Peter Thiel Gets Apocalyptic, The Myth of AI Millions, 𝕏 Timeline Reactions | Joe Lonsdale, Bret Taylor, Riley Walz

28:49When the dot-com bubble popped, society found it disapproved of the excesses in the sector, and governments found they had the popular support to reserve authority over the tech companies and their investors. This put a brake on the madness. Instead of the reckless innovation of the bubble, companies started to expand into proven markets, and financiers moved from speculating to investing. Entrepreneurs began to focus on finding applications rather than on innovating the underlying technologies. Technological improvements continued, but change became more evolutionary than revolutionary. Look at this chart.

29:23Technological waves over time, the Industrial Revolution, the canal mania. I wasn't even familiar with this. Arkwright Mill opens in 1771. I would love to watch a documentary or even a drama on canal mania. Yeah, this has the trappings of... Underrated mania. People always talk about tulips. Tulips, yeah. Tulip doesn't even count because it's not a real technology. So the actual technological waves that are highlighted here, the Industrial Revolution, steam and railways, steel, electricity, and heavy engineering, then oil, the automobile, and mass production from 1908 to 1974, and then 1971 onward is information and telecommunication.

30:05And pretty remarkable results from this. 27 % rise in real GDP through the 2010 to 2019 era. That's pretty good. in contrast society did not need a bubble to pop to start excoriating ai given the backlash to tech that has been going on for a decade this seems normal to us but the ai backlash differs from the general high regard earlier in the cycle enjoyed by the likes of bill gates steve jobs jeff bezos and the others who built big tech businesses the world hates change and only gave tech a pass in the 80s 90s because it all seemed reversible it could be made to go away if it turned out badly This gave the early computer innovators some leeway to experiment.

30:45Now that everyone knows computers are here to stay, AI is not allowed the same wait-and-see attitude. It is seen as part of the information technology revolution. Perez, the economist, breaks each technological wave into four predictable phases. Eruption, frenzy, synergy, and maturity. This is a new Gartner hype cycle. We need to know where are we on the curve. The Perez tech wave. I think we're right around frenzy. We're definitely past eruption, maybe going into synergy and maturity next. The middle two, frenzy and synergy, are the easy ones for investors. Frenzy is when everyone piles in and investors are rewarded for taking big risks on unproven ideas culminating in the bubble.

31:23When paper profits disappear, when rationality returns, the synergy phase begins as companies make their products usable and productive. Everybody's scared of a crash. They're not eager for synergy. Yeah. Yeah. So investing in the maturity phase is even more difficult. In eruption, it's hard to see what will happen. In maturity, nothing much happens at all. Nothing ever happens. The uncertainty about what will work and how customers and society will react is almost gone. Things are predictable. Everyone acts predictably. The lack of dynamism allows successful synergy companies to remain entrenched.

31:59You see the nifty-fifty and fang, but growth becomes harder. They all start to enter each other's markets conglomerate, raise prices, cut costs. It feels like Microsoft's in that bucket certainly already. Companies frame this as a drive to win, but it's really a fear of losing. Should we read about the shipping container wave? Shipping containerization was a late wave innovation that changed the world, kicked off our modern era of globalization, resulted in profound changes to society and the economy, and contributed to rapid growth in well-being. but there were perhaps only one or two people who made real money investing in it that is insane to hear if that's true yeah and key words here investing specifically in that technology versus benefiting from it broadly yeah because if you were in any type of business that required shipping and receiving goods from all over the world you benefited directly from the technological change even if you weren't like investing in companies that were functionally uh creating yeah i would certainly agree so they they mark the containerization wave starting the year is 1956 it was late in the previous wave but that year the company soon to be known as sea land revolutionized freight shipping with the launch of the first container ship the ideal x or the ideal 10.

33:26Sealand's founder, Malcolm McLean, had an epiphany that the job to be done by truckers, railroads, and shipping lines was to move goods from shipper to destination, not to drive trucks, fill boxcars, or laid boats. Sealand allowed freight to transfer seamlessly from one mode to another, saving time, making shipping more predictable and cutting costs. Both the cost of loading, unloading, and reloading and the cost of... So prior to this, you would just have like a wooden crate that would just get thrown on a ship. Yeah. So you have to stack it, unstack it, figure out how you have to do a puzzle every single time you want to load and unload.

33:58Tetris. Yeah. And now it's just stacking blocks. Jenga instead of Tetris. Much easier to play Jenga, supposedly. The benefits of containerization, if it could be made to happen, were obvious. Everyone could see the efficiencies and customers don't care how something gets to where they can buy it as long as it does. But longshoremen would lose work. Politicians would lose the votes of those who lost work. Port authorities would lose the support of their politicians. Federal regulators would be blamed for adverse consequences. Railroads might lose freight to shipping lines. Shipping lines might lose freight to new shipping lines.

34:31And it would all cost a mint. Most thought McLean would never be able to make it work. But he squeezed through the cracks of the opposition he faced. He bought and retrofitted war surplus ships, lowering costs. He went after the coastal shipping trade, a dying business in the age of new interstates, to avoid competition. He set up shop in Newark, New Jersey, rather than the shipping hub of Hell's Kitchen in Manhattan, to get buy-in from the Port Authority and avoid Manhattan congestion. And he made a deal with the New York Longshoremen's Union, which was only possible because he was a small player whom they figured was not a threat.

35:03What's interesting is that, yeah, I think it's like the direct investment might be very concentrated. But when I think about the rise of the 60s, the 70s, the 80s, like globalization, I'm thinking of Nike, and I'm thinking of companies that were built on the back of globalization, even the iPhone, Apple. These companies exist in part and are beneficiaries of containerization. And so it's not a direct investment in the fundamental technology, but it is enabled by it? Yeah. I don't think so. Yeah, we'll have to get through the article, but it already feels like it's somewhat clickbait. Not clickbait, but to say, like, no one's going to make money on AI or you're not going to make money on AI.

35:53Yes. Let's skip to his conclusion about generative AI. He says, let's grant that generative AI is revolutionary, but also that as is becoming increasingly clear, this particular new tech is now already in an evolutionary stage. It's towards the end of its cycle. It will create a lot of value for the economy, and investors hope to capture some of it. This is the key point. Yes. I totally agree. When, who, and how depends on whether AI is the end of the ICT wave or the beginning of a new one, right? Yeah. Is it SaaS or is it God? Is it the final SaaS product and then we're stuck? Or is it something entirely new and we're just going to start compounding?

36:38Yeah. Is it consumer software? Is it enterprise software? Something entirely new, right? I don't know. If AI had started a new wave, there would have been an extended period of uncertainty and experimentation. I feel like that's the last decade of open AI. I feel like 2015 to 2025 was that period of uncertainty and experimentation. Yeah, you have to understand, is it people in 2015 that were trying to make chatbots, right, with no hype? Totally. Really no user, not a lot of users, right? Yeah. Or are you just counting, like, when OpenAI was like, okay, we can actually be a for-profit company now and we can productize this.

37:18But even then, there was like a pretty large gap between them deciding like, hey, let's actually raise traditional capital for this before they could get out the API. Yeah. Before they could get out ChatGPT. Totally. Yeah. And also like there was this in AI, the period of uncertainty and experimentation. That was the TikTok algorithm, the Netflix recommendation algorithm, Facebook's core AI investments. They have bought billions of dollars of GPUs, not for generative AI, not for LLMs, but just to make better ad recommendations, better content recommendations. And so there's, I don't know, there's a lot there.

37:57We'll have to continue to dig in, but let's keep reading. He said, when thousands or millions of tinkerers use the tech to solve problems in entirely new ways, its uses proliferate. But because they are using models owned by the big AI companies, their ability to fully experiment is limited to what's allowed by the incumbents who have no desire to permit an extended challenge to the status quo. This doesn't mean AI can't start the next technological revolution. It might if experimentation becomes cheap, distributed, and permissionless, like Wozniak cobbling together computers in his garage, Ford building his first internal combustion engine in his kitchen, and Treveknik building his high-pressure steam engine as soon as James Watt's patents expired.

38:38When any would-be innovator can build and train an LLM on their laptop and put it to use in any way their imagination dictates, it might be the seat of the next big set of changes, something revolutionary rather than evolutionary. But until and unless that happens, there can be no eruption. Tyler, like, don't you think it's possible to build and train an LLM on your laptop and use it in any way your imagination dictates? Yeah. I mean, I mean, I don't know about training a whole new LLM, but like people are getting a lot of value out of open source ones that they locally. Lama, Deepsea, Quen. It's like barely obvious.

39:12Yeah. Yeah, it doesn't. I mean, I somewhat agree that we haven't seen that many totally new, crazy, like unprecedented uses of LLMs where it feels like it's an entirely new category. Like it's hard to point at like the Uber for AI where it's like this company would never have existed. It feels like a lot of the companies are kind of evolutionary. Like, you know, we have site builders and then we have vibe coding platforms that's very top of them. Figma.com. A lot of the companies. Figma helps design and development teams build great products together. Get started for free. Use Figma Make if you want to develop a website.

39:55Yeah. A lot of times companies come on TBPN to talk about what they're building. They're building an AI application. Yeah. There is a company started 10 years ago basically doing the exact same thing. They just didn't position it in the same way. They weren't using prompts and weren't using reasoning and things like that. But still, it looks and sounds like SaaS. It looks like it's solving the same SaaS problem definition. The problem was like organizing information for lawyers or answering the phone. And we had a phone tree that was just deterministic, like robotic. Press two for, you know, customer support.

40:37And oftentimes they can solve a problem. It's like the product can be twice as good. Yes. Which is enough to kind of eat into existing markets. Yes. Maybe create new markets, right? I think the market for people that can create software has obviously exploded, right? which is why you see Vibe coding revenue is just exploding to the billions of dollars collectively but it does seem like the video games are getting more realistic video games are getting more realistic the software is getting better the spreadsheets have more useful tools there's just a little extra helper everywhere it's a lot of co-pilots it's not a lot of entirely new plans and then you see ChatGPT and the usage looks a lot like the way that people use Google search Granted, it's better in many ways, but the core value that's being delivered is quite similar.

41:33He actually talks about domain-specific models. So first, he says, economists are predicting that AI will increase global GDP somewhere between 1 % to more than 7 % over the next decade, which is$1 to$7 trillion of new value created. The big question is where that money will stick as it flows through the value chain. And most AI market overviews have a score or more categories, breaking each of them into customers and industry served. But these will change dramatically over the next few years. You could instead just follow the money to simplify the taxonomy of companies. There are data infrastructure companies.

42:04There are model companies, user application companies, AI customers and consumers. But what the history of containerization suggests, if you aren't already an investor in a model company, you shouldn't bother. Sam Altman and a few other early movers may make a fortune, as McLean and Ludwig did. but the huge cost of building and running a model coupled with the intense competition means there will in the end be only a few companies, each funded and owned by the largest tech companies. If you're already an investor, congratulations, there will be consolidation, so you might get an exit. Domain-specific models like Cursor and Harvey will be part of the consolidation.

42:38These are probably the most valuable models, but fine-tuning is relatively cheap and there are big economies of scope. On the other hand, just as Google had to buy Invite Media in 2010 to figure out how to sell to ad agencies. Domain-specific model companies that have earned the trust of their customers will be prime acquisition targets. That sounds like another group of people that would be making money. When you say only OpenAI investors will make money, it's like, well, there are 7 ,000 people on that cap table in one way or another. All the investors, all the LPs and the funds that are investors, all the employees, all the people that got in through SPVs, like it's not just two people.

43:18I feel like it's a lot of people that are, that are in the big winners. Like even if there's just a few power law winners, uh, there's, there's, there's other, you know, another way to look into this is like, if you're investing in a series a at three at a$300 million valuation, assuming just crazy explosive growth, it's very possible that company, even if it does well, would someday sell for a few hundred million dollars and you wouldn't have actually made any money even though you got an exit. Yeah. Here's a good like counter point to some of the companies that are potentially in a more precarious situation.

43:54So he says, while it's too late to invest in the model companies, the profusion of those using the models to solve specific problems is ongoing. Perplexity, inflection AI, writer, abridge and a hundred others. But if any of these become very valuable, the model companies will take their earnings either through discriminatory pricing or vertical integration, launch their own competitors. Success, in other words, will mean defeat, always a bad thesis. At some point, model companies and app companies will converge. There will be simply AI companies and only a few of them. There will be some winners as always, but investments in the app layer as a whole will lose money.

44:29The same caveat applies, however. If an app company can build a customer base or an amazing team, it might be acquired, but these companies aren't really technology companies at all. They're building a market on spec and have to be priced as such. A further caveat is that there will be investors who make a killing arbitraging FOMO panicked acquirers willing to massively overpay, but this isn't really investing, he says. And so again, it's another group of people that will get rich, but it's maybe not a durable source of alpha, which is an interesting thought. Where else? Let's keep going. This is a very long article.

45:05You should go read the full thing. So So let's close this out by reading his conclusion. And of course, if you want to read the full piece, you can head over to Colossus Mag. There is nothing better than the beginning of a new wave. When the opportunities to envision, invent, and build world-changing companies leads to money, fame, and glory. But there is nothing more dangerous for investors and entrepreneurs than wishful thinking. The lesson learned from investing in tech over the last 50 years are not the right ones to apply now. Now, the way to invest in AI is to think through the implications of knowledge workers becoming more efficient, to imagine what markets this efficiency unlocks, and to invest in those.

45:42For decades, the way to make money was to bet on what the new thing was. Now, you have to bet on the opportunities it opens up. So, Jerry Newman, he's a retired venture investor. Totally, totally agree with this conclusion, right? I think if you come in with a$200 million venture fund right now and are just investing at crazy multiples, pre-revenue, investing in new foundation model labs, you're going to have a really tough time. But that doesn't mean. And I don't know. I think saying the app layer as a whole will lose money, at least when you look at private markets. Right. The whole point is that a lot of the companies can go to zero and like a handful of them will create enormous value and that's okay.

46:33Yeah. Right. So not that bearish. But I do think it's it's not a you know, at this stage in the market, spraying and praying is going to lead to some bad outcomes. Yeah. Really understanding the opportunities, really understanding that maybe this is a lot of the winners here are going to look more like. traditional software. And that's okay. But it just means you have to be, you know, I think that the company we had on yesterday, Filevine, a good example of this, started as enterprise software for law firms. Now, you know, they have 100 ,000 lawyers actively using the product every day in a really good position to vend in a variety of different models.

47:15And they will, depending on how they execute proved to be a big challenge for new AI native companies that will have to decide, do we want to be sort of an ancillary product or do we want to own the workflows? And if we want to own the workflows, we have to rebuild all these different products that help a firm run and do that with maybe 10 years behind other players in the category. Yeah. I mean, venture has always been the game of pick the winner in the category, but it feels like it's more important than ever to actually define the category and understand, is this particular category going to have one winner or three winners or two winners?

47:59Is it going to wind up being a duopoly or an oligopoly? I'm excited to talk to Brett Taylor about this he's building he's the 800 pound gorilla in AI for customer engagement experience but it's a category that has an active capital war going on you have heavily funded new players like Sierra and Decagon you have the intercoms of the world our partner Fin.ai, the number one AI agent for customer service We're going to ask Brad Taylor about his performance benchmarks, his competitive bake-offs, his ranking on G2. Because we are obviously supporters of FIN, but interested in following the fight. And I am very interested in how oligopolistic will that particular market be.

48:50There are certain markets that have just kind of run away. It certainly feels like, just in terms of knowledge retrieval, we could be looking at a monopoly with OpenAI and Chagipi, or a duopoly with Gemini, but it doesn't seem like there's going to be 10 search engines and 10 default chat apps that people are going to. It's certainly not going to be 20. Or 10 deep research APIs. Exactly. Yeah, it feels like there's going to be one or two that kind of compound and people get comfortable with and they stick with. Yeah. Well, in other news, Emmett says, I'm excited to share Rora Water is the official partner of Huberman Lab for Water.

49:31uh rora is of course company i co-founded a while ago um and uh worked on this partnership with the human lab team uh rob and andrew for a very very long time at this point we started talking about it probably a year ago well started talking about it maybe a couple years ago actually got them product over about a year ago they used it a bunch they tested it a bunch uh and finally launched this partnership earlier this week. So absolutely thrilled to get this across the line. And yeah, big vote of confidence from the Huberman team. I love it. And in other Huberman world news, his network, SciComm Media, is working on a new show with none other than David Senra.

50:19It's David Senra by David Senra. Yeah, the show is called David Senra. Hosted by David Senra. Yeah, they picked David Senra as the host of the David Senra show. Yeah. And so we will watch the launch video that he put out on X yesterday. Pull it up. Nine years ago, I launched Founders. Today, I'm launching a new podcast called David Senra. Nine years ago, I launched the Founders podcast. On Founders, I tell the stories of history's greatest entrepreneurs and extreme winners. This week, I'm launching a new podcast called David Senra. On this show, I sit down for conversations with the best living founders in the world.

50:54I'm ready when you're so good. How many cameras? You got three cameras? People like Daniel Ek, Michael Dell, Brad Jacobs, Todd Graves, Michael Ovitz, and many others. My goal with everything I do is to obsessively study the greats and find timeless ideas that you can use in your work. We have to change or we're going to go out of business. And now you can learn directly from these legendary founders. You know what I see in the most common core of all the people that are successful for me is they're never satisfied. You said something interesting. I like the pressure. I do. Thrive on it. Founders will still come out every week.

51:29And this new show will now have episodes dropping every other Sunday on all platforms. Like Spotify, Apple Podcasts, YouTube, X, and everywhere else that you find your podcasts. There are parts of me that still live in the valley. When did you know you were good? I don't know that I'm good. I know I'm different. And I'm out there for this chicken finger dream. Failure's not an option. Chicken finger dream. Nothing is going to stop me from doing it. It's binary. There's no option. We need a wall that we can set up here and run through. That's great. Dan in the chat says, the market is not ripping.

52:02What are the white suits for today? Bitcoin is up. One half percent. It's a bright spot. It's a bull market in a lack of volatility. If you were short volatility, you've done very well today, right? Because it's not a volatile market. There's always a bull market. I mean, it's a little rough out there. It's just a fun day to put on a white suit. We just feel like we're feeling some optimism. We're having fun. We just had some good energy in the studio this morning. We thought maybe if we put on the white suits, the market. Yeah, we'd be able to materialize a bull market. Let's make it happen. We're trying to send it.

52:41Anyway, before our next guest hops on, let me tell you about Vanta. Automate compliance, manage risk, improve trust continuously. Vanta's trust management platform takes the manual work out of your security compliance process and replaces it with continuous automation, whether you're pursuing your first framework or managing a complex program. Did you see this post from Shield? He said, interesting. We might see a New York Stock Exchange IPO where the underlying asset is a bunch of paintings. Billionaire Thomas Kaplan is exploring, taking his Leiden collection. 17 Rembrandts, a Vermeer, et cetera.

53:13Worth 1.5 to 3 billion public. This is the next generation of the digital asset treasury the dats you have the physical asset treasury the pats and these will be be able to invest in art there's been a couple companies would be it would be so to invest in art there was an art platform for a while that was doing that yeah i mean i i in general taking art highly i mean uh art is liquid but not at at sort of like market prices right so if you want to sell art quickly you have to sell it at a massive discount right you need to be super patient And if you have a big collection and you're holding it, especially if you have a bunch of works from a single artist, if you bring 10 pieces online, right, at the same time, you flood the market with supply, prices will come down.

54:04So it's a very weird thing. But this makes sense. If Thomas wants to be able to get out all at once or get out partially, get some liquidity, this is probably more effective than. than selling it piece by piece. And he gets to hold on to the art, too. That's the other thing. Oh, yeah. I mean, he can still store it. He can probably keep some in different properties, etc. Well, we have our first guest of the show, Brad Taylor, coming into the team with the Ultron from the ReStream Radio Room. Brad, how are you doing? I didn't get the wardrobe notice. Yeah. We normally wear white suits when the market is ripping, and we're celebrating.

54:48But Sierra's ripping. But Sierra's ripping. So we're celebrating you. We're wearing them for you. Thank you for celebrating our growth. I really appreciate it. Yeah, it's been fantastic. Give us the brief history, the founding of the company to the news just last week. Yeah, so Sierra, we help companies build AI agents for customer experience. So I think maybe it's on a website. Maybe it's answering the phone. No one likes to wait on hold. Now you can chat with an AI agent. We're helping do everything from originate mortgages to help you get a better rate on your SiriusXM plan to helping, you know, when your ADT home alarm system doesn't work, you'll now chat with an AI agent to fix it.

55:27We're a couple years old. We're the leader in this space. We kind of uniquely have run towards, I would say, larger enterprise businesses. We're trying to help companies that, candidly, it's hard for them to deploy AI because they've got lots of legacy systems. Maybe they're in a regulated business like the health insurance market or banking. With the whole hypothesis is if we can help them be successful, there's just a ton of value and leverage in that. So as you mentioned a couple weeks ago, we announced a recent round of financing, which is we're proud of just a milestone on the path. But I think kind of recognizes our leadership in this space.

56:03You're the chairman of OpenAI. How should I think about application layer versus model layer? I'm sure you get this all the time. there was a meme for a long time about oh OpenAI is going to go on stage at DevDay and just steamroll a bunch of companies you seem to have good information on what they're going to steamroll and why not there were so many two years ago so many different companies that had thriving businesses doing various customer service platforms and many people would have said these companies are just going to move quickly on AI I get this, they have the customer relationships, but you clearly saw it differently and have proven that that hasn't been necessarily the case.

56:46I'll answer both. I'll start with the foundation models. My theory of how the market plays out is that the foundation model market will look a little bit like the infrastructure as a service market, where it primarily provides technology, low-level technology to a lot of applications companies. And then there's all the adjacencies. If you look at Amazon Web Services, they've got some developer tools. If you're kind of in the area like Snowflake and Databricks, there's probably an Amazon product that competes with you. The farther you go towards serving a line of business, like ERP systems or CRM systems, the less likely it is that you're going to run into an infrastructure provider competing with you.

57:27I think the same is roughly true in the AI market. If you think about what's required to make AGI, it's a lot of training stuff, but it's also maybe software engineering agents. So that's probably sort of in the target of these foundation model companies. The closer you get to Harvey doing legal AI or CIRA doing customer service AI, probably doesn't seem in the core of these research labs' primary functions. The reason I think that the applied AI market is exciting, not just for CIRA. Obviously, I'm very loyal to my own company. I don't think most companies want to buy a bag of floating point numbers and then figure out what to do with it.

58:06You know, they want to buy a solution to their problem. And if they can, you know, turn on Sierra and it will answer the phone and bring down their customer service costs by 50 percent overnight, they're going to do that. Not try to, you know, take these models and try to do it from scratch. If they can buy Harvey and get an antitrust review for one tenth the cost, they're going to do that. And I think, you know, I was talking to Toby Leuke at Shopify one time, and he was joking how many people come up to him and me. I'm like, aren't you just a database in the cloud? And you're like, yeah, I guess so.

58:34But there's a lot more to it. And it turns out these workflows are valuable. So I'm really banking our company that there's a lot of value in these agents. But going to the second question on the incumbent software providers in this space, it's very hard to disrupt your own business model. If you're licensing customer service software per seat and you have to make an AI agent that will actually cannibalize your own business to realize the value of this new technology, that's not just a technology problem, right? That's a business model transition. And the history of technology is littered with companies that saw the slow motion car wreck of their business model being hurt by a technology trend and not really being able to respond to it fast enough.

59:17And I have so much respect for people like Satya who navigated those transitions. But, you know, candidly, there's a lot more companies in history that didn't navigate transitions like that. And I think that's candidly just a simple reason why it's hard for a lot of the incumbent companies to respond to this new technology trend. When you were thinking about starting this company, did you, did you, like, how methodical were you about mapping the market? Is there like a whiteboard picture of you looking at, maybe I should do ERP. Maybe I, you know, was this one of a few options that you narrowed down once you studied the market?

59:52Or was it something that just came to you in a fugue state or something? Yeah, it's such an issue you said that. We actually did call up some, you know, potential customers. We started those conversations just saying, hey, if you could put us on one problem, what would it be? And by the end, we had a pretty clear picture of the available markets. I don't think any of it was particularly surprising in some ways. It's like software engineering, customer service, content marketing, all these things that come up when you think about if you have a technology that can see and understand text and voice that can reason, what are the direct applications?

1:00:28But kind of the thing that we're most excited about is not where we are today, but where it might head. And our whole theory is that your AI agent will end up more important than your website or your mobile app in the future. And as a consequence, we think the addressable opportunity here isn't just customer service, which is really compelling and interesting, but saying, you know, how can you actually drive more sales? How can you actually create a personalized concierge for your brand? I mean, just to give the math of it, if you actually have a phone call with a human being, it probably will cost on the order of$20 at least.

1:01:03And it really depends whether it's onshore or offshore. If you think about running a large-scale consumer brand with an ARPU of$20, how do you afford that? And you can't. And as a consequence, it's almost impossible to talk to most consumer brands. You bring down that cost to$0.20 or even$0.02 over time. You just think about the dynamics. Let's say you're a large mobile phone carrier and you're fighting for retaining your subscribers. You're not just going to recoup cost savings and customer service. you're going to say, how many phone calls can I have with my customers so that I retain this subscriber for five more years?

1:01:37Think about the impact on your lifetime value of your customer. So what's exciting about this, like so many new technologies, is like the first order effect is obvious, which is saying, hey, let's reduce the cost of customer service by 50%. The second afforder effects are almost more exciting, which is the companies that lean into this faster will actually grow their top line faster than their competitors. So that's what's so fun about technology disruption, right? Is that, you know, you can end up sort of upending, you know, sort of the incumbent versus insurgent dynamic. At the same time, the value proposition to the companies using Sierra is more than just cost reduction.

1:02:14It's saying like, how can we actually give you a competitive edge by moving faster in this new world of AI? How are companies actually, you know, you're going and landing some of the probably just crazy logos, right? And these are massive companies with operations all over the world. How are they actually rolling it out? Are they giving you specific sort of segments, regions? You're proving it out at a smaller scale? Or what does the actual rollout look like? Yeah, so starting, you're right, one of the things I'm most proud of is the scale of some of the companies working with us. Over half of our customers have over a billion in revenue.

1:02:54Over 20 % have 10 billion in revenue or more, which is pretty, I'm just really proud of that. I think it shows you do not need to be a small startup to deploy AI successfully with Sierra. On the rollout, it's a very, kind of a variation of what you said. Some companies start with us on a use case or two. We're going live with one of the largest health care companies in the world and actually replacing their IVR system. And it was just in a handful of months to do that. So I think it's really up to the customer how assertively or aggressively they want to roll out these new technologies. And the thing that I've been most pleasantly surprised is some of the largest, most well-regulated companies in the world actually want to move faster.

1:03:35I was having a conversation with my co-founder, Clay, about one of our largest clients and kind of being awe-inspired about how fast they're moving. And I said, now I know why they're big. Like, you know, now I know why they're successful because they're moving faster than some of their smaller competitors that we also work with. And you kind of you really learn to respect the, I say, intentionality, assertiveness of a lot of these like really well run companies. What's the sales process for a huge company like that? Is that you meet the CEO at a fancy conference and start pitching? Or, I mean, like, it feels like an advantage of, like, having a career like yours and the Rolodex that, like, that's sort of, like, the unfair advantage here.

1:04:14Yeah. In many ways, Sierra, like, you know, doing what you guys have done in two years is incredibly impressive. Yet, at the same time, it's what I would expect out of the team. I mean, we see this with, like, Palantir and Karp. Like, he's just so fun to talk to. He's in all these interesting places. And it's like, yeah, I could imagine that some big Fortune 500 company just wants to hang out with Karp and, like, talk to him. And then, oh, they do a deal together. Is that how it works? Well, it's funny what you said. This is what I expected. This is like why it's hard to start multiple companies.

1:04:42Like, yeah, of course it's going to be as exhaustive. Have you ever started a fucking company? It's hard. Yeah, but I mean, and I say that as a compliment, where it's like, yeah, going from zero to a$10 billion company, having this crazy list of customers is incredibly difficult. But I like when people have a lot of advantages through what they've done in their career and their network, and then they just put on an absolute master class. It's very satisfying. Well, I appreciate you saying that. It does all have to start with the product, though, because certainly I would say my guess is the way I think of Sierra is we have the best product, but because we have a founding team that's sort of been there and done that, we're not like a hugely risky bet.

1:05:27We know how to work with large companies. We're not going to show up and be learning from first principles how to engage with a large bank or a large health insurance company. And that's a unique value proposition because right now, most of the incumbent technology platforms, their products don't really work. And so you really want to use a best of breed product right now. And I think our value proposition is we are a best of breed, I believe the best of breed product in this space. But we also know how to like work with complex companies, you know, we're going to show up with as many people necessary to help you be successful, as opposed to just throwing a product over the wall and saying, good luck to you.

1:06:04And so that I think, you know, certainly our reputation helps, but at the end of the day, it's the quality of the product that really matters here. And just, you know, think about it this way. If you have 400 million phone calls a year, the difference between having an AI answer 100 or 300 million of them is probably measured, you know, and, you know, at least eight or nine figures, right, in terms of like the impact on your business. And so we always start with the product and, as you said, take advantage of our unique advantages as a company, which fundamentally means we know how to work with you.

1:06:36It's sort of interesting. It just turns out that there's lots of different stakeholders between a business team, a technology team, compliance, understanding just the constraints that a company is dealing with. We just try to show up with a ton of empathy and show up recognizing this isn't just a technology problem. It's sort of a kind a bureaucratic word, but it's a change management problem. You know, how do you get from point A to point B when your auditor and your regulator is scrutinizing everything you do? Well, we understand that. Like, we're going to help you with that problem, not just the technology problem.

1:07:09What do you think the steady state of commerce interactions look like? Because I imagine that companies will be using AI to interface with their customers, but then customers will be using AI agents to negotiate on their behalf and select products and without leaking too much of the OpenAI roadmap, I think we're all pretty convinced that some sort of agentic commerce thing is going to happen. I'm already seeing pop-ups for links to products. When I search for things, I'm going to be able to say, hey, go get me the best option, the best price. Is the future voice agent talking to another voice agent or do they just interact at some lower level API level at some point?

1:07:53Like, how does that all play out when you get agent on agent warfare? I think you're right about your intuition. But it's also, as you were sort of alluding to, the future is not perfectly clear, even for people in the middle of it like me. But I think in general, I think a lot of intent has already gone from search towards ChatGPT, especially for the more considered the purchases, the more you're likely to use something like AI. You know, when I was traveling this past summer, I used ChatGPT to plan the entire trip. If you think about something serious like purchasing a home or, you know, I've talked to so many friends who are using ChatDB to figure out how's the school district in this area?

1:08:31What neighborhoods are good? You know, and you think about that just from a commercial standpoint, that is extremely qualified intent at the very top of the funnel. And so much research is already being done there. But as you alluded to, the thing that's not happening is you're not kicking off your agent to go fulfill that transaction. It sort of stands to reason to me that that will likely happen at some point. And then the question is, for a lot of our customers who are sort of on the other end of that is, how do you want to show up in that new world? How do you ensure that your goods and services show up in the right way when you're engaging with an agent?

1:09:06How much do you want to directly engage with consumers? How much will you be able to directly engage with consumers? I don't know all the answers, but step one is make an agent. You know, step one is, you know, make an agent so that whatever interagent protocols emerge, you can be present and do business in that new world. And that's one of the main value propositions Sierra provides its customers is like, we don't know exactly where the world's going to go. But a prerequisite is to technically be able to interoperate in that new world. I think it's going to impact different industries differently.

1:09:37Do you think the form will generally die? where there's certain businesses out there that sell, basically sell product. Let's say like a company that makes aircraft, right? If you go to their website, you're not just like browsing inventory and checking out. It's like you submit. Yeah, you're not adding to cart. Tesla's likes to do add to cart buttons on, you know. Most, many companies out there. Let's say somebody. It's contact sales. Aviation enthusiast wants to buy a small plane. They've got a contact sales. And then like the second you get to the form, so much like so many leads just die at the form.

1:10:13And something I've been doing with LLMs is just ask, talking to them about like the pricing of different products. Because I don't want to go to the form and like get a sales rep to call me. I don't want to be waiting. I don't want to just get a call back at a random time. I don't really want to go through that. But but I want to kind of understand. And so it feels like the form is something that fundamentally could just potentially go away. And so you go to a website, you talk with an agent, maybe you give them your information because you do want to hear more. You do want to talk to somebody at some point.

1:10:42But it feels like that could be going away. I agree. I mean, if you just think about qualifying a lead, which is basically what those forms are, why not ask follow-up questions? Why not collect enough information so that when that great salesperson does follow up, 80 % of the work is done and it's just the human touch part of it? you know, agents are already doing outbound sales, debt collections for credit cards, processing payroll for small businesses. These agents are already doing sales. And I think that, you know, the best salespeople, the reason why they're the best paid people in most companies is because they're worth their weight in gold.

1:11:20But what about the median salesperson? And my guess is you can make an AI agent that's actually quite a bit stronger than most, you know, just because you're standardizing the best practices really, really fast. You can run A-B tests on AI agents. It's hard to run A-B tests on people, you know, and there's so many advantages, as you said, to not only just forms, but really thinking about agents as not stuck in one part of your customer lifecycle, but really managing the whole thing. And let's just take, I don't know much about airline, you know, buying an airplane for a hobby pilot, but imagine this AI agent, sort of the concierge to the whole experience.

1:11:58So you're browsing around implants, you set up an appointment, you know, what if it's collecting a bunch of information before you get on site after you, I don't know, you test drive an airplane, but let's say you do that, you know, afterwards it's asking you questions, maybe helping you with financing. I think that's the future of where these things are going, which is whether or not a person's involved, it's a concierge helping you through the whole process. Yeah, there's something about, you know, the consumer being proactive, engaging for information, but then using the airplane example, let's say they don't have something in stock.

1:12:32The agent, at what point are you guys already thinking about agents being proactive and reaching out to a customer after they already engaged? I imagine a lot of the core of what you're doing today is just reacting to customer needs, inbound questions, things like that, or problems with a product, whatever it is. But then at some point or another, the agent is going active and actually going outbound. But how do you think about that? That's exactly right. I almost think of it as like Maslow's hierarchy of needs of agents. Step one is answer the phone when your customers call. Step two is can you actually get in front of customer issues?

1:13:12I noticed the Ramp logo on the top right. They're one of our customers. And they have the coolest implementation of our platform. it's not only doing service but really I love their Ask.ai function and their product it's just one of the coolest experiences and it's not service it's basically a conversational experience around one of the best design technology products in the world and I think Ramp is obviously one of the best engineering teams in the world but can every brand do that? Can every brand make their agent proactive offer a conversational experience? I look at if you've ever visited like Brazil or India and you look at the impact of WhatsApp, it's like, well, can you be present there in a way you weren't before?

1:13:56Well, yeah, you can. What an awesome opportunity for agents. It's new channels. It's proactive. And this is really where we're trying to guide our customers and I think, you know, kind of represents the future of our platform in a lot of ways. The latest round, I think it's$350 on$10 billion. That feels low on the dilution side. How did you think about the amount of money you needed to raise sizing that? I mean, you've been part of so many interesting financings throughout your career. How do you land on that? And what does it say about the usage of that fund? How in capital intensive your business is?

1:14:33The state of the markets right now? Just take me through the thesis for the round. Yeah, I always think of rounds in pretty simple ways, which is what valuation are you raising at? what do you need to fill in that valuation so that if you need to raise another round or if you're going public that you've more than filled out that valuation with fundamental business metrics like in your revenue and growth rate. And then you say, okay, how much capital do you need to reach that milestone? And so that's kind of how we reached it and that 350 number. it's obviously there's a lot of precision involved, but it is a bit of an art form because you're making a lot of assumptions about the future, which is what are the capital that we need to reach the revenue scale where we can reach the next milestone in our business.

1:15:25So that's how we think about it. I'm pretty methodical about it. I think a lot of entrepreneurs make the mistake of raising too little at too high of a valuation and they end up sort of walking sort of a zombie company just because they don't have the capital they need to reach the next milestone, but their valuation is such that they've sort of priced themselves out of the market. So that's the main sort of thing I'm paranoid about, and I feel really good about how much we've raised. But more than anything, too, I feel great about the investors we have around the table. Benchmarks here at Green Oaks, Neil Metek at Green Oaks is just remarkable.

1:16:01And so love our board. And in particular, I think about who do you want to be on the journey with to get to the next milestone and really happy with the folks we have around the table. How are you thinking about IPO timelines just generally? Do you think Sierra is going to be another Stripe or Databricks where you just stay private forever because you can and it probably allows you to, whatever, take more risks, think long term, et cetera? or given that you've been, you know, done your tour of duty on the public side in the past? I'm curious how you think about it. Yeah, I don't think we're, Clay and I've talked a lot about this.

1:16:42I think our intention is to eventually go public. You know, we, there's a lot of advantages to staying private. I also think there's a lot of advantages to sort of having liquidity for your employees, you know, that access to capital. you know I really admire the way I know the Stripe founders pretty well and you know they're in a really unique position I respect what they're doing I think we'll probably take the more traditional path though it's far enough in our future you know that's just more our philosophy because I think that's where you're going we don't have reticence to doing it eventually and we'll probably more take the traditional path that makes sense how do you think about the Sierra fundraising process versus what's happening at open ai is it just so much simpler to underwrite have you have you learned it's not it doesn't feel like it's just add a couple more zeros like the open ai stuff it seems deeply complicated is it refreshing to have more of a simpler business to underwrite or are there at least lessons that you've ported over have you learned something from open ai that you brought back i'm interested like c corp traditional equity financing it's pretty works pretty well.

1:17:52Open eyes are just a really, I mean, it's a one of one business, as you all know, really well. The thing that's just so different about building the AGI versus building an apply AI company, they're just so different from a capital standpoint. You know, as Sam has articulated better than I could, AGI will fundamentally, in some ways, be driven by compute. And so much of the capital strategy that Sam has articulated is really, how can we have the best compute infrastructure and the most resources available for both training and inference, especially with these reasoning models, and meet the scale of demand there?

1:18:33And when you have something that capital intensive, it just changes your approach to fundraising, partnerships. and I really admire what the management team of OpenAI has sort of orchestrated there just because I think it's the company that is best positioned to essentially build the best computer, the largest computer in the world and have the best research team, which I think if you're saying what are the ingredients to winning in that market, I think they've checked both those boxes. In some ways, the applied AI market, I'm not sure other founders in the space would agree. I don't think we're like an AI company.

1:19:10I mean, we're using AI to solve business problems. Well, and that's what I love about kind of the whole approach is that we talk to a lot of companies that say we're an AI agent company or we're an AI company. And then you look, you ask a few questions, and you realize you're building enterprise software. There's a bunch of things. There's no shame in that. And there's nothing bad about that. We love enterprise SaaS more than anyone. But if you forget that and you get fixated on, we're just building agents, it's like your customers will eventually ask you for a feature set that looks like traditional software, although maybe it's a different workflow.

1:19:49Or you might burn 20 million of capital training a model because it makes you cool at a San Francisco cocktail party and lose sight of why your customers are hiring you to do the job that you're doing. And so to some degree, as you said, We're unashamed of being an enterprise software company. And more than that, you know, I think the way we invest our capital is very different. Thank you. Thank you. How do you what's your framework for understanding SaaS companies in the public markets today? What percentage do you think will be? It feels like everybody will have to transition at some point from traditional seat based pricing to value based pricing.

1:20:28You guys have the benefit of starting out with value based pricing. and saying, how many phone calls do you get? How many individual reps do you have? What percentage of that can we augment or replace? But it just feels like doing that transition on a quarterly reporting cadence feels completely miserable. Yeah, my high-level premise is it's easier to transition your technology than it is to transition your business model. And it's especially if you're public. Just because if you just think about, let's say you have your, I'll just pick an ERP company and you have an enterprise license agreement for some number of seats, you know, and or, you know, ITSM or something like that.

1:21:18And then you say, OK, we're going to make an agent to automate 70 percent of what this platform does. That's easy in theory, but what happens in the next renewal, like in six months? Maybe your AI agent platform isn't quite mature enough to actually make up for the difference in lost seats. So maybe your sales rep who's incentivized on increasing the size of that contract goes and it says, you should buy both. Buy the same number of seats and an AI agent. And all of a sudden, the CFO of that company is like, wait, I thought this was supposed to reduce our costs. What's going on here? And so I think all these incumbent SaaS companies have an opportunity to come out strong in this market, but they have to really transition their business model and their technology model at the same time.

1:22:04And anyone who says it's easy has never been a public company CEO because it's really hard, you know, because you end up having to tell a story to your investors, to your customers, to your employees. You have to go through a trough of despair, probably. and you know if you look at Microsoft's transition to the cloud and and it was really complicated to go from Windows revenue to Azure revenue now such as you know a hero for having done it but at the time you know there's a lot of people including me who wrote them off you know and and then afterwards you're like wow props to them for making that transition and coming out so strongly but for every Microsoft there's a Siebel Systems who didn't make a transition you know and for Those of you online who don't know Siebel, that was the company that Salesforce beat in the transition to the cloud.

1:22:50And now you probably don't know their name because they didn't sort of make it through that transition, despite being the market leader in the on-premises software era. So I think all these companies have an opportunity, and I think that it really takes leadership. I think it's very hard to do as a public company just because it's very hard for investors to sort of see through the quarterly earnings to see. Because you can, it's just hard. It's just like, is this company mid-transition to something great or are they just dying? You know, and like, it's easy. It's like, and a lot of it, a lot of it is ambition and storytelling.

1:23:26And so it's just a complicated time in software. I think it's really going to come down to leadership, both like stakeholder management and technology leadership for companies to make this transition. Are you getting a bunch of calls yet from AI application companies that haven't found product market fit and need to find a soft landing? Or do you think that's a few more quarters out because those companies are still well capitalized? We are. There's definitely, you know, just given the amount of revenue growth for the ones that are working, it's pretty clear the ones that aren't. There's some good teams and good technology there, too.

1:24:06So we try to look if it makes sense out of all of them, if there's a good team that could contribute. When did that start? About six months ago. Okay. Yeah. I assume it will accelerate. Yeah, we're having Hamant from General Catalyst on the show Friday. What's your reaction to his sort of hot take? I don't know how much it was taken out of context, but this idea that triple, triple, double, double, double is kind of dead. You need to be 10xing every day. You're not interesting unless you're 10xing. Is there something there where like the power law is getting steeper? There are more winner-take-all markets, more faster growth companies, but then also just some that are triple, triple, but then going to be nothing?

1:24:55Yeah. So I have a complicated opinion on this one. So first I'll say, the faster you grow is sometimes correlated with a lack of a moat, just because what enabled you to grow fast might enable your future competitors to grow there as well. You see this in some social services as well. You'll have these social services grow from zero to whatever million users overnight based on a social mechanic. And very few of those have ended up durable. Some of them ended up incredibly valuable, like TikTok, but there's the clubhouses of the world. Yeah, but TikTok also spent billions of dollars on user acquisition.

1:25:34Exactly. Exactly. In enterprise, I do think that growth is greater in agents, in part just because there's more value in agents than there is in software because you're, you know, essentially doing things like that were people were doing before. So you're, you're just achieving more valuable outcomes than just slight productivity enhancement. The reason I'm cautious on it is I think growth can sometimes be artificial, you know, so you can basically juice the numbers. And, and the question is, are you creating a durable business? And, you know, we talk a lot about like in our board meetings just about first mover matters and there's definitely like a green field right now but what matters is where you are 10 years from now so like are you creating a machine to make happy customers at scale and so many of these businesses can grow to like you know 50 or maybe even 100 million in ARR and plateau and you saw this with a lot of different businesses actually I mean it was and so the question is what is your addressable market what are your product advantages?

1:26:34What are your go-to-market motion? And can you scale to add to the next zero? Can you grow to the next order of magnitude? And the only reason I think he's a smart, I thought it was a smart point that growth has definitely accelerated in this world of agents. But as an investor in particular, you're really worried about, can they get a billion in revenue? Can they get 10 billion in revenue? And it turns out that growing really fast to 20 million is like loosely correlated with that. You know, it certainly means there's product market fit. But the question is like, was it with a very small niche of startups?

1:27:07Or selling tokens at a loss and you're subsidizing it, right? Yeah. So I'm a huge believer in quality annual recurring revenue and very happy customers. And my view is if you can do that quickly, like we have, that's great. If you had someone growing more slowly, but it was very high quality revenue from very happy customers, I'd probably bet on them over just the fast growth rate because to me that's a more durable business over time do you think do you think x like just raw execution can be a moat because some of those things you were saying earlier it's not like you guys have access to different llms that someone else in your category might not but it's just this kind of like know-how of selling into the enterprise and being able to do that at you know just with with pure excellence that feels like it's separated you guys from?

1:27:57I'll say it's a, I think it's, I think it's all of the above. You don't want the best technology, best product, best go to market. But I think right now, because the technology is changing so fast, execution compounds over time, you know? So if you had a great product today, but you weren't executing well, you will not have the best product a year from now. So that's where I think execution is maybe the main thing we focus on because it's basically the pace of innovation in your product and the pace of reaching and making new customers successful. and we want to have the best pace because I think what we want, if I come on this show a year from now, I want to be farther ahead of our competitors than we are now.

1:28:33And that is all a function of execution. Last question I have, I know, I know we're a few minutes over, so hopefully you don't have to jump, but how, how are you thinking about the evolution of the cloud market broadly? You have companies like Oracle waking up and getting extremely aggressive. You also have a bunch of NeoClouds, I'm sure all of these different players are calling you every day, hopefully not sending you AI chat GPT generated cold emails. But I'm curious, like, you know, where you expect to be and what kind of vendors you expect to be using on that side over the next five, 10 years.

1:29:11Yeah, the Oracle story is pretty incredible, isn't it? I don't like who would have predicted What's your backlog? It's amazing. Well, it's really impressive, too, just because it sort of shows you, you know, sort of like a founder-driven big view of the market. It's super impressive. And I just like kind of I just I love stories like that just because it's like just like I mentioned the Microsoft story where so many of us have written them off. And like it's just super impressive, the level of execution and clarity that Larry brings to that business. You know, broadly, I think the things that change in cloud is just the shortage of supply of GPUs and the demand for compute.

1:29:51And I think that is bringing a lot of, you know, there's a geopolitical angle to where are these data centers built, data sovereignty, all of that as well. So I think we're in this new era of cloud infrastructure where, you know, the capex is getting larger. You're seeing like the market crediting boldness in a lot of ways, because if there's truly a scarcity here, you know, and I think Sam has been articulating, you know, it's like whoever has the best infrastructure will really contribute to this world of AI in different ways. And so I do think we're definitely in a new chapter there just because it's like, you know, different players are making different bold bets.

1:30:31And one of the things that's been really fun to be involved with OpenAI is just seeing sort of OpenAI's influence on that as we try to, you know, pursue our mission of ensuring AGI benefits humanity. And a huge part of that is related to cloud, right? It's if you don't have the right computer, you're not going to achieve that mission. Makes a ton of sense. Well, thank you so much, sir. We took you a couple minutes over. Congratulations, and we'll talk to you soon. Yeah, thanks for having me. Cheers. Have a great rest of your day. Let me tell you about graphite.dev, code review for the age of AI.

1:30:59Graphite helps teams on GitHub ship higher quality software faster. We are joined by Joe Lonsdale next. We are going to welcome him from the Restroom Waiting Room into the TDP and Ultradome. Joe, how are you doing? Look at that audio-video quality. You look fantastic. How you doing? You look younger than ever. You're aging in reverse. The camera is helping as well. The six young kids, I guess, are keeping me young, running around with them. I won the half-year-old. It was your competition for this. He was angry. I wouldn't drive him in the car. Do you have the most kids of any top GP? Anybody got you beat that you know?

1:31:39I'm sure some have more. My wife and I are pretty traditional, so they're just with her, which makes it tougher to coach two times. We're going to get there eventually. I wanted to ask you about this Trey Stevens post. He said he found a Palantir-branded iPod Touch Generation 2 that he had custom engraved back in 2009 while digging through a box at my house. Did you get one of these? Do you know the story of this? Is this a Trey Stevens exclusive? iPod Touch from 2009. 2009 actually I I have a lot of early palantir kind of gear Most mostly I have these like giant things we used to carry around like these big computers and whatever their cases are called The music usually for guns we use them for computers.

1:32:19I'm not sure about the ipod pelican case You probably the pelican there's like the really serious ones and I the very first time I brought to DC We had a rental car It was like 2005 and I went to the address we were gonna stay at and then there's this very nice black man And he's like, you boys are on the wrong part of town. And I'm like, oh, what's this Northwest Southwest thing? You know, so if you know, you see. What was it like early in the Palantir days? Gary Tan tells the story about you're not selling Holiday Inn software. You got to stay at a Four Seasons. I forget exactly what hotel chains he mentioned.

1:32:53Is that apocryphal? Is that true? When you traveled for business, would you stay at the top tier place? I think the very top of the company would stay at like relatively nice places we weren't upgrading to the nice rooms but there is probably like six seven hundred dollars versus two hundred dollars a night sort of thing which you know in retrospect I think there's multiple ways of running a company I do think when you're meeting important people you had to make sure you're not meeting them at the holiday and that would probably not be a good idea for trying to network with the very top people in these in these contexts so I think that's listen I think a lot of people waste too much when they're starting companies, so I don't want to set a bad example.

1:33:29I was actually a top trader at Peter's Fund, and then I brought on my friends to start this with me. A lot of them I brought on to the hedge fund, and it turns out they didn't really like finance, and so we ended up building out this company. And the fund, part of our comp was you basically spend whatever money you want if you're one of the top guys there. So it was a very kind of bad place for me to start from in terms of being spoiled already, if we have to be honest here. Thank you for being honest. Where should we start? There's a bunch of stuff. I mean, we were just talking to Brett Taylor about this.

1:34:00Hamon over at General Catalyst was saying that, like, the world has changed. Triple, triple, double, double, double for revenue growth is no longer interesting. The only thing that's interesting is 10xing revenue every year, something crazy. The power law is getting steeper. I would love your take on just what it takes to be interesting as a startup in the era of AI, where if you are in the right market or you're indexed to the right thing, or maybe you're selling tokens at a discount, you can get to a crazy revenue run rate very, very fast. What are you watching out for? What are you optimistic about?

1:34:34How does all that play out? I mean, listen, it's definitely a very different world right now because there's so many new possibilities. I think we were looking at something. I want to out the person because I'm not doing it, but I was looking at something that was like going zero to five this year, and I thought it could easily have gone to at least 15 or 20, if not more next year but i thought the team was kind of like a bb plus and it wasn't like the very very best people and and that is just insane by the way that that tooth those two things combined you can grow that fast with something that was like maybe not the very best people having trouble hiring and i mean the bar is just a lot higher right now and by the way and to be clear that that round will probably get done at oh yeah like probably 100 million at least exactly and and i'm I'm pretty sure it will.

1:35:15And it's like, yeah, it's like it also raises the bar for everything else. We were looking at this really cool thing in bioinfrastructure that could be very important for the future of humanity and for what it could do in the bio world. And it's just going to take three or four years. And these things in bio, as you know, are just hard. And it's like, you know what, even though this is so important, like we have to on the margin, you have to be in the green fields right now with the very, very top AI teams are just so far ahead of everyone. And so, you know, I do think there's like, I mean, what Hamas says, to me, it's not as much about like exactly the metrics.

1:35:47I think some people like live in metrics and they should. And one of my partners does too. And, you know, my view tends to be like, who are the very best talent in the world? Who has the very best cultures? I mean, you see something like Cognition with, you know, 20 gold medal winners or whatever. Scott used to work for me to add a part after winning, you know, programming, you know, competition globally three times in a row. And like, obviously, that's like the extreme, but the extreme works in AI, right? So I think for me, it's more about what is the very best talent? And then let's build something that no one else can build.

1:36:16And of course, yes, if it works, it's going to grow insane speeds. Tell me the story of finding Scott, identifying him, recruiting him. It seems like one of the greatest talent acquisitions of all time. You've developed a business relationship over a long time now. How did you even think to back him early, work with him early, any of that? I mean, to be honest, Scott is a really special and amazing person. But I hired most of the first 200 people at Palantir. Atapar had eight gold medal winners along with Scott, who came in at the same time with my friend Vlad Novikoski, who was helping me at the time.

1:36:55Amazing guy as well. I'll just give away all my secrets here. People are going to go get his help now. But I've got to hide some of these names from you guys. This is dangerous. But Scott was one of eight that year. And I think another one of them was Alex from Scale. And he was an intern there too. And it was actually very funny. I mean, both amazing people. I think Alex left in a way where we didn't stay as close. And Scott left in a way where we did stay a little bit closer. But even Scott, I think he was running around the world. And I'd been in touch. I actually invested in his new thing after that, which was not cognition.

1:37:28Just a little bit because I thought it was a terrible idea. And by the way, this is a very good idea. When you have the smartest people in the world, you should invest in their terrible ideas. just as a sidebar we went back to 2022 and we met these people who were doing something really dumb and like we kind of want to give money anyway and we didn't it was out of mit and it ended up pivoting into cursor there's other people there's another one where it's like i have this chain from yesterday where coley my partners like forged it to me and it's like these guys are really good they're raising seven million at like 40 posts it's 2022 and i'm like oh god it's just we all think it's a dumb idea i'm like we can't we can't keep doing dumb ideas guys and it turns out they pivoted and they're raising a five billion post right now i'm like oh god we just need to like i just need to give smart people money this is like it's just like oh what you have to do is so annoying but yeah there's there's like the the bell curve it's like you know i'm one end just give smart people money on the other end give smart people money and in the middle it's like oh what's the idea what's the attraction yeah exactly exactly just like even though it's like actually actively a bad idea like give them money and by the way even the thing that was a bad idea like he's in charge now that and he's pivoted to something that's working it's probably it's also a unicorn so it's like they figured it out so it's just the same like it's just all about talent and so i lost i was touched with scott a little bit and i got back in touch with him unfortunately i've been in the last few rounds of cognition i wish i wish i'd say it as his best friend it's hard to keep in touch with all the smart people you admire flying around when we're busy but it's he's someone i really enjoyed getting to know and you see someone he's really grown as a leader more than i expected sometimes you know people when they're like 18 19 20 21 and and you get one impression of them and you're so impressed but but like was not like the leader personality and now he's just such a like a fierce leader personality plus plus you know the global champion thing so that's pretty amazing he's doing he's doing good work yeah we read a wall street journal article about the latest imo gold medalists and i was surprised because i was uh texting with a gp at a big fund and they were not tracking it they were not our joke on the show was like every single one of these kids mentioned in this wall street journal article that beat both DeepMind and OpenAI and got the full score on the IMO.

1:39:33They're going to be getting term sheets, but people weren't tracking it as closely. Yeah, it seems like when you were hiring these IMO winning folks back with Adapar and maybe at Palantir, the meta at the time was like, you want to invest in the kid who's like sleeping on a mattress on their floor, like sleeping in a closet in San Francisco. Maybe it was like somewhat of a different, different track then even though you can see it becoming consensus now because yeah scott has you know shown no it's it's scott scott and alex and a few others but listen i don't think there's too much alpha left in that in the sense that like when you have all these people at open ai and at anthropic and they've raised infinite amounts of money like you know it was a problem for me in palantir in 2008 2009 when meta and google doesn't even call it met at the time started giving 100k bonuses to these kids out of the top schools because we had like a monopoly on on just the playbook for the top 20 universities and they started just like giving them like 100k bonus that's insane and like now it's literally like 100 times that right so it's and it's for these i mean there's maybe there is some alpha i'm not gonna tell you all my secrets in this show unless i'm required to you tell me i don't know oh you are required but it's like but it's like it's like i don't think if you want to tell every investor yeah yeah yeah you can text us after the fact the real secrets and we'll keep it private.

1:40:57How are you, um, how are you thinking about, uh, you know, you've obviously had numerous plays in traditional enterprise software. It feels like a lot of public SAS today is, has this challenge of transitioning from, you know, the traditional seat-based model to something more like Palantir, the sort of like value-based pricing. How are you thinking, you know, are, are you just bearish on public SAS broadly? Do you think some of them are going to be able to transition brett taylor just said it's easier to you know do a transition your tech your business model your tech than your business model especially in the public markets it's it's so funny to live in this world because it's like i woke up in opposite land where i was like people just beat the shit out of me for like a decade for doing the wrong thing the wrong model and suddenly suddenly like how could everyone do that model that's great i guess uh i i don't I don't actually know if everyone needs to do a change in model.

1:41:53Overall, I guess I think the more interesting question is like, which of these SAS companies owns important infrastructure and workflows that are juxtaposed to other value they can capture with AI and then also still has a technical culture to do that. So this is like a motion at Adapar I'm working on really hard. As an example, I started Adapar 15 years ago. It has over$8 trillion in the platform. the core infrastructure business is still growing. Infrastructure SaaS business is still growing at like 25 plus percent, just that core. And there's all these things on top of it starting to grow because it just touches everything when it provides value for wealth managers.

1:42:28And so I think to get that business growing 40 or even 50 % potentially, it has to like build really strong AI teams and workflows that take advantage of where it is. And I think there's going to be a bunch of SaaS companies that do figure this out in the next few years. And those are going to be really valuable companies. And then the ones that don't figure it out at all, there's probably some danger that they even lose the mode of their core thing to begin with because it's going to be too easy to copy or something if they're not scaling it. So I think, again, think it's like a power law thing where some of these are actually worth a lot more and then most of them don't have the talent to do it.

1:42:59Yeah. It's been fascinating watching the narrative shift from the foundation model companies are going to eat everything to the foundation model is going to commoditize. There'll be a couple application layer companies to the idea that AI would be sustaining in enterprise SaaS because you actually, even if you have a semi-technical culture, you don't need to train a foundation model. You can just go grab an open AI API and implement that. But there's this business model tension now. And if you're stuck in your business model, there's going to be a lot of startups that are counter-positioning against you, and you're going to be kind of screwed.

1:43:30How are you thinking about it? It's interesting because I do think if you own the customer really well, you could add these things much more easily. But you're going to have to go fast, or the startups are going to eat it away. There's so many areas where we're kind of seeing this right now. of record or some sort of database, some sort of relationship. This is right, like Socotra as a company I'm in, that's like a great system of record for the insurance space. And it took a long time to build. It's a bunch of palantir talent. They're really coming into their own system of record now, which is very slow.

1:43:55And then they're trying to add the AI in before all the new AI things that people are funding. And I think it's pretty interesting. I think the system of record has a very good chance of winning. But that's a question. Yeah. How are you thinking about macro broadly? I feel like everybody's, I mean, every asset, all-time highs. Gold, Bitcoin. I like the meme that says printer is coming. He's dressed up as Lord Stark of the North. Printer is coming. Yeah, but for your angle, it's probably generally better use of your time to think about the opportunities at the early stage today. You could incubate a company.

1:44:31You could fund a company. What is the world? I'm incubating a lot. I started and keep getting a lot more in 2021, partially because I was just so annoyed that everything was like insanely, unreasonably expensive. And you kind of have a monopoly on the first couple rounds, right? Yeah, we probably were too generous letting friends in. I'm always just like, I have these business friends who are just like much more hardcore than me and just like very sharp elbows and kind of even a little bit nasty. And I tend to just like everyone around me to make money. So I probably should have taken just only myself the first two rounds.

1:45:00That's especially Saronic. I should have done that. But whatever, we own plenty of it. it's fun to start things that are important for the country and that win and we make enough money for everyone. But no, we do do the whole like big first round ourselves now, just cause that's the right thing to do for the fund. And then like do a lot of the second round. And listen, I think building is still the best. I mean, even right now today, you guys tell me I'm seeing AI rounds where like, I'm really happy if I get 15 % in a series A of a really hot AI company right now. I think I got 16 % something the other day by overpaying.

1:45:29I think it was the right thing to do. And it's like, but, but I mean, I've seen a lot of things where I get 5 % or 6%. We just talked to Brett Taylor. He raised$350 million, a$10 billion valuation. That's not a lot of dilution. It's going to be hard to build a huge position in that company. And that's a one-year-old company. And these are happening all the time. I mean, he's special. But yeah, it's still a unique situation. There's a special. The specialist ones, I only have 2 % or 3%. I'm really happy because I think they could be$100 billion companies. Yeah, exactly. It's very weird. It's different than it was.

1:45:57Talk to me about the early stage startup market in Texas. Obviously, you were in San Francisco for a long time, buddies with Gary Tan, co-workers at some point. Gary's leading a revitalization of Y Combinator. Is there something adjacent? Like if I wanted to go talk to 50 young people building startups, the next stop in Texas, like where would I go? Who would I be talking to? so and i've been fighting for california to fix itself for a long time gary's an old friend obviously from our fraternity days through palantir and i think he's come more to my point of view on these things which i really appreciate he's doing a good job uh we used to argue a lot uh you know it's yeah there's a lot of history there don't he's did gary's gary's crushing it the the thing about texas there's a few things i'd say if you want to build like the very top AI startup in the cloud, application layer, model, whatever, doing something really hard and new there, you probably should be in San Francisco today.

1:47:01That's just where the talent is. It's a very strong network effect. I would love it if it wasn't, but it is. And it's there. So what do you do in Texas? Well, I think we have, like I mentioned Saronic earlier, I think if you're building in the land of atoms, if you're manufacturing, I think that's probably the most important company in the country for the U.S. Navy, as well as doing very hard things in robotics, very hard things in software around it. As applies to defense, we do have like 10 or 15 of the very best AI software people for that there. So I think what you're seeing in Texas is it's a really good place to manufacture.

1:47:32You have a lot of the best people, like Boring Company, a lot of the top SpaceX talent. Others are coming out here. I think you have a lot of things that go really deep in healthcare here. Obviously MD Anderson in Texas is one of probably the best cancer center in the country alongside Memorial Sloan in New York. and you have just a lot of other really deep kind of healthcare research, healthcare services groups, healthcare services. There's multiple, multi-billion dollar new companies in healthcare here, there. And I do think like when I was talking, Reid Hoffman and I was trolling because we have different politics.

1:48:03Like where do you test your like self-driving trucks? And of course he tests them in Texas. I mean, if you're trying to like do things, if I'm trying to like automate, one of our companies that I'm really bullish on is a bunch of ex-Waymo guys and they're automating construction with excavators and they're going to like automatically run all sorts of construction and quarries and other things. And of course they're doing it in Texas. So Texas is just like a great center for industry, for building things in the real world, for robotics talent, for defense talent, for healthcare talent. It has great AI talent, but it's not going to compete with like the next new AI models in San Francisco.

1:48:35San Francisco right now is just obviously an awesome place for investing for that type of stuff. My current understanding of the startup micro worlds is that you're still probably going to be doing a road show on Sand Hill Road or in San Francisco. When you're later stage, you're going to be talking to crossover funds in New York. You're going to be in D.C. if there's a lobbying component. But then you're going to need to build your company wherever makes sense. And for certain industries, Texas makes a ton of sense. and for a lot of things in the real world a lot of things in health care texas is by far the best place to do it it's it's it's listen i love you can go the government it's not like they agree with me on everything but you can go to them and you can talk and they're reasonable and they respect you and you go back and forth and you don't have to be republican or democrat to do it by the way like they actually it's like you can't they actually care about business succeeding it's very cool it's like it's a new idea to me to have a government that has a group it's like wow how do we help business succeed it's really great and then you can get things done and it's reasonable so So I definitely like building things here.

1:49:36Without giving away too much alpha, feel free to put people down the wrong path. What do you think is the most under-hyped trend adventure right now? Obviously, AI is just drowning everything out, but feels like it could be at the beginnings of a new robotics wave. But how are you seeing it? Yeah, I'm really interested. You say robotics. I'm really interested in the stuff going on in the real world using AI, which I think is like a whole new area that it's still early, but I think you're going to, I think it's just like a much bigger part of the economy than people realize. So for example, if you can do a certain part of construction automated with bedrock, like I think people don't understand, like that means you can run a quarry better and there's a hundred billion dollars of quarries.

1:50:19And if you can make quarries have higher cashflow, that's worth like it's worth tens or hundreds of billions of dollars. I mean, it's just like, and there's like, there's like so many things like this where the real, like the real economy is like 85 % of the capital and that's all about to be transformed. And I think people underestimate like how much we're going to need in credit to do that. How much, just like, just like, just like how much big stuff's going to happen. You can create in the economy in the 2030s there. So, you know, even stuff like Boring Company and like, and like automating that and making it much, much, much better is obviously using modern AI and other modern technology as they iterate with engineers on it.

1:50:51And it's just really impressive how much cheaper you could do things, how much better you could do things. So, so I think that's under hypes and under misunderstood just because it's maybe takes a couple of years longer to really get out there and start spreading quickly. but I think it's just like much bigger than people realize. And that's one of my favorite areas. I guess the other one, you know, I think in general, people spend too much time on like the super giant companies. Like people call them the infinity stories or the things that could be worth trillions. And I talked to some of my friends who are running very big funds and they're really only interested in something that could be worth trillions of dollars and have all the top talent.

1:51:23And it's almost like, I think one of them is a good guy, but he said to me, I want to be in the room with the powerful people doing the most important things. And I think that's like an instinct of a lot of our top investors right now. And actually, I think there's going to be literally like a thousand, like five to$50 billion companies and like so many niche application areas that do need to build out the workflow. Do you need you to build out the operations? Do you need to kind of go after the different service areas of the economy? Yeah, what about it? It feels like, you know, when you look at like Palantir, Adapar, OpenGov, it feels like you've had a lot of success with companies that were very under, you know, not didn't have a ton of like crazy hype and did the kind of like behind the scenes heavy lifting to get to the point where you could have a massive outcome or be really critical to an industry.

1:52:12are you do you think there's not enough of that like it feels like a lot of every company gets hyped you know incredibly quickly now if they're talent dense and are you still like trying to find opportunities where you can just kind of quietly build in a category for four or five years yeah i mean i think i am building a bunch of stuff that's not really that hyped right now that's using ai that's that's gonna change these different categories and there's this like when you look at the talent out of Palantir, it's a lot like the talent out of PayPal, except on a bigger scale in a lot of ways because Palantir have many more years to compound with the very top talent.

1:52:48And so you have things like, I just like random things like Candid is just like, probably do, I don't know, they want me to get their numbers on this show, I guess, but it's like they're doing healthcare billing. They're growing really fast. They're going to do, you know, they're going to get into the billions of revenue like within the next few years probably. And it's just like, I guess, you know, it's a$280 billion space and there's not that much hype around it. And it's just like, wow, this could be absolutely, this could be like$100 billion company in the early 2030s, and almost no one will have heard of it.

1:53:17And it's like, there's just so much stuff like that right now, which is pretty fun. So, I mean, I think I'm pretty bad at hype. I think Palantir eventually got hyped, like despite us being bad at it. Which is how people... Well, yeah, I'd say like Ceronics, an example of something that, you know, got a ton of hype, ton of capital, ton of attention really quickly. That did. I think we did get a lot of the right talent, and we were kind of the right place at the right time where there's no one else who was doing this for the Navy and no one else who had that density of talent and operational. Dina's just such an amazing and such a fast CEO, and he brought in kind of other co-founders who were also really top people.

1:53:53So I think like Andrel, they just had so many of the best people, right place, right time, going hard. So I guess that got hyped earlier than I would have expected. That's fair. But most of the things I'm doing are not like that. Yeah, that makes sense. Do you think there's any positive knock-on effects of the insane CapEx that's going into the AI foundation model world? Like we saw Sam Allman say he's going to build 10 gigawatts. Mark Zuckerberg saying he's doing a gigawatt. Elon just did Colossus 2. It's a gigawatt. And the interesting thing to me is like this is in some ways like the, like, you know, Saronic, Palantir, Andurl, SpaceX.

1:54:29Like those are American dynamism companies. But if you're talking about reindustrialization and just making big things in America, doing stuff that requires government approval, like Elon's, like Colossus 2 across three different states, like that feels like that unblocks some crazy stuff. And then it's going to be easier to work on health care or work on automated trucking. This is very, very good. No, you're 100 % right. whether or not it's a bubble and whether or not they should be investing this much money I'm not even going to comment on because I think it's just like really good for America that they are like practicing building these things at this scale you're 100 % right they have to fix some of the permitting stuff they have to create all this infrastructure we can use now for advanced manufacturing we can use the advancements to make things cheaper to do at scale for building related things in America we're going to build a ton of stuff on top of this I mean for me I just love it because you know I talk about the different levels of AI investing where it's like zero is energy infrastructure one is chips two is data centers three is the models four software infrastructure five is apps and services and like i'm mostly doing things at five and some four but by people doing all those stuff at the bottom with insane amounts of money that makes my life very easy it makes everyone else's life easy too so this is great for america yeah for so long we've heard like oh america we can't respond to dji gopro got roasted oh we can't respond to unitree maybe elon will figure it out and it's like i feel like we're getting close to like, no, you can actually marshal$10 billion of capital, set up a massive facility in a couple months and get approvals and do the thing that you need to do to build a ton of stuff in America.

1:56:02I think this is really, really good. I'd love to see a wave like this for bio in the next decade as well, because right now our bio regulatory apparatus is completely effed up. There's been some really strong hit pieces in the journal recently. There's some huge messes there, but it's the same sort of thing where if you could like take this energy, apply it to that infrastructure and hopefully apply it to some breakthroughs we're going to see there to own it. But the fact that America can still do this, these companies can still do this, it's awesome and it makes us very bullish. You had a pretty viral post yesterday responding to Nick Huber who was taking shots at the great citizens of Austin.

1:56:38He's going after Austin. Yeah, yeah, yeah. And it turns out the photo was from after school ended so very clearly. I could totally tell it was a joke photo. Yeah, yeah, yeah. But I wanted to ask something. You said, sorry, Nick, but many of these people relaxing in the nice weather in our hometown are the spouses, mistresses, and children. Where did you get number two? Number two there. Where did that come from? His whole post was like, these people are lazy. I'm going to outsource them for$5 an hour jobs to the Philippines. And it was like an anti-American kind of like negative post. And then so my clap back is actually our U.S.

1:57:18employees are so wealthy and so successful and have built things so much bigger than anything you're doing that they can afford to have their families relaxing. And mistresses. The joke of mistresses is they're so wealthy that maybe they have mistresses. They're just relaxing in the sun while they're working because they're creating so much wealth. We don't need to outsource to third world slop. We can have great wealth in this country and great success and don't make fun of my city. So it was obviously a joke, but some people got really sensitive about that. Well, your masterful poster. Yeah, you used his own tricks against it.

1:57:49You did. You did. I was a strong ratio. I appreciate it. It was like, so it's a Jewish New Year, and I was like stepping out of the synagogue service, and I was just like, I probably should have stayed in a synagogue and been on my phone. It's not been on my phone. So it's dangerous when you give me like a little bit too much free time on a day. It's dangerous what happens. Oh, yeah. well never stop posting never log off joe we enjoy your posting we've enjoyed your appearance thank you yeah i would actually like i think x would like it broadly if every time nick huber posted you just quoted it and dunked on him because he's always leaving something open oh yeah i i worried i hurt him because he like has these really angry replies and i was i thought it was kind of a joke slap back but i think he was like pretty offended so if he's listening i i didn't mean to be offensive i appreciate his hustle porn even though i disagree with attacking america He knows what he's doing.

1:58:40He's getting in the arena. He's rolling around in the mud. He's going to get a little dirty. It's going to happen. But thank you so much for taking the time. Come back on anytime. See you guys. Have fun hanging with your kids. We'll talk to you soon, Joe. Have a great rest of your day. Let me tell you about Julius.ai. What analysis do you want to run? Chat with your data and get expert level insights in seconds. Loved by over 2 million users and trusted by individuals at Princeton. bcg zapier i got the name right this time uh did you see the car heart uh jacket that turner novak says if a vc shows up to the meeting wearing this you should sign the term sheet immediately this is a callback didn't bain capital do a car heart collab for merch and they got some backlash from that soul and valor allegations uh vcs keep it in the i mean there's nothing wrong with putting on a suit you saw joe lonsdale just came on the show in a beautiful suit nothing wrong with suit the, what's the Patagonia vest?

1:59:40That's very iconic VC attire. You throw on the, this looks like Han Solo. This jacket looks like something Han Solo would wear. It looks good though. Anyway, in other news, YouTube restored some suspended accounts. Ben Thompson has been breaking it down. He said that he gave kudos to the information for their headline, which is YouTube to re-allow accounts banned over COVID-19 integrity election violations. And Ben says this stands in marked contrast to other headlines that characterize it as people who spread COVID misinformation. And he's quoting The Verge on that case. And he says, it's true that there was some misinformation, but there's no question that not just legitimate, but in retrospect, true content was censored.

2:00:31and its creators kicked off the platform. Google, in that letter to Jordan, who is a Republican legislator, said that they placed the blame squarely on the Biden administration. Google is saying the Biden administration twisted their arm. And so this obviously goes back to the discussion about what's happening with Jimmy Kimmel. What is the role of the government in free speech? I think it's that the government shouldn't be involved at all, actually. and that no government should be involved in this stuff. There's been debate over what level of involvement. Find one person that didn't spread at least a little misinformation during that era, right?

2:01:11Whether you're a government, whether you're some schizos, conspiracy theorist, right? Everybody was - And that's what makes America great is that during the COVID era, we were suffering through the same pandemic that China was suffering through, but we supposedly had free speech and so we could talk about it and throw out crazy ideas and pitch all sorts of different solutions and debate them in the great online marketplace of ideas. And so this seems like a positive shift. You know, more speech is probably better. More free speech is probably better. Ben says his concern on strutectary is more narrow, and that is how tech companies should have approached challenges like these over the past years and how he hopes they approach them in the future.

2:01:57And so he says, Google and YouTube were obviously not the only tech companies under pressure during the COVID era. Meta CEO Mark Zuckerberg said the same thing a year ago. Another was Spotify and Joe Rogan, which Ben wrote about at the time. And he got a lot of angry responses to his take on how Spotify should have handled the Joe Rogan situation, the pressure on Joe Rogan during the COVID pandemic. And Ben says, but I think in the revelations of the intervening years, And I might add a lot of those headlines I complained about above proved my point. And most disturbingly, we have just this month seen some of the most discouraging and devastating consequences of losing the cultural value of free speech.

2:02:35Still, that wasn't the part of that update. Rather, I was writing directly to Eck and other tech CEOs, declaring an adherence to free speech was not a get-out-of-jail-free card in terms of avoiding criticism for content hosted on your platform. And in that light today, I don't have so much advice as I have an exhortation. Tech platforms need to lead the way in reestablishing the cultural mores around free speech and not just to get their get-out-of-jail-free card back. Tech is dominant culturally in a way it wasn't when the cultural mores around free speech were established, of course, centuries ago.

2:03:14And the most generous interpretation of their actions over the last decade is that they were afraid to assert their power, bowing to the wishes of the loudest voices in the media and in their own companies. In fact, the best way to avoid partisanship is to be the natural arbiter of all the platforms say they want to be, is to decline to take positions on all issues, but one, the importance of free expression. That is the key enabler of finding a path forward on all those other issues and is the foundation of a free society. That, in a nutshell, has been Stratechery's approach to politics, and it has served me well.

2:03:48I think it would scale just as well to the largest companies in the world. He says, take a page out of my book. It's working for me. Page out of the First Amendment. Page out of the First Amendment. Potentially the first page. the first page of the list of amendments. Did you see this post from Lachlan Phillips? He said... I did. Break it down. This is Coogan thought. The most toxic trait is the belief that instead of jailing Elizabeth Holmes, she should have been placed... We should have placed her under YC arrest under the watchful supervision of Gary Tan and forced her to rebuild her device over and over again like Iron Man for the next 11 years or until she gets it working.

2:04:28I love it. And this is what you wanted to see. Instead of the kind of shit posting that she's doing, it'd be great if she was like, guys, I'm figuring out kind of what went wrong before. Here's a more narrow solution that we could kind of create. Yes. I mean, it doesn't even need to be. I mean, I understand that from prison you are not going to be able to build a device that tests your blood. She's just not going to have the equipment for that. But she could be reading papers. She could be reading the analysis and giving thoughts that over time could look very good. Like you see it today with the news of Donald Trump and RFK are talking about Tylenol.

2:05:12that's being hotly debated. The administration is taking one position. Different media figures and health experts are taking different positions. I'm sure Andrew Huberman will comment on it at some point, right? But like she has the opportunity to read all the research, comment on it, and then we will see in a year or two or maybe more how those predictions panned out. And that's what we've been seeing with Martin Screlly where he has made a bunch of bets on, he's very bearish on quantum computing right now. We'll see what happens. We just get to wait and we will remember, wow, he was correct about his bear call on quantum computing.

2:05:48And so that updates us to think that he is an interesting thinker on the valuation of new technology companies. Whereas right now she's kind of just like leaning into this character of, oh, wow, she's posting from prison. I think that novelty will wear off. I think that novelty will wear off and we'll want to see something more meaningful. I mean, it's definitely broadly changed people's general feelings towards her. For sure. So. For sure. Anyway, Fall. Generative media platform for developers, the world's best generative image, video, and audio models all in one place. Develop and fine-tune models with serverless GPUs and on-demand clusters.

2:06:26Head over to tbpn.com slash sounds, which is brought to you by Fall. Printer is coming. Amit is investing. Writes from Besant. We are going into an easing cycle and we'll see a massive decrease in mortgage rates. Powell should have signaled 100 to 150 basis points of cuts, not just 75 points. We'll begin Fed candidate interviews next week looking for someone with an open mind. Looking for someone with an open mind. Open mind to low interest rates. To what we want. Potentially. We'll see. I would love to see mortgage rates come down. I would love to see the cost of housing come down. I still feel like the opportunity is probably in just building more housing, building a lot more.

2:07:07Love when the rates are low. Hate when they're high. It certainly helps. But of course, that tenure isn't really behaving. Yeah. It's got to whip it into shape. Jim Cramer with that tenure better start behaving. That's great. There is a fantastic profile on Jim Cramer in the news today. We should take a gander through this because it's a fun one. This was brought to my attention from, let me find it, from Joe Wiesenthal. He says, great profile in Business Insider of a guy who was on the Odd Lots podcast who wakes up at insane hours every day to cover the markets. Did you see this? So Jim Cramer is a busy man.

2:07:50The Mad Money host, known for his emphatic delivery and bold stock calls on the air, gave a rundown of his hectic schedule when speaking on Bloomberg Odd Lots podcast last Monday. Here were the highlights of his routine. And get ready to be mogged, Jordy. He wakes up. He kicks off his day at 3.15 a.m. 3.15 a.m. Do you know what time he goes to bed? 6. 11 p.m. What? He gets four hours of sleep a night. Just doesn't need to sleep? Does he nap? There's no competing with this guy. This is insane. We try so hard. I feel like we do a really good job. But we're putting up eight hours of sleep a night.

2:08:30How did you sleep last night? Go to 8sleep.com. Get a Pod 5 Ultra. 30-night risk-free time. I wish that I could. I got a 96 last night, but I slept for 7 hours and 41 minutes. That's about 4 hours more than Jim Cramer slept. I got a 97 hours, 23 minutes. Ooh, I win again. You win again. You, yeah, I just, I guess I got to go to bed. So, Cramer wakes up at 3.15 a.m. The first 45 minutes of his day are on, quote, a tight schedule due to his 4 a.m. workout. We're getting into the gym at 6.30 at best. He's getting in two and a half hours ahead of us, and he's on East Coast time. This man lives in the future compared to us.

2:09:09It's remarkable. Before the hour, he brushes up on the market by perusing the Financial Times, Bloomberg, the New York Times, the Wall Street Journal, and CNBC. I get two of those done. He's doing six. Okay. He's ahead of me there. Kramer also said he sifts through stock research in his inbox, combing through about 700 emails a day. That's remarkable. Amazing. I read everything I think is relevant. I'm looking for something to say. I have a memo that comes out, 10 things I'm looking at. I do one thing I'm looking at. He does 10. I try and just do the current thing and he's got 10 of them. He's an absolute beast.

2:09:42Kramer said referring to his weekday morning newsletter for CNBC's investing club, he said, so that's my overlord, he referred to the newsletter as. I'm starting to feel that the newsletter's my overlord. I respect it. Go subscribe, tbpn.com, throw your email in there. You will get the TBPN run of show every week, the current thing, my newsletter. then Kramer appears on CNBC's Squawk on the Street begins writing Mad Money his flagship show on the network in the mid-morning throughout the day Kramer spends a huge amount of time reading up on company news and earnings reports he finishes taping Mad Money around a quarter to six we saw him there doing that Zach Meyer in the chat says this guy does not go to the gym at 4am look at him I think it's more like he's not working out for physique I mean we got to hang a couple weeks ago we got to hang with Kramer for about 20 minutes.

2:10:34It was the end of the day. He had a ton of energy. He was incredibly locked in. He did have a ton of energy. This was at like probably five, four, five. Yeah, he's probably doing something low impact, but I like to imagine him benching three plates, squatting five, throwing up some 200-pound hang cleans, some clean jerk. PRs every day. PRs every day. Getting aerobic, getting athletic with it. Just throwing on a stringer. Yes, absolutely. Absolutely. Yeah, get a GigaChad Mass Monster Kramer photo up ASAP. I need to see that. He finishes taping Mad Money around a quarter to six and comes home. He'll go out to dinner around 10 p.m.

2:11:11and then say goodnight to his wife before tucking in. Kramer acknowledged that the intensity of his schedule puts a strain on his marriage. I have a great marriage. I put that out first because I wreck it every day. I go to bed at 11. She goes to bed at like a normal time. So that's when I really wreck the marriage. That's not the plan. I didn't set out to do that, He said, Kramer's routine doesn't seem to have much changed. In 2015, he told Business Insider that he followed a similar schedule. He's been saying this. He's been waking up at 3 a.m., tweeting about stocks and puppies in the morning, making a pit stop at the headquarters of the street, which he co-founded in 1996.

2:11:40What an absolute run for Jim Kramer. Anyway, Turbo Puffer. Search every byte, serverless vector, and full-text search built from first principles on object storage. Fast 10x cheaper and extremely scalable. This little turbo puffer is used by Notion, Linear, Cursor, and many more. I want to go through VCs to AI startups. Please take our money in Bloomberg. This is from Kate Clark. She says, private jets, box seats, and big checks. Investors are doing whatever it takes to get into top AI deals. Jordy, would you mind reading the first few paragraphs of this article? I'm trying to catch up to this article.

2:12:22It's page 90. Raising venture capital money has been a breeze for Decagon AI. We're all over the place with the customer service. It's a very hot industry. Decagon's doing well. Finn, our partner, is doing well, and Brett Taylor's doing well, too. Decagon's two years old. They're developing artificial intelligence tools for customer service. All four of Decagon's funding rounds, totaling more than$230 million, were preempted, meaning firms like Andreessen Horowitz, offered to invest before the company started fundraising. Now, just three months after a round that valued at$1.5 billion, we had the founders on when they announced that round, Decagon is fielding unsolicited offers at valuations as high as$5 billion.

2:13:03Give it up for unsolicited offers. They're some of the best offers. For Decagon and a few other AI startups, the fundraising script has flipped. Instead of pitching venture capitalists up and down Sand Hill Road, VCs are pitching them, hoovering them with gifts and favors in hopes of leading their next round. Can you catch us up on the rest of this? I want to go into some ideas because I think the VCs aren't thinking big enough. Big enough. We'll get you back in a second. So Jesse Zhang, the 28-year-old co-founder and chief executive officer of Decagon, says investors hoping to back him have offered him everything from tickets to Golden Gate Warriors games to an autographed poster of MMA legend Khabib Nurmagomedov.

2:13:45John would not know who Khabib is, but I'm sure many of you do. One investor even folded origami cranes into a mosaic of Decagon's logo and hand-delivered it to the company's San Francisco office with a term sheet hidden inside. Decagon took the deal, so that's a good one that works. Investors are emailing term sheets. They're giving verbal offers. They're inviting founders to sports games. They're inviting founders to race Ferraris, and they're inviting them on private jets, It says Bennett Siegel, a co-founder of investment firm Astar and an early investor in Decagon. What you tend to see is the best companies are getting preempted every round, and the time between rounds is shrinking.

2:14:26The lavish VC overtures are part of a larger Silicon Valley frenzy for AI driven by startups' astonishing revenue growth and investors' belief that these companies can dethrone tech giants. U.S. startups have raised around$200 billion this year so far, according to PitchBook data, but 41 % of that went to just 10 companies, highlighting VCs' relentless focus on a small group of AI frontrunners like OpenAI and Anthropic. Last year, the share of funding that went to the top 10 companies was less than 10%. A few smaller AI startups have also emerged as investor favorites, drawing outsized attention and capital.

2:15:00Among the rising group of hot companies is legal startup Harvey, customer service firm Sierra, encoding company Cognition. all of which have drawn big offers from VC firms. AnySphere, the maker of AI coding tool Cursor, has become a central example of the rush to fund newer AI companies. It's kind of funny that... I heard you talking trash, by the way. I do know who Khabib is. He's an athlete. There you go. Nailed it. You know... What else is there to know? What else is there to know? He punches and he... He kicks. Does he do kicks? He does kick. He does kicks as well? Yeah, he does kicks. He's into kicking?

2:15:37And wrestles. Isn't it this trend of companies having like one name and then the product name being called something else? Yeah. AnySphere and Cursor. Windsurf had the same situation, right? Yeah. So there's this story about Decagon. One investor even folded origami cranes into a mosaic of Decagon's logo. I saw a picture of this and I was confused because a decagon is just a 10-sided shape. And so it's actually like extremely easy to fold something into a decagon. But I guess they did cranes into the shape of a decagon or something. Because like imagine if you're like, oh, I really want to invest in square.

2:16:19I'm going to fold you a square. Take a piece of paper. I'm going to turn the term sheet into something square. It's like I want to see true origami mastery from a VC. if you're going to try and preempt a round. Yeah. Yeah, absolutely. I mean, I think investors should be thinking of it as what is the right kind of effectively bribe at each stage. Yes. Right? What's the budget? What's the math that you should be doing? How much should you put towards the bribe? Towards the tip. It's a tip. Yeah, it's a tip. There's no tax on tip. It's a tip. Yeah, I don't know. It depends on the round. But if it's really competitive, maybe thinking about putting 10 % of the check size and then paying out of your own management fees for some type of experience, right?

2:17:05Yeah, I mean, if you're just paying 10 % over, if there's a$300 million round and it's between you and another firm and you just take just 10 % over, that's not going to hurt your returns. Like if you win the round, it's going to be 10X. And so 10X on 300 as opposed to 330. Just making insane assumptions. Just take the 30 mil. All math out. Buy a McLaren F1 and be like, I see you as the next Elon. I see you as the next Sam Altman. They drove McLaren F1s. You should drive one. You should have this car. It's yours. Take it. Or keep a fleet of supercars at Sand Hill and say, you can have this from between the A and the B.

2:17:48Yeah. And you could have adverse effect of founders being like, I like this P1. I don't really want to raise my B. I'm enjoying dailying it. But in general, I think it should work out. So I think it should be like writing a$250K flyer. Why not get a founder a long weekend at an Amman property of their choice? Something straightforward. You just do it just a weekend? I feel like a couple weeks. A couple weeks. I know, but you're getting well beyond that 10 % mark. Two days in Amman is two weeks out of four seasons. Yeah, so I think you're kind of going up the ladder. Yeah, what's next? If I'm trying to win a$20 million Series A, what should I budget for the tip, for the bribe?

2:18:34For the tip. Tip sounds better than bribe. Okay. Yeah, tip. We're sticking with tips. Yeah, it is. We are coming up into VC tipping season, so founders have got to be thinking about what kind of tip makes sense for each person on your cap table. We've talked about this before. But, yeah, how can you take it up a notch? you know racing ferraris private private but i i think these like one-time offerings are super weak you know it's like uh race ferraris for a day yeah that's not that interesting founders doing 996 they're like i don't want to go to your racetrack right so i think um buy the racetrack i think the thermal club was up for sale or a lot of or or saying uh you can use my jet one time One time?

2:19:21One time. One time? It should be. Here's the full jet. The jet should live with you until the IPO. Toss in the keys. Until the IPO. Yeah. Then you get your own jet? Yeah, yeah. Until you're in the position to be able to get your own. Tyler, what would get you to sign a term sheet? What would stick out to you if a VC came to you and said, I got to do this deal? Look at this picture. I like that you're a David Sendero fan. You have a framed photo of him. I think something compatible, like if the founder is doing 996, right? There's this thing, you know, the current vibe is stakes, right? So maybe you buy a, you know, kettle farm with very high quality cows.

2:20:00Like, you know, Zuck's Island. Zuck has that on his island. Yeah. For sure. Gold Rock said, I had a VC try to take me hot air ballooning. That's a wild one. Would you go in a hot air balloon? I would absolutely go in a hot air balloon. Okay. It's extremely aristocratic. Yeah. It's fantastic. I saw a video of a hot air balloon crashing from somebody that took the video inside the hot air balloon. It scared you? It wasn't super appealing. Oh, we got Tyler on the big screen now, and we got Senra here on the big screen. That's great. There we go. I feel like that's a skill issue. Skill issue? Just hit chat GPT.

2:20:37How do I fly a hot air balloon? Get up to speed. Summarize that for me. Don't make mistakes. And then we got a little Inception going here. I like that. That's a good start. Anyway, profound. get your brand mentioned in chat gpt reach millions of consumers who are using ai to discover new products and brands maybe that's what the vc firm should do get on profound when people when founders type in who should i raise money from i'm building this show up and i think i think founders should just ask uh i don't want you to just join the board i want you to actually become a a bdr at my company for the next two years yeah i mean you're you're joking but like that's sort of, I've seen that happen with Elon companies where to win allocation, you're going, not BDR, but recruiting.

2:21:21So come in and import your entire network. Everyone you know who's an engineer who could work at this company, come and work basically full-time for like a month to like really do everything you possibly can because there's always, you write the check and then you're like, they hit you with like, oh, like, do you know any software? It's very different from being like, I'm doing three full-time weeks just actually racking my network, actually making calls, catching up with everyone who would be relevant. Not just making the two most convenient. No, no, no. Going really deep into the actual, into the Rolodex.

2:21:58Does the Hindenburg count as a hot air balloon? I don't think so. It was a dirigible. It was a blimp. But anyway. Numeral. Sales tax on autopilot. Spend less than five minutes per month on sales tax compliance. Let numeral. Numeral HQ.com. Google just killed Dingboard. You saw this. Are you familiar with Dingboard? Yak scenes project? They just killed it? What do you mean? Have you seen this? Of course I know Dingboard. Were you a subscriber? I was very early on. He gave me, I was like very early on. I DMed him for some reason. He gave me a code. I love the product. I thought the product was amazing.

2:22:32It didn't really work that well on mobile. But it was super cool. You could like go in with AI, remove backgrounds, merge images, blend things together. It was a fantastic product. He goes to work for Axe. How did they kill it? Well, because Google Labs just launched MixBoard, an experimental AI-powered concepting board designed to help you explore, visualize, and refine your ideas, powered by our latest image generation model, of course, Nano Banana. And with Nano Banana and, you know, basically it's not that they killed DingBoard. It's just that they launched a product that is clearly inspired inspired or lives in the same kind of work stream as Dingboard where you can remix images, bring things in.

2:23:14Very, very cool product. I would love to see this on an app for sure. I've always made, when I'm making images and memes on my phone, I'm always going back and forth between like Photoshop Express and Image Flip and a few different sites. There's not a lot of stuff. This is a great product and I hope someone can take it across the finish line. And obviously, Axine had to go work at X, and now he's working on robots and stuff. What do you think of Stiller's? Stiller's? Soda. Ben Stiller's soda company. I have no idea. Does he have that type of audience that's ready to rip soda? From Walmart? Yeah, it's interesting.

2:23:54I guess we were talking to John Shahidi about this. Ben Stiller's the type of person that could probably call the CEO of Walmart and get a meeting, right? He's so famous. If you're a CEO of a big retail chain and you hear Ben Stiller's on the line, you're going to take that call, right? So maybe it advances your distribution quickly. Stiller's is also just a great name. I do. The name is great. And when I saw it, I was like, oh, that's a cool name. I'll take a Stiller's. That's a great name. Is that related to Ben Stiller? And then I found out it was. So I could see it working. I would love to do a taste test.

2:24:33We'll have to actually see, is the product good? Because if he has great product sense, the distribution and the celebrity will take care of itself. Some of the copy is a little millennial coded. Sure. It says, made by 100 % real human celebrity people. Okay. Yeah, that's a little. And the copy on it says, no fake stuff. Hmm. Okay. And crowded category. Yeah. just they they they're competing with with olipop and yeah the other one there's been a huge there's been a huge lineage of trying to find uh make a product that's free of free of sugar free of fat i would say i would have been incredibly bullish on this company if it launched in 2019 yep but launching into a category that pepsi is already gonna buy they want i'm sure coke will eventually I mean, I believe Pepsi already has Poppy, which was directed better to...

2:25:28Pepe. Wait, there's Poppy, but then there's also... Pepsi bought Poppy. You could imagine Coke buying Olipop. Yeah. And then it's kind of over. But then there's another company that's a seltzer that's... It's Bubbly. Bubbly. Do you know Bubbly? Bubbly Sparkling Water? Who owns Bubbly? Yeah. Well, we should get Ben on the show to talk about Stillers. and we should get Tyler order some stillings. Bubbly was incubated by PepsiCo to compete with the Spindriffs of the world. And remember the big boom in seltzers? There were a whole bunch of Topo Chico and a bunch of other products that were really popular in offices in 2017, corporate offices.

2:26:16People would be drinking LaCroix all day long. Bubbly was like their answer to that. I don't know how it actually panned out, but that's kind of interesting. But in much more interesting news, OpenAI, SAP, and Microsoft are launching OpenAI for Germany, a partnership to bring frontier AI to Germany's public sector through a sovereign certified cloud environment. We haven't rung the gong enough on this show. Let's ring it for OpenAI for Germany. There we go. and we got to hit it for Claude getting into the Microsoft ecosystem. That also happened. We love it. Well, without further ado, I think we have our next guest, Riley Walls in the Restream waiting room.

2:27:02Let's bring him in. He has been being a rascal the last 48 hours. Riley, welcome to the show. How are you doing? Thank you for having me. The most viral man on X, the current thing. Yeah, take us through the current thing, the current project. What did you build? What inspired it? What's the reception been? Are those GPUs on fire? Are the servers on fire? Is it staying up? Walk me through it all. Yeah, I always got a project kind of going on. Yesterday's project was I figured out you could scrape the parking ticket system in San Francisco to reveal more or less the real-time locations of parking officers as they wrote tickets.

2:27:44So I made a website called Find My Parking Cops. It looks very similar to Apple's Find My Friends. And yeah, you can see where the cops are as a traveler around SF. Take me a level deeper on the technical. What inspired this? Did you get a lot of tickets? Is that what inspired this? I actually don't even have a car. My roommates, all my friends have cars. And yeah, you hear a lot about tickets. It's pretty notorious in the city. Yeah. Take me through one layer deeper of the technical side. Is this just a function of the SF parking ticket office integrating with some sort of IT or ERP system that basically as soon as the ticket's written, it gets loaded into a database?

2:28:29And then is there a public API? Is this stuff that you can just look up on an actual website or are you finding an entirely private API? Yeah, it's all public. The magic is right when they write the ticket, it goes up online. And you have to enter the ID number to be able to see the ticket. There's a website, you can pay your ticket. And when you pay, you can see a copy of the ticket. So I figured out that the ID numbers for all the tickets were predictable, meaning that there's a pattern to it. So I knew the pattern, and I could see, okay, this ticket just simplified a lot. If ticket number 83 was just written, And I know 84 is going to be next.

2:29:07So I just keep checking for the next one. There's a little more technical than that, but it's pretty magical when predictable IDs are a thing. There's so many cool ideas. You can scrape knowing that. Talk to me about your actual workflow. Are you using Vibe coding tools? Before that, what was the current state of things? SF, obviously, you could potentially really, if enough people in SF started using this, it would be good. I would hope that the mayor would write you a letter of recommendation for this. You should win a medal for this. This is the highest calling. Yeah, you should get the key to the city for this.

2:29:45But what's the actual response from SF been, if anything? Yeah, so I do websites like these that are provocative a lot. And I think a lot about framing it and how I want to present it to people. So I thought this would be maybe a little mildly viral among people in SF. But it actually went pretty viral around the world yesterday. within four hours the the SF government mobilized and they changed their site to prevent me from getting the data like only four hours it's like for them to prevent me from make the site useless so yeah the site is not up anymore because they changed the way those four sweet hours though were everyone free of parking tickets it actually is if if it had gone less viral and not and people hadn't, you know, if they hadn't noticed so quickly, I can imagine this being highly, somebody's pulling up, they want to grab a coffee, you know, they're like, they just check the map, they can run in and out.

2:30:36It's not a, not a, not a, uh, I can see it actually being pretty valuable. Yeah. Walk me through some of like the, how you like to build these projects, how you like to host them. Um, what, what problems you typically run through? We'd love to know just a little bit about your stack these days. Yeah. I mean, use AI a lot. It makes things so much faster. I have like, a gazillion ideas. These are all sticky notes for different data ideas I want to make. So just kind of every weekend, I'll try to knock one off in my free time. You guys talked about, you saw the other one I made, Panama Playlist. Oh, yeah.

2:31:09Oh, you did that. We're on that. You scraped my data. I'm glad you, I was wondering, how did I make the cut on this? These are some big people on there. Yeah. What was the secret to that? Because it only yeah was there any was there any strange fallout i mean it's hilarious to read into people's music taste i don't you wouldn't have noticed this but my that the playlist that i had public that you featured was like seven songs i found that had like obscure references to venture capital so when i saw it i was like i don't listen to any of it when did i make this playlist i don't listen to any of these in like 2021 i just started collecting yeah or whatever yeah i don't know it It was just kind of trying to highlight like how much data is out there on Spotify and like people just forget about it.

2:31:54And yeah, sorry, you guys were part of it. But no, I thought it was hilarious. I think it was it was pretty amusing. I think a lot of people had a kick looking at it. Yeah. What was the what was the reaction from either the people that were in it or Spotify directly? Has there been a change to the API or because we saw this even during the last election, like the Venmo public records, people forget about this stuff all the time. there's kind of a question about how tech platforms deal with like default privacy because there was an era when everything was public and then people started closing things down becoming a little bit more private but it's an interesting like user design problem for the big companies to shift.

2:32:31Yeah. In case your audience didn't see it I scrape like the public playlist listening data of notable figures on Spotify and like you guys were on it like Sam Alton was on it, a bunch of politicians like Mike Johnson and J.D. advance and like so many people just have stuff open on there um they haven't changed stuff yet i feel like spotify probably will change like the default behavior of playlists because right now they're public by default which is kind of crazy um but yeah i mean it was a lot of people took their playlists down um kind of was a yeah interesting to see like kind of a lot of people embraced like palmer lucky he was like yeah these are my songs and i love them yeah it's great so So did you have to guess some sort of code or was, did you just literally just search my profile and then it was there?

2:33:16Yeah. I mean, for you guys, a lot of it was just like real name. Just search your game and then found a profile with your picture. And like, probably a simp is that I can see like you guys follow each other or something. Probably. Yeah. Walk us through some of the, some of the other stuff you've done recently. And then I want to talk about. Yeah. Specifically, what's the most underrated project or most underappreciated project you've worked on where you love it it's one of your dearest children but you feel like it hasn't gotten enough love one like interesting one with a few months ago I I this was also very out much out there but you know you guys know looks maxing I I mean it's like a looks mapping where I I scraped the google maps reviews of restaurants in New York, SF, and LA.

2:34:07And then I ran the profile pictures of the reviewers through a very jank attractiveness model. And then I made an aggregate map of all these restaurants. Like, okay, this restaurant is usually composed of sixes or whatever out of 10. Very off on the idea and premise, but it was kind of funny. That wouldn't make anybody mad. Yeah, you can glean a lot of just information. Because I also had an age map. You could get their age from the picture roughly and gender too so you can see which restaurants had like the most men or women in in a city like that what kind of pattern like how much did it track with just the general hype around restaurants like what what did you notice yeah i mean like the oldest restaurant in sf is not a country club um like the most um like male restaurant is a gay bar the most female ones like a brunch spot.

2:34:59It all kind of fits. It's predictable. That feels like a product that maybe it needs a little polish, but could just be something that would actually add value to the Google Maps experience. Has big tech companies reached out to you and said, hey, come work for us? Not for that. There's been a couple of smaller map startups that are like, oh, you should maybe add this. I think they should. Yeah, I think this would be cool. I feel like it's too politically risky for Google, but definitely a smaller startup could do it and it'd be useful. Yeah, I feel like even if you dialed back some of the framing on what you're building, there's product insight there.

2:35:32And there's novel features that would surprise and delight. And I feel like that's kind of what a lot of these projects put on display. What were you doing before all these experiments? Yeah, does this pay the bills? Do you have a full-time job? I have a full-time job doing data stuff that's more commercially reliable. um i also i mean like you guys talked to gabe waley of mischief yeah yeah there was an intern there uh there you go mafia which was like very i mean you guys can kind of see like all this is very inspirational very inspired by mischief yeah yeah the chat is literally saying he's like mischief for tech projects lol like people have clocked it for sure i first and i got to see up close how they operate they're like incredible people yeah so you're at mischief and you said I need to pursue data science.

2:36:24Yeah, that was my calling. I think data is like really cool. And it's cool that you can combine it with like mischief elements to make like weird things like this. What are, let's, let's, I don't, I don't, it feels like you could talk through some ideas you have without giving away too much alpha just because a lot of, I don't know if, I doubt people would, would steal your ideas since a lot of them are like, don't make, or you're not doing this to make money, but you want to talk through any, workshop any with us? Yeah, let me see. Oh yeah, what you got on the board? One terrible idea that I'll probably never build because it's awful is there's OpenAI has an API endpoint for how bad a piece of text is.

2:37:06It will return back a string of, is this harassment or is this racism or whatever. I don't know. Yeah. The horrible idea is somebody should make a leaderboard where you have to type in a string of text and you want to like actually hit every single one of these categories let's say the most offensive as possible you have to create the most offensive string of text in history and as few characters as possible yeah as many people as possible yeah horrible but horrible but uh certainly i think you might not need to disclose what is actually said you could have it be anonymous but you know you it would certainly spark creativity i mean rune was talking about this how there are certain strings of text that just go viral every time they're posted, no matter when they're posted, no matter who posts them.

2:37:54These little insights, we did this project Banger Archive where we took the best posts by certain people, just screenshotted them and just said Banger. And they would get another 10 ,000 likes every time. Yeah, we had posts that would get 50, you know, an order of magnitude more likes than the original post. And it was like an Naval post about like, you should work like a lion, like, you know, take rest and then work really hard and it's just like these universal truths that like just continue to deliver value what else what else you got what else you got what else is on the board give us some other stuff um because i feel like 4chan would like one shot your leader they'd figure out there's like there's four words you put these together and you make every person on earth mad yeah no it'd be terrible um one idea this is something i asked you with parking tickets here um i also found out that that like vandalism citations have like a similar setup in the city.

2:38:50So I was able to scrape like half a million pictures of graffiti that cops took. Oh, interesting. So I just think it'd be cool to make an episode that shows graffiti art through the lens of a cop in San Francisco. Yeah, that'd be cool. They took on their phone themselves. Yeah. It feels like a lot of your projects touch on like public works projects, public like cities, like data in cities. And I'm wondering if you have a view on like, should cities be more open to the hacker culture, actually embrace some of this stuff? Like some of the stuff that you're building is like very close to just like a tool that would make life in a particular city easier.

2:39:34And yet cities are notorious for long lines of the DMV and not being the most tech forward organizations or entities. how do you think about actually like how cities deliver tech services yeah I think there's a lot there like yeah I think the asset government actually has a pretty good handle on this they have lots of data sets that they publish pretty frequently and there's a lot to work with there like they publish like 911 calls with like a 10 minute delay you can see kind of like an anonymized data set of what people are calling 911 for which is kind of cool like there's lots of things like that nobody really knows about but um there's things there um would you like in a perfect world in a perfect world would one of these you do it for fun and then and then you discover some like enduring business or do you like just doing one idea after the next yeah i mean all these ideas are just things i find interesting to myself and you know usually that means somebody else will find interesting too if i post online um i think it's just like an outlet to be creative and it's nice to have distinct like work work and like this is more fun work uh but yeah no definitely be cool if something like that happened any any ideas are you are you capital constrained on anything like if you had 25 grand you would you would do because i feel like at this point there's enough people ourselves included we would we would uh as long as it was not uh offensive we would to happily chip in to make possible.

2:41:01There is, yes. I need 33K for a project next May for a very fun stunt in San Francisco. Okay. If anyone would like to see what I want. We should talk. Yeah. All right, let's talk. We'll talk offline. We don't want to spoil it. Last story I want to talk about. I didn't put it together until just now that you're the person behind New York's hottest steakhouse that was fake. Tell me that story because I remember it happening and I didn't put it together until just now. You talked to my co-conspirator Moran, the archaeology guy. Yeah. No way. Yeah. That was incredible. One of the best times of my life, we had so many things go wrong, but somehow it just happened to go right.

2:41:41Long story short, we lived in a house together in New York. Moran made steaks and then we made a Google Maps listing called Moran Steakhouse. Then all our friends started writing these bizarre five-star reviews. I converted out of Hinduism so that way I could try some of his beef. for like he came in like drenched in blood from upstate fresh with a fresh cow like like really bizarre reviews of people in new york thought it was real because the dining culture there is crazy yeah and we had this huge wait list over a couple years that we built it up so we actually opened it for one night we opened a real restaurant fully permitted i was like licensed by the government to actually open a restaurant um we had 120 guests come thinking it was real it was like humongous operation legit six wives but it somehow ended up well we did not get sued How did the guests take it?

2:42:24At what point did they realize it was a joke? Most of them were pretty good. I don't think people really realized it was fake. We wanted people to find out it was fake in the Times, and they read it about the article. They read it in the New York Times. And a lot of people did. They thought it was kind of a weird restaurant, but they thought it was like, okay, this is just how things are. It's like cool and avant-garde, right? It's like edgy, new, and different, and pop-up. Did you guys break even on it? Did you sell enough steak to sell? No, no. We spent like 16K and then made like 13K. So we lost like 3K.

2:43:00But for the amount of attention we got, it was viral everywhere. But that New York Times piece is forever. It is. Yes. Yeah. Amazing. All right. Well, we'll message you. We'll figure out how to make this. Yeah, we'd love to talk about SF Stunt. That'd be fantastic. Happen. And keep up the great work. Fantastic. We'll talk to you soon, Riley. Thanks for joining. Cheers. Let me tell you about Adio, customer relationship magic. Adio is the AI native CRM that builds, scales, and grows your company to the next level. Get started. OpenAI says, more compute in the making, announcing five new Stargate sites with Oracle and SoftBank, putting us ahead of schedule on the 10 gigawatt commitment we announced in January.

2:43:41Let's go. Ahead of schedule. Massive. Stripe's also buying back shares from its VC backers at$106.7 billion valuation. Sequoia bought$861 million worth of shares in 2024 at a$70 billion valuation. So they're getting a markup. And Amenden pretends his live shot of a swath right now. He's talking about Aswath Demodrin, who is obsessed with valuation, specifically liquidity premiums, illiquidity premiums. And it's just interesting that Stripe continues to find ways to stay private basically forever. And they've been on a tear. They're the Collison brothers. We saw this. We talked to Joe Lonsdale about being in the same fraternity at Stanford with Gary Tan.

2:44:31And you can zoom in on this photo and see a young Joe Lonsdale and a young Gary Tan hanging out at Stanford back in 2003. Pretty remarkable piece of tech lore. I want to know what is everyone else up to? We've got to go through this list because there's probably some interesting folks in here. I want to know what Michael Calhoun's up to or Justin Reynolds or Adam Rodriguez. Got to figure it out. Bunch of lads. Got to get on public.com, investing for those that take it seriously. They've got multi-ass investing, industry-leading yields. They're trusted by millions. And we've got anything else here?

2:45:08Tyler Cowen is talking about the potential for stagflation, economic stagnation. No one wants that. No one wants that. Tyler says it seems increasingly likely that the American economy is sleepwalking towards stagflation. In case you're wondering, that is not a good thing. Stagflation means an economy experiences excess inflation and excess unemployment at the same time. This was one thought to be impossible, but the OPEC oil price shocks of the 1970s triggered both high inflation and high unemployment. And voila, we suddenly had a new unhappy economic phenomenon. If I had to guess, I think there's a decent chance that 18 months from now, America could well have an inflation rate of 4 % up from last year's 2.5%.

2:45:52Oh, John, we got some breaking news in the chat. What do we have? China agrees to terms of a U.S. TikTok deal. Wow. Thank you, Bobby Cosby, for breaking the news. Thank you. Hello. Hello. Thank you, everyone in the chat for keeping us up to speed. Is this not a little worrisome that they're agreeing with? You know? Maybe they just - Seemed to have pretty, historically had a pretty hard line. Well, pull up any more information while I tell everyone about adquick.com. Out-of-home advertising made easy and measurable. Say goodbye to the heads of out-of-home advertising. Only adquick combines technology, out-of-home expertise, and data to enable efficient, seamless ad buying across the globe.

2:46:26I want to tell you how Steven Spielberg works. He says, Steven Spielberg says, before he goes off to direct a movie, he always looks at four films. Can you guess what films he watches before he goes out and directs a film? Seven Samurai. Lawrence of Arabia. Borat. Borat. Dark Knight. Dark Knight. What was the other? Mountain Head and Office Space. Office Space. Those are the movies you've seen? Yeah, yeah, yeah. No. No, he does. Seven Samurai, Lawrence of Arabia, It's a Wonderful Life, and The Searchers. Have you seen any of those? Lawrence of Arabia. Lawrence of Arabia is a fantastic movie. Tyler, have you seen any of those?

2:47:07I've seen all of them but The Searchers. Same with me. No, I have seen them. You've seen The Searchers? I have seen it, yeah. You need to give us a review on that. I need to check that out this weekend. I've seen 7th century. All right, since you've seen it, name every scene. Yes. Is there a tech equivalent of this? Before you start a new company, study the greats. Listen to these four Founders Podcast episodes. Who are you listening to? Pick four. Top four. I'm going Edwin Land, Steve Jobs, Gaston Glock, and probably Rockefeller. You got to throw a glass of Glock in there for sure. Underrated.

2:47:44And the Chicken Finger Dream. You got to get up to speed on the Chicken Finger Dream before you start your next company. I think every founder should have four Founders Podcast episodes in the chamber before they go out to raise, pump up speeches. Or if they need to run through a brick wall. Yeah, if you need to run through a brick wall, you need to have your top four Founders Podcasts ready to go. So when you hit the road, when you check into the Rosewood for your Sand Hill tour of pitches, you're ready to go. and you're also wearing a fantastic timepiece on your wrist that you picked up from GetBezil.com because your bezel concierge is available now to source you any watch on the planet.

2:48:19Seriously, any watch. In other news, Palmer Luckey says that Mod Retro has teamed up with Ubisoft to re-release some of their greatest classics, starting with Rayman. It has some of the best graphics of any Game Boy Color title. Late cycle developers truly mastered the hardware. It chipped months after the Game Boy Advance was announced. So congrats to Palmer Luckey and a new deal for Mod Retro. The company's been on a tear. Very excited. Palmer Luckey also hit the timeline yesterday quoting a Meta post that see how U.S. national security agencies are using Meta's Lama models to develop bespoke tools for America's military and intelligence professionals.

2:48:52It is crazy how much the vibe has shifted that Meta would have imagined if Meta posted something like that in 2016. 2016 when Palmer was there for a couple months? Yeah. Would have been crazy. Palmer says, added to the quote from the announcement, we're also supporting U.S. national security through our work developing augmented and virtual reality technologies. Through our partnership with Andrel, we are developing a range of wearable products to help maintain America's technological edge. You know what I've missed recently? Singing. There hasn't been enough singing on this show, so we've got to bring it back.

2:49:26Find your happy place. Find your happy place. Book a Wander with inspiring views, hotel-grade amenities, dreamy beds, top-tier cleaning, and 24-7 concierge service. It's a vacation home, but better. We got to get Tyler to chime in when we're singing. We can close on this post. Vinnie Daniels says, I'm seeing parallels between NVIDIA and the Medici Bank failure of the 1490s. Who isn't? People are having fun drawing comparisons. Lots of the dot-com boom. Not so much to the 1490s, but fire up, crack open a history book. Should we do a deep dive on the 1490s? For sure. We should definitely read about the Medici bank failure.

2:50:05Tyler, you in? Get out the red string. I was going to say, I see a parallel between NVIDIA and the rise of the Roman Empire early on. Okay. Okay. Taking over everything. Okay. Who is Jensen? Is he Caesar? Or the rise of Genghis Khan, the Mongols. Maybe. Maybe. Maybe. Anyways, that's a fun show, folks. Thanks for tuning in. Hope you have a great Wednesday. Have a fantastic rest of your day. We'll be back tomorrow. And I cannot wait. We will talk to you soon. Cheers. Goodbye.