This Week in Startups - Liquidity Summit Talks: Antonio Gracias and Gavin Baker | E1990

Starting point is 00:00:00 Hey, everybody, welcome back to this week in startups. This is Alex and I have a treat for you today. Earlier this year, we hosted our liquidity summit in Napa, California. I flew out. It was absolutely fantastic. And then a couple weeks ago, we shared two of the talks here on the podcast. You all absolutely loved it and asked for more. So who am I to disappoint? I had two more talks from that event for you today, and they are amazing.

Starting point is 00:00:25 This week in startups is brought to you by Squarespace. Turn your idea into a new website. Go to Squarespace.com slash Twist for a free trial. When you're ready to launch, use offer code Twist to save 10% off your first purchase of a website or domain. Command Bar. Seamlessly integrate an AI-powered guide into your software, making navigation intuitive and interactive. Visit commandbar.com slash twist to get a custom live. demo. And Brex. The financial stack founders can bank on. Brex knows cash is king for startups,

Starting point is 00:01:03 so they built a banking experience that takes every dollar further. Get the business bank account trusted by one in three U.S. startups at brex.com slash twist 24. So first up, we're going to hear from Antonio Gracius from Valor Equity Partners. You may know Antonio for his time on the boards of SpaceX and Tesla. But more recently, Valor put a lot of money into the $6 billion X-A-I round. And he sat down with Jason to talk about that and a host of other topics. Valor is a big name. This is a great chat.

Starting point is 00:01:36 Sit back. Enjoy. Antonio Rosses is one of my best friends. And he runs Valor Equity Partners. It's been a VC for, if you consider some, we'll talk about it. But what most people would consider a VC, but a very operational firm that takes very big swings. in some amazing companies you might have heard of. I really wanted him to come here today because he has such a unique perspective on the landscape.

Starting point is 00:01:57 He just literally got off making his largest investment ever as a firm in a company called X. a.I. Elon's large language model AI company. And then you helped with the fundraise, I assume. And this was a very large fundraise, a couple of billion, six billion dollars, I believe. Six billion and maybe counting. And counting, yeah. So this hasn't existed in our industry before.

Starting point is 00:02:24 I'm trying to think of $6 billion raises. It's a very rare occurrence. Take me through what the thesis is of this investment, why you're raising so much capital, and how it's going to be deployed. I know that there's a data center that's been acquired already, and you've got to see it recently. But let's talk about what Elon's trying to do with XAI and why you place that. I think you put in $600 or $800 million. Six or $800 million. Yeah.

Starting point is 00:02:48 Big number. I think it's, the answer to that in pieces, this is the largest series B that I'm aware of. There are companies obviously like SpaceX have raised more money over time. But in a series B, the largest series B I'm aware of because the most of its capital is going to build a data center. And as the way we think about this investment, but I'm speaking for myself and my firm, not for XI or Elon anyone else. It's just my personal opinion in the firm's opinion. There is a, we didn't invest in any model companies, any of the foundational model companies, because we believe they were basic commodities.

Starting point is 00:03:20 The models are commodities, the capacity to train them, the data centers are commodities. The only thing that's not really commodity is the trained data and the reinforcement learning. That's it. So as we looked across the landscape

Starting point is 00:03:33 most of these models, if you weren't inside the big platform companies, you didn't have real prior to data and really great reinforcement learning. And even inside the platform companies, you have some prior data, but you have reinforcement learning, enforcement learning that's being done many ways with human intervention. So the example I'll give you,

Starting point is 00:03:52 you know, why does Bard give you some strange outcomes when you ask questions? Number one, the train data itself is driven by advertising clicks. It's sort of an innovative dilemma here. If you want to have a source of truth on data, it can't be driven by advertising. It's pretty much something else. And then you have human intervention that is biased, so it's human bias. If you believe what we believe that these models are one of the most important technology innovations in human history, more akin to electricity than, say, the internet, then you want to make sure that they end up with the least amount of human bias. And really, they are in view the values of creators.

Starting point is 00:04:30 So if the people that are inserting the bias have a certain set of values, what those are they will be inserted into the system. The X-trained data, we believe, has the least amount of bias in the system because it is a debate. is the modern version of the forum, the Roman forum, right? Where all sides are being heard. Will you like it at or not? That was a lot of controversy about this,

Starting point is 00:04:53 but the reality is, fundamentally, X's data is driven by free speech. You can say it in the street corner, you can say it on X. Now, you may not like that. That's what the Constitution says, right? I have a lawyer by training, and I have a real belief in constitutional free speech. that train data plus community notes makes, we believe, the most effective train data for a system

Starting point is 00:05:18 for an LLM because now you have all sides of debate and the system itself, just like training a kid, can start making its own decisions what the right answer is with the minimum amount of implication of human bias in that system. That's why we believe this is going to win.

Starting point is 00:05:33 Then if you add the other parts of Elon's ecosystem, let's say vision system data from Tesla, obviously communications with Starlink, the brain-compan interface with Neurrelink, all of those parts of the ecosystem don't exist anywhere else. That's why we believe this will be the most valuable company in the space, winter in the space.

Starting point is 00:05:50 And in addition to that, adding the very final piece of this is that the company is reconceiving from first principles what a large data center should look like. The largest data center to date today is about 25 megawatts. They're building one that is much larger

Starting point is 00:06:08 to 100 megawatts, and this is public, in a very dense configuration, and it's playing to exactly what Elon and his engineers do so well. They're rethinking the entire system from first principles and trying to make it cheap or better, faster. And so I've seen this movie at Tesla, SpaceX, other companies, where it's Elon, but also dozens of engineers being led in a mission to create the very best, most effective system possible to allow for the best training possible. fastest training, best training,

Starting point is 00:06:41 along with the, I'd call it the clearest view of truth, whether it's complete truth or not, the train data and the ecosystem altogether. This is why we made the charge investment. What is the product that comes out of this, do you think? Because right now there's a big debate of, hey, and you mentioned it, the LLMs are commodified,

Starting point is 00:06:59 like asking a question to an open source one like LAMA, asking one to a closed system like closed AI, I mean, Open AI, Close AI, closed AI, you know, chat GPT, if you look at the results that are coming out, if they just leapfrog each other every couple of months with each release. And it does seem like there's parity already, even though Open AI had a pretty big lead. So what is the business model going to be in these data centers? Who are the customers and are they going to be paying? Like, is it a competitor to AWS in your mind?

Starting point is 00:07:30 How do you frame in your mental model what this will do? or is it enough to get the data center up and running, get the queries and the reinforcement learning going, and then figure it out later? So I would say yes to both, actually. I mean, the one way we thought about this was, if for some reason it doesn't work, which I think it will work,

Starting point is 00:07:48 the data center innovations alone will be exceptionally viable. Like, just like the asset value of that alone is exceptionally viable. But we have much more than that. Right. So the first part today, any of you are using X, just go by the premium service. I suggest you use it. It's great.

Starting point is 00:08:04 I use it for you is like one of my most enjoyable moments. There you know that. I said, you still all the time. Yeah, we trade our favorite memes back and forth all day. It's pretty funny. Yeah. But the GROC button is there and you can use it. And so in the Brock presentation itself, the day before they made some investment presentations,

Starting point is 00:08:22 the Iranians had launched missiles at Israel. And they asked kind of in real time, hey, what's going on in Iran? And you had the other models say, don't know. We have some nonsense, nonsense, nonsense. And you had GROC say, well, they're just missiles launched Iran. Why? Because it was a real-time source of data. Now, that's very important for many commercial applications.

Starting point is 00:08:42 So as an example, if you go to any trading desk in America today, this is just, this is my opinion, not the company. Sure. If you go to any trade-dust today in America, you will find several screens up. There's a Bloomberg terminal that someone's paying several thousand dollars a day for because they want the best, best information in the world. You know what they also have? Twitter. Yeah. They're all up on Twitter.

Starting point is 00:09:04 If you want the best, most update information possible of any of these all of thems today, forget about the inference speed, which is the speed which it takes a token. But the information right now, you have to use GROC today. And there is a monetization strategy right now, which is a revenue share with GROC on the X app that's occurring today. That'll be the first stage. You know, the next stage will be agents. So, you know, I've looked at a couple companies already that have hardware that are being integrated with whatever LLM you want that provide an agent, they'll be like a, you know,

Starting point is 00:09:35 whether it's seeing your ear or senior desk, whatever, that agent will help you interact the LLM. Again, do you want a source of truth, that something that's close to source of truth that's got the best real time information in the world, or do you want something that's influenced by the advertising clicks from the platform companies? And you just, you choose as a consumer.

Starting point is 00:09:53 You all will choose. Do you care about source of truth, or do you care about advertising? Right. When you sold bullshit or do you want to actually have the right answer? I mean, you guys choose what you want. No one knows your product like you do, but knowing how to build something customers want is only half the battle. You have to sell it to them, and that's where Squarespace comes in.

Starting point is 00:10:12 Squarespace is the e-commerce tool that you need. We all know Squarespace builds beautiful websites, right? I say Squarespace, you see a beautiful, stunning website. But did you know Squarespace payments makes it super simple to earn online, or that Squarespace has the marketing tools you need to reach new customers, and that Squarespace can help you with email, analytics, and even design. Just like you're the best at building your product, Squarespace is the best at websites and e-commerce. There's no doubt. They're the best. They've been at the longest. They make the best product. They're relentless at listening to their customers and making great features so that your business can

Starting point is 00:10:45 grow. So check out Squarespace.com slash twist for a free trial. And when you're ready to launch, go to Squarespace.com slash twist. And they'll give you 10% off your first website or domain purchase. That's Squarespace.com slash twist. You know what? It would be really helpful. Go ahead and tweet or put on LinkedIn. Thank you to Squarespace for supporting. this weekend startups. If you do that, just like five of you do that, they see it. And they're like, yeah, you know, cool people listen to the pod and we should support that pod. They're the longest running partner we've ever had on this weekend startups. And for that, I am so grateful. The next layer, which is quite important, is how do you get to AGI? A machine that lives inside of a

Starting point is 00:11:22 computer is not embodied. And I believe, and I think many people believe, that to actually reach human level intelligence. And I don't mean, look, right now the systems are smarter than humans. Machines could add and subtract and multiply and build bottles faster than me two years ago, five years ago, didn't matter. What does it mean to be a human? What does consciousness mean? Right? What kind of values do you want those systems to have? You get that when you actually have an embodiment of the system. This you will have inside of robotics. Taz already has a robot, the atom system. It's functioning today. It functions today using the vision systems that Tesla built with millions and millions of miles,

Starting point is 00:11:58 a huge unassailable moat, how Tom's driving, if you integrate an LLM with that, you get an embodied system. And that is how I think we achieve an AGI that is more compatible with a benign future for humanity. Ultimately, our brains,

Starting point is 00:12:14 at the center of our brain is Imigula, which is a seat of both compassion and anxiety. We want the machine to think like we do in the sense that it cares about us, the way we care about each other. How do you do that if you can't, in touch. And these robots, when do you think, having seen the experience with Tesla, how long before a robot like that could be in our, you know, in the world from Tesla, humane, from whichever company,

Starting point is 00:12:43 but we would start interacting with a robot at an event like this and it would be refreshing the coffee or, you know, be the valet or carry your bags to your room. Carry your bags to your room. How about that as the that. How many years till a bellhop? It's using my job, carrying your bags to your room. Yes, well, you do it well. I tip well too. You too well, that's true. Look, I retire from the test of board in 2021. So I have, again, the opinion here is my own. I have no information. Sure. Dislamer, disclaimer. Yes, disclaimer, disclaimer. I have had a chance to see the system some months ago. And what I think is going to happen is there are several robots with lots of companies, some small, some large, some very specific use.

Starting point is 00:13:24 The Octus Robot is, it has to be 5-8, weighs about 150 pounds. It's the same size as me. So maybe on a perfect human. Yeah, maybe. Well, certainly average. Certainly average. Sorry. You put it up there for me.

Starting point is 00:13:42 No, they built it to be average. In terms of size, because if it's the average size, then it can navigate the world correctly. Exactly. When will one of those be, you know, a bellhop, do you think? You had to pick a year. So I would say, right, I'd say today it's deployed in a factory, I think, doing simple tasks.

Starting point is 00:14:00 What's going to happen is it'll do simple tasks in the factory, which is a huge advantage for training data, learn that. And if I had to bet, I would say three-ish years, two, three years. I don't think it's five. It could be a bellhop at a hotel. Yes, I think within, I'm not sure that they will go down at the use case, but I think that if I had a bet, I would say inside of five, probably three. Wow.

Starting point is 00:14:24 I bet a dollar. It's within five. So this, I think, leads to some conversations you and I have had about the pace of all this. We've both been with the same age. We've both been in the industry of the same amount of time. And the pace now is very different. Something has fundamentally changed with the pace of innovation. So what do you think?

Starting point is 00:14:43 how do you reconcile what the next 10 years will look like when compared to the last 10? It feels different. It feels way different. Yeah. I mean, the pace of change is definitely change. It's definitely the slope. So the second derivative is very, very high. And the challenge we have as capital allocators, like I think about, as you know, we think

Starting point is 00:15:06 about ourselves not as venture capitalists, but as risk allocators because of how we run our business, how we think about developing asymmetric information over time. the number of opportunities coming in that look amazing are just through the roof. The problem is, it's a little like the internet in this way, in that it's very hard to separate signal from noise. Most of these companies will die. And so, and there might be five working the same thing. You know, if you look at, we look at our strategy and AI is both its infrastructure. So call it data center chips, behind the chip, software and hardware that makes to go faster.

Starting point is 00:15:41 and then verticalized applications that have proprietary data and reinforcement loops. The number of opportunities we are seeing is, I mean, it's mind-boggling. We've had it both software, as you know, just keep track of all this stuff, right? Right.

Starting point is 00:15:56 And to be clear, you've built inside your firm software to analyze companies and opportunities. Yes, yes, and help us triage our top of on a pipeline and then keep track of how we're tracking the information on our companies. can no longer do it with human brains.

Starting point is 00:16:12 We can't scale horizontally fast enough to keep up with all the information coming in. We're using, and we're using our intelligence. So you are correct that the pace of change is dramatically different, of technological change, and the number of companies being created is finally different.

Starting point is 00:16:27 And very importantly, especially I think for the folks here and for you, is that this is the first cycle I've been in, and I've not been in this that's changed, it's 20-plus years. We existed in an industry that was an allegation. They were like, you know, five firms, you can name them.

Starting point is 00:16:45 You can probably throw, I don't know, a rock, not from here, but from San Hill Road. And hit all five of their offices. Yeah. And these are great firms, by the way. We do business all of them. They're amazing. And occasionally one will break in, you know, founders are breaking or someone break in. But something happened with this technological revolution and with COVID that I think has

Starting point is 00:17:05 disrupted that al-Gopoli. The venture oligopoly. Yes. The early stage venture oligopoly. I think that that oligopoly has been disrupted. I think even why comers being disrupted by firms like yours. And because there is so much happening and so many new people into the industry that and they don't have the same brand affiliation that you might have had.

Starting point is 00:17:28 It's kind of like, I have a son who goes with fancy school on East Coast. Got a great brand. It doesn't matter as much today as it mattered three years ago or five years ago. And the same thing's happening. What matters is, are you adding value? And so firms that are, I think our reason our deal flows up so much is, you know, our whole view of the world is, our customer is the company we're investing in. We design our scale group products, our operations, products around serving those customers, our LPs, our partners. That's how we think about it.

Starting point is 00:17:58 I think about it the same way. You had value of companies the same way. And entrepreneurs are getting really smart about this. They're like, yeah, the brand was great, but you know what? Capital is kind of like available. But I really need some of the help. Right. And it's, I can give you a distinct story about this.

Starting point is 00:18:13 We had a founder, Marco and Jonathan, who created a company at Thumbtack. I was happy to be the first investor. I met them when nobody would invest in the company. And they, Jonathan came to me, I don't know, a year ago, maybe six months ago. And he said, hey, you know, I just wanted to let you know how much you helped me at. Can I take you to dinner? How much you helped me with Thumbtack? I said, yeah, that's great.

Starting point is 00:18:30 Thank you. It makes me feel great. And he said, well, I started a new company. I want you to be the first investor. I said, great. I said, what is it? said, Athena. He said, what I really need your help with is picking a venture firm to go with. And I said, well, who you're talking to? And he gave me the list. And it was like all the blue

Starting point is 00:18:46 chips as well as yourself and, you know, a couple of upstarts. I mean, everybody wanted to be on the company. I'm a brown chip, by the way. What's that? I'm a brown chip. Yeah. I got it. I get it. But I said, he's, you know, I said, well, before I told him, you know, that we know each other. And I said, well, what's your ranking, whatever? And he said, well, I just like Antonio and Valor the most. I said, why? I said, just look at what they do for companies. And I said, well, that's the right answer. You know, that's somebody who's going to really help your firm. And he made the right decision and you invested in Athena as well. Tell me with that company, which is doing virtual assistance in the Philippines, matching them. Their ARR is going through the roof.

Starting point is 00:19:25 It's been pretty incredible. Maybe you could tell us about the non-obvious thesis for that company that the public doesn't see yet. Well, I'm going to tell that story, and I'm going to finish your story with the end of how the value add has come around, right? Which is this company has, it started out really as a lifestyle business where the founder wanted to create outsourced EAs to the Philippines in a call center because the thumbtack was a call center, basically. And then realized that the LLM's occurred that, okay, this is a great way to develop train data to build LLN that could be an LLM assistant. like this is the idea of having the virtual assistant they'll get to know you. And it's not that all the A jobs will go away, but if, you know, if there's one EA per one executive, you might have one A for four executives. You might use a couple of these.

Starting point is 00:20:14 I know yourself. I have two. You have two. Servicing four people. Yeah. So it's really great. And they are using an LAM to train the A's to make them more effective over time. And this leverages them dramatically.

Starting point is 00:20:25 But the interesting part of the story is, so Jason kindly did help us invest in this company, they then came back to Jason and said, okay, Jason, you've got a media brand that's very viable. Will you do deal with us in marketing? And Jason started a website that's called Athena Wow. He has this really cool, kind of funny, who's voice is it, man? It's a, yeah. Christopher Walken. Wow. Wow. Wow. It's, I mean. Well, guys. Yeah. Virtual assistant. Yeah, I wish I could be, I wish I had your ability to be funny. But it's just a quite But it's really funny. Actually, it's funny, and it grabs people.

Starting point is 00:21:03 And they turn the thing on, and it blew up the lead-gen system. It's like the fastest growing thing we've seen probably faster in Uber. It can't even process it. So you think about how you add value. Well, what does an early-stage company need? We have a whole system of corporate relationships. We help our relationships get revenue from. That's one of the things they wanted from us that they got from you was actually your brand

Starting point is 00:21:24 and your ability to reach hundreds of thousands or, I don't know, millions of consumers. Millions, yeah. Yeah, millions of consumers. In business. Yeah. Yeah. Who need this product. Yeah.

Starting point is 00:21:32 And so your ability to help them in that co-marking program has been, I mean, I saw the numbers recently. It's like off the charts. They have like two years in demand. They can't feel or like scrambling now to fill the demand. Yeah, they had 1,400 people sign up. Yeah. And they were getting whatever it was, a 15 or 30 people a day, steady state and then boom.

Starting point is 00:21:51 And now they have no inventory. And I think there's a, you know, 100,000 person plus backlog now. Yeah. It's something, it's some insane number. Yeah. Founders, I know a lot of you listening to this podcast, you build software for a living, right? We love doing it, but we all know it's hard. You know the pain of trying to onboard new users and get them up to speak quickly.

Starting point is 00:22:10 Worse, most chatbots and guides built for the task are annoying as heck. Users tune them out because we all hate random pop-ups. Thankfully, there's one company that uses generative AI to help users onboard without annoying them, and it's called Command Bar. It has a chatbot that gives personalized responses to you. user questions instead of a basic Q&A. And it shows users around your product like a live guide, cursor included. Even more, CommandBar can detect when a user needs a nudge, giving them a product hint or a special offer to close that important sale. CommandBar is used

Starting point is 00:22:45 by Unicorns you know, HashiCorp, Gusto, Sixth, Angelist, and others. So here's a simple call to action. You've got to check out this product. Integrate an AI Power Guide into your software so your customers can navigate your product intuitively and quickly. Visit command bar.com slash twist to get a custom live demo. The same thing happened with Uber to a lesser extent. I had a smaller following then, but I got probably the first thousand people for the L.A. Yeah. When Travis launched L.A., I was in L.A., and I tweeted it and da-da-da-da-da.

Starting point is 00:23:16 And the rest of history. But this is, I think, very important to sort of meditate on in this room, which is the field is open right now. there's something that's happened where, you know, we talked about your history a lot privately, but you weren't a darling, to say the least of like the classic endowments when you started raising your funds. You had to go to other sources. You were pretty pioneering in that, especially in the Middle East as well and other sources of capital. You don't need to follow the traditional adventure playbook. Things are different. The game on the field has changed. Yes. Yes, for sure. Look, and there's a lot of people at Calcutta here are. view is just as we want the most viable devout of our companies, we also want investors value to us. What does that mean? That means they're easy to deal with. They commit easily and quickly. If we have a question, they're always willing to help us and they're good people, right? And so as we went out in the world to raise money, we realized that there were certain segments of the population that liked us. You know, they just like, we did very well entrepreneurs, family of people that had

Starting point is 00:24:18 built companies and understood our operational talent. We did a little less well. We have some now, but as you know in the other days, with kind of foundations and downments who were operating in a certain playbook. We did not spin out as some big firm and had a bunch of guys fancy degrees, right? We started out as operators. We built like a connector company. My first real job after Goldman Sachs was running a plating line as a plan manager, right? So this was not a traditional background. It was harder for some of those investors to see that and see the value in that.

Starting point is 00:24:49 That's changing now 20 years later. But in the beginning, you're right. I mean, it was hard to get through some of those systems, but the clients it saw through it and had been loyal to us over the years had been well rewarded. It seems like today it is extremely, extremely cheap to get started and then to win it's capital intensive. And this requires different levels of capital, different levels of focus. Maybe you could talk about how easy it is for founders today to start businesses and then

Starting point is 00:25:21 these inflection points that occur. And what's the proper governance and funding system? Because it seems like there was an anti-governance, anti-k vC vibe that was going around for a little while. People didn't want to have board meetings. We just saw a lot of during peak Zerp noise. Yeah. And it's very hard to do the job when there's noise and actors behaving irrationally. We would see a deal want to be in business with that founder. And they'd say, well, I need an answer today. And we said, well, we haven't even talked to any customers. We haven't done due diligence. And they said, well, these two other folks didn't do due diligence.

Starting point is 00:25:55 I said, I can't really make this investment. I guess I'll talk to you the next round and I'll meet the customers next week and then you can let me know in the next round is. But there was a lot of chaos. And now it seems like the game's very different in terms of the pace. So maybe we could talk a little bit of that. Yeah. Yeah, look, I think in the zero rate environment, there were crossover hedge funds came into the business. and fundamentally in the early stage business,

Starting point is 00:26:22 not the late stage business. They were doing, you know, series A's, series Bs. And this fundamentally changed the speed and pace at which these investors were happening. And on top of that, you had COVID, so people weren't traveling, and they could do meetings over Zoom, which, by the way, I never did.

Starting point is 00:26:36 Like, I refused to write a check. Even during COVID, I would fly somewhere and see somebody. I wouldn't do it. I can't. I'm too old and too old fashion without looking across the table and someone say, hey, I trust you or not.

Starting point is 00:26:47 But why were they doing this? This is very important. And anyone who's a capital allocator here, I'm going to get in trouble saying this to my friends for your probably going to cross-dressions, but this is actually volatility washing. So what happens is if you're running a public portfolio with a private sleeve

Starting point is 00:27:02 and you're being marked up and down every day, your volatility numbers, so your short Tino and sharp ratios, is how we think about paying portfolio managers in addition return. They get better if you add privates that mark less often. It looks like the volatility is lower. So, oh, this was a little game's been,

Starting point is 00:27:18 of putting a foundation and lowering vol. Yes. So when you lower volatility, you, by definition, even with the same return, will increase your sharp ratio and your shortino ratio. That's the way the math works.

Starting point is 00:27:30 This is what drove it. In addition of that, they thought it was easy to write checks in early-stage companies because they all seem to win. Well, I got news for you. They don't. And I mean, look,

Starting point is 00:27:40 I remember having a conversation with one of our friends in the least. And they'd ask me about one of these funds, and I said, look, I wish I was smart enough to hire Bain to do my work. That would be great. I mean, I've got, you know, 30 guys in operations, 30 public investments. I mean, I just feel like a fool because I've got to manage all these people and take care of them and pay them and make sure they have good careers.

Starting point is 00:27:59 Yeah. Boy, it'd be better at Ossesda, Bain. It's great. Turns out it doesn't work so well. Yeah, so management consultants can't do the bespoke work of entrepreneurs and venture capitalists. Yeah, go figure. Go figure. Yeah. I think it's, look, no offense to the guys at Bain and they can do they're very smart, but they're not doing what we're doing. And they're not. not at a risk the way we are. So outsourcing your fiduciary decision and most your research is somebody who doesn't have the same capital risk. That's just an agency problem. You're trying to drive capital through at high speed because of this weird dynamic of what I call volatility washing. Let's talk about the state of, I don't want to get into politics, but the state of the balance sheet in America, we have really over the last two, and this makes it super not partisan,

Starting point is 00:28:45 the last two administrations just really went fuck wild. It's. spending, whatever their pet projects were, COVID, whatever, tax breaks, student loans, but the country feels like it is reckless in how it is managing its finances. And that seems to me to be something that is really going to be harmful in, I don't know, it's the near future, but certainly in our kids' future. How do you think about the debt we've been running up, the interest rates, and what seems like a bit of chaos in how. we're running the enterprise that is the United States and then let alone putting on top of that recklessness, this anti-entrepreneurship, you know, kind of vibes in certain pockets of the culture.

Starting point is 00:29:36 It feels like a very dangerous combination to me. Yeah, so I think this is correct. Let me give you a few additional data points. So the federal data numbers are through the roof. I know those numbers. and they are, it's okay to have debt if you're investing in productive capacity. The problem is here we're investing

Starting point is 00:29:52 in just giving money to people, right? And what happens? In about, in the, call it the mid-70s when women, better the workforce, the employment population ratio, which is the total number of people that can work that are working

Starting point is 00:30:05 was about 63.8%. If you look at the curve, it went up, peaked somewhere in like the 2000s. Yeah, 69%. Yeah, just under 70. Exactly. And that it came back

Starting point is 00:30:17 down during COVID, went to about 62 and change, and now it's back at 63.8%. What does that mean? That means that about half the population came to the workforce, but the number of people working are actually the same as it was before they came in. This means that there's a lot of fewer people that could be working that aren't working. And so I think that, look, the debt is a real problem, okay, but it can be solved. We're not at a place where debt to GDP ratio is like 110%, I think, something like that. We're not like at 140, 150.

Starting point is 00:30:47 We're not like Latin American numbers. So the dollar can definitely be at risk at some point if you're going. But the real problem is we don't have enough productivity. GDP is very, very fundamentally, a function of number of people working, pounded productivity, productivity of capital and productivity of those humans, right? It's driven by capital. If that doesn't go up and we don't have more people working, then we will have a real problem because the dollar will be devalued by a function

Starting point is 00:31:15 of our low productivity levels relative to our competitive trade partners. That's the real problem in my mind. And this is where the culture of, I'd say, anti-entreprenealism you're talking about and anti-work is a problem. And we have to work. What area around the world excite you? If you think about entrepreneurship, you know, we had Ibrahim from Mubadali here. We're talking about UAE, very exciting. What regions make you excited in terms of capital, capitalism, entrepreneurship, and people who want to change the world for the better? So, look, I would say, I'm still a huge fan of America. I mean, I would not move anywhere.

Starting point is 00:31:52 I would not go anywhere. I love this country. I'm raised my children in this country. They're going to work in this country. So I would say still number one is America. You still believe in American exceptionalism. Yeah. I believe in, well, let me actually go double-click on that.

Starting point is 00:32:04 Sure. I believe that we were in a bipolar world. We're now in a multipolar world. And we'll have to come across from the idea that somewhere around Gulf For one, we started exporting American values as opposed to American interest. We need to return to a policy of exporting American interests. Are we exceptional? We're not exceptional?

Starting point is 00:32:24 I don't really care. What matters to me is we have the ability to live well, live freely, and respect other people's cultures. That's what matters, which is where we were prior to Gulf War I. I believe we'll return to that over time. So I wouldn't say it's exceptional necessarily, but it's the American ingenuity. We can have partners that are equals to us. That's okay. No problem.

Starting point is 00:32:43 We just don't have to fight with them. can be equals. The second places I would go in the world, the biggest delta between what is happening on the ground and was being reported in the U.S., I think, is actually in the Middle East. You and I have both gone there and spent time there. It's extraordinary to me

Starting point is 00:32:58 how much these countries are sort of running the Singapore playbook, except they have resources and they have land. So if you think about if Lekwan-U had lots of capital and lots of land,

Starting point is 00:33:15 and a very, very young population that was well educated, which is what you have in these countries in the Middle East. Culturally, they are different than us. However, I think we should respect these cultural differences and find the commonality we have to work together. The second area, which is obvious to people, more obvious than the Middle East, is India. Again, demographics are destiny. You look at the Middle East in India. They have a very high population growth. Let's end on immigration then, because I think it's super important topic. A couple of million people coming to the country every year. They seem to come in illegally, and we don't seem to have any thoughtfulness on either side of the aisle of a strategy or a discussion of who's coming in,

Starting point is 00:33:59 where the number of people coming in, or how those people coming in match to the needs and the interest of the country, country, rather, today and in the future. If you were president, how would you architect our immigration policy for this century? I mean, I would run it exactly a company. I'd say, okay, we want the very smartest people. Let's figure how to find them. Just open up the top of funnel for the very smartest people in the areas we need them and give them H1B visas to come in and turn off the nonsense at the other end.

Starting point is 00:34:34 So we do need, I do believe we need a guest worker program for labor that is 15, 20, an hour. And I think that's probably a smart thing to do in a controlled way was where we were, you know, 20 years ago. The H-1B program is totally broken. We need reformation there. I think it's become a political football between the left and the right, which is not good for America. And I hope it gets resolved soon. But for sure, I would, we would never have a company that just brought people in and hired them without any qualifications. Right. Which is what's happening right now, right? Just walk in and start working. I think we need to reframe it from immigration to recruitment, because companies do recruitment. Yes. And recruiting what we need and the highest quality people would give us

Starting point is 00:35:15 a distinct advantage because if you were to recruit the best of India, the best of Japan, the best of China, the best of Russia, they lose that person. You get a great player on your team. Yeah. The other team loses the player. You win twice. Yes. The advantage doubles. And people want to come in, but we've created this bizarre architecture where you can come to the border, and 80% of, this is not a political issue, 80% of the country's in agreement. We should have some control down there. No, look, I just hope to show my parents and my father actually volunteered for the Vietnam draft to get to this country. To get this country. Yeah. They get a, they get a visa, to get a, a Greek card. And I'm sure, you know, your grandparents,

Starting point is 00:35:54 your parents, someone did something similar. It's hard. And I still believe that if we allow, I created a system where the best people could come, they would come. We'd have recruit them. We just open up and have, you know, however you want to qualify, then they come. Yeah, I mean, Canada, New Zealand, Australia, all point-based systems. Like, it's very obvious that this needs to shift to a recruitment-based one. What gets you excited these days, just generally speaking? I mean, I flew here to talk to you. That was a pretty, yeah.

Starting point is 00:36:22 It's been pretty exciting for me. Yeah, even wore my collar boots. I'm pretty excited about that. Look, this is the most exciting time of my career. It is, it really is a moment where I feel like there is, you know, we have this investment these, which is called proentropic right. Lots of chaos going through the world and we want to invest companies with the world better.

Starting point is 00:36:41 These two things are more important than ever. It's investing companies made the world better that are good at helping us manage through the chaos that's going to come. And the chaos is getting worse, not better. So it's very exciting. And the number of young people I'm dealing with daily, both inside the firm, outside the firm,

Starting point is 00:36:57 the systems we're building, scale which are operating, all this really does excite me. And it excites me partially because it is scary. Like, it's not just exciting because it's all benign. It might be benign if we make it benign, but it could also be dystopian if we don't really work to make it a great outcome.

Starting point is 00:37:18 And I think we, and the people in this room, we have the opportunity to do this way. We live in a system in America, and I'll moralize for a moment, which is in many ways it's ethical, but we drifted toward being amoral. We think about what the legal thing to do is,

Starting point is 00:37:33 not always the right thing to do. And I think this is changing. I think people are actually thinking more about not just, hey, is it legal, but is it right? And if I can lead with one thought, it's just think about that as you analyze managers, analyze people you work with. I know Jason, you look for it that way. It's not just man, is this inside the law, but should I actually do this? It's the right place to put my energy, my life's energy, and that on my team on this company to see succeed. Yeah, and we make mistakes, and we change the world and everything in between,

Starting point is 00:38:01 and it's just such great work, and it's great to work with you as a partner. Everybody give it up for Antonio Gras. Thank you. Okay, most of us in the startup game have heard of Brex. Not only have they perfected corporate cards for startups, they've also built the financial stack that founders can bank on. Nearly 40% of startups fail due to running out of cash. So, to help founders, Brex has built a banking experience that takes every dollar further.

Starting point is 00:38:27 It can help protect your cash and extend your startups runway. How? By combining the best things about checking, treasury, and FDIC insurance into one powerhouse account. Brex's banking solution offers 20 times the standard FDIC protection through program banks, industry leading yield on your cash, and the ability to send money worldwide at lightning speed. So, here's your call to action. Brex is used by one in every three U.S. startups. Amazing. Get the financial stack founders trust at brex.com slash twist 24. At brex.com slash twist 24. I hope you enjoy that as much as I did. I always love hearing what Antonio has to say.

Starting point is 00:39:07 Now, next up we have Gavin Baker. He's the chief investment officer over at a tradies management, spent nearly two decades at Fidelity, running their OTC portfolio, and also helping lead their venture capital activity. He talks to Jason about investing, humility, and which big tech companies are moving the fast and doing the best work in AI. Enjoy. I try not to do many fireside chats here. I like to make everybody sing for their supper and really put a deck together.

Starting point is 00:39:34 But last year, Gavin Baker, come on. Gavin, from a treatise, sang for his supper. And most people said, you, when I, I asked people, who's your favorite speaker? Good to see you. Appreciate you coming out for me. And Gavin and I have become, you know, colleagues and then friends over the last couple years. And I said, hey, you know, I have this event. Probably not worth your time. But I would love to, you know, like, introduce you to, you know, people in my orbit. And he came last year and he gave a talk. And almost universally, people said that was the best talk of the event. But it was, you were a lot of good talks last year and even this year as well. And I said, you know,

Starting point is 00:40:10 this year, just come back for the fireside. You don't have to do any work. If you would be willing to come back. And he said, Jake L, for you, I support everything you do because I see how relentlessly supportive you are of Antonio and our friends. So thank you for coming back. And we'll do a fireside chat here. We just get to wrap out and you didn't have to do any prep. By the way, I didn't prep last year. I didn't have any slides. You did have any slides. I don't. If so, that's used to me. You didn't. I had no slides. You just spoke extemporaneously. Yeah, I have a strict, no preparation ever for anything rule. Okay.

Starting point is 00:40:43 I did just want to reflect on some of the things Megan said. I thought that was one of the best conversations I've ever heard. That was amazing. I don't know if she's still there. The lights are quite bright. She's one of my favorite Twitter followers. I learned from her weekly email. And I am rarely starstruck when I am introduced to someone.

Starting point is 00:41:03 And Brad is a good friend. Yeah. But Brad introduced me to Ben. I had in Miami, maybe, I don't know, three months ago. And I was actually starstruck. So Megan, that was great to meet you. They were very busy. They had to run off.

Starting point is 00:41:16 It was great to meet Megan. Also love Brad. And I loved her talk. And I would just say a few other things because it feels like maybe there's some people in the audience who are trying to raise funds, like just really quick riffs and reflections on what she said. The first thing is, is it all feels so personal. Like I know, I'm sure it feels personal to you.

Starting point is 00:41:36 and it still feels so personal to me, but it's just not personal for the people on the other side of the table. You know, it just may be like they love you, they don't have an opportunity, they really like you, they don't do. They've got a lot of other things going on and just recognize it's not personal,

Starting point is 00:41:55 and that really helped me because I still find it takes a big emotional toll because the rate of rejection, you know, is incredibly high. You know, I would say like 99, probably everybody in this room, you succeed at most things you do. And probably a lot of people were good students.

Starting point is 00:42:16 And it's like, there's like a 1% success rate raising money. And so the 99% of the times you get rejected, they take a big toll on you and just understanding that it's not personal was really helpful to me. I would also say, just not that into you. Maybe they're just not that into me.

Starting point is 00:42:38 No, no. I mean, to your point, I think the other side of the table has their own multitudes of issues they are dealing with. And I had some people pull me aside. This is when I actually took this to heart. Because to be totally honest, I am really not used to any kind of rejection. It never happens. Like, it's a very strange thing.

Starting point is 00:42:59 I particularly understand it after seeing how good you looked in white last night, Jason. Thank you, sir. No, I'm not even talking about romantically or anything like that. I'm a happily married man, but I'm talking about just in, generally speaking, I am so used to making friends or get, you know, in a seat space. Like, you know, there's always room for me in a deal. I have people who reopen rounds. I've had people reopen their rounds six months ever. To then go meet the LP community.

Starting point is 00:43:26 And then this is when I realized, it took two or three of these for me to realize, oh, this is not about you, Jago. When two or three different LPs said to me after the call, they did a follow-up. And they said, hey, I really enjoyed our call. I just wanted to talk to you real quick. And I said, yeah, I don't know how long I'm going to be here exactly. So what's it? I mean, venture, like, what's it like on that side of the table? And I was like, you're looking for a job?

Starting point is 00:43:52 It's like, well, I'm considering some options and critics. I'm like, oh, I'm meeting with people who their situation is so acute that they're expecting to be fired or laid off. 100% on their side. So that changed my whole dynamic. Like, oh my God, they're taking the meetings because they have to take meetings. That's the job. They have to take 200 meetings to make one investment. Just like we take 200 meetings to get one.

Starting point is 00:44:15 To get one investment. Oh, this is just what founders go through every day. And once I had that empathy and I was talking to a friend who was raising his fun and he was catatonic that he couldn't get any yeses. And I said to him, now I guess we know what it's like to be founders. 100% and it's actually I was always surprised before I did this I generally show up to meetings

Starting point is 00:44:41 with company CEOs or founders knowing their name knowing the name of the company having done some work on them having thought about some questions prepared mind yeah prepared mind to quote our friend Brad who loves to say that

Starting point is 00:44:54 and I was just always shocked by how much it seemed to mean to the founders but it's like sometimes you'll go to a meeting you know with a Let me not speak ill, but just, it is, you know, it's not a, every once in a while, every once in a while, you get, who are you and why am I meeting with you? And it's like, okay, let me, let me set the stage. Who are an LP? I deploy capital. I've done well. But have you said that, we have, you know, lots of LPs are amazing and great. But I did just, I do have a lot more empathy, like you say, for founders. And, you know, just things happen like before. before my firm at Trades launched, I ate people from a very prestigious university endowment came to our office.

Starting point is 00:45:43 They spent five hours with me and I was just like, oh my God, this is amazing. They all followed by Twitter. They were asking me all these questions about tweets. I was so happy. And then the next day they're like, oh, we're definitely not going to invest. And then I went to lunch with one of them and I was like, oh, you know, what was that? I thought there was great meeting. They're like, oh, yeah, we love you, but we're just not going to.

Starting point is 00:46:02 employ anything into your sector for many years. And we just want to meet with you because you're interesting. And we thought it'd be a good use of time for us. And I was like, well, did you think about like I'm trying to watch a firm? Like, but anyway, it's just some other things to riff on things Megan said, which I thought was great. But just it really is a journey. It takes a lot of, it takes a lot of time. You want to invest in those relationships.

Starting point is 00:46:25 And there are, you know, truly world class LPs who will, who I think will go along with that journey with you. If you kind of follow all of Megan's rules, they'd be really supportive through thick and thin. And karma is also very, very real. That just resonated with me so much. Like one of my rules is just always be super nice to everyone. And if someone is going through a hard time, be really nice to them. Not only is it the right thing to do, but it often, yeah, I try to bounce the ball to a lot of people. the people who consistently bounce it back

Starting point is 00:47:00 are the people who I reached out to or tried to be helpful when they were going through a hard time and I think that's a great rule for life but also for raising money. And the last thing, and then we can talk about AI or software or whatever you want to talk about is just also something Megan said is like,

Starting point is 00:47:16 you have to ask. Like, you know, like just, even with people who, you know, it's very rare that somebody is going to say, hey, just here, take my money. Yeah. You have to. ask. People want to, people want to be asked. Um, and I think getting comfortable doing that,

Starting point is 00:47:33 uh, it is important. But anyways, those are just some thoughts inspired by Megan's epically awesome talk. So we can, I, you know, I think this is the talk that matters actually. And by the way, in terms of the rule for life, I, it's very strange where you pick things up, you know, like childhood actually like really is formative. And, you know, one of the things I learned was like, when somebody gets their asses kick, like, you got to go like stand up for your friends and, and, like, That's like, that's the moment in time of friendship. 100%. Not the moment of time when like, you know, they win the lottery and you go pop champagne

Starting point is 00:48:03 bottles with them and you're like, hey, you won the lottery. This is fucking great. You know, it's like when they actually fall on their face. And so it just very naturally came to me when people got their asses kicked that I should just immediately call them or come to them and say, hey, let's get a, let's get a dinner or a drink or go for a hike or something like that. And I kind of just expect that everybody had that in their DNA. That's not in people's DNA.

Starting point is 00:48:24 The reaction is, oh, I just read a horrible news story about this person. Something terrible happened. I probably should give them some space. Yeah. And then that person is sitting at home alone on a Friday or Saturday night saying, I am a fucking loser. I have no friends. I am falling on my face and nobody has called me.

Starting point is 00:48:44 And they interpreted it as, I've just failed that life. Yeah. Well, I will share where it came from from me. And I'll admit it was not intuitive to me. But one of my best friends and mentors, and I was like 25, 24, 23, and somebody had gotten fired in a terrible way from the firm where we both worked. You know, not for anything unethical, but just, you know, probably bad performance. And I said, you know, hey, you know, I really like that person. You know, I want to reach out, but it just doesn't feel like the right thing to do.

Starting point is 00:49:18 I should give him some space. And she said, no, you should definitively reach out. And then I'm sure she's very close friends. So hopefully it's okay with her. And if it's not, I'm very confident she will forgive me. But she told me about how her dad had died when she was very young. And she'd been very close to her dad when she was very young when she was 30. Her dad, they'd been best friends.

Starting point is 00:49:42 It was really, really hard. You know, even 35 years later, it's hard for her to talk about. But she said, you know, when my dad died, there were all these people. who are good friends of mine, and they reached out, and I was like, oh, thank you. I really appreciate that. Then 80% of people didn't reach out. And it took me a long time to understand why they didn't reach out and to forgive them. And I just, I had to understand they didn't know what to say and they felt awkward. He said, but the one group of people I have never forgotten are people who weren't good friends and reached out and just said, I'm sorry. I don't know what to say, but I'm sorry.

Starting point is 00:50:17 I remember their names 35 years later. And so it's just... And they weren't good friends yet. They were good friends yet. And now they are. Yeah, but just kindness. Like I just think in the short term, sharky, aggressive people often get ahead. But in over any long term time frame, being kind, being loyal, being ethical, doing the right thing, standing by people when no,

Starting point is 00:50:47 else will, even if it's really important. They just matter so much. And if you superimpose this on what we do for a living, we make investments. Not bets. We make investments. By the way, I have that problem, too. We did. Actually, what are you betting on now? What can we double down on? Exactly. Fuck it. Let's ship these chips all in. And the LPs are like, why are these people so crazy? Yes. And why are we giving them money? Yes. Yes. I mean, I had one time, one LP just decided they were like double clicking and they were like, tell me about this company that, you know, like failed. And I was like, wow, it's interesting you bring that one up.

Starting point is 00:51:28 The founder absconded with the money, never talked to us again. And they just literally took the 150K that we gave them and like just burned it. And they never responded to our emails and we're still chasing them to try to dissolve this company so we can get attacked by it. Yes. They're in a, they're in a non-extradition country. treatment. And I was like, they're like, what was the diligence process like? I was like, what do you mean? Like, do diligence. And this is back in the day when like, when you gave a $100K check, you would just take a flyer and like, you know, like diligence. That's not what

Starting point is 00:52:01 you do now, Jason. No, no. That is not what you do now. There's a very careful process. Oh my God. The diligence we do, we have founders complain like, you're putting in $250K and the person who's leading the round didn't do as much diligence as you. I said, what does that tell you? And they're like, ah, oh, you've made stupid bets before. Yes. But if you superimpose this discussion we're having about friendship, about loyalty, and karma, and then stuff to tell it with Megan's incredible presentation, legendary presentation, what we do is place a series of bets.

Starting point is 00:52:32 And we humbly place those investments and hope for the best. It's going to take time. Did you do that on purpose? Did you do that on purpose? Kind of. I go for a cheap laugh sometimes. Okay. That was just like a soft.

Starting point is 00:52:44 Yes. It's like, wow, it's just hanging up there for me to smash it. But you have to be humble because the truth is we don't know if they're going to pay off, to what extent they're going to pay off, and most importantly, when. So somebody who is just getting smashed by wave after wave and getting sucked under could hit it. You've had this happen. You've gotten your ass kicked and you've hit huge waves. So I want to talk about maybe that experience for you because when you were at Fidelity,

Starting point is 00:53:16 man, you beat 99% of your contemporaries, then you made your own fund, and there have been bumps in the road. There are always bumps in the road in public equity investing. I was actually just reflecting, if you're a public equity investor, you have no issue staying humble because I would say the world's the world's best investors are actually wrong 60% of the time. And then I would say your average, good professional investor is probably only right low 50s percent of the time like I've seen batty averages on a lot of investors so you're wrong a lot and even when you're having a great year all you're consumed with is what you what you could have done differently and then when you have a bad year yeah it's public equities are very very humbling and I do think particularly for public

Starting point is 00:54:09 equity investing, but in venture, you move back to the comments about being a journey. Resilience and grit and tenacity are the most important characteristics if you want to have a career, because no matter how good you are, you're going to go through really hard times. And I've, I mean, I've had many hard times in my career as a public equity investor. What was the worst? What was the worst? The worst was probably, yeah, that one. The worst was probably, I see it in your eyes. The first was probably 2011 or 2012. And I made two of the biggest mistakes simultaneously. And, you know, there's good bets, bad bets, winning decisions, bad decisions.

Starting point is 00:54:57 But I had been a, I had run a telecom fund earlier in my career. And for a long time, a really easy way to make money in telecom is whenever in almost any market, but particularly in emerging market where cell phone penetration was still growing, you bet on the company with the newest network, you always want. And the reason for that is the newest network is the best product and ultimately the best product

Starting point is 00:55:18 wins the reason it's the best product. It has the latest and greatest technology. And also it's empty. No one can get a cell phone signal, you know, at Grand Central or PIN station, you know, even, you know, at airport's cell phone. And that's just because they're crowded. The networks are crowded.

Starting point is 00:55:33 A lot of people are there. And so if you have an empty network, it's an amazing experience. and then, you know, word of mouth is the best way to sell anything, whether it's a firm or a cell phone network. There's a company called Nextel International. They launched, if anybody knows it went bankrupt, just to cut to the end of the story. They had launched a brand new, amazing network all over kind of South America and Mexico, and it was by far the best. And they were coming into the market.

Starting point is 00:56:02 They were a little levered, you know, three times levered. and then things just started to go wrong that were out of their control. A price war broke out between two of their biggest competitors. The exchange rate went the wrong way and you have to pay for telecom equipment in dollars and the revenue comes in and other things. All these things went wrong. And at each point, you think, hey, maybe I should reduce the position. But it felt like, hey, there's been an overreaction to this.

Starting point is 00:56:31 And I'd so consistently made money on this. and I'm embarrassed to say, I think, when the stock was maybe $8 or $10, I wrote a letter to the board of directors demanding they buy back, not demanding. You know, I was not an activist, you know, walking through the merits of buying stock back. I think the company is bankrupt 15 months later. So that was a pretty bad mistake. But then I actually think the next year I had my, I had the best year of my career. a lot of things came through and just kind of have to have the tenacity and the belief that, hey, I've been doing this.

Starting point is 00:57:10 I have a process. I have a framework. I have a philosophy that works. That has not changed. And I mean, I will, and just, you know, you have a lot of people, people are not shy about sharing their opinions after you have a really bad year. And you're human. And that impacts you.

Starting point is 00:57:28 And you just have to have the resilience to continue making a high quality decision. But yeah, that was the hardest year. It was a really hard year. I went from that two, I had two moments in my prior career. When I was very young, I was, you know, the highest rated out of 200 analysts, just to make, you know, I was 24 years old. Ah, this is amazing. The next year, they told me, they're assigning me a new group. And if I don't get the stocks right quickly, I will be fired.

Starting point is 00:58:01 And that terrible. terrible year that I had. The year before, it was, you know, we think you're amazing. We believe absolutely in your, you know, process and everything, you know, you're off to an incredible start. And then one year later, hey, if you have another bad year, you might be fired. And there are some great lessons in this, which is, you know, there are highs and there are lows, and then there is your process. You kept referring to the process. And I think that's critical, because the outcomes are in some cases out of your control. They're in public equities.

Starting point is 00:58:34 They're almost, they turn or they are. You just have to have a discipline. Right. And then that discipline and process, you can actually improve every day. Absolutely. You can get better every day at the process. Even if you're missing the ball.

Starting point is 00:58:46 And I will say, quote another great for the mentor of mine, Steve Weimer, you know, has a Warren Buffett like track record on $150 billion. Steve has two things always stuck with me in investing. There's only two things, numbers of excuses.

Starting point is 00:58:58 And if you don't have the first, generally nobody wants to hear the second. Wait, wait, wait, wait, it's what and excuses? There are only two things in investing, numbers and excuses. Numbers and excuses. You don't have the first, nobody cares about the second. Love it. The other thing you would say is the reason to never have a really bad year is that people

Starting point is 00:59:16 either try to help you or they put pressure on you, and both of those are the enemy of excellence. And just, you know, you do have to as a professional investor manage the downside. it is, what is it, the disposition effect? I can't remember, but people value losses between two and five times more than gains. Fascinating. Yeah. I do want to talk about, last year you talked about, hey, if you don't get your AI strategy correct, you're going to be roadkill, essentially, like the gap between the people who get it right and the people who don't is going to be like the spread rate, the dispersion. It's just, it's going to be brutal.

Starting point is 00:59:56 So here we are exactly one year later. AI has had quite a year in terms of the velocity of product releases and it being incorporated at least what I see on the ground inside of startups who are the first to adopt these kind of technologies because they're resource constrained and they always look for an advantage and an edge. And what I'm seeing is 100% adoption of anything that can make their firm, you know, their three or four person or 30 or 40 person firm, more competitive, eliminate jobs, make the people who are working them more efficient. It is a ruthless, unbelievable, you know, sort of process.

Starting point is 01:00:34 Now, what are you seeing in the big enterprises and with the companies that we were talking about last year, whether it's Google, Apple, Amazon, let's just go through the top 10 companies or so. And did they embrace it? Who embraced it the most? Who did the worst job embracing it? And who kind of kicked the can down the road? Let's put them into three buckets.

Starting point is 01:00:55 Yeah. Crushed it, fucked it up. Or like a sleep at the wheel. Well, this is not a commentary on stock performance. Just specifically about execution on AI. To me, if you think about Google, a year ago, it was at dawn we slept, December 7th, 1941. They were, you know, they're, you know, there's, they're this company with a country club culture,

Starting point is 01:01:20 um, and which I'm sure they still have, who hadn't really shipped anything, maybe in a long time. They have this, you know, search is probably the world's single best business. You could probably run that and generate all of the revenue with no human beings. It could be a 10-person company. It could be a 10-person company, literally. And Open AI comes out. And it was a question, will, you know, will the sleeping giant awaken, you know, Yamamoto after Pearl Harbor wrote a letter where he said, hey, you know, I've given.

Starting point is 01:01:56 given Japan and the Imperial Navy, an incredible victory, and I'll continue to do so for a year. I'll drive America across the Pacific, and it will be an unbridled string of victories. And he said, but if America doesn't give up going back to grit and resilience, then eventually the steel mills of Detroit and the oil wells of Texas will overwhelm me and Japan. And after this first year, it will be continuous defeat. And that's eventually what happened because America had resolve, which is awesome. Yeah, America.

Starting point is 01:02:31 Yes. By the way, I'm super paid. We also immigrated all the great scientists around the world. 100% to have us win the most important race that ended that, you know, tragically, you know, ended that savagery. Bringing all the world smartest Jewish people to America in World War II, many of them from Russia and Eastern Europe was an enormous win for the country. Perhaps a lesson we need to meditate.

Starting point is 01:02:54 On the United States? Yes. Recruitment of, we talked, I had Antonio Grasas here. I think you've met him. He was here yesterday and we talked about immigration and just, why are we not recruiting the smartest people we can find? No, it's the biggest advantage of America has. I mean, outside of, you know, we have two oceans, incredible natural resources,

Starting point is 01:03:10 but all the world's smartest people want to come here and we should take them all. Not only does it help us, but in many cases it weakens our enemies. But I'm very pro-American. I didn't say, yeah, American lightly. I'm super pro-American, very pro-national defense in investing. I'm very grateful to be a citizen. I think America's greatest country. Yeah, you can be patriotic.

Starting point is 01:03:29 You won't get canceled. Yes. We've moved past that insanity. Now you can fly an American flag. Yeah, yeah, right side up. Yes. Fly an American flag right side up. And by the way, you could always do that.

Starting point is 01:03:41 Just, you know. And you can do it upside. It's your right. It's your right. Certain people, which is one of the great things about America. But anyways, coming back to Google, which is also rich in resources, is well defended with two oceans,

Starting point is 01:03:54 you know, those oceans. But asleep at the wheel. But a sleep at those wheel. And they're, you know, America, we have the Pacific of the Atlantic Ocean.

Starting point is 01:03:59 They have Android and Chrome, which are really dominant distribution systems. But they woke up. They put Demis Hasibus in charge of AI. They've started to really leverage the fact that they do have their own compute infrastructure that is really unique and really differentiated. To this day, large language models have only been trained on three kinds of chips.

Starting point is 01:04:20 Invidia, Cerebrus. and Google's TPUs. And this guy, Demas, is ruthless, effective, started Deep Mind, which was the original OpenAI. And, you know, he's, you know, he is a wartime general, and he is in charge. I think from my perspective, Jim and I was the first time that Open AI was ever passed, and it does have a really important advantage.

Starting point is 01:04:50 It can do context caching in a way that, No other LLM can. I think it's going to take maybe people time to appreciate how powerful this is. What that basically means is it remembers questions that it has been asked. And if it has in a given use case, and if it's given a good answer, it just goes back to that answer. So for all of these enterprises that you're seeing these startups that are trying to save money, the AI has to think again every time it's asked a question. because of Google's TPU architecture,

Starting point is 01:05:22 they can do context caching, and we will see whether that is in the next generation of GPUs from Nvidia or AMD. You know, Jensen has seen around so many technical corners. He's an exceptional CEO. So Google's back. Google's back. Okay, so they got it right this year.

Starting point is 01:05:42 Nvidia obviously is crushing it. Invidia has continued to crush it. Then we look at Apple asleep at the wheel. Or somewhere in between. I think Apple is about to wake up. And I think. So that could be a parallel to what we saw with Google. Yeah.

Starting point is 01:05:56 And I do think Apple's strategy is generally not to be the first, but almost to be the last and the best. You know, the Palm pilot came out in 2001. And I think at one point, Palm had a bigger market cap than like Apple, Dell, Hewlett-Packard, compact, all these companies combined. And, you know, now where's where's, where's, where's, where's, where's, where's, where's, palm nowhere um it's an operating system for LG TVs that is literally where it is um so being first isn't always most important but if you're going to be last you need to be best and Apple makes more money off search than you one but Google because they own iOS which is another toll booth on the

Starting point is 01:06:38 internet just like Android and Chrome and I think what you will see them do this is this is what I would do if I were them. I would have a small on-device LLM that is privacy safe and can access all of your information and you trust it because Apple has built a brand around privacy. And then whenever that LLM doesn't know what to do, it asks OpenAI. Permission to go to the cloud. Yeah, permission to go to the cloud. I think you're right because I noticed in my iPhone library, or which is just Apple Photos now, there's a little AI wizard button and when you press it, it's like, that's a bulldog and, you know, whatever, a Tesla behind it

Starting point is 01:07:21 it and you can see the power of exactly what you're saying. On your phone, every 10th photo has it, and they're just subtly going to make it. So when you go into photos and say, hey, I need pictures of my bulldo when it was a puppy and it would be like, which one? Torres or fondue or Toro or Macs? Well, even more important, instead of doing searches, you'll say, instead of, you know, doing searches for, you know,

Starting point is 01:07:41 best vacation in Italy. You'll just say, hey, book me a vacation. You know I will like. Give me three choices in Italy. And then why don't you pick another country that you think I'd like that I didn't suggest? Then we'll say here, four options, press one, boom. And this is, you know, agents and actions and transactions replacing search. And I think this will happen.

Starting point is 01:08:02 I think agents are going to be massive. Before I leave that, I will just say, this is a two-year-old iPhone. It's the first time I've had a two-year-old iPhone until it's cracked. Same with me. You skipped a generation. You skipped 15? just because there was nothing that mattered. The camera's good enough.

Starting point is 01:08:16 But to have that local AI model running, have my, have my Jarvis, have my Gavin AI that knows me and likes me, and is friendly to me, and, you know, will protect me in a, you know,

Starting point is 01:08:30 the Terminator world where we're, worth upgrading. I can't remember if we're Gavin's Angels or Jason's angels, but we're, we're in the resistance. This process, this has enough compute power to run that LLL, It doesn't have enough memory.

Starting point is 01:08:44 And this is very important because right now, the two rate limiting factors for AI, they're no longer computed in GPUs, it's power, finding places to plug the GPUs in, and to get to AGI, we're eventually going to have to have a one gigawatt data center and then a five gigawatt data center.

Starting point is 01:09:04 In our lifetime, there will be data centers that are bigger than Manhattan. Forget the Pentagon. Whoa, whoa, whoa. like vertically every way. All the largest buildings, like in our lifetimes, the hundred largest buildings in the world

Starting point is 01:09:19 will be data centers. The world is now power constrained. There's only three places in the United States that you can put in one of these data centers. There's a massive competition to have them. It's got to be somewhere near. Nuclear reactors. Nuclear reactors.

Starting point is 01:09:34 You need multiple nuclear reactors that are unregulated within 50 miles of each other, given the current state of technology. But that's one constraining thing for AI. The other is something called HBM memory, high bandwidth memory. This is the primary axis of competition. Who makes that? And most compute.

Starting point is 01:09:51 Micron and Hynex today. Samsung has lost their way. And Micron and SK. Hinex, it's a Korean company, it's a great company. To make highbend with memory, you need 4x more wafer space than make the DRAM that goes into this iPhone. So if Apple doubles or triples the amount of memory in the iPhone and then Google will do the same for Android phones. At the same time, to enable AI on the phone that this is happening in the data centers, you could have the first real memory cycle since 1996.

Starting point is 01:10:21 And in 1996, and we'll see this probability, and it's not going to happen. But in 1996, Micron was a 50-bagger. And the prices of DRAM, if it is the rate-limited thing for selling iPhones and AI, we'll see where it goes. So I think that's important to talk about in the context of Apple. But please, what else? I think we got it. Okay.

Starting point is 01:10:47 You gave it to us. Awesome. I mean, lightning around Amazon, sideways, up or down? I mean, I haven't heard of people. I would say they have been sideways. They're trying hard. A core belief of mine is if you're a foundation model company and you do not have unique data and Internet scale distribution.

Starting point is 01:11:07 you are the fastest depreciating asset in human history. And I think most of these companies are zeros. And there's like 10 of them. Yeah. And I think the only hope most of them have of getting the preference back is for Amazon to acquire them the way Microsoft did with inflection. Yeah. And then when we're on the other side of that,

Starting point is 01:11:25 and we have AGI, I think those few companies that make it are going to be immensely valuable. But that's Amazon, sideways. Sideways with AI. That's the question here. They could be crushing it in other places. Exactly. But on that case, and then Microsoft, obviously, top of their game.

Starting point is 01:11:40 Yeah, I would say top of their game, but everybody else has raised their games. Right. So the relative advantage has probably slowed down. And meta is in a much better position than they were. And just, you know, to me, true greatness for Mark Zuckerberg would be to rename the company. You know. Again. Yes, again.

Starting point is 01:12:00 To what? Call it Facebook again. Call it whatever. Yeah. Call it big blue AI. Call it AGI. just rename it. And I give the guy a lot of credit

Starting point is 01:12:09 for being so publicly in on the Metaverse, which by the way, I think will still eventually happen. That's going to be a legendary misadventure in CEOing. I think in 15 years

Starting point is 01:12:21 it'll probably look okay. But it was just the Metaverse, we just need many more cycles of Moore's Law so you can have AR glasses that work. You know,

Starting point is 01:12:31 the meta ray bands are good. And then ultimately we need brain computer interfaces to work. And then you will We'll have the Metaverse. Nobody wants to shit on their face, man. It's hard enough to wear ski goggles for a couple hours.

Starting point is 01:12:42 I'd be so into, like, I mean, I have prescription glasses. For glasses, yes, but for goggles, no. Yeah, not for goggles, but I mean, I would be very into if they can. That's what I mean. You need, you just need six years of Moore's law making these chips smaller, more energy efficient, be able to fit in, you know, whatever kind of glasses. They'll get it. But he's pivoted hard, hard to AI.

Starting point is 01:13:05 Putting a search box on the top of every app is gangster. It's gangster. That is a shot across the bow of Google and anybody else in the industry. Like, I am willing to take my top real estate, the top of the search bar on every fucking app I have on three, four billion phones, whatever he's got. And I'll just divert all your traffic. Absolutely. What's up? Yeah.

Starting point is 01:13:26 And we're going to have an epic competition. What's up? What's up? Yeah. He's like, what's up? Gold chain. I'm going to MMA. By the way, I thought that that was real, the one of him with the, uh, with the coat team,

Starting point is 01:13:39 I mean, yeah, I was like, he's like, I fucking, I'm hiring it. You guys, you guys want to get him my grill about meta? Yes. Fuck it. I'm going to, I'm going to MMA gold chain. Yes. Fuck it. This guy's coming out for Trump next.

Starting point is 01:13:50 Yes. No, he doesn't care. All right. Give it up for Gavin Beck. Awesome. Thanks, everybody. Thanks, thank you. Thanks, Jason.

Starting point is 01:13:57 Good stuff. I hope you enjoyed all of that. I took a lot of notes myself when they were talking, but I'd be remiss to not give three more shoutouts before I let you go. The sponsors of the event were fantastic. They helped make it happen. So a big thank you to Eventus Advisory. They are a leader in on-demand finance and accounting support. Then there's Vin-Shirt, V-E-N-S-U-R-E. They do end-dain solutions for payroll, HR benefits, and even risk management. And then there's Forge Global, a very well-known secondary marketplace for startup shares. So if you're looking to buy stock in that company you don't work for,

Starting point is 01:14:31 well, Forge might have the answer for you. I'm Alex. I'm at Alex over on Twitter. I write a cautious optimism, and I also write the Twist 500 newsletter here for launch. I'll talk to you soon. Goodbye.

This Week in Startups - Liquidity Summit Talks: Antonio Gracias and Gavin Baker | E1990

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.