E17: Elad Gil on AI, Crypto, and What’s Next

25:43I would go for the product market. I think it's the intersection. Yeah. When you're talking to brilliant people who are pursuing what you think is a, unfortunate idea or a really tough market. What are examples of those ideas or markets where you're like, when a founder is telling you that they're going for it, you're like, I don't know, man, maybe you should do something else that seems like a really brutal, brutal one, but just seems to attract founders' imaginations. Yeah. And I should say, I really care about product market. It's that intersection of like the product and the market, but you know, you can proxy it as market.

26:18you know I'll give you an example and maybe AI changes this one is ed tech in the U.S. and in other markets people will pay for ed tech in the U.S. it's actually very hard to find people who'll pay for it right and so school districts can't afford it teachers can't afford it parents often don't want to pay for it I'm thinking like k-12 software right and so it's it tends it's historically been a very tough market because even if you build something useful there just aren't a lot of people will pay for it and that's why a lot of the ed tech funds that existed 10 years ago don't anymore and why there's a lot of non-profit funding of ed tech and simply because it isn't commercially viable in many cases and there are undoubtedly great counter examples to that in terms of ed tech companies that have become very valuable or useful um but it's a tougher market.

27:09Now, AI may change that where now you're going to have an agent which tutors your kid from the time they're three and gives them personalized learning for their entire life, et cetera. And I think there's very exciting things that are going to happen there. When I look at AI, I think it's the single biggest motive force for global equity in education and healthcare that I've ever seen. And a lot of the calls to regulate it, I think, are not thinking deeply enough about how important this is in terms of health equity or educational equity. Now, does that create viable companies in ed tech? I don't know.

27:44It definitely will in healthcare. But, you know, I think ed tech would be one example where I say, you know what, that's usually a tough market. And why are you so confident in healthcare given sort of the regulatory challenges and how there's been so few, you know, Deca corns or, you know, of healthcare is effectively what we would call services. It's a lot of manual labor to do billing properly. It's a lot of manual labor to reconcile an insurance claim. It's a lot of manual labor to write down physician notes, whatever it may be. And so you just have technologies that seem to be very well suited to healthcare services.

28:25I'm more skeptical about drug development and a few other areas where a lot of money is flooding in right now. And I could be wrong in those areas. but for healthcare services, I think it's actually like pretty useful and I could be wrong, but it just strikes me as that. And you mentioned in another podcast, how in, in, you know, bio in general biotech, we haven't had like a$50 billion company in like 35 years or something, or since the late eighties or something. Do you expect that to change, you know, in the next decade or so, or do you, do you think it's just harder to have massive outcomes, even though, you know, it's such a big percentage of our economy.

29:00Yeah. I mean, it's kind of crazy, right? Healthcare is 20 % of GDP. And then I think biopharma is like 20 % of that. And, you know, the biggest market caps in the biopharma world, including the giant pharma companies, they're now about half a trillion, I think, or maybe there's some that are encouraging on a trillion. And that's largely due to the fact that there's these anti-obesity drugs, the GLP-1 pathway drugs, like a Zempic and a few others. And that's really what's driving the market cap there. And so if you add up all of the global biopharma market cap, all of the pharma companies like Pfizer and Lilly and all the biotechs like Genentech and all the public companies and private companies and all the rest, I think it adds up to like three or four tech companies in terms of the biggest tech companies in the world.

29:45So that's crazy. An entire sector that's 20 % of GDP, at least the drug part of it. I mean, again, it's a 4 % of GDP or whatever you want to call it. That's insane. right and so then you start asking what are the reasons and why is it so hard to start good biotech it has to do with like the founding structures right you are often incubated by a bc instead of the free market kind of determines what happens so bc's open a huge chunk of it up front they hire in a hired gun operator from an amgen or a pfizer or whatever to run it from day one often with an anti-dilution clause in their contract so they can dilute the crap out of it along the way.

30:24They are often built to flip into the arms of other pharma or biotech companies. So they only focus on a handful of small areas, the five areas or whatever it is, you know, cancer and cardiovascular and certain forms of neuro, et cetera. And it's very trend driven by these big pharma companies. And so they determine much of the R &D pipeline of the world. And then you have regulatory constraints that cause these very long lead times for development you know 10 15 years uh and so you just end up with these very screwed up like market structures um and founding models and everything else and the tech equivalent would be imagine and i should say all these all the main companies are really old 50 years 100 years you know these are old companies right so imagine if every um big company in tech was ibm and every company was founded by a set of academics at Stanford or Harvard.

31:18And they didn't actually leave to start the companies. They just kind of spun them out. And then you hired somebody from IBM to run them. And the whole goal of the company was to flip into the arms of IBM or some IBM equivalent. There's five of them say IBM and HP and whatever. You'd have no innovation. You'd have no AI and transformer models. You'd have no iPhone. You'd have no cloud services. You'd have nothing. You just have these eighties era mainframes all talking to each other. Right. And you just be adding onto that. And so that's how biopharma works. Yeah, that's a great overview. I want to segue a little bit into your practice.

31:53You've invested in what seems like dozens of unicorns, and not just unicorns, but some of our biggest and most important companies of this era. You've scaled your individual investing practice, but at any point you could have decided, hey, I'm going to try to build the next Sequoia or the next Andreessen or the next sort of big firm, but you've decided to stay solo. Why is that? I'm sure part of it's personality driven, although you've built companies too, and obviously you can build teams. But is it something about how you thought about what it takes to win in venture? Or how do you sort of think about your strategy as it relates to the broader market?

32:36Yeah, this is a really good moment to announce my new firm. It's called Aladkoya. You heard it here first. I'm trying to take the very best of Sequoia, but make it more solo now. So I think, you know, fundamentally, a lot of my investing started off very organically. And so I started off just as an angel and, you know, I'd help people with their companies and they just asked me if I wanted to invest as part of rounds and things like that. And so it almost happened accidentally and then things got bigger and bigger and then I started doing bigger investments. And then eventually, you know, I ended up in the situation I have now where I can either do things personally or I can do large checks out of a fund or, you know, through some combination.

33:19You know, I think from a future forward perspective, the most important thing to me is just working with the best technologies, the most important founders, the most important companies in the Western world, at least in technology. And I don't know the exact form that'll take in the next five to 10 years. You know, I'm very tech forward and optimism forward and all the rest of it. And so that's, what's important to me. Now there's lots of different forms that that could take. And I've been considering or thinking about those different forms for years. And I think there's some really creative, like interesting things to do.

33:53And so it's possible in a couple of years, I'm doing something radically different. The main thing I want to do is just something that's interesting and useful. And I think that's, that's the real question is what's the most interesting and useful form of all this within a couple of years. Yeah. Um, and you know, like what are examples of things that, that, that could be, or even like things that you find is, is it like incubators? Is it like accelerators or what are kind of either innovations on the model or, or things that, that seem interesting? Yeah, I think there's some innovations on the model.

34:22I think there's some areas of investment that everybody's ignoring right now that I think are super interesting. Oh, like what? Um, well, I mean, if I told you maybe people wouldn't be exactly like two or three areas that are like really exciting that um I may not pursue them but if I decide to pursue them then I'd probably have to hire a couple people to help with them right um because in some cases I may have specialized knowledge but I think that I think there's really cool stuff to be done um over the next five or ten years I'll let you keep your your alpha without uh without having to having to share it but we'll look out for that um you know you mentioned YC earlier I see at a moment in time where it was really kind of like dominating market share at sort of, you know, pre-seed or seed and at special economics.

35:05And the terms got, I think, got slightly worse for them, although still pretty good. But now they're just kind of competing with pre-seed and seed. And it feels like sort of the market is more fragmented. Do you think that there will be any major player or any player that sort of dominates market share in the way that YC did at their peak, if you accept my characterization? Or do you think that, you know, venture is just always going to be fragmented and lots of the good players in a commoditized market? And how do you think about that? Yeah, I think YC will continue to be one of the most important sort of platforms in the startup world.

35:45And I don't think that it was I think it was always the most important sort of early stage platform. And I don't think that's changed. um and i do think perhaps i also think the market was more fragmented before than you did um and that may be a difference in how i view the world um and that doesn't mean it's correct that's just kind of how i perceive it so you know my belief is basically what happened is we had a three to four year period where capital was abundantly available and everybody lost all sorts of control discipline interests whatever and um it also meant a lot of people who were effectively LARPing came into the startup ecosystem and started companies.

36:25And they were kind of doing it for the status of doing a startup or because the money was easy to raise and the secondary was easy to take out, right? You had people having life-changing exit equivalents with companies that weren't working because they're able to sell secondary at every round. And so it shifted the nature of people coming in to venture to start companies. And I think that really shifted the overall entrepreneurial ecosystem in really negative ways. And big chunks of that is going away right now. And my hypothesis, which may be incorrect, is that as that continues to shrink down, the good people will continue to flock to both YC and non-YC options.

37:02But perhaps there'll be a higher concentration everywhere of good people, and that will be within YC and that'll be without YC. I think things got kind of diluted for a couple of years for everyone. I want to articulate what some people describe as the bearish case for venture sort of the next few years, which is the idea that, hey, rates are going to probably stay high for a while. There's going to be a lot of capital that's going to leave the system or a lot of less dollars in venture overall. So it's going to be really hard to fundraise. IPO window is not going to be super strong. M &A is not going to be super strong.

37:34A lot of founders will get

37:40disillusioned with venture after having raised so much money and then having their companies fail. and then just there's a way oversupply of funds that's going to take a while for them to get out of the sort of wipe themselves out. Are you sympathetic with sort of the bearish? Do you agree with that bearish case or anything you'd edit or add or how would you counter that as an asset class? Yeah, I think if anything, like venture should get better as there's fewer funds competing, which is true of almost any market. But we do have a very fundamental ongoing technology shift. And part of that is a shift to the cloud still.

38:18A lot of enterprises are still making that transition. Part of that is the AI shift. Part of that is new forms of vertical software. There's a handful of companies that I think are doing very exciting things in the vertical software areas right now. There's still a lot to do in consumer, as discussed earlier, although obviously that's more saturated. So if anything, I think there's really huge opportunities still. and the hope is we just clear out the larpers right the people who are kind of in it because it seems like a quick easy buck right that kind of happened in crypto a bunch of those people flood crypto and then they went into ai and maybe they'll stay in ai for a bit and leave but um i do think that if you look at it very broadly the biggest shifts in the economy continue to seem to be technology driven and i mean technology in a broad way right it's biotech and it's software and it's various forms of technology maybe some things in energy and then it's regulatory Right.

39:09So probably the two places to make money are or make lots of money in the next 10 years is technology related. Again, broadly stated in terms of technology and then maybe things that take advantage of distortions from different aspects of regulation or alternatively shifts due to regulation. Yeah. Are you speaking of regularly? That reminded me of 8VC, which has sort of a practice of investing in sort of regulated industries and they have strong relationships with the government. So they can sort of help those, and you know, founders fund, of course, as well. Are you bullish on incubations in general, whether it's venture firms, incubate?

39:47I know you've had maybe one or two incubations, you know, over the past couple of years or just, you know, vehicles that specialize in incubations like Atomic and others. What's your perspective on incubations? Yeah, I think the hard part with incubators, and you mentioned a few who are good, right, is that most ideas that anybody have are bad. And there's no reason to think you're going to have better ideas in a small group in an incubator than the entire entrepreneurial ecosystem. Right? You have 100 engineers working on stuff inside your company and there's in your incubator and there's 100 ,000 smart people eventually starting companies over the next 10 years or whatever.

40:28It's just a numbers game, right? And so the question is, what do you have as your core advantage that really differentiates it? And there's three types of differentiation. There's differentiated insights. We're spending all our time in defense, so we know what actually is needed there. That may be an insight. Or, hey, we're spending all our time in healthcare services, so we know what's really needed there. The second is differentiation in terms of network and access. And often that translates into customer acceleration of things. Right. It's the ability to sell effectively into a specific customer type.

40:57So General Catalyst, for example, does a lot of incubations in healthcare. And I think they're strongly supported by the fact that they have this amazing network. Sutter Hill, you know, does very hard technical incubations like Snowflake, but then they're amazing at enterprise sales, right, or B2B sales. And so that's part of their big advantage is the sales side of it, not just the technology development side of it. And then the third is the degree to which controlling investments or controlling companies gives you unfair advantages. And so you see some of the folks who do large scale incubations pull off mergers that nobody else could pull off.

41:34Or they take companies public in certain ways. And so it's almost like you have the financial engineering aspect of it under your belt. And you see that actually quite a bit in biotech. but it also crops up in certain areas of technology and so i think if you have those differentiations in terms of insights in terms of core networks including the ability to sell effectively into specific market verticals and then that controlling ability of companies so you can pull off interesting forms of financial engineering then you can be incredibly successful at incubation but that's not the vast majority of people who do incubations yeah that makes sense I want you to reflect on the solo capitalist trend that you helped, you know, the pioneered sort of, you know, not trying to do so, but people looked at your success and said, hey, maybe this is a model.

42:19On the one hand, I could see the argument of, hey, people trust individuals more than institutions. And so that trend just in general across the board has become truer. And so that would be in favor of sort of, you know, Elad Gale Fund versus, you know, some just sort of brand name that has a lot of turnover. and there's more accountability when you put your name on the thing. At the same time, you could also say, or the counter would be, hey, there's a lot of expertise out there. No individual person can have it. And not just on the investing side, but also on the supporting side, all these big firms have massive platform teams.

42:53And some people were thinking we'd have a barbell where there'd be like the big sort of aggregator firms at sort of the later stages. And then the individuals sort of on the other side of the barbell, sort of the specialists with their own firms. Now you've done both in that you're, you know, you could write small checks, but you could also write really, really big checks. So there are times where you're competing with some of these aggregators, even if that's not what you do all the time. Can you reflect on that? Yeah, you know, it's interesting. I think that the solo capitalist trend is overstated.

43:28You know, fundamentally, if you go back to the 80s, you had people doing solo capitalism like Arthur Rock, who was an investor in Intel and Apple and all these great companies. So it's not like it's a new thing. Maybe the new thing is that people have gone multi-stage. But even then, if you look at it, there's a lot of firms where there's a small number of people who are either clearly in charge and or are driving returns for that firm. And there are some firms where it's really one person who's a decision maker, but it's dressed up as a firm. It's dressed up as an institution. They have five people supporting staff, but really they're the decision maker on everything.

44:04And I don't want to necessarily name them in case these firms don't want to be called out that way, but there's some very large firms that function that way. And so I think they're effectively solo capitalists. They just are branded as an institution. And in some cases, they also tend to do later stage investing. And that's the reason the institutional brand is helpful to them, right? Because early on founders, the earliest stage is often prefer to work with individuals. And then at later stages, they just want another brand name to slap on the thing because the capital is less differentiated, right?

44:35So I feel like it's a little bit overstated in multiple directions. I do think there was a trend where people who normally would have joined a fund have gone off and done their own thing. And I think part of that was capital availability over the last couple of years. I think part of it is, you know, in some cases, we thought they could drive good returns or whatever it may be. And then I think in some cases people just had an independent streak and now some of those same people that i see are considering i think in many cases what to do next do they want to join something do they want to start something else like and so i think we'll see ongoing people becoming solo capitalists and some of them will join firms some of them will start their own firm some of them will join a company some of them will start a company and i just think it's going to be ongoing turnover every five to seven years of cohorts of people which in general is how um silicon valley works right it it tends to be these five to seven year cohorts of which a huge subset end up leaving, right?

45:27Every five to seven years. Yeah. I'm just experiencing perhaps my second, you know, sort of transition. And I'm like, wow, who are all these new people coming into Silicon Valley running stuff? It's fascinating how many cycles there are. And I'll talk to some young people and I'll mention someone like Bill Gurley and they'll be like, who's Bill Gurley? And I'm like, what? It's funny. What name stays, you know, super relevant and most don't, right? You have to reinvent yourself every cycle? Yeah, every five to seven years is a huge shift in terms of the founders that all the other founders look up to.

46:01The companies everybody looks up to. Today, everybody looks up to OpenAI. Five years ago, everybody looked up to Stripe. And I think people still look up to Stripe as an example, right? And then who are the main investors in these networks shifts every couple of years? And the big platforms tend to stick around, but everything else tends to shift around. and so relevancy changes very rapidly in silicon valley and rapidly is five years but five years is not that much time for a generational cycle of something right and so i think people um misunderstand the degree to which these waves happen and which to which there's enormous turnover with each wave why don't you have a platform team or you know recruiting a market all these you know services that venture firms offer to help founders you you think it's kind of overstated or not the way in which you want to add value or how do you think about that No, I mean, I think those things can be hugely valuable.

46:55You know, I've started two companies myself and as an investor, excuse me, as a founder, you know, I think what every founder needs help with early is they need help hiring a few people. They need help for a couple customer introductions. They need help with PR when they first launched the company. and um you know later on they need help hiring executives they need help with mna they need help with a variety of other things so i do think there's a set of services the angels or advisors or other founders who are friends or um sort of peer mentors can really help each other with and so i do think there's all sorts of services that are very useful um and valuable even things like um you know i had somebody join my team recently from ramp this guy shreyan where he was an engine manager at Ramp before.

47:40And one of the things that he's helped put together is just kind of cheap or free access to everything from cloud compute credits from Azure or AWS or others on through to, you know, special deals with Vanta for SOC 2 compliance or, you know, other programs like that. And it's just meant to, again, help make people's money last longer and stretch out and all the rest. And so, you know, as a former founder, I always thought that kind of help was really useful. And if you can save somebody a couple hundred thousand dollars and they've only raised a million, that's very meaningful. Right. So, um, I think the services stuff is useful.

48:17Totally. Um, but in terms of you not investing in it, because it just doesn't, it doesn't, uh, you know, make sense for how you can differentiate. You know, I think, um, I was doing just a lot of it myself in terms of helping founders with it. And I think at some point it's always possible I've decided to add those sorts of things. I think you have to be really thoughtful about how to add it, because if you look at it, I'll give you an example. The most internal recruiting for venture funds aren't that useful. They're very useful if you're hiring executives, but they're actually not useful for ICs, which for early stage companies is most of what you need.

48:51And the reason is you're kind of shopping the same person to the seven different portfolio companies who are all hiring an engineer at the same time. You know, it's kind of like the reason that you bring an in-house recruiter versus use an outsource one, which is working with four companies at once, is you never really get all the best candidates just for yourself. So there's all sorts of like agency and other issues that come up or incentive issues because they want to keep rotating a good candidate to different companies because they want to make sure each company feels like they're helping them.

49:17And that kind of dilutes the help. Maybe the right service would be you have two or three recruiters on staff and they go in and they help staff up a whole company for three months. And then they go on to the next one. Maybe that's the way to do it. Or maybe there's other ways to experiment with services that would be really useful for people. Do you think investors should think about macro, you know, like in terms of whether it's fund size or investment strategy or deployment pacing, or they should just focus on the micro. Hey, are the companies good, etc. Yeah, I think there's two opposing and equally good points of view on it.

49:49What are they? Right. The case for not focusing on macro as well, does a macro really impact the underlying technology shift or regulatory shift or other shift that's really going to drive adoption of this thing right if interest rates were at 10 instead of you know five or six percent does that impact whether generative ai works better not really right it may impact all sorts of other things in terms of buying behavior or you know how much capital companies have to deploy on new technologies but fundamentally it doesn't impact the technology shift and you have to remember in the 70s you had iconic companies founded right in the middle of a very high inflationary high interest rate environment right what were interest rates in the in the late 70s like it's like pretty high 12 percent 15 percent at its peak because it got pretty high right 10 percent um and that's when you had apple founded and microsoft founded and atari founded and um oracle and all these iconic companies got started in a terrible interest rate in macro environment we had stagflation and we had all these things right so should they not have started should bill gates not have started microsoft because of inflation right um so that's one viewpoint.

50:58The other viewpoint is, well, obviously macro impacts things and it means there's less capital availability and maybe people become meeker and they don't want to join a startup. But maybe these things are features, not bugs, right? Maybe some of those things are good. But if you're talking about fund managers, investors, VCs, they probably have to pay a little bit more attention to macro more because it impacts capital availability for them. But in this case, I think, or in this era, that's counterbalanced by the fact that there's a lot of people with lots of money who never had exposure to tech who want to get more exposure.

51:30And so I think we continue to see these waves of capital come into technology and it just doesn't abate. A friend of mine worked in Hollywood for a while and his view was that every time they thought a capital wave was over in Hollywood, a new wave would come in. And those waves actually weren't motivated by financial return. They were motivated by the status of owning a studio or being involved with Hollywood or the glamour or whatever it may be. Technology isn't quite that, right? I mean, none of us are very glamorous people, maybe with a few counter examples, like Jack Dorsey or something is, you know, a very charismatic guy.

52:06But, you know, fundamentally, I think that these people do want to participate in the new economy and they do want to own ownership in the future. And, you know, that's why, you know, I think it's very wise, for example, for a number of sovereign wealth funds from the Gulf to say, okay, we're going to effectively be trading petro assets for technology assets over time by investing through venture. Like that makes a lot of sense to me. Shifting gears, wrapping up a couple of grab bag questions that I've selfishly wanted one of the answer to, because I've been thinking about sort of, you know, LinkedIn competitors for a long time.

52:38You have sort of a pitch for a desire request for startup. That's a web three or sort of LinkedIn with some crypto incentives there. Do you want to briefly unpack it? Sure. I mean, if I think about a natural home for identity, the blockchain actually works incredibly well for that. Again, you have a trusted system, you can fork identity. And people for a while have been talking about either soulbound tokens or ways to have non-transferable NFTs that you can associate with a wallet address where somebody could cryptographically sign that you did work on their behalf. So, for example, I could say, Eric, or this wallet address represents a person who spent four years at, name the company, Stripe, working as an engineer in this capacity.

53:22And here's some of the work they did. And then Stripe could sign it. And then you have a token that is associated with that account that can't be moved that says whoever owns this wallet is actually behind this work. and if you're at the same time working at night on Monero or Zcash or something that you may not want your name directly associated with or a DAO or whatever it is you could have the DAO issue a token to another wallet address saying wallet XYZ contributed to the DAO in these five ways so you can branch identity in ways that still represent what you've really done and then you could build a block explorer on top of that which is effectively LinkedIn to show off if you put in this wallet address here's all the things that person has done That means you never have to do a background check again.

54:02It means you can prove that you have certain things that you've done, but you can do that anonymously. You could do it pseudonymously. You could do it with real identity. So it creates a really powerful mechanism to represent identity and accomplishments and other data effectively in the clear. And so I think that's very powerful and it's very useful and it's useful both in the context of employment, but it's also, I think, going to be very powerful in the context of AI. Because as you move towards an agentive or agent-driven set of AI systems, which is probably many years from now, I actually think there's a lot of stuff to build still for that.

54:40You need some provable form of identity that the agent can point to, to prove its association with you and share certain aspects of your data in an anonymous or reasonably anonymized way, et cetera. And I actually think the blockchain could work very well for that. it doesn't have to be the blockchain, but I'm, you know, it would be an elegant solution. Yeah. The, um, the idea I've thought about in LinkedIn is basically like, if you could figure out a way to get reputation on the internet, LinkedIn doesn't really have that. Like, what do people actually think about Eric Tornberg or actually think about Elad Gil?

55:10Um, and there's two ways I thought about that. One is sort of more of a privatized service, like references as a service where if I, you know, I'm thinking of working together, You call this firm, they do like 50 references on you and then they have that data forever and they just keep reselling it. It doesn't seem like a, it's not a LinkedIn size company, but it seems like a massive or important service. The other is more if you could figure out LinkedIn endorsements done right. But yeah, it was a lot of controversial idea. Yeah, it sounds like you're very interested in the sort of endorsement and reputation system side.

55:44And I think that's really powerful. And then I think in addition, there's the accomplishment slash activity driven side of it. It's like the reference letter versus the resume. And both are useful. And then the question is, what is easy to represent in a non-gameable fashion? And there's all sorts of ways you could game either of them. Thanks so much for coming on a lot. This has been great. All right. Turpentine VC is a podcast from Turpentine, the network behind Moment of Zen and Econ 102. If you liked the episode, please leave a review in the Apple Store or rate us on Spotify. Thank you.