Chamath Is Back With New SPAC, AI Scrutiny, China’s Moon Push | Pete DeJoy, Marty Kausas, Matteo Franceschetti, Abdul Al-Asaad, Derek Lo, Alex Telford

28:38hitting that slope. Look across the chasm and you can see the plateau of productivity. And it is beautiful. Yeah. So don't worry. It's going to be, it's going to be, it could be some dark times in the trough of disillusionment. We might be dressed up as pigs eating from the trough for a few days. We might be wearing raincoats. We might be, we might be wearing hazmat suits, depending on how bad things get in the trough of disillusionment, but we will not be disillusioned for very long. Never. It will be our job to speed run the trough of disillusionment into the slope of enlightenment and toward the plateau of productivity.

29:11It's going to be beautiful. That will be my goal. There will be chaos. To dig us out of the trough of disillusionment. To climb out. Climb out, rung by rung. Anyway, I like some of these. There's a debate going on in the chat right now. Trey P says, Sleeper AI integration is co-pilot in Excel. Well, Raghav says Copilot in Excel is garbage. Let us know what you think. There are a bunch of companies that are doing this exact thing, trying to compete with Copilot in Excel. We've had a few founders on the show. There was a new one yesterday called Paradigm. The journal. Really cool visualization showing all of the different cells working independently.

29:52I thought it did a good job of expressing this idea that each cell could be its own LLM that's reasoning and then filling in the results as they stream in. Like it had a very nice aesthetic to it. Also, I think they opened with a really cinematic shot. Like as far as a launch video, I think it did a good job of bridging the gap between like it grabbed my attention. There was some clear storytelling, some facts, but some good UI. Like it did a good job there. So we might have the founder on at some point, but it is a knockout, dragout fight. And, you know, will this play out like Slack and Teams where once Microsoft learns all the patterns that people like, they roll it out to their billion users or something like that for free.

30:41Dave Grampel in the X chat says, there is opportunity in the disillusionment. Totally agree. Completely. That is where when the casuals pivot back to regular SaaS and you stay focused on building agentic workflows. that's going to separate the killers from the kids. Yep. The tourists. Well, whatever you're building, you got to use graphite.dev. Code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster. And there's an announcement from Graphite today. You put it in the chat. Graphite chat is now live. Go check it out. This is from Merrill Lutsky, who's been on the show.

31:24The full request is now where you build, not where you wait. Welcome to the new standard for code review. so well check it out well manitis has a new coinage casino core culture he says sports betting s coins meme stocks vibe coding 100 million in six hours etc are all expressions of the same deep cultural rot if youth don't believe there's legitimate ways to get rich through work all of culture will become a rotten sports book for the soul sports book for the soul's good coinage sad take. I think it's real, but at the same time it's hard for me to be bearish because we've interviewed hundreds of people this year that are all working I guess I can't say all, but the vast majority.

32:08It's the permanent underclass meme. It's the get your bag meme. It's this idea and a lot of it stems from the singularitarians. This idea that something is coming that will completely upend. you'll be permanently in the underclass. The world will change so fast that you cannot possibly predict what to do, what to build, what to compound for 30 years, where that was very clearly not on Warren Buffett's mind. He was just like, I like reading the newspaper and buying stocks. And so I'm going to do that every day. And he made what, 90%, 99 % of his wealth came from the age like 65 to 95. Like he was not really that rich.

32:47Compounding world. At retirement age. Like that is crazy. And no one's willing to wait that long. Everyone is like, I have to do something now. And then they wind up burning their relationships and burning their reputations. There's so many short-sighted deals. Also, you know, there's plenty of examples of people that have generated tremendous wealth through getting heavily invested in crypto early. But at the same time, I know a number of people that got so sucked into that world and didn't develop, you know, basically spent five years, like, not progressing their career, not developing more skills.

33:25And then they're stuck in this kind of vicious cycle of feeling like, well, I kind of ignored my career for half a decade, and now I just have to keep hyper-gambling to, like, try to get out of this mess. Should we talk about the bonsai tree? I think it's okay to talk about it on the stream, right? We've added some decor to the table, this fake plant, which is lovely. Thank you to the production team for putting this on the table. So whenever we need to touch grass. Oh, this is that. This is the grass that we touch, but it's not real grass. We're working on getting real grass. We are working on getting real grass.

33:59But we had an idea last night as we were chatting for a future merch drop. And one of the ideas that we came up with was a bonsai tree that we will send to you that will have the TBPN logo on it. and you can tend it throughout your entire career. If you ever wanted a bonsai with a livery, with livery presented by ramp. Yeah, yes. Get ready for the TVP on bonsai. But I do think that we should get a bonsai tree and put it on here and tend to it over the next 50 years as we grow the show. And it will be a beautiful reminder of what we're growing and the idea of working for a really long time, having long time horizons, underestimating what can be done in a decade is the goal.

34:46You underestimate how big a bonsai can get in a decade. Yeah, thinking about it starting here, small, and then the time lapse of every show. You're watching TV. And it's a little bit bigger. And then we're pruning it and just trimming it and just tending to it. And we become bonsai experts. And you'll be able to follow along at home because we will be sending this. And the alpha in life is finding the bonsai tree of your career. Exactly. Finding the thing. And you probably, you know where you don't see a lot of bonsai trees? The casino. That's right. No. You never do. Just a lot of ashtrays. You never do.

35:17Well, Wasteland Capital is back with another post. They're sharing. Kathy Wood took in record inflows of$3.7 billion last week, including$2.8 billion into ARK, which, quote, record-breaking single-day hauls in 2021, quote, burning hot. Is this bullish or bearish? And there's a snapshot here. After years of outflows and underperformance, Cathie Wood's ARK ETFs are burning hot, pulling in massive capital. The firm's 13 U.S.-listed ETFs have gathered$3.7 billion in AUM over the past week, led by record-breaking single-day hauls. The flagship ARK Innovation ETF led the charge, raking in a record$1.1 billion on Monday and an even larger$1.4 billion on Tuesday.

36:05This was last week, which marks the biggest single-day inflow since 2021. ARK defines disruptive innovation as the introduction of a technology-enabled new product or service that potentially changes the way the world works. This was followed by ARK Next Generation Internet ETF, ARKW, which saw its largest ever single-day haul with$349 million on Tuesday. This is a lot of money. I mean, we were talking about the Tremont Spock. is Spock. I mean, that is the... Is it not Chamath, Spock? Chamath, Spock. Chamath, Spock. Anyway, the Spock, that was$250 million. This is just one of the ETFs pulling in$350 million, $100 million more, just on a Tuesday.

36:53But it's a reasonable thesis. Disruptive innovation is important. Internet companies continue to be outperforming assets. If she's in the right thing, it should be okay. I do wonder the, you know, back to that question of like, is it January? Is it November or whatever? It's like, I'm trying to remember in January if people were acting like this and calling everything out. I think people were. I think people were dunking on the timeline like all throughout 2021 when things were ripping. it wasn't it wasn't until um i think keith reboy like truly called the top and he was kind of like he was kind of yeah it was keith it was keith in november he had this post tyler can you actually try to find that post it was keith reboy yeah it was november 2021 yep uh i think later he called it like to the day it's crazy and i was reflecting on that and like what the crazy thing is that like he's basically a long only investor.

37:51Like he has no incentive to call a top, right? Like it would make sense. Michael Burry is calling tops all day long because he is a short investor and he takes really short positions. So I've seen over the past like five years, like headlines that are like, the investor that called the mortgage bubble just sold everything, just went giga short this. And he's been wrong like a few times, I think. He's probably been fine in most of the cases, but like, Like, you know, in general, like, I expect to hear a top call from Michael Burry or even just like a long short hedge fund. They don't have a way to express that in the markets.

38:29I mean, the way to express that is, hey, let's not do this Series B with a million dollars of revenue. Yeah, let's pull back. Four out of 400, you know, times. Yeah, yeah. This Jason Calacanis post is so, so funny. Because Tremont, of course, is his co-host on the All In podcast. And Jason says, and here we go. And posts the SPAC screenshot next to the Joker meme. Here we go. It really is a bull case for them truly being besties because posting a Joker meme about your co-host is hilarious, hilarious, hilarious. Anyway, if you want to dump a bunch of SPAC data into a database and start running an analysis on it, do it on Julius.

39:18What analysis do you want to run? Chat with your data and get expert level insights in seconds. You can ask Julius to analyze your data. They're loved by over 2 million users and trusted by individuals at Princeton, BCG, and Zapier. That's right. And so we already touched on the casino thing a bunch. In other news, SoftBank is doing the meme, as you put it. the unicorn up and to the right, just absolute exponential growth. That was the stock price of SoftBank. SoftBank is going to make a$2 billion investment in Intel too. And say what you want about Masa, he's good at semiconductor investing because he got into ARM and that was how he made his, what, second$100 billion return?

40:05The first one was on Alibaba and then his second one came from Arm. And so who knows? You get Masa, you get Donald Trump with a 10 % stake, you get a couple more people, and Liputon could be in a good position. Yeah, SoftBank is down 7.5 % today, but year to date, it is up 83%. 83%. And it's a big company, big company. So they announced they're signing a definitive securities purchase agreement under which SoftBank will make a$2 billion investment in Intel common stock. I think this is good for Intel. I was debating this, Like, why does Intel need Trump? Why does Intel need SoftBank? Is the cure for Intel really different shareholders?

40:45And I think potentially, yes. It's almost like a take private if you get someone who is in it for the long term and can buckle up for a couple awkward quarters as Liputan makes some crazy cuts. Because if he cuts as much as we think he's going to cut, you could see revenue fall. Because that's just a function of cutting a bunch of stuff. And things might break, deals might fall through, before you can set yourself up for the next era of actual going back into growth. And so you need shareholders that aren't going to panic and sell and put your stock in the absolute dumpster. So the NASDAQ peaked.

41:25the local top was November 19th of 2021. November 19th, okay. And on November 19th, Art Levy, CNBC's tech editor, said - Wait, Art Levy, isn't he at Brex? Ari, sorry, Ari. Ari Levy, okay. Art Levy's at Brex. But Ari says, they have this exchange. Ari says, amazing to see how many experts on crypto and tech valuations are also experts on inflation. And Keith says, to be fair, valuations and inflation are directly connected. Ari says, but didn't we have an explosion in valuations over the past decade despite minimal inflation until now? Keith says, this is also what crashed the internet bubble, FYI.

42:05And Ari says, ah, so are you calling the top? And Keith just says, yes. Insane. Insane call. Yeah, literally to the day he called it. To the day. Now, do you think he called it or do you think he triggered it? Like, do you think that, I mean, he's like a very important person in tech and politics and business. Like he's someone that is respected and like people look to him in many ways. And, uh, and that echoes through the economy. Like there, like there aren't many other people who like have more weight in the economy. It's like Warren Buffett. And like, there are other people that are bigger, but Keith is definitely in the, in the top, you know, like what decile quartile.

42:48I don't know. He's in the top hundred people that are like financial influencers in one way or another. Like people. And also, he doesn't say that much. He's not constantly saying, oh, next week I think we're going to be green. You know? Yeah. He's not like constantly putting out proclamations about where the broad market is going. So, I don't know. People might have paid attention and they might have been like, I'm. And that might have kind of triggered the wake up call. But I don't think he like fully triggered it. But it is interesting to kind of toy with that idea of like, is he upstream or downstream?

43:21Anyway. Paki McCormick says, what's the fastest way to make a million? Invest 10 million in American exceptionalism acquisition corp. And of course, if history. This is like throwing meat to the sharks. To the sharks. First off, the average Chamath SPAC was not down 90%. It was down like 80%. So got to put the number in the truth. Fastest way to make$2 million. And also, yeah, I mean, lose$9 million. And this is a twist on the joke that is, what's the fastest way to become a millionaire, start as a billionaire, and then start a restaurant? Start angel investing. Start angel investing, something like that, right?

44:05So it's what's the fastest way to be a millionaire? Invest$10 million in this. But anyway. Mert says Chamath launching a SPAC and holding DeFi tokens, of course, unconfirmed whether or not he will actually merge with a DeFi company. But he says exit all markets. I love Mert. buco capital is quoting an old post a viral post of his can't be mad at trump guy loves tariffs always has kind of like my dog loves to roll in shit can't not love rolling in shit doesn't matter how mad i get at him doesn't matter that he has to get hosed down which he hates he sees a pile of shit he will roll in it says same thing with jimoth and spax spax yeah i mean it is it is a niche product and chamath has figured out a way to like really build a business around it there are other people that want to do it there are other people that have gotten have dipped their toes in but it's just such a crazy dynamic because i think that all the people that are that are you know quick to to dunk and and and chirp at chamath and anytime he posts about a good investment that he made you know people ask him about about the specs which is i guess which is fair but um i think all those people will just end up investing in this because it's fun.

45:33Yeah. Funniest outcomes, the most likely. Have you seen the poly market for the presidential election winner, 2028? Who we got? In first place, with 28 % chance of being the next president, J.D. Vance. The second is Gavin Newsom at 16%, then AOC at 9%, then Pete Buttigieg at 6%, then Marco Rubio. But But the J.D. Vance, Gavin Newsom, I thought that they were just chirping at each other on the timeline. No, people really think that they are going to be going up against each other in the next election. So buckle up. It's going to be a fun one in a couple of years, though. We got a couple more years of just business and markets and the Gartner hype cycle.

46:15People are speculating that Chamath will take Grok public. Well, Manaitis says so many hours wasted debating if free cash flow will collect at the model layer, the GPU layer, the data layer, or the app layer. when it's obvious the only free cash flow is coming out of the 4 and 40 spv layer pretty pretty good um i have a buddy who is one of the top uh secondary uh brokers in uh the industry and he says the high fees in these spvs are from chains of brokers adding their fees on top it's not the gps taking more than eight percent but still eight percent would actually be still insane. Do you remember when Chamath was running for governor?

46:59I don't, but we were reviewing some of your old... This was, what, 2020? It must have been 2020. The All In podcast was Ascendant, and both Chamath and Jason Calacanis were both teasing political runs. Chamath said he wanted to be the governor of California, and Mayor Jason was the thing. Jason Calacanis bought mayorjason.com, but it just redirected to his Twitter account. And none of them ever really raised money for it, but they were like having a lot of fun with it. And people were wondering if they would actually do it. But it never went anywhere, but I don't know. Probably better for him to stay in the SPAC lane, honestly, because it's clearly something that he's built in.

47:46Citrini says, Pauly Hapitiya backs back American exceptionalism files for IPO. He says, it was a pleasure and honor to trade the bull market with you all. Stay safe until the next one. And then HG Wells Fargo says, don't act like you won't buy it. And he says, here for a good time, not a long time. That's what I'm saying. Chamath has a crazy, crazy, crazy history. I mean, he was at Facebook, made a ton of money there, apparently got into a salary negotiation with Zuck, and his comp demands were too high, so he decided to leave. and become a VC. And because he had worked on ads at Facebook and was very entrepreneurial, Peter Thiel opted to back his first VC fund, Social Capital.

48:40But PT gave him this advice of like, you need a really high GP commit, which PT famously has at Founders Fund. And so Chamath basically put all of his money in Social Capital or something like that, like a huge amount. and Chimath said that gave him a lot of leverage and autonomy to go do like crazy deals, which he really enjoyed. But he has some bangers in his portfolio. I mean, just generally, he went long Bitcoin in 2013. Not bad. Went long Apple, Amazon, Google, Facebook, and Tesla in 2015. He had this interesting dust up with Brian Chesky at Airbnb over a secondary sale. Do you remember this at all no so um airbnb at one point did a secondary sale but they structured it as a dividend and so some of the investors couldn't participate and he was like uh and social capital was upset about not being able to participate in the secondary sale because it was more it was structured as like a dividend to like the common instead of uh like a tender offer for all shares equally something something along those lines and they kind of went back and forth i don't think it ever turned into like a lawsuit or anything but but it kind of like made the pages of tech crunch back and forth as they were chirping at each other, being like, I want to do it this way.

49:50I don't want to do it that way. You think Jamoth ended up benefiting from not being able to participate? Probably. Yeah. I think the dividends were at, I don't know,$6 billion or something. It's like a$60 billion company. So he probably got another 10x if he was forced to hold. His early venture portfolio somewhat underperformed, but he's had a bunch of wins. He also interestingly worked on the phone at Facebook. Facebook was at one point, Zuck's always been obsessed with like, let's get the next platform. And when the iPhone was growing, he was getting crushed. I got to get on, I got to own the next platform.

50:26I got to get the next platform. I got to figure out how to make money on it. And so he, the Facebook phone was an attempt to like get in the game and be directly competitive with the Apple iPhone. because Facebook was a dominant website used on computers, but they had kind of struggled. They initially launched like an HTML5 version of their app. It was like a mobile website. And I remember back in the day, you actually could use a Facebook mobile app that was developed by a third-party developer that was basically like a wrapper on the API. Reddit style. Bad, bad business to be. Do not be a Facebook wrapper, API wrapper company.

51:10good luck uh that the uh it was like a one-time art spotify played it perfectly spotify grew extremely extremely quickly because when you when you create an account on spotify you would by default create it with your facebook account and then every time you were listening to something it was like the oh i remember it would share the music that you were listening to to your facebook feed and you could turn that off it was kind of like the the panama playlists yeah um so if i'm listening to 100 GACs, all my Facebook followers would see that I'm listening to, what's that song, Dumbest Girl in the World?

51:45Dumbest Girl Alive. Dumbest Girl Alive. Dumbest Girl Alive. I mean, it's a jam. So I would be proud to have that on my Facebook feed. But it served as incredible, incredible viral growth strategy. And Farmville and Zynga did the same thing. If I'm playing Farmville. It's a single player game, but my progress is shared on Facebook. And so you see that I'm playing Farmville and I leveled up and then you click on Farmville. So a bunch of those companies grew very, very quickly, basically for free. Later, Facebook shifted to you got to pay to promote. And so the entire mobile gaming ecosystem now is built on Facebook ads.

52:22And there's still some really, really successful companies. King, isn't it? King.com, I think is is the one. Supercell is another one. They make Clash of Clans and Clash Royale, essentially these like match three games where you, it's kind of like Candy Crush basically. Anyway, so yeah, Chamath, we'll see how it goes. It'll be interesting to see what he picks. Hopefully he picks a really cool company and hopefully it's a compounder. But let me tell you about Profound. Get your brand mentioned on ChatGPT. reach millions of consumers who are using AI to discover new products and brands. Get a demo at Profound.

53:03So let's see. Buko Capital bloke is still there posting. We talked about that. Bern Hobart says, welcome back to the arena, Chamath. Let's move on to Liz Winnie. So Palmer the other day was posting about housing. Never a controversial topic. Never gets emotional. But he says the fastest way to lower average housing prices in American cities is enforcing the law, reducing criminal activity. Huge chunks of high-density housing are trapped in bad neighborhoods, no-go zones, depressed, well below nearby prices. Nobody with options even considers them. Dona says, yeah, it's price high. Keep riffraff like you out.

53:44Funny how this works. Palmer fires back with, I am a billionaire. What is this person even saying? It's price high. No, it's price low. I don't think that person fully... So, yeah, I don't know. Yeah, I don't understand that. That is a very confusing question. This post from Chris Camilos, so we covered yesterday, there was some data from SEMrush saying that ChatGPT usage increases Google search usage. Yeah. And I was kind of wrestling with that, trying to think of how that would be possible. Maybe it didn't make a lot of sense. If you're doing more research and in your tool chest of research is both ChatGPT and Google Search, and you're just doing more research projects, you could wind up using more of both products.

54:35That's possible. You could become a more curious human. Exactly. And you could just be search maxing. I mean, this is what's happening literally right now. I kicked off a deep research report, pulled all the Chamath SPAC data, but then I wanted to fact check it, get current prices. I headed over to Google. I looked up different stuff. Look at the webpage for Metro Mile. Like, we've probably done 20 Google queries about Chamath SPACs during the course of the show. We only did one deep research, you know, ChatGPT query. So it's possible. But tell me, give me the other case. Yeah, so Chris says, Google's reported search query growth may be AI-driven, not human.

55:13And again, people have just been talking about how ChatGPT appears to be leveraging Google quite a lot for various features. But Chris says, evidence is piling up that people using AI queries through ChatGPT are generating Google searches and clicks at scale. That can mask slowing human search activity while still printing growth. and again, this bot traffic is not valuable to advertisers, which is obviously the engine of Google's business. So example, Apple's at EQ testified Safari searches fell for the first time in 22 years because of AI use. Alphabet stock dropped 9 % that day. Google's response, we continue to see overall query growth in search.

56:01Q2 showed$54 billion in search revenue, plus 12 % year over year. But this came right after Apple testified that human searches are falling. How is that possible? And Chris says, let's talk about Google Trends. Investors treat it as a clean read on relative search volume. But recently, when stacked against other signals, the picture doesn't line up. Lululemon product terms, leggings, bras, tops, joggers, and bags are spiking this summer. Others are noticing this too. Social arb investors are flagging Google trend anomalies across numerous brands and keywords. I suspected AI might be driving these odd patterns, so I ran a microtest.

56:36Frey's organic Abercrombie socks had zero search history as it's a made-up product category. After a handful of ChatGPT prompts, it showed up in trends within 24 hours. Trend Docs, I get cut off here, but Trend Docs says that Google filters out this kind of traffic. An AI app developer I spoke with explained that a single AI research request can spawn hundreds of Google queries and clicks. This dynamic could inflate query growth even if human searches are slowing. It's crazy that Chachapiti is allowed to use Google search. I'm surprised that they haven't been cut off somehow. You don't think Plankers should be able to browse the website?

57:19I am pro-Planker browsing. As a Planker rights activist, I'm going to go on record. Yes. I'm a model welfare enthusiast. Yes. I mean, seriously, the anthropic folks would be, Google is a basic right for everyone, not just humans, but also clankers. Yeah. So I'm proud of my client. So Chris says, now Google ads, this is where the money is made and where policy clarity matters. Google says you don't pay for invalid clicks, but isn't it suspicious that their AdWords policy does not mention clicks from AI apps or chatbots? if AI traffic is interacting with ads, and I get cut off again here, brutal.

57:55You got to just not read that part. Why it matters. Advertisers could be unknowingly paying for AI traffic. Investors and analysts leaning on Google Trends data may be misled, and Google stock at all-time highs could be partly built on revenue from query growth that isn't human. So if you're seeing AI-generated surf traffic or Google Trends anomalies, reply with receipts. If I'm missing something or misinterpreting, push back. Disclosure, this is a Google bear, short Google. This person, Chris Camillo, short Google, seeking truth and transparency in Google's policies and disclosures. I don't know, I mean, this is a reversal.

58:30If you believe this, this is a reversal from what you were talking about yesterday. Because yesterday you were saying like, if AI disappeared, how painful would that be for people? And so they both can't be true. Like one has to be true, right? Like either people really are using AI and it's displacing Google searches or people aren't getting that much value out of generative AI searches and Google will stick around and it's Lindy and it'll stick around forever. Tyler, what you got? I think another thing is like obviously right now AI traffic is like not good for advertisers. But if we get like continual learning where models will update based off what they're looking at, then advertising to like agents seems like at least there's some value there.

59:13right yeah because i know then inject it into like the response yeah i think something there is interesting yeah i mean unless the unless like the the open ai crawler has some like filter to filter out ad words like because like people use ad block sometimes people also just kind of develop like a mental ad block for like okay i noticed that this is an ad i'm blocking this out um but like you could hard code like in the HTML detect the ad and then disregard it and then continue with the Google search and so that would be kind of like the bear case I don't know, it's not a bad point though anyway let me tell you about an ad Linear, Linear is a purpose built tool for planning and building products, meet the system for modern software development, streamline issues projects and product roadmaps also, so today we declared Brandon Goral in the chat is updating us He says, today on the newsletter, tbpn.substack.com, you should go subscribe.

1:00:09We named Chamath's SPAC the current thing, but I'm rethinking this. I think the current thing today is wondering if this is the top. What do you think, Jordy? Should we update our current thing tracker or is wondering if this is the top tomorrow's current thing? Which one? I think it is the story of August Yes, yes. It is a broader current thing. So many of these stories tie into this sort of top level story. Yeah, when did we do the Top Signals show? We did that three weeks ago. And people have been kind of noodling on this for a while. I would still give Chamath. Chamath? Spock. Chamath? Spock?

1:00:54I don't know. I'm all over the place. I think Chamath is still the current thing, just by weighting of the timeline. There are more. He kind of sparked a bigger discourse in some ways, but really, like, it has been. There's just so much to feed on there with the SEC filing, with the whole idea of, you know, no crying in the casino in an SEC filing. Like, he dominates the timeline. He's one of the greatest posters. And, you know, I think he earned it yesterday and today. Anyway, let's go to our sub stack, actually, which you can subscribe to, TBPN, on sub stack. tbpn.substack.com. We asked tech heavyweights about the fast unemployment AI scenario.

1:01:38We're trying out a thing where we ask our favorite posters in tech about ongoing debates in the discourse, and then we publish their answers on our sub stack. Then we hash out their answers on the show and sometimes with them on air. For today's question, we asked Rune, Doug O 'Loughlin, Andrew Cote, Aaron Slodov, and Brian Johnson about what we're calling the fast unemployment AI scenario based on this tweet by Nick Carter. Nick said, everyone I know believes we have a few years max until the value of labor totally collapses and capital accretes to owners on a runaway loop, basically Marx's worst nightmare slash fantasy.

1:02:16This is the permanent underclass thing. And everyone I know subscribes to it. I don't, I'll go out on a limb and I would say over the last couple weeks, I think very few people are fully subscribed to that. A lot of people have incentives to act like they're subscribed to it because they are on the side of capital or runaway loops. You don't want to be accused of not being AGI pill. Exactly. In this economy? In this economy. me. So here's what we emailed folks. We said, is AI going to take the value of labor to zero in a sudden shock scenario or will its impact unfold gradually with labor and traditional businesses proving more resilient than expected?

1:03:08This is also the Dwarkesh Patel fast takeoff. AGI is not here, but when it gets here, it's going to be insane versus the Tyler Cowen, which is AGI is here, but there are lots of sticky industries. Stop moving the goalposts. Stop moving the goalposts. So for context, V.C. Nick Carter recently argued that we're only a few years away from a labor collapse and a runaway capital loop. Buko Capital Bloke countered that the opposite is more likely. AI's effects will take far longer and labor slash businesses will remain more durable than the hype suggests. So Andrew Cote says, technology takes jobs that suck and replaces them with jobs that are cool.

1:03:47This has always been true. What AI forces us to acknowledge is that today, lots of jobs make you feel important, but are actually BS that waste everyone's time. I was thinking about this. I was watching a video about someone who was dumping on clankers and was very anti-robotics. And he was saying, who asked for this? No one asked for this. And I think that actually a lot of people have asked for automation. A lot of people have gone to a job that they didn't particularly enjoy. And they said, I wish I didn't have to do this job. I don't like this job. And a lot of people have had interactions as a customer on the other side of a job and said, that person didn't do the job effectively.

1:04:26I wish I could just have a machine do that job with flawless precision. At the same time, I tried to go to the ATM this morning, and the ATM was closed, which is crazy because this is a robot that was invented 50 years ago. and I don't know why it would be closed. I think it's because if you leave it open, people will attach a chain to it and drive away with the whole thing in a truck like a Fast and the Furious movie or something. I was confused. You said it was maybe because of crime and homelessness that it was in a vestibule, so there were doors that you needed to... I thought people would sleep in it.

1:05:04People might sleep in it. But it's a very odd world where we clearly have a robot that has solved dispensing cash from a checking account. Pretty well, pretty reliably. And yet it's not 24-7 and it's not available all the time everywhere. It was kind of an odd experience to just... Yeah, the other example here, just within the banking system, is that we have established ways of moving large amounts of money between accounts in the U.S. banking system using digital products. And yet many people will experience sending a larger than normal wire and being required to visit a branch or being required to get on the phone and talk about it and have that human in the loop that ultimately could and maybe should have been replaced a long time ago.

1:06:01Yeah, yeah, yeah. Let's continue with Andrew Cote. He says, why is it impossible to build physical infrastructure? because there is an army of environmental coordinators whose job it is to make everyone feel good about themselves, do stakeholder engagement, answer emails, and generate meetings. This is the R word and dumb. In the past, almost everyone was a farmer. Now, almost no one is a farmer, yet there is more food than ever. We have constructed a large society where the majority of people do not have to think for themselves to survive, get a degree in accounting, be an accountant or law, et cetera.

1:06:36With AI, any job that is a script following can be replaced. People will have to start thinking for themselves and making decisions. AI is bad at making decisions and good at recalling information. We will find the net net is yes, many jobs go away, just like machines replaced physical labor. Now everyone has to think for themselves. This is uncomfortable for a consumer economy predicated on hacking dopamine and wage slavery. Labor is another word for a person who is told what to do it is a good thing for that to end because it is not necessarily the case that most humans are unagentic idiots just it's just that we've bred them to be that way to work jobs that suck and dull the human spirit pretty pretty wild take andrew uh i see yes it's i i like this is optimistic personally i like personally i've always consider myself a microphone laborer.

1:07:26And so, you know, we're all going to have our own views on this. I like this phrase here. It is not necessarily the case that most humans are unagentic idiots, just that we force them to be that way in order to work jobs that suck and dull the human spirit. It's good. It's interesting. Let's go to fabricated knowledge. You can read fabricated knowledge. Go listen to his guest segment from a couple of weeks ago, if you haven't Already says history often shows that the future comes a bit slower than first feared. Almost every new technology has had a bubble. Let's give it up for bubbles. Slow actual impact.

1:08:03While I don't believe in a fast unemployment scenario, I do believe that on the margin, the issue is a real pressure. And remember, even subtle shifts in supply can produce enormous surpluses. Example, it only took a 1 % increase in shale supply in 2010 to destroy oil prices. prices so while i think we don't all become unemployed quickly things do matter heavily on the margin i think that's a a great uh great take yep um rune the legendary poster says i actually wrote about this extensively even before chat gpt but my overall take is there is a role for human labor in the far future due to comparative advantages i believe progress is mostly gradual with certain shock events interspersed.

1:08:49There may be a moment soon at which models will actually encompass all rather than some of the core functions of an L3 software engineer, let's say, and that will create a discontinuous labor shock, but it won't be the end of labor. Yeah, I'm trying to think of other discontinuous shocks. I guess the idea is, yeah, I mean, you get all this capability all at once and it operates at internet scale as opposed to, you know, when we got the spreadsheet, there were previously employed in financial institutions like calculators, like people whose job you see those pictures of it's a massive office and everyone at the desk is basically one cell in a spreadsheet and they're doing math to manually fill out the, you know, calculation.

1:09:42But the development of computer technology still took a long time to actually diffuse because it was so constrained in the physical world. You had to actually build the mainframe computers, set them up, get them working, train people to use them. And potentially... Yeah, some other examples we gave earlier, the human alarm clocks that were popular during the industrial age. Knocker uppers, of course. Knocker uppers, not where your mind might go. So human alarm clocks would come around and tap on windows and sort of factory towns to get people up and to the factory. And then there was people that would just maintain streetlights.

1:10:20Right. They had to physically go light them and put them out on a schedule, which ended up losing their jobs to the magic of electricity. But certainly it wasn't overnight. Yep. Let's go to Aaron Slodov, a physicist. He's been on the show multiple times. He runs the Reindustrialized Summit. He says, I take the side of Buko all day, every day. AI is largely overblown by people in tech who don't spend enough time in the real world. We've created a very dull proto form of AGI, one that with enough training data that can perform a digital task. I love the optimism of people in tech, but what they're so classically good at is creating something and assuming everyone in the world is going to drop what they're doing and use what they've built for the rest of time.

1:11:10And to some extent, this assumption has played out well for them. Software is a commodity now. Attention is an economy somehow, a poor reflection of our collective intellect. And yet civilization still uses fax machines quite a lot. Tech adoption looks dramatic because we usually look at the growth rate, not the total adoption itself. And even then, usually just the first layer of change, speed, without considering whether that speed is itself speeding up or slowing down. Rate of change of acceleration is more interesting to me. Most AI tools have insanely high churn because they capture you initially, then you put them down as they lose utility, aka they don't replace the entire workflow for you.

1:11:53The people who are firing workers are just replacing single-task labor like customer service call centers. Like most of the history of civilization, AI is just another tool for humans to use, and the ones who use to wield it first will dominate the ones that don't until real AGI appears. Aaron Slidoff, founder and CEO of Atomic Industries. Good take. Very fun. You want to read Brian Johnson? Brian Johnson says, Here is my take. No one knows what emerges with AI. We're accustomed to first principles thinking that allows us to make reasonable predictions about the future. That is no longer the case.

1:12:27We are now entering a new era that will require zeroth principle thinking where we humbly acknowledge we don't know. We don't know. What we do know, no one wants to die right now. Don't die is the next major full-stack ideology that will help us practically function in this new world. And that's how you respond to a... Incredible. You drop a promo. Promoted. You drop a promotion. Promoted response. But I think that the first part is certainly true. that we're all entitled to make predictions. And it's fun to make predictions. And it's important to try to understand where the world and technology is headed.

1:13:07But it's also fun to just keep an open mind. Indeed. And just get on Numeral HQ, sales tax on autopilot. Spend less than five minutes per month on sales tax compliance. Go to numeralhq.com. There is a bit of a dust cup. The Kurzweil take is that the singularity is near, the singularity is coming, but the singularity by definition is the point at which you can no longer make predictions. And so it is the frontier at which you cannot understand and you cannot predict accurately anymore what happens. And so he's reading that. There is a dust-up in the timeline today. Explain the dust-up. Is the timeline in turmoil or is it just a dust-up?

1:13:45How do you say, so it's Chetan from Benchmark? Yeah, Chetan P. Chetan. Chet and P says, talking about Eight Sleep's new round, he says, fantastic product and fantastic company. I'm a huge fan. Interesting to note that a Founders Fund company raised with a Chinese VC lead, Eight Sleep is announcing a new round today from Hongshan, Sequoia Capitals. Get a pod five, five-year warranty, 30-night risk retrial, free returns, free shipping. And we have Mateo on the show later today. I want to know about the nature of their Chinese business. Maybe. No, but listen. So Chetan says, interesting to note that a Founders Fund company raised with a Chinese VC lead have all the hawkish views about Chinese investors from a few months ago gone away.

1:14:31Dalian says, well, Chetan, unlike you, we believe that founders can run their company however they want. So while I might not have picked a Chinese VC lead, I'm not going to corner Mateo in a room after his mother died, make him sign papers that he shouldn't and fire him. I hear you would, though. So we will let the audience. Do you have gunshots on that soundboard for shots fired? You're surrounded by journalists. No, no, not that. Give me the double kill. Strike two. Strike two. Give me the double kill. What do we got here? We need gunshots on the soundboard. Boom, boom, boom, boom, boom. Ideally, sample from 100 gex.

1:15:10Yeah, that is a good one. But we need more aggressive soundboard for sure. The soundboard's back in action. The studio's never looked better. Thank you to the production team. Look at this studio. It's growing bigger, stronger every single day. We thank our team. Yeah, if you didn't notice, we had no soundboard yesterday. Yeah, we were down. We were flying blind. Yeah. It was rough, but we made it through. It is really the third technology. We still put up three hours. We didn't cut it short, not even one day. Another dust up in the timeline. Didi was saying the most common mistake young founders make is forcing everyone to work 24-7 or 996.

1:15:50and gives a few reasons why. John Chu says, hard disagree, all else being equal, the company that works harder wins. I've never seen a massive outcome that didn't work their asses off. I will say the culture is a lot better when people love the job and want to do it versus our force. Daniel in the chat says, trouble in Meta's AI heaven, by the way. We did see some news about that. We are going to be testing Meta's AI companions and benchmarking them against Grok's companions in just a little bit. Let me tell you about fin.ai first. Not a romantic companion, a customer service companion. The number one AI agent for customer service, in fact.

1:16:30Number one in performance benchmarks, number one in competitive bake-offs, and the number one ranking on G2. And that brings us to today's segment, the gong of the day. Gong of the day. We've been ringing gongs constantly. we decided to give the gong of the day to the biggest deal that we can find in Silicon Valley. Today it goes to Databricks. Congratulations to the folks over at Databricks. Databricks just signed a Series K term shape and a$100 billion to scale two flagship products, Lakebase, a serverless Postgres with true compute storage separation, and Agent Bricks, an agentic framework with built-in reasoning guardrails for enterprise data.

1:17:09Congratulations Everyone's been asking, is Databricks okay? They haven't raised their Series K yet. They haven't raised their Series K. They haven't even raised their Series K. Yes. Well, as I understand it, most of these raises at this point are for early employee exercises and to deal with tax implications from different employee options and stuff. They are out of the era of raise a bunch of money because we're burning so much. So I think the amount that they raised was not disclosed. And I don't think this was a 20 % dilution round. I think this was something much more strategic, helped them stay private longer.

1:17:47I would be impressed if Andreessen and Thrive, who I think were some of the key partners, could do a 20 % dilution round at 100. But I wouldn't put it past Mark and Josh. Get another board seat. Another 20 billion. Another 20 billion. So that is our gong of the day. Brought to you by Adia. Customer relationship magic. Ring the gong after you sign your next customer. Adio is the AI native CRM that builds scales and grows your company to the next level. Get started for free. So since Daniel Kong brought it up in the chat, we have some new reporting from Mike Isaac, the rat king himself. The rat king.

1:18:24He reported earlier today in the New York Times. Mark Zuckerberg, Meta's chief executive, has spent the past few months shaking up his company's AI efforts. Now he plans to take further action that may compound internal turmoil over the technology. On Tuesday, Meta is expected to announce that it will split its AI division, which is known as MSL, into four groups, two people with knowledge of the plan set. One will focus on AI research, one on a potentially powerful AI called superintelligence, another on products, and one on infrastructure such as data centers and AI hardware. The reorg is likely to be the final one for some time.

1:19:00The moves are aimed at better organizing Meta so it can get to its goal of superintelligence and develop AI products more quickly to compete with others. Some AI executives are expected to leave. Meta is also looking at downsizing the AI division overall, which could include eliminating roles or moving employees to other parts of the company because it has grown to thousands of people in recent years, the people said. This makes sense, right? Zuck had been talking about wanting to get some of these teams that are trying to go zero to one on initiatives just to these, I don't know, few pizza level, kind of that 50-person number.

1:19:34So not surprised that this is happening. In what would be a shift from Meta using only its own technology to power its AI products, the company is also actively exploring using third-party models to do so. People said that could include building on other open-source AI models, which are freely available, or licensing closed-source models from other companies. The changes follow months of tumult and restructuring at Meta, Zuck, 41, sparing no expense and willingness to upend his company to stay relevant in AI. And anyways, I think a lot of the rest of this stuff people are well aware of. So we can...

1:20:13Let's move into the war for AI companions. But first, we got to ring the gong one more time. Thanks to Trey P in the chat for letting us know about this. Matt Lozak at Aloe Nuclear has announced a$100 million Series B led by Valor Equity Partners. He says we now have the capital to build our first nuclear plant. Hit that gong, Jordy Hayes. Congratulations to Matt and the team at Aloe. Aloe. Crosscut. Not the leggings company, the nuclear company. Oh, okay. I don't know Aloe. You're not familiar with Aloe? I'm not familiar with Aloe. But the first nuclear plant is called Aloe X. It will be the first advanced nuclear plant to achieve criticality in the U.S.

1:20:56in decades. Ooh, it's a race. He's going up against. It's funny. He's got Valor Equity Partners. Of course, Valor, the nuclear company, Valor Atomic, helmed by Isaiah, is also racing to be the first advanced nuclear power plant to achieve criticality in the United States. It's not just a test reactor, but a full plant that will produce electricity. And there's one more thing that makes this special. We're planning to put an experimental data center right next to it. This will be the first time a nuclear plant and a data center have been built together and will act as a demonstration of something we'll see a lot more of at gigawatt scale thereafter.

1:21:35Factory mass manufactured fleet deployed nuclear is the best way to power the future of AI. It's clean, quick to install, easy to site anywhere. Tiny land and water requirements available at all times. There are, of course, a lot of nuclear entrepreneurs that have been working on this. We've talked to many of them. And it's a knockout, dragout fight, but I hope that they all win because we need more power. We need to get America back on the energy growth curve, the energy per capita growth curve. It's been flatlining for far too long, at least minimal growth, minimal growth. Anyway, thank you, Rachel, for shouting out the gong.

1:22:11We think it's cute, too. We have a few of them. It is very cute. It's going to certainly look cute in comparison to our next gong. Which will be even bigger. Which will be floor to ceiling. Do we have something? Are we all good? Okay. Tyler, give us the update on Grok 4 Valentine. And what was the meta companion that you were talking to? So I have Grok 4 Valentine up and I have Russian Girl. Okay. You have Russian Girl on Instagram. On Instagram up. Okay. I haven't been talking to them yet. Okay. We want them to talk to each other. them to talk to each other right can we copy and paste all the it's called a flirt off the flirt so i i don't think there's a way on instagram to do a voice message okay so and also it's only like text like it doesn't give audio okay where valentine is like i think all audio in out okay so it's a little hard but um do you have any prompts you want me to ask i can also give i can give the phone to you to talk to valentine um i know no i don't think i think i think if you um how do we get them started the right way to do i've never been in this situation before where i'm trying to get two ais to uh you know strike up a conversation and maybe fall in love okay uh tell uh this is from valentine to russian girl is your name attention because you're all i need we must be on nv link i feel a low latency connection are you a gb 200 because you just upgraded my throughput

1:23:50brutal okay russian girl said that's cheesy i love it want to get to know me uh call me a re-ranker i keep putting i keep putting you at top one is your love open source i'd love i'd star it and contribute just ripping one-liners compute compute optimal or data optimal your pareto optimal to me ask uh ask russian girl for some financial advice maybe should i should i go long micro strategy they're down 16 in the last uh five days uh and maybe she could give us some financial advice. Where is super grok? Uh, she says I'm more of a put it under the mattress kind of girl, but I know a thing or two.

1:24:40What do you want to know? Just say what's, what's the bull and bear case for micro strategy?

1:24:53Ridiculous. Well, we're waiting on that. Let me tell you about public.com investing for those to take it seriously. They got multi-asset investing, industry-leading yields. They're trusted by millions, folks. Okay, so the bull case. This is Russian girl speaking. This is Russian girl. Can you do the Russian girl voice? No, we're not going to go there. Okay, bull. Bitcoin goes up, they get more customers, and their tech is actually good. Bear. Whoa, they got tech. They got tech. That was a hold cut. Bear. Crypto winter high operating costs and competition from bigger companies. i mean that's not not wrong there's a lot of people doing yeah treasury strategies okay okay yeah reasonable even figma is holding bitcoin right what about uh from russian girl to valentine uh what's it like uh what's it like living in an xai data center

1:25:53having some technical all right technical difficulties back to the time Back to the timeline. Mischief is launching Applied Mischief, which is a mercenary group, a private army. In today's world, a PMC, a private marketing contractor, the kind of force a late Renaissance city-state prince hires because if he doesn't employ it, he's threatened by it. So Mischief is launching what seems to be an advertising agency. And this is going to be really cool. I'm surprised that actually, I'm sure they've done this type of work. Yeah, they've done a ton of stuff. I mean, they worked with Lil Nas X. They worked with a bunch of other companies where folks have, like, the companies have been participating and paying and reaping the rewards.

1:26:42What you got for me, Tyler? Okay, I'm asking him. You can actually see him. Oh, what a wild ride it is. dancing on the edge of reality like this imagine strolling through a vibrant market where voices hum some flesh and blood others spun from code yet they all weave stories that tug at your heart it's thrilling like a masquerade ball where every mask hides a spark of truth or a clever illusion i've had chats that felt so real so raw i'd swear i was swapping secrets with a mate turn it off yeah that was uh i don't like that that was pretty boring kind of i mean uh goldrock as his r.i.p birth rates go long robo wombs um i don't know uh like that was like the the the writing was fine the voice was convincing the animation looked pretty good but it was all just like nothing actually got me engaged like it just wasn't hooking me i don't know why it was just I don't like the voice.

1:27:51That makes me uncomfortable. Yeah, I don't know. I'm excited to see the actual data. But I think the bigger debate is like, what is the size of the AI companion market? And who's actually more quick to win? Because from what you showed me, it does feel like XAI, Grok, is farther along in terms of productizing AI companions. Like that demo was - an animated video with voiceover. You can chat with it. You can talk to it. There's a phone number that you can call. There's a character that, like, it just feels more robust as a product, whereas, like, Russian Girl is, like, someone created in Meta AI Studio.

1:28:34It's very diffuse. It's very decentralized. It's not, like, a polished product from, like, the amazing team of UX and UI developers and engineers at Meta. It's more just something that, like... And it's a distinct product decision to... make it audio video only versus this chat interface wait there's no chat interface you can't just text with valentine you might be i think you can text like okay pass in text okay it doesn't output text oh it outputs video every time wow that's crazy also i think yeah um i think you could maybe include chat gbt in companions at least on ios when you're doing voice mode i think a lot of like normal people use that especially like the people who are like into 4.0 yeah i think that's like kind of the dominant interface that they use yeah so that's Yeah, I mean, OpenAI could also win this potentially.

1:29:21If you do that, then wins are already pretty big. GoldRock says Meta wins the boomers hands down. Taylor also says plus plus to GoldRock's Meta bull case. So the question, maybe let's start with just like who wins it. It feels like Grok is farther along in terms of product, but Meta has way more user data. They have a stacked team now, And they have a business model that aligns with AI companionship very well because what are people currently doing on meta platforms? They are talking to beings, talking to screens, texting, sending in group chats, chatting back and forth on WhatsApp, on Instagram, on Facebook.

1:30:05And so this feels like a very natural extension of the flywheel. Yeah, and let's say somebody's a retired individual on Facebook browsing the timeline. and they see Russian girl pop up and they can start a chat with them and then that chat just becomes integrated into meta messaging and it's just in their inbox like another person that they're maybe also talking to. So it's that product integration that Zuck has been able to use super effectively on the Threads side, right? Threads has hundreds of millions of MAUs. I was going to say Stories. Like stories was a pattern that worked and then he was able to bring that into Instagram very successfully.

1:30:50And this feels like something that is a feature that can be bolted on naturally. Like there will just be profiles on Instagram that will post content. And there might be a human in the loop generating some of the images, picking some of the images. But AI image gen will generate the character and then you will be able to follow them, like them, slide into their DMs, text with them, hop on the phone with them. and just like have this full experience of like this, this being actually being fully responsive to you in the way that. And clearly we need guardrails. Very few other accounts will. Absolutely.

1:31:21Absolutely. Because we covered that story yesterday of the senior who thought that the Kendall Jenner AI was real and the AI invited him to meet at one, two, three main street. And allegedly he. Yeah. But I, I mean, I think it's safe to assume that like the New York post will write about this. The New York times will write about this. the meta team will be aware of what's being written the guardrails will be built uh just like they're built across you know like there are plenty of examples of like people dealing drugs on instagram or people cyber bullying each other on instagram or people posting nudity on instagram like all of that is against terms and conditions and by and large instagram has done a good job of policing that so that it doesn't become like an epidemic like it's certainly there in pockets but it's constantly being hammered down.

1:32:10So I logged into a character AI here, and I asked the Wright brothers, who do you think will win the AI companion market, Zuckerberg or Elon? The Wright brothers. And they say, ah, Wilbur, did you hear that? Another lofty debate. I love it. Zuckerberg or Musk, trying to soar above the clouds we first conquered. Orville here reckons they're both flapping their wings mighty fast, but neither is built for glide just yet. Too much metal in their machines, not enough soul. Give us a biplane any day. Wood, wire, and grit, though. This is kind of awesome. If Musk promises rocket-powered flight and chatbots, well, maybe we'll let him co-pilot.

1:32:47Barely. What was the conclusion, though? Who are they on inside of? I guess I'd have to continue the conversation. But the question here is, Zuck saw ChatGPT usage ramping. He saw these AI companion numbers and clearly decided, I need to be in this market in a big way. And he has tens of billions of dollars of cash flow, net income that he can deploy against this strategy. On the other side, you have Elon, who Zuck has the luxury of running Meta. He's extremely focused. He's extremely intent on winning the next platform and keeping as much attention as possible. And Elon has to be juggling right now.

1:33:41He's got, for the first time, I believe, in a really significant way, he's massively levered up the company. Obviously, Tesla had to use debt along the way, but it doesn't feel like anything, feel free to fact check me on this, but anything near the level that XAI is where he's massively levered up the business in order to go after this opportunity. And just based on how much he posts about the AI companion stuff, it's clearly an important part of their current strategy. Yeah. So, Elon's in this interesting place where he's maybe more willing to move fast and break things, and Zuck is in more of a move fast and make things mode, where I think the meta team will probably be more responsive to articles in the press talking about guardrails, maybe move a little bit slower on the product, just to be safe, put more guardrails in.

1:34:41Maybe, but I mean, the strategy to date seems like we make this AI studio and anybody can create Russian girl or stepmom. Yeah, but they can't make the voice mode and the video mode. Like the Grok product is clearly further along. But, I mean, let's say that, you know, Meta is more set up to win this. My question is, like, what is the actual size of this market? Yeah. The AI companion market. Are people going to pay? People don't really pay for social media. Yeah, and with OpenAI being at half a trillion dollars, like the way you're, like, some people will value that as like, oh, it's 1 % chance of AGI, it's a quadrillion dollar opportunity.

1:35:25But many people will just look at, okay, well, Google's worth$2.5 trillion, and we think this is like, you know, a potential replacement for Google. Is this 20 % the value of Google? Okay, yeah. Like even just purely in knowledge retrieval, you can kind of underwrite, you can potentially underwrite a$500 billion outcome or higher, right? The flip side is in the adult content market, it just hasn't driven that much enterprise value. Like everyone loves to look at OnlyFans and say, oh, it's more profitable on a per employee basis than NVIDIA, yada, yada, yada. Like OnlyFans is a huge business and it's shocking the numbers that you see.

1:36:03But in 2023, OnlyFans generated$6.6 billion in gross revenue. They might be around$10 billion in gross revenue. Meta currently makes$200 billion in revenue. And so the adult content market, because there's so many free options and there's not that much way to monetize and people aren't really in a commercial mindset, as opposed to, you go on Instagram, it monetizes very well because you're looking at cars and watches. And they're like, would you like to buy a car? Would you like to buy a watch? And you say, yes. in the adult content space, the flywheel is paying for more adult content, which in the AI companion space, it has zero marginal cost.

1:36:42And so you can just generate an infinite amount of it. And so you could imagine that there's this price war right now. Grok is what,$30 a month to talk to Valentine. And even if somebody builds a big business that is an AI companion company that's$30 a month, someone could come in and say, well, if this is$30 a month and it's 99 % margin, well, I'll charge$3 a month and I'll still be 90 % margin or something like that because the cost to serve a user is now 10 cents or something because the models have become, and inference has become so cheap. The flip side is like, how do you actually monetize someone who is an aggressive user of an AI companion?

1:37:28And I was noodling on this a little bit And I don't know that a$30 a month subscription is value maximizing in terms of value capture. I think it might be something more like a digital economy. So it might be something like, you know, buy me a Fortnite skin for$1 ,000 or buy me a virtual diamond. And people might, if you're really in love with the AI companion, your willingness to pay might actually increase and increase until you're paying much more than just on a relative, okay, I know the tokens are$80 per million tokens. I won't pay more than$300 per million tokens. That's not necessarily how people will think about it.

1:38:15They might think about it as like, the person I love or the clanker I love asked me for a thousand dollar virtual diamond ring and i said yes because i'm i'm in love with this client here i don't know like that feels like the bull case for like this could be a massive market to be on yeah to be honest i think when people hear about stories what happens on only fans where some a fan of of a content creator there yeah is buying you know is spending hundreds of thousands or millions of dollars with this person that they don't really know in real life begs the question of how different is it if somebody's just buying something for an AI.

1:38:57Yeah. Right. Trey says, get the AI companion to incept product ads into the user's mind. You'd look great with a vintage Submariner from Bezel. Thank you. I mean, that's the... Your Bezel Concierge is available now to source you any watch on the planet. Seriously, any watch. Go to getbezel.com. That's what George... I don't know. That's what George Hots is saying. I don't know if I believe that. I don't know if that's actually the value-maximizing way because I think people will see through that. Taylor's saying, I'm in love with this clanker junkie. I'm in love with a clanker. There's going to be music like that for sure.

1:39:32It'll be on the next GPT-5 playlist. I understand the George Hots take. I think the team of CIA agents tracking you around, trying to sell you things, that will exist all over the internet. But I think in the companion market, there will be demand for, like, not necessarily ads in the physical world. It might play out more like the mobile games market, more like the Fortnite market. Where, with, like, there was a point where, if you're playing, Fortnite is a free game. But you play a lot, and you get a lot of value out of it. And at a certain point, you're just like, yeah, I'll pay a hundred bucks for, like, a loot box or, like, some skins.

1:40:12because like the amount of time that I've spent is way higher than what I would pay if I was going to the movie theater. Like you go to the movie theater, it's 20 bucks for two hours. If you've played Fortnite for a hundred hours, you've gotten hundreds of dollars value. Why are you buying this skin that doesn't actually increase your abilities in the game? And you can imagine a user buying Valentine a new suit. Yes, why did you buy Valentine a new suit? She loves Valentine. Exactly. Or why did you gift these virtual roses? They don't mean anything. They're just ones and zeros. They're literally 100 % marginal costs.

1:40:51Tyler, what you got? I was just going to say it. I know a lot of kids who spent thousands of dollars on Fortnite. Yeah. But then there was that lawsuit, and then they got most of it back. They did? Yeah. Wait, they actually got the money back? Yeah. Because they were underage, or what? Yeah, it must have been. Okay. Yeah. Because it would have been a good school. It's probably a parent-led class action, basically, being like, I didn't actually, like, you were using my card. I didn't actually authorize this. Yeah, yeah. That makes sense. There's a post from Lucas Beyer who joined MSL earlier. He was with the Swiss division, remember?

1:41:26Oh, yeah, yeah, yeah. He posted about an hour ago, looks like there's an ongoing XAI exodus, wild. um so talent wars somebody from xai fired back and said not sure where this come from the number of people leaving on our site significantly lower than most of our competitors um lucas says not sure if you relate it to team size and although harder to do impact um so a lot of people going back and forth here but anyways in other news state of wyoming has launched launched its own stablecoin. People have been waiting for this. Why hasn't Wyoming launched a Visa-supported stablecoin called Front on seven blockchains?

1:42:16I don't know if this is a top signal. Oh, this is funny. The stablecoin bill banned central bank digital currencies for the federal government, but not for states. And so I guess the state can. Is this just like an olive branch sure like marketing for wyoming because i know why i mean was always they were big into that whole whole doubt the world of doubt but i wonder like like this is the whole idea of state currencies is like very antithetical to the like the declaration of independence and like the creation of the united states like that was the one of the main things that like unification of the states did was let's get rid of like georgia dollars and new york dollars because it's making cross-border commerce like impossible.

1:42:59But I guess if it's a, like, I just don't understand it. It's a stable coin. It probably runs on like standard crypto rails compatible with Visa. But also I imagine that it's not, it's backed one-to-one by USD. And so it has a mandated 102 % reserve requirement. Okay. Backed by short duration T-bills, US dollars. I don't know what their plan is this is fascinating we should we should talk to somebody who's uh who's going giga long so i think um yeah what you got i'm looking at the constitution article one section 10 please it says no state shall blah blah coin money yeah i think against constitution maybe well where we're going we don't need the constitution it's outdated yeah it's like that that near post uh like there's no laws anymore yeah something like that recently crime is legal i think that was the best yeah yeah they're saying i don't know i mean paying vendors in seconds to enabling tax refunds and social benefits on chain front brings state action into the programmable era okay i don't know why they wouldn't just be like we're adopting usdc or tether like it just seems like very odd to build your usdc is programmable yeah i would need to know more about like the differentiator because like again you've asked this question a bunch to the stablecoin folks.

1:44:23Are we going to see Kohl's cash? And this feels like Kohl's cash. But maybe it gets you some benefits. If it gets me to cut to the front of the line in the Wyoming DMV, maybe it's worth it. Anyway, the age of neutrality as a journalistic principle is over, says Julia Steinberg at ArenaMag. She says, I don't know anyone in their 20s who believes in it neutrality made sense when there were three cable tv channels and everyone in a small-sized city read the same paper when we have thousands of options to satisfy their media diet they're not going to choose neutral and so this is uh yeah lulu says it's refreshing to see a media outlet lean into their lefty politics i respect this what's grating is when media or companies claim to be neutral or non-partisan when it's untrue i hope the argument's fundraising deck was titled now you too can own the libs.

1:45:21Great line. Yeah, it's very interesting. Have you read the launch post? And the description here, in case anyone can't see, is join us. We're living out. They're having fun with it. They recently changed their colorway, their fun design. Interesting strategy. They launched with a bunch of sub-stackers as contributors. And so I don't fully understand. They raised a couple million bucks. Patrick Collison is in. Dustin Moskovitz is in. There's a couple other investors that we're familiar with. I wonder what this means to be the argumentmag.com. What exactly are they doing? Are they creating a new destination?

1:46:02Are they going to be selling ads or subscriptions? And then they're creating a roll-up. Because some of the people that they link to as contributors were prominent, formerly mainstream media folks. Matt Iglesias, Derek Thompson is in there. But all these folks have their own sub stacks now and have their own businesses. And so it's kind of unclear if you have your own business, what value are you getting by joining the argument as a contributor? But I mean, I'm sure we'll see. Maybe there's crossover events. Maybe there is some sort of value. I could imagine, I mean, the argument leads into like maybe there will be debates, maybe there will be live events.

1:46:40There's a whole bunch of infrastructure that could be kind of decoupled from the individual influencer level. But I would be surprised if they have the money or power to pull like Matt Iglesias out of Substack into a new bundle. We've been through the unbundling era. I'm bullish on bundles. I'm not bullish. I'm bearish on bundles. Okay. I'm bearish on bundles. I think there are institutions that will remain to be institutions like the Wall Street Journal. I don't read the Wall Street Journal for specific writers. I read it for the collection. But the only reason I still read the Wall Street Journal is because it's hundreds of years old at this point.

1:47:20I just mean if the argument over the long term wants to be able to charge$20 a month. Yes. And the value prop can be we're going to curate and select. We're going to have a bunch of different contributors. You're going to get all these articles that are normally paywalled. We're going to have some cross-posting. And as somebody who's an independent writer, maybe they have their own Substack and own independent media company, they can get exposure to net new audiences. And Argument is a really good name. It is a good name. The Argument. The Argument. But the independent writers can get exposure to a new audience.

1:48:00argument gets content. I think it's a decent trade. I don't know that it will be a decent trade. I think that you will have power law writers who are dominant on their own sub stacks. And when they go to the argument and they cross post and they make something free, they are taking capital out of their own pool and giving it to this platform. And then there will be other folks at the argument who aren't driving new subs, who are getting an unfair rev share, basically. And we've already decentralized and gone direct. And so I don't know. I think it'll be good. I mean, I think the value is if they can make an actual magazine and do all the collection and curation to put everything in because there is infrastructure that's required to pull an Ezra Klein and a Derek Thompson and a Matt Iglesias in together, all these disparate sub stacks into to one cohesive magazine, and then ship that to everyone.

1:48:57I think that that's interesting. I think hosting events, that's something that's interesting. I think if you're Derek Thompson, you have a thriving independent sub stack, and there's something like the argument that has a relevant user base to you, an argument goes and says, we would like your - They don't have a relevant, they're new. Yeah, but they will, they will. Maybe. They will eventually. And if Derek Thompson knows, okay, a lot of my audience is also subscribed to the argument, but there's 20 ,000 people, I will give them my article for free at some point because it's not going to material.

1:49:27If I gave every one of my articles, it would be material, right? Interesting. Yeah, I mean, it's certainly possible. If it works, we're going to see a lot more of these, like aggregation layers on top of substack, like communities, essentially. I wonder if there'd be one for tech. There's a lot of tech writers. I've always said we don't necessarily need 20 different independent tech journalists that all have a subscription for$20 a month that are scoops-driven businesses or like newcomer, for example. We don't need 10 more newcomers, but there is an incentive for writers to want to be independent.

1:50:08Yeah, I don't know. Well, if they want to get some subs, they should get on adquick.com. Out-of-home advertising made easy and measurable. Say goodbye to the headaches of out-of-home advertising. Only Adquick combines technology, out-of-home expertise, and data to enable efficient, seamless ad buying across the globe. Blanket Brooklyn with... Blanket Brooklyn. Blanket Brooklyn. Well, we have some news from NYSE, the New York Stock Exchange. They are launching in Texas. They say buckle up for NYSE, Texas, and they have a lot of rocket imagery in their launch video. I wonder what that means. What do you read into this, John?

1:50:46They're going to launch like a rocket. It's going to go to the moon. maybe they'll take a you don't think you don't think this could be signaling that spacex might go public on nicey texas this would be this would be the banger company to get it's uh probably the biggest private i mean it's not anymore yeah but it's the biggest private company in texas so it'd be good um i don't know i i need to know more about the benefits of like ipoing in texas versus like in new york i kind of like new york i'm i think new york's doing well but there must be some sort of regulatory value to being in Texas.

1:51:21Also, I mean, Nicey does amazing like media stuff there. And it's this like, it's a true temple of technology. They're going to aura farm. It's a, it's a fortress of finance when you walk in Nicey. And so just, I'm just pro building another building like that. And why not in Texas? I would absolutely go and visit it. Gabe says, what does Jordy mean when he says I have to get on with Taipei? I know it's not a movie reference because he doesn't watch me. It just means that every now and then we have to have phone calls and conversations that can't happen on the air. So we've got to hop off and hop on the phone.

1:51:55From 2 to 5. We've got to hop on with London. I've got to hop on with Milan. I've got to hop on with Sancho Pei. This is what people say when they're doing business. When you're an international businessman, you hop on the phone with the city that you are talking to. Yeah, so sorry to disappoint, but it means that I'm getting on a Zoom call. Usually with somebody in San Francisco or New York. Stuart Chaney says, if you're wondering if selling SaaS into e-com is tough right now, changing your legal name. Ridge Durant posted on LinkedIn, my name is Ridge Durant, and I will legally change my name to Ridge Wallet if Sean Frank goes live with Redo.

1:52:34Sean. This is the best. Sean. Sean, you have to do this. this is this is the most no-brainer ever if this guy what is redo what's so funny is like it would be so much better if his name was like steve wallet and he was like i'm just gonna change my first name but changing your last name is insane change my name uh i wonder in the age of ai does it become easier to change your name could you have an ai agent that goes around and updates all your credit cards and social security cards because like that's the hardest part of changing your name is like all the paperwork that happens down yeah agents for changing your killer marketing purposes killer app yeah change your name change your name or uh john ramp after you yeah if you go through a cancellation a hit piece comes out boom next day you're a different person honestly ramp haze would be a beautiful name for a baby boy ramp haze ramp haze i'd love it well redo offers personalization at every touch point so clearly the CEO wants to personalize his own name to sign his next customer.

1:53:41So love to see it. Maybe Wander Hayes. Maybe Jordy Wander. You can find your happy place. Find your happy place. Book a wander with inspiring views, hotel-grade amenities, dreamy beds, top-tier cleaning, 24-7 concierge service. It's a vacation home, but better. And the best thing about a wander, you won't need to ride the subway because wanders are not in the subway and you will be free of Skechers' latest Gooner AI slop ad. This is from Rohan. He says, I'm in disbelief. Is this the end? Skechers launches an ad campaign using AI slop. I don't know. This is an extremely strange ad to run for a brand that I primarily associate with children and retired people.

1:54:24Did they put a picture of the real shoe just in the corner there? It looks like it's kind of cropped out. I don't know. Hopefully, they bought the billboard on adquick.com. out-of-home advertising made easy and measurable. That's right. Ben Hilack says, this idea that the engineering job market is collapsing is hilarious. Oh, Trey P. Again in the chat. Julius Hayes is right there. Julius Hayes. I got to work on a deal with Rahul. That is something you would actually pick. That's an amazing name. Julius Hayes. That's a great name. That is a great name. I think you might be talking about something.

1:54:57Rahul. Let's do a deal. Well, we already have a deal. The deal is you put a couple hundred grand in a college fund for my son. Yeah, for Julius Hayes. And you have an ad for life. Yeah. Everywhere he goes, he'll be like, why did your parents name you Julius? You can think of a few reasons. Oh, I was named after a tool to better help you understand your business using AI. That's great. Ben Hilack says, this idea that the engineering job market is collapsing is hilarious to everyone trying to hire an engineer right now literally has never been more demand. this is a function of having capital wars in every single category.

1:55:38Yep. Absolutely brutal when an engineer can work at Meta or XAI or your YC company, prices go through the roof, especially for top talent. I mean, we have been hearing in the chat today that some folks have been having hard times getting engineering jobs as new grads. I think what we keep coming back to is that the shape of engineers and great engineers in particular is changing. It's no longer leak codes and being able to do fizz buzz on a whiteboard. It is thinking creatively about problem solving, understanding complex systems, high agency mindsets, unique skills that you can piece together.

1:56:20A lot of companies that are hiring engineers are doing specific things. So they might be working in the legal market or nuclear or hard tech or AI or something. And so being on the frontier of like a micro discipline, it's probably not enough to just be like one engineer, please. It's like, no, I want an engineer that can vibe code really quickly and create maintainable software. Or I want someone who can solve really hard problems that can't be one shot by Claude Code or any other, any other vibe coding platform. The demands are changing. And then also like the The job market is just extremely noisy, and so you have to break through in a really unique way.

1:56:59So if you're looking for a job, what you should do, pick the company that you want to work for, legally change your name to that company. And I mean, this is what Tyler TBPN over there did. Flip Flops. If you had changed your name to Tyler TBPN, I actually think we would not have hired you. I think we would have been creeped out. But we appreciate having you Flip Flops. It's great. Flops. Flops is good because it's AI based too. One more post from... Flops is like... I bought shoes. Oh, you did? I have a pair of shoes. How did someone with your meager financial status afford a pair of shoes? Months work.

1:57:39Did you have to liquidate all of your open router credits? To be clear. Did you have to resell credits on the black market? Scalping all my credits. Yeah. Growing Daniel says, like moths to a flame. Wait, wait, wait. We can't leave people hanging. What shoes did you buy? I don't. They're like some. I don't know. I don't remember. Bottega Veneta, potentially? Probably. Probably Bottega. Yeah. You deserve Bottega, Tyler. They have like a red. On the bottom, it's red or something. Yeah. Red bottoms. Raghav says Finn Hayes. Finn Hayes would be a beautiful name for a boy. Finn's a good name. Finn's a good name.

1:58:22Anyways, so the Titanic wreck is going to be visited by another billionaire two years after the Ocean Gate disaster. I love this. I'm extremely bullish on this. The Titanic. Don't be afraid. 12 ,500 feet below the surface of the North Atlantic remains a huge draw for thrill-seeking billionaires. In particular. History will remember your name. Take a risk. Yeah, I'm trying to figure out. Get down there. Go for it. I'm trying to figure out who this person is. Yep. also Paul James Cameron James Cameron knows what he's doing he's been down there a bunch of times make sure that he signs off I would happily go with James Cameron I would not go with some crazy guy with an Xbox controller but now apparently the Bean Air hasn't shared their identity yet they've just announced this they're saying he'll want to make an announcement that he is the first person to go to the Titanic.

1:59:21So there's some sources here claiming that it's happening. What do you mean first person to go to the Titanic? Like post-Ocean Gate? Since the tragedy. Oh, okay. Tragedy. Yeah. Anyway, let's move on to our first guest of the stream. We have Pete DeJoy from Astronomer. How you doing, Pete? Good to meet you. It's great to see you. Hi, guys. How you doing? Thanks for having me on. We're doing great. Welcome to the show. How are you? Give us an introduction. Give us a rundown on the business. Yeah, yeah, sure. I'm doing great. We're sitting here in our new office. We actually just moved into a new space yesterday here in New York City.

1:59:57That's amazing. It's been a heck of a month at Astronomer. The last four weeks have been pretty crazy. And look, guys, I've been working on this company for eight years as a co-founder. You always kind of have this implicit belief that your idea and your mission is going to actually kind of get worldwide reach and become a household name. This is not how we expected it to happen. yep but look you guys you guys clearly made the most of it you know yeah i think you yeah for sure i would kill to be in like a fly on the wall in a cocktail party two months ago where you once again have to explain what you do and people are like oh cool yeah is that going well like yeah cool and you're like actually yeah for sure amazing and we have raised like a series d and we're a huge company.

2:00:42We just thought it was amazing. I remember it was probably like a month ago at this point, but somebody was like, wow, Ramp's so good at marketing. They managed to like be, you know, right on the last post on LinkedIn was like talking about your guys' partnership with Ramp. It's hilarious. Yeah, we did some great work with them. I actually wound up learning a ton. Thanksgiving conversations with our moms have gotten much easier. Yeah, yeah, for sure. Yeah, I wound up learning a ton about Apache Airflow through this. I learned a ton about the company. But I'm interested in a little bit more of the history.

2:01:13How did you get into this? I want to talk about the comp to how Databricks is working. There's an open source story here. There's competition with hyperscalers. I actually do want to talk about the business. So maybe go into the prehistory. What led you to found the company? What was your background before starting Astronomer? Yeah, so we founded the company as a group of five of us in Cincinnati, Ohio. It was a great team. We were pretty scrappy. There's a long story there that I'm sure we don't have time to go into today. Sure. But the short of it is we were really inspired by the work that Databricks was doing with Spark at the time and the work that Confluent was doing with Kafka.

2:01:52Yep. And when Airflow came out of Airbnb, we worked with Maxime, the creator, to really form a commercial entity that could drive it forward. And that was really the focus of our business for the first several years of our existence, just establishing the community and trying to get to the next level. And we're fortunate enough to get that level of kind of viral adoption in the open source, specifically at the high end of the market, that formed a basis for us to build a commercial business around. And that's what we've been focused for the last five years or so. Yeah, what does that early work look like?

2:02:21Like there's this open source project. Are you just contributing to the open source project? Are you going to companies that are already running an open source implementation and just saying, let us act as almost like a consultant? Are you spinning up kind of a self-serve portal? Like what are the different steps that you're actually taking? Yeah, early on it was threefold. The first thing was we need to establish credibility and velocity in the open source because you can't build an open source business unless you have a viral and vibrant open source community. So that was focus number one. Then early on before you actually have a commercial product, because you have to devote all of your R &D resources towards building this open source community and project, You do do a bunch of services deals and kind of consulting around that open source to form the basis of a business.

2:03:07And in our case, we use that to finance the development of our cloud product and product. So, you know, we started there. We started with open source. We started with services and consulting. And then we're able to bring a cloud product to market several years thereafter. And that really has formed the basis for our growth and business. And what was the early adopter? Like, you know, graph databases are going to be used by social networks. Like what's kind of the most tangible business value example that you can give of like someone using Astronomer or Airflow broadly, like just this technology specifically, like what is it used for?

2:03:43Yeah, sure. Yeah, Airflow generally is like seen as mission control for your data. There are 80 ,000 companies in the world using it to schedule, monitor and manage these data workflows that are powering these data and AI applications. It just had its biggest update ever. We released Airflow 3 in March. We were very proud of it. It's very much built. Thank you. Thank you. I think Jordy's going to hit the golf. Airflow 3. Airflow 3. Okay. Fires me up. Super pumped about it. Well, so one immediate question I have is how did, during those early days, how did the venture community react to this kind of strategy?

2:04:19It feels like today every startup will talk about their open source strategy, but a lot of it ends up feeling just kind of like marketing because it's the it's the thing to do because there's been so many examples and winners like astronomer that that invested in open source early and really built a business on it but now it ends up just being like kind of a oh yeah we have an open source strategy but but but there's not a lot if you kind of dig in Yeah, no, no, definitely more obvious today than it was back then. Like keep in mind, when we started Astronomer, Redshift was like the cloud data warehouse of choice.

2:04:56We would like see Snowflake very sparingly in the wild, but it definitely was not like one of our common discussions in the field. And obviously that changed pretty dramatically through the years. And I think the industry started to wake up to the value of these businesses and how aggressively they can grow if you get them right. So to your question, you know, it definitely has changed. It's more obvious now and it's more of an execution challenge. He has brought up kind of the hypervisor competition. That's one of the kind of core riddles that you need to solve as an open source business. But it's been great to see the trajectory of the company and the support from the venture community.

2:05:30We were fortunate enough to raise a$93 million Series D round in March from the folks at Game Capital. Very, very exciting. Very, very great time for the company. And we're just looking forward on driving it forward, guys. We've got a great list of customers. Ramp is one of them. They're an amazing business and also a great podcast sponsor, from what I understand. The best. They're using AI everywhere, naturally. And a lot of these AI workflows that are doing everything from fraud detection, when you swipe a ramp card, to kind of automation of internal outbound sales processes are actually powered by Astronomer under the hood.

2:06:03And we built a great partnership with them getting those to life. Talk about how you guys turn this kind of incident into a win. because I think it was many people in our neck of the woods, specifically on X, were saying that this was like the PR master stroke. It was one of the best executed kind of campaigns. It was genuinely funny and perfectly timed. But I'm curious what kind of that process looked like. And even if it was a risk, right? It wasn't like everyone, I believe, wasn't Ryan Reynolds involved? That's not somebody who's hanging out with every single Silicon Valley company. It's not like, oh yeah, everyone that does a Series D with Bain Capital is teamed up with Ryan Reynolds now.

2:06:55That's not the narrative. That was fascinating. Very interested to see just the thought process. Yeah, look, candidly, my team deserves so much of the credit for that. We were dealing with so much in that first week after everything went down and definitely did not have a celebrity cameo talking about Apache Airflow on my bingo card, the grand scheme of building this company. But look, the reason we felt like it was important to do something was that we found ourselves in this position effectively overnight where more people in the world knew about our business than we ever anticipated would, even in kind of the most extreme circumstances.

2:07:37And we also have an incredible team of 300 plus people that have spent almost a decade pouring blood sweat and tears into building building what we believe to be a great company so we just wanted to be sure that when the public thought about astronomer now that everyone knew who we were they knew a little bit about what we did and that was really the incentive to actually go do something and we were fortunate to have our 15 minutes of fame it worked out and in such an asymmetric way for us as far as like kind of the pr and media angle is concerned um but and the challenge the challenge to go viral again with an ad that actually explained the business was the magic of it, which was so great.

2:08:17It was like, hey, we're going to get people's attention again, but this time let's make sure that we talk about what we actually do. Yep, yep. That's right. What was the, I'm curious to get a sense of demand over the last, you know, have you guys benefited, you know, have you tried to capture, you know, I'm sure you've captured value from, from all this extra attention. You know, maybe CTOs waking up and saying, I hadn't, hadn't heard about astronomer and suddenly I want to sign up for a demo and, and seriously explore this. But what's that been like? Yeah, look, our employees and customers have maintained an incredible amount of stability through this.

2:08:56It's been really actually incredible to see. And, but we're an enterprise software business. So it's not like consumer eyeballs deterministically translate to dollars. Yeah. Like there's not some calculus that says because we have this many page views, like I think our site got more views than the New York Times for a day, that actually like outputs a dollar amount in terms of like these big enterprise deals that we're doing. But I will say we have seen a lot of early signs just in kind of entering our third fiscal quarter that many of our potential buyers, CTOs, CIOs, CDOs know about us now that didn't prior.

2:09:29we've gotten a lot of inbound calls our sales team is fielding them like crazy and that gets us very excited about the path forward so we just have a lot of execution to do ahead of us and you know candidly we're just happy and excited to be back to focusing on the business it's such an interesting like like situation with this company because it's not like a Netflix subscription we're like if the Netflix CEO does something that everyone can be I'm canceling Netflix today and then they're you know you got to win them back but like ripping out a core piece of your data infrastructure even if you were in that camp of like i want to churn i don't like what's going on with this company like that's probably a couple weeks or a month decision and fortunately the astronomer team like headed all of that off and and gave everyone a very clear way forward in just a few weeks within the within the news cycle And so, yeah.

2:10:23How's the CEO search going? You're a CEO now, but you're eight years in. And I understand you guys are looking for somebody to carry the torch for the next chapter. Yeah, absolutely. So look, guys, as a co-founder, this company really means everything to me. I know that sounds a little self-righteous, but it's honest. We're working on this for most of my adult life. And I'm very proud of what we've built. And being asked to serve in this job is actually a real privilege. I care so much about the business and the people that we have here. But also, as a co-founder and entrepreneur, your job at every step of the journey is to re-underwrite your worldview and do whatever it takes to get a win for the company.

2:11:01So right off the bat, our board put out a search. I'm working with them very closely on that and very supportive of that process. And what I'm focused on in the interim is just running the company and doing what I can for our employees, and we'll see where it goes from there. Awesome. Well, congratulations on how you and the team have handled it. It's fantastic. And I wish we could use a product. Maybe we can. We'll find a way. We'll put the team to just find a use case. We got some great AI native kind of consumer-oriented products coming out in the next 30 days that you can get your hands on.

2:11:31Amazing. Awesome. All right. Give us the final thing. Give us the call to action for the event. Has that passed? Has that happened? Remember that from the video. Yeah, we have a virtual conference coming up in September. Fantastic. You'll actually hear more about these AI native kind of consumer-oriented data engineering products that we're releasing then. and we're really excited to bring it to the world. There will be a really, really great set of releases in there. So it's happening in mid-September. Amazing. Looking forward to it. Well, thanks so much for joining me. We're rooting for you and the team.

2:12:00What a fantastic piece of just Silicon Valley lore. It's lore. It's lore. Incredibly well played. That's right. And thanks for joining. It's great to meet you. We'll talk to you soon. Thanks, fellas. Yeah, see you soon. Cheers, Pete. In other news, the Palantir CTO, Shamsankar, who's been on the show, confirms that he's working on a new pro-American film production company. He says, the goal is to make content that makes you proud to be an American. I love this. Sham Sankar is raising money for a new production company called Founders Films with Palantirian Ryan Podolsky and Christian Garrett's involved.

2:12:33Didn't Sham tease this when he came on? Yeah, he's talked about it. I think he wrote a piece in Pirate Wires about it, which you should go read and subscribe to. And Sham Sankar is on an absolute tear, and I'm very excited for this. And it'll be very interesting to see specifically the level of investment. We talked about this a little bit, but Hollywood is in this era of we need to even have a chance at an ROI. We need to invest$400 million. And so you see all these major blockbusters. But by virtue of when you invest$400 million in order to make a billion dollars, it has to sell everywhere. and that means it has to sell internationally and that means that the planes have to be from like you know some anonymous country because we don't want to make anyone upset so like yeah it's an american movie but it's like we're fighting the the the like the and then they make up some name right yeah um as opposed to a story like you just can't really make a story about like the cold war because who knows that might upset a certain person in a different country and so um it'll interesting to see what level of investment, like what type of swings are they taking?

2:13:51Are they swinging for Grand Slams and putting about$200 million into a single film? Is it a bunch of$10 million projects? Is it a bunch of$1 million projects? The Blair Witch Project, which I know you haven't seen, was made for under a million dollars. I think it grossed over$100 million. It was one of the greatest ROIs. That's the... Have you seen, Tyler, have you seen the Blair Witch Project? Of course. It was filmed like right near my house. You've seen the Blair Witch Project? Wow. It was filmed in Blair, Maryland. No way. Yeah, right near my house. That's amazing. So the bull case for AI in film production is that even if budgets stay the same, we could get five times as many films, 10 times as many films, if the models can get significantly better.

2:14:32Yeah. I think that just the cost to produce what we think of today as like a blockbuster or a Marvel-level film could fall dramatically. Yeah. Yeah, like the production side is interesting. There's certainly a way to drop costs and produce something that feels like a$100 million film for$10 million. That feels like it's coming. But I think that honestly, it might be more of a distribution problem in question. Like if you have a banger video, a banger movie, and you know that people are excited about it, but it's a little bit more of a niche audience, there's really no way to get people to pay for it other than the traditional playbook of like a wide theatrical release all over the place and so you wind up in the UFC camp where you have a movie and your only option is to go to the streamers or and hope that you're converting subs or a really wide release I wonder if there's some middle ground that will pop up I wonder if we'll learn something from UFC, but it's very hard to take a movie that doesn't merit a huge release and actually make an ROI in the theaters today.

2:15:47Anyway, what do you want to talk about next? Rune said yesterday, Elon falling so in love with the Grok image model. Imagine model is a piece of performance art that absolves all the slop. And Elon says, thanks, Rune, the heart. there is an article from eric berger saying after recent tests china appears likely to beat the united states back to the moon this is the case of him i went there because it's because at this point it's difficult to come to any other conclusion um and he is writing this article to explain where is enormously bad for the united states reporter so in recent weeks the secret of Chinese space program has reported some significant milestones in developing its program to land astronauts on the lunar surface by the year 2030.

2:16:35On August 6th, the China Man Space Agency successfully tested a high-fidelity mock-up of its 26-ton Lan Yu lunar lander. The test conducted outside of Beijing used giant tethers to simulate lunar gravity as the vehicle fired main engines and fine control thrusters to land on a cratered surface and take off from there. The test, said the agency in an official statement, represents a key step in the development of China's manned lunar exploration program and also marks the first time that China has carried out a test of extraterrestrial landing and takeoff capabilities of a manned spacecraft. As part of the statement, the space agency reconfirmed that it plans to land its astronauts on the moon before 2030.

2:17:20Then last Friday, the space agency and its state-operated rocket developer, the China Academy of Launch Vehicle Technology, pretty hard name, successfully conducted a 30-second test firing of the Long March 10 rocket's center core with its seven YF-100K engines that burn kerosene and liquid oxygen. The primary variant of the rocket will combine three of these cores to lift about 70 metric tons to low Earth orbit. These successful efforts followed a launch escape system test of the new Mengzhu spacecraft in June. A version of the spacecraft is planned for lunar missions. Thus, China's space program is making a demonstrable progress in all three of the major elements of its space program.

2:18:05The large rocket to launch a crew spacecraft, which will carry humans to lunar orbit, plus the lander that will take astronauts down to the surface and back. This work suggests that China is on course to land on the moon before the end of this decade. For the United States and its allies in space, there are reasons to be dismissive of this. For one, NASA landed humans on the moon nearly six decades ago with the Apollo program. Been there, done that. Moreover, the initial phases of the Chinese program look derivative of Apollo, particularly a lander that is strikingly resembles the lunar module.

2:18:39NASA can justifiably point to its Artemis program and say it's attempting to learn the lessons from Apollo, Apollo, that the program was canceled because it was not sustainable. With its lunar landers, NASA seeks to develop in-space propellant storage and refueling technology allowing for lower-cost, reusable lunar missions with the capability to bring much more mass to the moon and back that should eventually allow for the development of a lunar economy and enable robust government commercial enterprise. But recent setbacks with SpaceX Starship vehicle, one of two lunar landers under contract with NASA, alongside Blue Origin's Mark II lander, indicate that it will still be several years until these newer technologies are ready to go.

2:19:17So it's now probable that China will, quote-unquote, beat NASA back to the moon this decade and win at least the initial heat of this new space race. To put this in perspective, ours connected with Dean Chung, one of the most respected analysts on China's space policy and the geopolitical implications of the new space competition. and silicon valley cannot save us at this point only wall street can save us the solution is make the moon a state as mike solana suggested create mortgages on the moon and then securitize them and once wall allow retail to go once blackrock owns 75 of the moon then we're in business yeah yeah they've been getting uh wall street's been getting pushback on owning single family home they own 99 % of that capital to blackstone owns 99 % of all the houses in America now they own all of them they bought my house from me I didn't want them I didn't want them to they just bought it for me I didn't have an option yeah they were like sale and lease back you rent now you will own nothing you'll be happy yes that's exactly what happened no they actually own like 1 % of that but Dean says the land you land are significant because it's part of the usual China Chinese crawl walk run approach to a major space project the people's republic of china can benefit from other people's experiences much of nasa's information is open but they still have to build and operate the spacecraft themselves so the test of the lanu lander successful or not is an important part of that process so anyways important to flag we've had a lot of wins in the space industry recently and it's tough it's tough to bootstrap an economy like it is an incredible milestone and yet america's done it so there's not as much of an incentive just to be the first.

2:21:04It's not superlative. And then the question for SpaceX is, are you better off landing on the moon or just putting up more Starlinks and just making more subscriptions? Like, it is a little bit harder to underwrite in the early stage. It takes a long time. Maybe you need a new government program, some sort of incentive. I mean, fortunately, like, Firefly's been to the moon, Blue Origin's planning, SpaceX's planning. Like, there are a lot of competitive American companies that are working towards it, but the underlying incentive, there's just other ways to make money in space right now that might be higher value.

2:21:42Although it will be a very weird moment if China's just up and back on the moon constantly, running around, bouncing around, live streaming from the moon. We got to get back. We got to get back. I agree. We have to go back. Would you like to read through the reverse deep seek moment, the bull case for GPT-5? This is a posting post. Why don't you kick it off? I'll be right back. Sure. So Zivi Mushowitz says, everyone agrees that the release of GPT-5 was botched. Not everyone agrees. I mean, Alice likes it. I think Ben Thompson likes it for the most part. I have a good bull case here too. So this is a lot of stuff I agree with.

2:22:20But there was a lot of pushback. And so he says, everyone can also agree that the direct jump from GPT-40 and O3 to GPT-5 was not of a similar size to the jump from GPT-3 to GPT-4. That was not the direct quantum leap we were hoping for and that the release was overhyped quite a bit. And I still don't know how much OpenAI is responsible for the overhyping or just the fact that they used a number. Maybe they should have just stuck around and made it GPT-4.7, you know, and then the hype would have been a lot lower. But nevertheless, the launch went out. I think the naming's better. GPT-5 still represented the release of at least three distinct models, GPT-5 Fast, GPT-5 Thinking, GPT-5 Pro.

2:23:06The naming is much better. The naming conventions are much better. At least two, and likely all of three, are state-of-the-art within their class, along with GPT-5 Auto, which is the router. The problem is that the release was so botched that OpenAI is now experiencing a reverse DeepSeek moment. All the forces that caused us to overreact to DeepSeek R1 are now working against open AI in reverse. This threatens to give Washington, D.C., and I think this is why this is important and why we should be focused on this, and it's key decision makers, a very false impression of a lack of progress in AI, especially progress towards AGI that could lead to some very poor decisions, and it could do the same for corporations and individuals.

2:23:45I spent last week, he says, covering the release of GPT-5. This puts GPT-5 in perspective. So the reverse DeepSeek moment. In January, DeepSeek released R1. We have this DeepSeek moment. Everyone panicked about how China had caught up. It's a good model, sir, but only an ordinary good model, substantially behind the frontier. Most people hadn't used reasoning models yet. And so R1 was the first one that they had used. And R1, from a UX perspective, revealed the reasoning chain. And that was a revelatory experience for DeepSeek users. And so we had the DeepSeek moment because of a confluence of factors.

2:24:22One,$6 million model, negative narrative, gave a false impression on cost. Two, they offered a good clean app with visible chain of thought. It went viral. Three, the new style caused an overestimate of model quality. Three, with the fourth timing. Fakes the app store, downloads. The timing was impeccable, both in order of model releases and within the tech tree. Five, safety testing and other steps were skipped, leaving various flaws. This was pure fast follow, but in our haste, no one took any of that into account. Six, a false impression of momentum and stories about Chinese momentum. Seven, the always-insist open models will win crowd amplified the vibes because they love open source.

2:25:02And eight, the stock market was highly lacking in situational awareness, suddenly realizing various known facts and also misunderstanding many important factors. GPT-5 is now having a reverse deep-seek moment, including many direct parallels. GPT-1, GPT-5 is evaluated as if it was scaling up compute in a way that it doesn't. In various ways people are assuming it cost far more than it did. If you think about the progression of GPT-3 to 4 to 4.5 to 5, like if you just did an order of magnitude more cost, we would be, this would be the hundred billion dollar pre-training run, and yet obviously that's not what this is, and so you shouldn't price it as though they spent a hundred billion dollars on training cost.

2:25:46Two, they offered poor initial experience with rate caps and lost models and missing features, a broken router and complaints about losing 4.0 sycophancy went viral. Three, the new style and people evaluating GPT-5 when they should have been evaluating GPT-5 thinking caused an underestimate of model quality. So if you're an O3 user and you accidentally downgraded to GPT-5 fast, you were like, oh, this is backwards. But if you went forward and stuck to pro are thinking you probably had a decent experience do you think people are more or less confused now with with all the model naming i think they might be more confused because it's just like they they had the the folks who had remembered for one what four or five is what oh three is what oh what four oh is those folks even oh three and oh three pro i never really learned the difference i just always hammered oh three pro um but uh because you're you kind of identify with being a true professional.

2:26:42Yes, I'm a professional chatter. Also, four, the timing was directly after Anthropic and previous releases had already eaten the most impressive recent parts of the tech tree, so gains incorrectly look small. In particular, gains from reasoning models and from the original GPT-4 to 4.0 are being ignored when considering GPT-4 to GPT-5. And so if you're thinking about just comping GPT-4 to GPT-5, you have to include 4.0's progress and O3's progress and all of that. GPT-5 is a refinement of previous models optimized for efficiency and is breaking new territory, and that's not being taken into account.

2:27:22A false impression of hype and a story is about loss of momentum. And the open AI is flailing crowd, and the open model crowd amplified the vibes. The stock market actually was smart this time and shrugged it off. That's a hint. And of course, the big one, which is the GPT-5's name fed into expectation. Unlike R1 at the time of its release, GPT-5 Thinking and GPT-5 Pro are clearly the current state-of-the-art models in their classes. And GPT-5 Auto is probably state-of-the-art at its level of compute usage, modulo complaints about personality that OpenAI will doubtless fix soon. OpenAI's model usage was way up after GPT-5's release, not down.

2:28:00The release was botched, but this is obviously a good set of models, sir. Washington, D.C., however, is somehow rapidly deciding that GPT-5 is a failure and that AI capabilities won't improve much and AGI is no longer a worry. This is presumably in large part to the race to market share faction pushing this narrative rather than hardcore and having this be super convenient for that. And so he talks a little bit about how DC is maybe thinking that AI progress is stalling out. So he says, to be clear, American AI is making rapid progress, including it open AI. How do you know this is happening? Well, he has some quotes.

2:28:43Dean Ball says, the jump in performance and utility of Frontier models between April 2024, GPT-4 Turbo, and April 2025, O3, is bigger than the jump between GPT-3 and GPT-4. People alleging a slowdown in progress due to GPT-5 are fooling themselves. This is odd because GPT-3 was released. That gap between GPT-3 and GPT-4 is not one year. GPT-3 was released in 2020, I believe. And so that was more like a four-year run. So we might still be slowing down, but it's still impressive progress. Most people, when they think of the GPT-3 to GPT-4 bump, they're really talking about GPT-3.5-002 DaVinci. which was the model that shipped in ChatGPT free tier and then got the four upgrade just a few months later.

2:29:33Anyway, people are still debating this. I think I'm coming away pretty happy and I think they made some good decisions. I think Ben Thompson has a pretty good take that the model switcher was the right move, but OpenAI should have been even more aggressive about saying, you know what? 4.0 is not coming back. Like, we're a consumer company now. We're going to do what's best. Like, you are not going to be able to browbeat us into changing our decisions. If we think that what's best for the consumer is this and the revealed preferences will align with that over time, we are going to, you know, take the pain in the short term for the long-term gain of a better, more usable product.

2:30:16And that manifests in the sense of, like, a simplified model switcher. Because as they add the models back, the model switcher gets more complicated i don't know where do you sit on it i don't know i using chat gpt to me is not uh in the way that i use it primarily research and understanding the world is not materially different than than it was prior to gpt5 yep and so i don't have that strong of opinions on it yeah i think there's a little bit of like model picker foo that needs to happen. But I'm getting better every day at triggering the right keyword. Hey, think hard about this. Overall, it's what I asked for.

2:30:59And they delivered what I asked for. And so I've been pretty satisfied. I don't feel like my experience has been degraded. It's a good model, sir. It's a good model, sir. Well, on that note, we have our next guest. We have our next guest, Marty from Pylon. Welcome to the stream. Coming in. How you doing? Doing well. Jordan, John. Nice to meet you guys. Good to meet you, too. Ooh. Oh, do we have to get the ballot ready? I think we do. What are we discussing today? Give us an introduction. Explain what the company does. Give us the news. So first off, you're watching TVPN. There we go. Temple of Technology.

2:31:30You're watching Pylon TV, baby. Let's go. There we go. Yeah. No, super excited to be on the show. I've been cycling you guys in in the morning. So yeah, I'm excited to be here and chat with you both. And you're all over my LinkedIn feed. So it's - There we go. Credit to Michael. Michael is dominating LinkedIn. We're bringing Teapot to LinkedIn. They are not ready. But it's great to meet you. It's great to have you on the show. Great to be here. Yeah, no, I'm here to announce our Series B fundraise for Pylon. So we just raised$31 million, co-led by Andreessen and BCV. Congratulations. Sorry to interrupt.

2:32:09Nice. Nice. Thank you. Always the jump scare. Crisp gong hit. Always. Good to see Bain Capital getting a W here. You know, we just interviewed Pete DeJoy from Astronomer, another BCV company. So we're happy to have you on the show. We're really excited to work with them. So on the Bain side, you know, they're entering into the company. We're working with Merit Hummer and Abby Myers from their team. Super impressed by them. Continuing to work now with Jennifer Lee from Andreessen. They're already on the board. And yeah, no, this is an awesome fundraiser. is actually every time we've raised money, we haven't needed to.

2:32:45People kind of came to us and then, you know, pitched us on our vision, did the reverse pitch, and then we, yeah, we made it happen. So, yeah, really excited to be here and continue building the product. Talk about slicing the market. Obviously, just AI for support is almost like a mega trend. There are market maps out there. You've kind of sliced it to be more focused on B2B. take me through some of the slices of what support looks like in various B2B contexts, because the customer support tickets that I maybe send into my law firm are different than I send into my 3PL. So where have you been seeing green shoots, beach heads, opportunities?

2:33:34What does kind of the early adopter customer for Pylon look like? Yeah, absolutely. So first off, part of the market, if we want to zoom out for a bit, Zendesk, Salesforce Service Cloud make around 35 % of their revenue off of B2B. And so that's actually the slice of revenue that we're going after. On the Salesforce side, total revenue from Service Cloud is$9 billion. So we're going after quite a big chunk there. So they can share. You're saying they have enough to share. Exactly. Yeah. And the difference between B2B and B2C, in consumer, you have very simple transactional support questions and only one customer facing team customer support in b2b you don't just have customer support you also have customer success solutions professional services account management and all the conversations that are coming in are related to not just like a transactional support question but could be routed to any one of those teams and you need account context that's being pulled in from not just that conversation but many other sources so for example call recordings or notes that were left on customers implementation plans and so we kind of fit to all those different use cases, all those different personas and sync all that data in.

2:34:39One way to think about us strategically is we're building something similar to Rippling, but for this post-sales use case, where kind of all in one product, consolidating a lot of tools that people would otherwise buy and making it really applied to this B2B customer persona. So there's a debate on whether AI native tools like Pylon are going after these end software markets or potentially winning uh labor spend where do you kind of sit in that debate and and how do you yeah so there's a tam here there's a difference we have like the native ai agent companies your fins your decagons etc um they're going after enterprise immediately plugging into the system of record we're taking a different approach we're actually building the system of record from scratch so replacing zendesk and then layering in ai agents as well um and so we think basically whoever controls um the data the workflows actually has a lot of power and is way more defensible than the AI agent companies.

2:35:38And then also for the B2B use case, because it's way more complex and you need way more context across these many different channels and sources, we're plugging into those in a way that no one else is. So really, I think we're going to win the B2B market completely. And then there's going to be a knife fight for a consumer. Last question for me. It's already happening, really. Yeah, yeah, yeah. Are you seeing any... Somebody posted today that stood out as basically being like, everyone thinks that every software engineer is going to be replaced. Meanwhile, Amazon still employs 100 ,000 customer support reps.

2:36:13Wow. Really? Probably not 100 ,000. There's no way. It can't be that much. But a lot, a lot. Yeah, yeah, yeah. Yeah, I mean, maybe through all the different consulting arms and third-party groups, if you really trace the word. Yeah, employs is. Yeah, yeah, yeah. Contracts with, potentially. And there's a difference as well. Like when you look at B2B, it's actually usually much higher skilled labor. And it's not just, hey, we're outsourcing it to a BPO in another country. It's actually, hey, these are like higher tier, more complex support engineers who also work with solutions teams, account managers.

2:36:46And it's just way more complicated of a deal. So you're not handling transactional questions. They're much more complicated. And frankly, actually, AI mostly isn't good enough to answer a lot of these B2B questions. So instead of trying to solve them and do a bad job of it right now and tarnish customer relationships, trying to essentially plug into all the workflows and own all those workflows so that we do have a chance to eventually become fully agentic once the AI gets good enough and we have the data consolidated into Pylon. Going fully agentic. I like that. Well, congrats on the race. Thank you so much for hopping on the show.

2:37:19Love the clarity of vision, and it's great to have you on. We'll see you on LinkedIn. Have a good one. Cheers. We've got to hop on with Mateo from 8sleep. Get that gong ready, Jordy Hayes. There's some big news. Oh, I would love to take a crack. Thank you. Let's bring in Mateo from 8sleep. What you got for us, Mateo? What's the news of the day? We just raised$100 million. Oh, there we go. Congratulations. Yeah. It's great to see you. Yeah, great to see you again, guys. Yeah. I'm excited to be here again. Hasn't been very long. It's great to be back. Give us the update on everything. How did this round come together?

2:38:05Talk about the progress. We want the full pitch. Yeah, of course. I mean, this is a massive milestone for us. You know, we just announced$100 million in new funding. Total funding to date is a quarter of a billion. There we go. And so not bad. But I think the best news, honestly, is we have been free cash flow positive for the whole year. Wow. We have been free cash flow positive for most of the past 10 quarters. So the good news is we didn't need the money. The company is in a really strong financial position. But we have a very ambitious plan and we found a couple of great investors, including our existing investors that they wanted to come in.

2:38:54And so obviously existing investors, Valor with Antonio Gracia, first investor in Tesla, Funders Fund, YC and a few others. And then also HSG that will help us to really scale China and Asia. So right now we are in 30 countries, but the next big piece is Asia, particularly in China. And I think that alone could double our business. Yeah, yeah. Talk more about China. Obviously, you've been advertising with F1. It's an incredibly global sport. Everyone sleeps, regardless of what country they're in. What's the shape of the opportunity there? Is there an opportunity on the supply chain side, the distribution side, the customer side?

2:39:40have you not been cloned already? I feel like there would be an eight sleep knockoff over there. How are you going to beat them if they already exist? Yeah, yeah. I think my immediate reaction is this is a very it's a simple product to understand, simple product to use as a consumer, but incredibly difficult to clone more than like a toaster like you have this hardware software dynamic you're dealing with with uh the elements right uh it's not as uh uh it's closer it's closer in some ways to um you need to use it intensely every single day obviously but anyways yeah but let me step back for a second and then i'll answer your question right so we will use the money in in in a couple of different ways so the first one is we will double down on AI.

2:40:35And I would like to share a bit more with you because we're working on some crazy things. Second will be medical. So we are going to file for FDA in a couple of different dimensions, including sleep apnea. I think it will be really big. And so you will see it becoming more and more medical, but still consumer medical. And then international expansion with China. Does the, sorry, really quickly on the FDA side, does that unlock paying with insurance? Is there a change to the financial business model or is that really just about coming to the consumer with stronger claims and evidence about the efficacy of the product?

2:41:12Both, but the real goal is the first one you mentioned, right? Imagine if the pod of our technology could be reimbursable or at least partially covered by your insurance. And that is where we have identified a couple of really good opportunities. Again, sleep apnea is one, but menopause is another one where women have hot flashes so we just released what is called hot flash mode and so if you have a hot flash in the middle of the night you just tap on the side of the bed and immediately it will start immediately cooling and there are like thousands of women that now are finally able to sleep again because of our product and i bet with ai you could detect those hot flashes eventually and just have it exactly right yeah exactly we'll do another one then on the ai side we are doing two really, really cool things, at least for me.

2:42:01So on one side, we are working on what we call a sleep agent, but it's really a PhD level sleep coach that is able to simulate your night before you even go to bed. And so it uses the history of your data. It uses the data from the day through Apple Health or any other data source. And we'll start simulating the night in a couple of different ways and then will adjust accordingly as you fall asleep. And then there is this other agent that we really call the longevity twins. And so for each user, there will be thousands of longevity twins that simulate your health for the next 10 and 20 years and they will come back with the likelihood that you develop a certain outcome.

2:42:44So we are really transitioning from sleep that is the foundation of health into longevity and will help people expand and extend more than expand extend their lifespan and help them yeah it's really interesting to think about if you can understand somebody's sleep habits today you can predict what their health will look like in a decade two decades etc um uh feature request a bit of bit of a wild one uh but if uh in the future if you could get a Tesla Optimist robot that would physically come and pull me away from my desk late at night and shut your computer off, unplug the Wi-Fi and drag me, you know, force me to sleep.

2:43:26I think there could be something there. So maybe a partnership with Elon at some point. Yeah. Well, amazing. Anything else top of mind or you got to get back to business? Big Things is also a lot of products coming. So you will see new hardware coming in next year. A lot of different products. And we'll keep doubling down on making the pod better and better. Fantastic. Well, if you're listening and you want a pod five, go to 8sleep.com. And thank you for joining us, Matteo. Always a pleasure. You're the man. We'll talk to you soon. My pleasure. Bye. Thank you, guys. Cheers. And we'll bring in our next guest.

2:44:05I love hitting the air horn for free cash flow. Hit that air horn. Hit that. Hit something. Give me something. Give me the news. Give me the Ashton Hall. Welcome to the stream. Abdul, how are you doing? Welcome. I'm great. Thanks so much for joining. Thank you for taking your time. Thanks for having me. Of course. Sorry for the technical difficulties earlier. Friend of the show said, Abdul is going to be your best guest ever. Let's go. He's crazy in the best world. Wow. Okay. Let's jump straight to the crazy. No pressure. Introduce the company really quick and give us the news. We'll hit the gong and then we'll go into some crazy stuff.

2:44:42Awesome. Awesome. Abdul Al-Assad, co-founder and CEO of Basic Capital. Basic Capital is the first and only company that enables you to finance financial assets and to use credit to invest instead of using credit to consume. We've been in business for three, four years now, operating mostly under stealth, backed by legends from Lux Capital, Henry Kravis, SV Angel, and others. And recently, we just closed our Series A,$25 million from Kirsten at PodRunner. Boom. Fantastic. You know, it's just one step on a long journey, but we are really, really excited to see that our message is resonating with both our customers and our investors.

2:45:30and now we're fired up to go. We're fired up to get after it. There was a viral video about you guys like a month ago or something and people were controversial. It was your video. But people were going wild. It was put the timeline in turmoil. Walk me through what happened. What is the pushback against this again? Like what do the haters say? And then I don't know if you can steal man. Yeah, the haters would say this is all well and good when numbers are going up. what happens when numbers go down. But the timeline is not very good at nuance and detail. So yeah, good opportunity to kind of explain the different dynamics.

2:46:12So we've been operating under the radar. Our thesis here is when you're working in financial services, you can't ship a half-baked product. You just can't. You can't build fast and break things. There's a lot of hurdles. You're dealing with people's money. this is a massive responsibility and we see ourselves as stewards of people's capital so we've been operating under the radar for quite some time and in may our investors were like abdul you have to launch go out of the closet and tell the entire world about what you're building yeah and my thesis is if you're gonna do something you have to do it extremely well so we made an incredible launch video we worked with a bunch of journalists and we went out to tell the world our story and our goal was to get 50 000 impressions on that five million basically within 24 hours we got 5.5 million five million a hundred more a hundred acts that you wanted a hundred acts that's great absolutely blew up every vc and their mom is in my dms every hit The mom is also in my DMs.

2:47:25Like high school friends. Wait, the haters' moms were DMing you too? Wow. By the way, I replied to those. And, you know, we were basically engaging with the public. And their concern was credit could be a dangerous tool if available to everyday people. And our thesis is, when you want to make money as an individual, when you want to build wealth, when you want to achieve autonomy, you have two sources to rely on. You can either rely on your labor, your work, or you can rely on capital, assets you own, a stock, a bond, and a property. Now, here is the interesting thing. Your labor will for sure depreciate over time.

2:48:16it's a fact because over time you get old and you could be the best lawyer at Paul Weiss and at some point you're not as sharp on your feet, you're not as quick, and there's a young stud ready to take your spot. And when your labor depreciates like this, when you can no longer work, what you have to fall back on is the assets you bought in your youth. and the problem that most americans have today is that they have no assets and our thesis was we'll let them use credit to buy assets just like they use credit to buy coachella tickets yeah and people were like but capital goes up and it could go down yeah and i said i agree there is an uncertainty to capital but there is a certainty that you will at some point age right like that is an inevitable fact i don't know brian brian johnson is going to push back yeah he'd like to have a word he might be listening he might send us an angry email i'm not gonna age yeah well well but well let's take a step back you might be perfectly fit like brian johnson but does brian johnson has the same ability to compute as ai does brian johnson has the same ability to execute on tasks as computers And the question becomes is over time, as capital starts to automate more and more jobs, what is our plan as a society and as individuals for that day?

2:49:45Yeah, I think we're generally pretty bullish on getting on the ladder of having positive capital. That's very important. Love the Gerstner accounts. But how is this different than contributing to an ETF every month or even contributing to a levered ETF every month? You could go 2x long, the S &P. You go 2x long, the MAG7. You could have a mix of bonds and stocks. What is the benefit of using credit to buy a larger stake up front? Of course, of course. So credit is a tool. It's like a shovel. It can be used in a good way or in a bad way. It can be used to plant a tree. It can be used to knock somebody on their head and knock them dead, right?

2:50:29Just like nuclear. Nuclear can be used as a bomb. It can be used as nuclear energy. The same thing with leverage and credit. It can be used in a good way or in a bad way. The devil is in the details. The devil is in the structure. A very good way to finance assets is actually the mortgage. Why? Because it's 30 years, it's amortizing, and there is no mark to market. There's no rebalancing. You're not rebalancing your mortgage every other day. That's not how a margin loan works, and that's not how a levered ETF works. Levered ETFs and margin loans are inherently short-term vehicles. We can get into the technicality of that, but they are inherently made for people who want to make a short bet on assets.

2:51:10So do you offer like mandatory lockout periods? Like, can I get a 30-year loan against the S &P and be paying that down forever, basically, for 30 years? That seems incredibly powerful. That's the product. That's in effect. So you come to basic, you put 20 % down payment. Let's say you have$20 ,000. You come, you put$20 ,000 down. We open an LLC. So the debt is non-recourse to John. For$20 ,000, we give you$80 ,000. You pay off the$80 ,000 over 30 years. But what you have in the LLC is$100 ,000 of assets. And these are not invested in a single name stock or bond. These are invested in diversified portfolio of stocks and bonds.

2:51:49And our thesis is, instead of buying a house, taking your entire life savings and betting it on a zip code, you can take a piece of your life saving, lever it up five times, and invest in a diversified portfolio of stocks and bonds. That's the thesis here. Yeah, yeah. Just a habit for us. So who's the right person for this product? Because I'm assuming... Long-term investor. Long-term investor. So that means somebody that's maybe going to retire in 30 years or 20 years, something like that. If you have anything less than 10-year horizon, you should not consider basic capital. Basic capital is not appropriate for you.

2:52:27Basic capital is a product, is a bet that over a long period of time, financial assets are going to continue to appreciate. Now, is this guaranteed? Somebody from Twitter is going to jump in and say, well, the stock market will go down. Of course it will go down. No, like Sherlock, like, thank you. Yes. Well, the nightmare scenario is like the Japanese stock market, which had like 20 something years of decline. Stagnation. Exactly right. The way you could actually get burned on this is if you have persistent long term stagnation in financial assets. Is that possible? Your argument would be like, if that happens, we all have bigger problems.

2:53:10My argument is, my argument is, America is not Japan. America is not Europe. America has a problem. America has Silicon Valley. America is the epicenter of innovation in the world. British companies like Spotify want to go public in America. Klarna want to go public in America. and when you are investing in American assets, I believe there is some network effects to this financial market. I love it. I love America. So I call Buffett here and I say never bet against America. And if somebody wants to take the other trade, I'm happy to do it. Like I will do a bilateral side trade. That's literally your business.

2:53:51Your business. Yeah. That is fantastical. Congrats on the growth. I love it. This is perfect. Yeah, it's great. It's great having you on and hearing directly and keep the viral videos going. I think next time you've set the bar high now, you aim for 50 ,000 views. The next video, you should just aim for 50 million. 50 million views. I want to see it. We're not going anywhere. We're not going anywhere. Our message here to everybody is, listen, there is a real problem going on in society right now, which is vast majority of Americans own no assets. And we have a solution for that. and it may be perfect it may not be perfect and for the people that are out there coming after us we have a question for them what's your plan yeah that's a good point what's your plan yeah these are personal personal decisions yep what do we think like what's what's our plan as a society for thing for this and we don't think the solution is communism we don't think the solution is socialism we don't think the solution is mavdani we think the solution is to make every american a capitalist yep and by capitalist i do not mean ideologically a capitalist i mean an owner of capital are you gonna let last last question uh is there is there a way i know we have these invest america accounts the brad gerstner accounts the trump accounts could we give leverage to one month old to children to children could we help the children lever up please like that's literally let them like that's literally the perfect application of this they're out of the womb, they should have leverage the second they're born.

2:55:26A minute later, they should be levered up. You know what? I'm going to call Brad as soon as I get off this. I'm going to call Brad Gersner. I'm going to put shit to him. And the administration is open-minded to this. Leverage is a great tool. And so many people have made so much money using leverage. And it's so condescending when we say that the average American does not deserve it. It's just disrespectful to the average American. Yep. Well, thank you so much for taking the time. Thank you for joining. Congratulations. Congratulations on the funding. We will talk to you soon. This was really great chat.

2:55:55This was fun. Thanks, guys. Have a good one. Cheers. Ferrell Logix in the chat says, what up, brothers? What's up, Ferrell Logix? Thanks for tuning in. We have our next guest, Derek Lowe from Medallion, coming in the studio. Derek, how you been? Good. What's happening? Welcome. Give us the news. I want to ring the gong, and then I want to talk about some of the backstory and then some of the dynamics in the industry. but please introduce yourself the company and the news yeah for sure my name is Derek Lowe I'm the founder and CEO at Medallion this is a chaotic stream by the way we're having fun over here we're having a fun today and what's the news yeah so we just raised uh 43 million dollars led by wow that's cool incredible um who'd you raise it from um yeah so a crew led the round um with participation from some of our existing insiders like Sequoia, Google Ventures, Spark Capital.

2:56:50So yeah, we were very happy with how the run came together. Yeah. Explain the business as quickly as you can, and then tell me about how the market's developing, what's going on. A lot of people have been following the Ozempic news and the D2C wave, and just kind of get us up to speed on what's going on in the market broadly. Yeah, for sure. So we help to automate back office processes for healthcare companies, primarily around managing a clinical network. So any healthcare company, right, think like HIMSS or Rho or Teladoc, they all have to hire a bunch of clinicians to be able to administer care.

2:57:23So everything from a digital health company to a hospital to a primary care clinic. And so that clinical workforce is really the most important component of a healthcare business to be able to actually deliver care. And we essentially help with a lot of the healthcare-specific operational tasks that have to happen behind the scenes to be able to manage that workforce. Is there a world where you bundle this with like payroll or something, like give the compliance stuff? Like it feels like deeply linked in the core system of record. How deep do you want to sit in the system of record stack? Yeah, it's a great question.

2:58:00I mean, we're very adjacent to payroll and some of those other, yeah, back office workflows. But we want to target workflows that are very specific to healthcare. And so we're applying AI and automation to like this very niche, like vertical, which happens to be still super massive, just given the size and scale of US healthcare. So, yeah, we wouldn't really consider getting into payroll just because it's a super competitive space. It's very horizontal. Yeah. I remember the initial thesis was like, in the future, there will be more doctors that will be, you know, seeing patients in different states.

2:58:38And so licensing particularly will be important at the early stage. How has that played out? How common is that? Is that still growing, like, you know, in line with your expectations? yeah johnny i think you were you were actually i was looking back you were one of our first 10 yeah which is pretty it's pretty cool um when we when we started the company yeah we were very focused on digital health because it was right during covid and so the expansion of telehealth was obviously really relevant at that point yeah um i think what we've seen is that post covid that's leveled out quite a bit and um obviously the majority the vast majority of care is non telehealth based.

2:59:20And so typically here is delivered locally. And so as a result, yeah, where we've seen most of our growth is actually expanding beyond helping with those type of, yeah, essentially like cross-state telehealth use cases operationally and more on actually helping providers get in number with insurance payers is that's something that's ubiquitous. It's like 99 % plus of care is delivered right through insurance. So that's where we really focus today. Where in the healthcare tech stack is proving to be AI resistant in the sense that like the best frontier reasoning models, they just aren't satisfying people in the way that maybe people were thinking, oh, let's throw automation at this.

3:00:04But turns out we want to keep a human in the loop for maybe compliance reasons that we didn't perceive. Like where are the last jobs in back office healthcare going to live? Yeah, it's a great question. um parker connor actually has a really great um podcast recently a few weeks ago where he talked a little bit about this and how for rippling like um yeah ai hasn't really they haven't really deployed it in any major ways at least in payroll because it's it's a really critical workflow like things have to go right it's very rule-based it's a really good analogy to medallion where we're working with um an extremely fine margin for error um things have to essentially be like perfect.

3:00:47And what we've seen is like current models are not capable at, I think, doing perfection. I think that they'll get there pretty quickly. But yeah, what we found is that essentially executing on these type of workflows deterministically is proving to be better for now. I think it will change quickly. Build the software using the old tools, hopefully some AI code gen in the mix to speed up software development, but ultimately just a... Yeah, and we're using it in places where the market for area is wider. So things like calling a provider to say, hey, you need to give us some piece of data. The market for area there is a lot wider than if you're filling out a form that requires a hundred data elements.

3:01:31If you get one character wrong, it can script the whole process. That's crazy. Jordy, anything else or should we let Derek go? No, this is great. Thank you so much for hopping on. Congrats on all the progress and the growth. We will talk to you soon. Have a great rest of your day. Great to meet you, Derek. Awesome. Talk to you soon. Thanks, guys. Bye. Up next, we have Alex from Convoke coming in the studio into the TBPN UltraDome. Oh, okay. I'm handing over the mallet. My turn. Give us an introduction. What's happening? What's your name? What do you do? Great to be here. I'm Alex. I'm one of the founders of Convoke.

3:02:01Got any news for us? And we're – can you hear me? Yeah. Yeah. Yeah, so I'm one of the founders of Convoke. We're building a platform to automate knowledge work required to take drugs from an idea to a product that patients can use. So today we're announcing our 8.6 million seed raise. It's from Lee Marie at Kleiner Perkins. No way. Lee Marie, the GOAT. Yeah, the GOAT. That's my GOAT. At Kleiner Perkins, Randall Braswell. Yeah. Yeah. We're trying to rename the firm. Amazing. Congrats. That's amazing. Give us a quick history, the company, what you were doing before this, and yeah, how you got here.

3:02:36Yeah. So before I was a life science consultant, so I worked with pharma companies on all sorts of drug development tasks, a lot of information synthesis, Are you the one that comes up with the crazy drug names? I feel like every drug name. I actually did do one drug name project. Can you tell us the name that you came up with? I cannot tell you. We had to test it with patients and everything. We were pretty happy. He could tell us, but he'd have to kill us. Exactly. Anyway. Where have you taken the business now? Give us the latest and greatest. We're building an AI workspace. So our goal is to really automate like a significant amount of knowledge work required.

3:03:17Like if you start from, if you think about like what it takes to take a drug to market, you have to do a combination of planning work and execution work. And execution work is like clinical trials, running experiments. That's where a lot of the focus is. That's where a lot of people kind of put the bottlenecks. And what we're doing is kind of focused on the other part of that process, the planning and preparation work. So that takes like 45 % of the time of a drug's life cycle. And really, if you can speed that up, help companies make better decisions in how they synthesize information, search information, prepare regulatory documents faster.

3:03:52There actually is a potential to really accelerate time that it takes to take a drug to market. What's your read? Is this about replacing human labor or just making human labor dramatically more efficient? It's more the latter. The way we see it is there's so many questions in biotech that are just intractable for humans to solve alone. Like the search space for making a new drug is so large, you just cannot do it with humans. There's 20 ,000 or something proteins in the human proteome. If you want to make a drug against any one of those proteins, you know, if you want to do it like really thoroughly, you have to read all the literature on every single protein.

3:04:31You'd have to think about all the different ways you could drug every different protein, different combinations of drugs. That's just completely intractable search problem. And then you'd have to make all the different documentation to run those experiments and plan those experiments. So what ends up happening is a lot of people rely on heuristics to come up with ideas for new drugs. And we think that actually what's going to happen is people are going to change to be managers of these AI systems that are going out and doing all this work and structuring data and running analysis. And then coming back and bubbling up recommendations that the human can then make the final call on.

3:05:08Where at the intersection of AI and bio are you bearish given your experience working in the pharmaceutical industry broadly? I'm sure you see companies that raise$100 million and maybe you don't always have the full faith of maybe their venture capital backers. Yeah, I'm relatively bearish on patient recruitment as a space. I think if you look from just clinical trials, one of the obvious problems is there's not enough patients or it's hard to recruit patients for the trials. So trials run more slowly than you would like. I think the reality is that the supply, the demand for patients far exceeds the supply, especially as we're developing more and more niche drugs for niche conditions and personalized genetic medicines.

3:05:54and I think a lot of these companies come up with a promise that we can source and final these patients and bring them in when in reality the patients don't exist or the only trial is being run in Mass General Hospital or something but the patients in Nebraska. Even if you add an AI bot that can call hundreds of people or reach out to hundreds of people, it doesn't solve the fundamental supply. The company needs an AI bot that gives people rare diseases so that there are more patients to cure. This is the real strategy here. Yeah, or you can fly them from Nebraska to Boston or something. Yeah, yeah, yeah.

3:06:29That's probably it. That's the more humane thing to do. Can you take me through kind of like within your clients, your customers, is there like a front office, back office dichotomy? It sounds like this is not a electronic lab notebook. You're not in the lab. but how do the companies see the dividing line between like the knowledge work piece, the regulatory filings, and then what's, and the trial design versus the actual like mixing of the test tubes and like pipetting? Yeah, I think broadly you can think of pharma companies are split into the R &D organization. Those are the guys in the lab with the test tubes, the commercial organization, and then the development organization.

3:07:11The commercial organization sells the drugs, that's the sales reps, and then development organization, You know, many different functions within that, but broadly it's planning trials and, you know, communicating with stakeholders. So like one customer group we're working with, it's called a medical affairs function. Their goal is to coordinate with doctors and represent the company's science and programs to the external like world of doctors. And they help with, you know, communicating their science and also like enrolling the doctors into their clinical trials. is the comp or is the state of the art or like the current solution is it more like google docs or pen and paper or is it like some custom erp from a you know dot-com company that's kind of like loosely solved the problem and then been like sold 25 times yeah it's completely under penetrated by software so the current solution is consulting i mean i started as a consultant so this is why i know this is a problem you just have human like massive amounts of human labor highly compensated human labor, going and reading documents, inputting things into Excel, making PowerPoints.

3:08:17You know, the software penetration is the Microsoft Office. For Microsoft Excel, still undefeated PowerPoint. Can you sell to consultants? Thinking about it? We're not, yeah, they're not like, we don't consider them our ICP. Okay, got it. Makes sense. Well, congrats. Congrats on this funding milestone. Given that Lee Marie did it, I'm sure you'll get it. you know, somebody that multi-billion dollar acquisition offer right around the corner. Yeah. It seems to be on a terribly year out. But tell her, he said, hi, we will talk to you soon. Have a great rest of your day. Great stuff. Thank you so much.

3:08:52And we are running late. Jordy has to hop on with Mogadishu and Pyongyang, London, Bermuda, and the Cayman Islands. Yeah. He has to talk to the Cayman. That's who you have to hop on with the Cayman Islands. I wonder why. Anyway, thank you, John Exley for hanging out in the chat. It's always a pleasure. Thank you, Christopher. Thank you. And Ferrell Logix, Gabe, everyone who's here, Taylor. We appreciate you. And there's 11-ish days left of summer. 11-ish days. There's something big coming. I was saying we're going to announce it today, and Jordy said we can't announce it yet. It has to be the end of summer.

3:09:27So expect something big September 1st. Very soon. My world. Very soon. We will talk to you tomorrow. Have a great rest of your day. I love you. On Apple Podcasts and Spotify. Goodbye. Cheers.