1 One thing I will say about all of this is that I think the most important thing we should focus on is how to address relative mental health. The idea is to look at people in relative terms, in terms of relative outcomes, and if you’re focused on relative outcomes, you’re unhappy. I think that’s the biggest problem we have in society. The quality of life is undoubtedly improving everywhere, and the promise of technology is that you can distribute things to make the quality of life better and better, but we just don’t feel it at all. ### Ads (Fidelity Crypto) Approach cryptocurrency investing with discipline with Fidelity Crypto, a platform built in-house with over a decade of cryptocurrency experience. At Fidelity, you can trade cryptocurrencies and stocks simultaneously with industry-leading security. When you see your accounts in one place, it’s easier to make smarter decisions. Get started at fidelity.com/crypto. Cryptocurrencies are offered by Fidelity Digital Assets and are not insured by the FDIC or SIPC and involve the risk of significant loss. Securities offered by Fidelity Brokerage Services LLC, member NYSE SIPC. ### Gambling, Inequality, And Financial Nihilism Yes, yes. If you were king of America and trying to solve this, financially, there are only 3 big levers. You can cut taxes for a specific group of people, raise taxes for a specific group of people, or print money to give away. So let’s call it UBI, tax the rich, and cut taxes for the poor. Michael Dempsey We’re going far from where I thought we were heading. My 2 things, based on what’s realistic in America, would be these. First of all, I’d say all gambling should be for people over 26, because I think the big problem is that young people are jumping on this nihilism train too early, and they don’t really perceive anything. They never even learn about the concept of folding anything. Speaker 1 18 or 21? Michael Dempsey 18. It depends on what you consider gambling, I guess. I don’t know if you can go to a casino if you’re 21. Maybe. I don’t know. Speaker 1 Yes, of course. Michael Dempsey So I’d say 25 or 26 and up. Secondly, I’d have a very simple, but not-so-small, wealth transfer tax. If you want to die and leave someone a lot of money, a lot of it will be taxed, and there’s no way around it. It would be very simple and transparent. I think those 2 things would create a lot of secondary and tertiary effects that would be quite beneficial. Speaker 1 Let’s talk about hypergambling or hyperfinancialization. What do you think about this in general? Michael Dempsey I hate this. Speaker 1 You hate it? Michael Dempsey Yes. I hate prediction markets. I made my own money in college—I don’t think I paid for tuition, but I paid for living expenses by betting on sports outcomes quantitatively. I loved it. I was fascinated by it. I was never so fascinated that it ruined my life or anything, but I really enjoyed it. Of course, I’ve been trading options since I was 18. I’ve been a very active participant in the markets for a long time. I think we’re so far from where we should be, and I think it’s incredibly destructive for young people. Sometimes I hear Haralabob on the podcast talking about all the money he’s made betting on sports, and then he talks about how terrible it is. There’s this weird situation where you’ve made a ton of money and now you’re climbing the ladder. I understand the idea, but I think it’s just terrible. I think this amount of apathy and lack of desire to create anything is really bad. Creation is one of the most important things for humans, and hypergambling is simply not like that. Speaker 1 By the way, without attacking, are you creating anything? Simply put, you’re just an investor. You invest in public or private markets. Is there a desire to create? That’s what I’m trying to get at, not as an attack. The point is, where do we draw the line? Prediction gambling is definitely bad, you would say? Michael Dempsey Yes. Speaker 1 Prediction markets are probably bad. Speaker 1 Robinhood’s zero-day options are probably bad. Michael Dempsey Yes. Speaker 1 Robinhood options are good. So where do we draw the line? Michael Dempsey I think that part of creation is actually more about human fulfillment. I write a lot, take a lot of photos, and paint poorly. I do a lot of things that are acts of creation, more about how you get the things floating around in your head and your heart out into the world. And of course, children are the ultimate act of creation, I could say. This is kind of a separate thing: I think it’s really important to think of life as something long, in a sense, and to think of it as a craft that you’re constantly working on, with some kind of complex effect on it. The moment you lose that, you lose so many of the fundamental pillars of society. It becomes really bad, and you get the apathy and nihilism that we see. With hypergambling, it doesn’t even bother me that there’s no creation. It’s just so short-term-oriented, and you don’t really put any effort into it in the normal way. Again, you could say being in the Pump.fun trenches is a huge effort, and honestly, I would never want to do it. I can’t do it. I think it’s terrible, but I think it takes a lot of effort. I think it’s kind of lacking in intellect. It’s funny how we can come to terms with this, and then these same people can think that using AI models will destroy our brains and we’ll lose the ability to think. Maybe Pump.fun actually damages your brain 100 times more than using ChatGPT every day. So I still think both are probably not great. Speaker 1 Are you interested in investing in companies that fall into this category? Would you go long on Robinhood or buy the Pump.fun token? Michael Dempsey No. No. Speaker 1 You wouldn’t do that? Michael Dempsey Yes. I don’t. I think Robinhood is—personally, I disagree with how they’ve advertised to their users throughout the history of the company. Speaker 1 Gamified investing? Michael Dempsey Yes. And also how they use push notifications to attract and trigger a range of behaviors. I think this is very bad. ### Has Tech Lost Its Soul? Speaker 1 Interesting. You should see the Figma notifications we receive. Tell me about this idea of romanticism. Michael Dempsey Yes. Speaker 1 In building a company. You mentioned Dylan from Figma. You said he had a romantic streak in building his company. I saw you tweeting about losing the sense of romance and humanity in building a company. Why are you thinking about this? Michael Dempsey I think it’s really important to see life not as a problem to be solved, but as an experience to be felt and gone through. There are a lot of words for this, but “romance” is probably one of the better ones, simply because everyone knows the realm of feelings they’re in. In romance, in its simplest sense, the clearest fact is that it leads to pseudo-rationality. I think this pseudo-rationality for something deeply felt is actually quite good in the world. You have to balance it, right? I’m talking about Dylan from Figma, and I don’t know him personally, but I think it says a lot about the fact that creativity and design matter, and designers matter, and so on. That’s really great, but the problem is that on that same earnings conference call, the next person who speaks up is their CFO, who says that the number of engineering positions has doubled compared to the quarter, and starts talking about these things. You don’t want to feel like companies have a mismatched level of romanticism in their management team, because it causes a lot of problems, especially in founder-led companies. Typically, founders are basically dictators of their organizations. On a life level, without the craziness going on around us, our lives are undoubtedly on a great trajectory, right? We have many amazing things every day. Sometimes we just can’t get out of each other’s way. I think the concept of romanticism eliminates that hyperoptimization. You don’t do something because it’s maximally a first-principles reason to do it. You do it because it creates some feeling that you know is either right or something you want to chase. If everyone had a little more of that, we could have a little more humanity injected into the way we think about these things. I think we’ve lost that, and I think that’s one of the factors that supports your argument. Speaker 1 What if people started building companies that were more interesting or more fun, instead of thinking about how to go from zero to $100 million in ARR in the next 18 months? Michael Dempsey Yes. I think I would change the way I think about it. I would slightly object to that: I think you’re looking at it from a very European romantic perspective, as opposed to the Gilded Age— Speaker 1 I feel like the Gilded Age was just the pinnacle of American business. Michael Dempsey Okay. They actually had a lot of romanticism, but it was a kind of technocratic romanticism instead of European romanticism. They felt an extreme sense of civic duty in building their companies, and I think that’s what’s missing today. Speaker 1 I think that’s true. I don’t think Dylan’s romanticism is right, though. You’ll probably need to change this as your business grows. Michael Dempsey Yes. Well, I think there’s a certain civic duty that some of the Zucks, Sams, and Darios of today’s world are lacking. Speaker 1 ### How Compound Finds Its Next Bets Yes. I think it is—I would say yes, definitely. Yes, yes, yes. Well, how do you find or choose founders? Do people come to you, or are you going to them? Michael Dempsey Seventy percent of our investments have historically been abroad. We spend a lot of time thinking about the world, trying to understand where we think it’s going. In that time, we talk to a lot of people, write a lot, and message a lot of people. Sometimes founders come back after 2, 3, 4, or 5 years. For example, I read the Wave founder’s PhD thesis two and a half years before he started the company, when he was still a student at Cambridge University. His name is Alex Kendall. Then I heard from another friend, who was working on autonomous vehicles at the time, that this guy from Cambridge was starting to take some pretty strange approaches to autonomous vehicles. I checked my email because I knew the name sounded familiar, revived our thread, and said, “Hey, I think you started a company. I’d love to talk.” A lot of what we do has the advantage of operating within a five-year time horizon for funds. The optimizations we make in the moment can be so long-term that we don’t worry too much about whether I made the best use of my time this week, or whether a meeting I went to with some random researcher was the right meeting in October 2026. Should I have postponed it to December, or maybe not gone at all because I really wanted this opportunity? This gives us a sense of freedom, and I think investing is increasingly about limiting information noise and being in a state of flow. A lot of our process is figuring out where the world is going and talking to people. We write, and that drives a lot of inbound activity. We also organize many very specific events, for example, research events on specific topics. But I think the most important thing is that we treat research as a priority in the organization, and everything flows from there. Compound’s existential situation is that we’re an overpaid research firm, and we’re not making a fortune for investors. We just have a lot of great ideas, but we don’t implement them. The hope is that we know enough about it that we can make sure all the resources are directed towards investment. But it’s definitely possible that one day we’ll look up and see a lot of great ideas, and we haven’t made any good investments in them. Speaker 1 What are the most interesting portfolio companies you have today? Michael Dempsey You can’t choose your favorite children. They’re all interesting. Speaker 1 Okay, let me rephrase that. There’s a tweet you had about writing a book. You said someone could write a book, but it’s actually more interesting to see someone write a collection of essays, producing an essay every few months, and to see their thinking develop and the ideas swirling around in their head. So, rather than asking you to pick your favorite child, what ideas are floating around in your head right now, and what are you focusing on? Michael Dempsey From an investment perspective, I’d say there are many ideas about how to shape companies. The company shape I really love is when we think about these long-term types of businesses that could exist. One of the things we’ve seen from Orbital Materials in our portfolio is building new types of foundation models for chemistry or other fields of science. That’s very valuable, but what’s really interesting is what new types of products might emerge. Orbital did it very well with the material they put into the product. The product is a modular data center, but the secret is partly in this material, which they synthesize in the lab based on an artificial intelligence model. It’s a completely new material that absorbs heat at a very, very specific temperature. They can design it to do that, which allows them to dissipate heat most efficiently at the temperature at which graphics processors operate. Michael Dempsey In particular, a certain class of GPUs that will soon be on the market will be able to cool very efficiently. Once you get the hang of that, you can do a lot of other things designed around the set of GPUs they’re in. This is a very interesting product. I don’t think creating material out of nowhere is interesting. That’s great for someone doing science, but not so much for someone who has to invest in companies that make money. The question becomes, “Okay, what other types of materials are there?” Many people have a basic understanding that we have a lot of bottleneck materials in the world, like rare earth elements, uranium, and all these kinds of raw materials. They know these materials are valuable because they’re heavily used, and about 90% of their supply comes from China or somewhere else. Everyone is going crazy about this. But there are other things, too, such as products. We were talking to a company that we didn’t actually invest in, but they were essentially taking a similar approach, less computationally driven, and creating new missiles. They said, “The projectiles are very important. There’s huge demand for them, and we can synthesize the properties and create something that’s essentially one-to-one, but with much cheaper materials and all these other things. So we can scale production.” I have no desire to invest in a missile company. It’s quite an offensive technology, not a defensive one. But this type of business, where new materials create new products, is very interesting to us. It also allows you to have greater control over your fate. I think the biggest problem with investing in high-tech companies is that at some point you become dependent on someone, and that someone can essentially determine the company’s success or failure. Space companies today are highly dependent on launch service providers. Robotics companies have historically been highly dependent on systems integrators or customers to take them from pilots to large commercial deployments. So, as long as we can control the path to market, it’s very interesting. Well, I would say yes. And then there are the areas we’ve spent a lot of time on in biotechnology: drug discovery platforms, new treatments, laboratory automation, and software. That’s probably where we’ve invested the most. At any given time, we have a public database of theses that contains, say, 100 items. We have a private database with a lot of research that we do, and that’s probably another 200 things. We’re still trying to figure out what the right set of ideas is that’s worth investing money in. Based on that, it usually makes sense to invest from T0 today to 24 months into the future. Speaker 1 I think I just realized why your tweets are so interesting to me: you’re semi-pessimistic about this moment. A lot of people are pretty fed up. It’s a big bull market, a generational bull market, and everyone’s making a lot of money. But you seem quite pessimistic—maybe that’s the wrong word. I’d say you’re skeptical about this moment. At the same time, because of the nature of your investments, you’re incredibly optimistic, because you see that one day we can cure cancer and, I don’t know, make human eggs derived from stem cells and things like that. I don’t know if that’s the right categorization, but I think I just always want people to be more nuanced. I think that’s my problem. It’s probably a lifelong problem. Michael Dempsey It may get worse before it gets better. Speaker 1 Yes. Michael Dempsey Yes. I think that’s it. I just want a little more intellectual honesty. Maybe these things work, maybe they don’t. There are different branches of the tech tree and parts of the world we don’t know about. It’s okay to have a system or set of beliefs, but we don’t have to be fanatics about everything. Speaker 1 When a company founder presents a pitch to a venture capitalist, they often shouldn’t hedge their bets. True? Michael Dempsey I actually like it when founders are intellectually honest about risk. For example, when they understand the complexity of their company—and even better, when they understand that complexity compared to historical attempts. It shows a certain level of understanding of building a company as a craft and as a problem to solve, or as a dark forest they’re navigating. They know what that forest looks like; they just probably don’t know all the pockets. They can still say, “I have very high confidence in solving these problems,” but I want to know they understand what they’re getting into. I find it very difficult when people talk about very complex things and act as if they’re inevitable and easy. Speaker 1 I know you don’t want to choose portfolio companies that are your favorites, so I won’t force you to do that. But do you see any ideas now? I keep nitpicking about the first human eggs derived from stem cells, because it’s such a crazy idea. I think people are really interested in investing in things like Ramp because the user interface is really cool or something. I don’t think people even think about it. I think it might have been you, or maybe Sam, someone on your team, who also tweeted that we’ve lost all creativity when it comes to building companies. So maybe we can use this part to inspire some creativity in building your business. What are the most interesting things you see people working on today? Michael Dempsey I keep coming back to the biological stuff, but I’ll try to move away from that. We’re getting to a few simple heuristics: How do we make people happy? How do we make people healthy? How do we make people sustainable? Sustainability is energy and food supply. I think mental health is part of health. Our latest investment concerns psychiatry. It’s basically psychiatry based on artificial intelligence. It’s very clear that if we could increase the number of mental health services in America 100 times, that would be a real benefit. We spent a lot of time in this industry, and we were one of the early investors in Talkspace back in the day because they had this crazy idea of doing therapy through video and text. I think energy issues are very interesting. It’s very crowded there now, so it’s hard for us to say. I have some personal investments there, and we invested in a company called Panthelozo, which generates energy using ocean waves. Another area we’re trying to focus on is biosecurity. It’s more interesting now, and it’s gained traction. We’ve spent a lot of time thinking about how we can make sure the problems of the next generation can be solved. These are very interesting issues, from food supply to animal health to deadly pathogens and so on. In software, I’ve spent a lot of time over the last few years thinking that if we assume this all evolves, there will probably be new forms of businesses using local compute that will free up a lot of creative space. But I’m not smart enough yet to know what that looks like. Speaker 1 You’ve made great investments in biotechnology. What do you think about this? I think there’s a rapidly growing consensus that the labs are going to become biotech companies, or that the labs are going to do something big with biotech. I think that’s probably—I said this on the podcast too—that the most important thing they can do to change public sentiment from antipathy to enthusiasm is do something like, “Anthropic saves people with cancer.” What do you think? It’s a two-part question: What do you think about the labs and their biotechnology strategy? And, from a biotechnology perspective, will this be the best decade ever? Will we cure cancer? Will we find a solution to Alzheimer’s disease and things like that, or will artificial intelligence not have as much of an impact as people like to think? Michael Dempsey I’m optimistic about this. I think it will be—I think your pithy quote to Everett is, “I think 50% of the major diseases that are killing us won’t kill us in about 20 years.” I think you and I will live to be over 120 years old. There are many things we will solve. I don’t know if the AI labs will be directly responsible for this. I think there’s a very clear reason why, narratively, it makes sense to bring biotechnology into the labs today, especially in the context of an IPO. This is probably a much larger market than the software market. Despite SpaceX’s beautiful S-1, where they talk about the whole SaaS thing being the biggest market ever, I think it’s very likely that every single person working in biotech at Anthropic—and maybe even more so in AI—won’t be working there in five years. The reality of what people are learning, the tacit knowledge of biotechnology, is still very real. It’s not something that can be recursively self-improved in the way people think. I think we’re going to have significant regulatory reform around all of these things, which is going to make the next wave of biotech companies very valuable and make many more resources available to solve these problems. There are still some things I find difficult. It’s funny, if you believe it—and I’m not someone who knows enough math—but let’s be honest: how did OpenAI solve math last week? They’ve solved all the evals, they have all this stuff, and everyone looks around and thinks, “Okay, so we’ve solved it.” What does that mean? If we have these things, our understanding of mathematics is probably still orders of magnitude better than our understanding of the brain. There’s a lot of complexity in all of this. If we can arrive at an answer rationally, we still need to know the full trace before allowing someone to inject something into their body that will affect their brain. But I think further regulatory changes are happening. The last downside to this nihilism might be that a lot of people will just say, “Screw it, I’ll do it myself,” and a large number of people will experience negative effects from their biological experiments in the next 10 years. Speaker 1 This is already visible with peptides. People want to take them and do it themselves. They’re pretty safe, right? I have a friend who had a pretty bad reaction to peptides. Michael Dempsey I think they’re pretty safe, but people will keep pushing. It doesn’t end there. Speaker 1 I saw that Eli Lilly has around 1,000 Blackwell GPUs, and they’re building their big LillyPod or something like that with Nvidia. I think it was the CEO of Eli Lilly on Cheeky Pint— Michael Dempsey Yes. Speaker 1 —with John Collison? Michael Dempsey Well, yeah—Collison, John Collison. Speaker 1 Will Eli Lilly win, or will the small biotech company win? Maybe Eli Lilly will be a multi-trillion-dollar company in the future. I mean, it’s a one-trillion-dollar company. Do you have long positions in Eli Lilly? Michael Dempsey Yes. Generally, this isn’t investment advice, but I think there will be a lot more companies that ultimately generate very valuable assets. Then the question is consolidation: Do people simply surrender to the massive infrastructure, economies of scale, expertise in conducting clinical trials, and regulatory apparatus that larger pharmaceutical companies have? Or will they try to act on their own and actually make a revolution? Lilly is very interesting because they’ve also started taking a more direct-to-consumer approach and taking on more of the distribution. It’s a little more difficult in the United States, but if you look at technology companies, you can vaguely see a path where some of these biocompanies can continue to grow and go through multiple inflection points over time. Speaker 1 Definitely. Are they multi-tasking companies or not? Michael Dempsey On the other hand, we have Novo Nordisk, which is trading lower today than the price it was at when they started investing in GLP-1. There’s a lot of volatility, but I think these companies, with a lot of cash and really smart people, are going to grow in a way we didn’t think GCP could—at 50% year over year, to hundreds of billions in revenue. Speaker 1 The cost structure in biotech is interesting because— Michael Dempsey That actually causes a lot of companies to go public quite early. Speaker 1 Yes. At a SaaS company, you might have a billion dollars in revenue before you go public. Biotechnology is— Michael Dempsey Yes, you have 8 people in Jersey, and it’s as if you were a public company. If Phase 2b goes well, it’s like you have a winner, and if not, there’s no company. But this is a public company, so you end up with all these shells. It’s quite an interesting capital structure. Will this change as AI makes drug development cheaper? I think what’s going to change is that AI will make access to medications safer. We’ll know they’re safe to give to people in a shorter time, and this way it’ll be cheaper and there’ll be less regulation. Speaker 1 Yes. Michael Dempsey This will also result in cheaper and faster market dynamics. Secondly, Montana has something called the Right to Try Act, which could be another interesting early regulatory change. If a drug is deeper into clinical trials and you’re in Montana, you can choose to take it. It’s a very interesting libertarian view of the world. Again, you can start getting evidence from human studies quite quickly. We’re very interested in human challenge trials as a concept. We believe they’re one of the most underrated endeavors in biology. In human challenge experiments, we have a drug, we give it to infect you, and I’ll give you that drug and see if it works. We’ll do a blind trial with another group of people and keep them in the center for 2 weeks. We think it can be done in a lot of different areas. Today, it mainly concerns respiratory viruses, but I think the apparatus will change. Most people in biology take the opposite position. It reminds me of very early AI, where we had people like Yann LeCun today, or even worse, Gary Marcus, saying, “These are just predictors of the next token. They don’t really know anything,” and so on. In biotechnology, a group of people think the problem isn’t discovery; it’s all these other things. It’ll never work anyway. It won’t change anything. All you AI experts are wrong. ### Why Still Invest In Crypto? One thing people often get confused about in biotech is that our biggest advantage is that we can sit between hyper-naive, very optimistic people and hyper-blasé, very intelligent and talented people. Our specialty is knowing when to ignore these people. If we hadn’t ignored them, I think we would never have made any of our investments. Speaker 1 Michael, this is actually a podcast about cryptocurrencies, so we should talk about them for 70 minutes. I just realized 70 minutes have passed. With all the things you’re doing in biotech, high tech, new materials, and so on, why are you still investing in cryptocurrencies? Michael Dempsey I think this is one of the most asymmetrically tilted-upward areas. If you think about it long term, and take the societal perspective we talked about, even if society is in some form of pendulum swing, a lot of the principles that underpin cryptocurrencies will survive. Then we just figure out what the use cases and areas are, starting with capital organization and then creating more efficient systems through programming, with smart contracts in a simple form. Speaker 1 What areas should we invest in? Michael Dempsey I think that’s it. We started investing in cryptocurrencies because weird people were really interested in them, and we should figure them out. Speaker 1 I would say so. It was really cool, but also a bit disappointing for Ramp. Michael Dempsey Looking forward, I still think there are a lot of things, like rare diseases, that are among the best uses for cryptocurrencies. Speaker 1 Funding for rare disease research: How do you allow companies that don’t really have any funding to raise capital and introduce amazing solutions? Do you have any thoughts? Shelby has— Michael Dempsey Yes, I looked at all the DCI tokens. Speaker 1 Yes, more investors in HairDAO. Michael Dempsey But yeah, I think those deep questions are super interesting, and obviously I’m a huge fan of Bitcoin. Speaker 1 When you say it’s interesting, you have a private fund and a public fund. Are you buying cryptocurrencies or tokens? What tokens do you have now? Michael Dempsey We don’t disclose all of them, but some of our larger positions are things like Hyperliquid. Historically, we had Helium. We have HairDAO and Kinetic, and probably 13 other crypto assets. Speaker 1 What stage do you think we’re at in cryptocurrency investing? You could call it a four-year cycle, a cycle, or just the phase of crypto we’re currently in. Michael Dempsey I think crypto ventures have been pretty much dead for 2 years now. Speaker 1 Yes. Michael Dempsey I don’t think we know yet why, each time crypto cycles repeat, the bar for reinvesting is higher. What’s cool is that people are starting to believe that one day they won’t be called before Congress if they create something in crypto, which is great. I think that’s the main limiting factor. The bigger problem is actually a technology problem. If you put crypto at the top of the tech industry rankings, it’s somewhere near the very bottom. All the technology has just been absorbed by artificial intelligence. Nobody wants to create anything but AI. Bio is this weird thing where people have a very specific set of skills they’ve spent 8 years honing, but everything else is dead, too. It’s not like we’re prospering everywhere. In 2021, the most interesting thing was that everyone was creating all sorts of things, and then they switched to cryptocurrencies anyway. Now, of course, no one will switch to cryptocurrencies because everyone is dealing with artificial intelligence. So the main issue is, first, how do we attract more talent? Secondly, how do we solve the many problems associated with tokens? Tokenization is great, but it obviously really hurts the long-term incentives that I care about. I think the regulations will sort themselves out. From a liquid-assets perspective, it’s very, very interesting for us right now, especially as we have a lot of fully distributed tokens that are starting to accrue value. You can understand these things fundamentally and narratively, and some of them are starting to work. From a venture capital perspective, unless you’re creating new perp DEXs, there aren’t many people making innovative attempts to hit the target right now. ### Which Crypto Companies Could Go Public? The interesting thing that will happen next year is that there will be a lot of IPOs. We should see a lot of them. If the market develops well, it’s going to be a really interesting market for cryptocurrencies because we now have good companies that own them. So we have real companies generating real cash flow that decided they didn’t want equity; they wanted a token. That’s okay. The market will also have companies listed on the stock exchange. So now, if you’re a public investor and you want exposure to crypto, it’s really something like Coinbase, BitGo, Circle, and Securitize. I’m sure I missed one or two. You could have a few IPOs next year and then over 20 in the next 24 months, let’s call it that. Speaker 1 That’s probably too optimistic. Let me say 10. Michael Dempsey Yes, 10 in the next 24 months. There are many of them, for example, Fireblocks and Chainalysis. They’re not ready for an IPO yet, but they could be. Ledger, too. Again, many of these companies aren’t at this stage yet, but they could be. Speaker 1 Are there any private companies you’d like to see on the public markets? Michael Dempsey I’ve been trying to buy Chainalysis stock for the past year and a half. I’m constantly under pressure by some investor. Speaker 1 Interesting. Why are you trying to buy Chainalysis? Michael Dempsey First of all, they traded at a very, very large discount in the last round. Speaker 1 I’m blowing up my Chainalysis transaction in real time. I wanted to ask why you’re sharing this. At this point, I capitulate. Michael Dempsey Yes. Secondly, I think they’re one of the three reasonable companies that have a right to win if we assume that cryptocurrencies will continue to become more important in the world and cybercrime continues to grow at a very rapid rate. I’ve been in cybersecurity for a very long time. They have some proof points, both in terms of cryptocurrency security and, from my outside understanding, they’ve put a lot of hard work into finding the toughest clients to try to win over. Speaker 1 I think there are two other companies that are directly competing. Did you do a round in TRM Labs, or maybe a round in TRM? Michael Dempsey No, no. Speaker 1 Michael, thanks for this. Michael Dempsey Thank you. It’s great fun. I appreciate it.