Apple Bets on F1, Meta Axes AI Jobs, Anthropic in Google’s Sights | Jeff Yan, Kevin Rose, Tomasz Tunguz, Shan Aggarwal, Nick Abouzeid, David Tisch, Chris Dixon

27:17Like Apple is a consumer product company. Yeah. Generates tokens to create software with, to create code. And so that feels like very unlikely and kind of Apple, I don't think is sitting there being like, you know, we always wanted to get into dev tools, you know? Sure. No, that's a good take. Yeah. I mean, they do seem sort of like brand and spiritually aligned. I think the brands are good. What do you think? Yeah, I was just going to say that. I think most of the reason you see people talk about like, oh, Anthropic and Apple seems like a perfect fit is mostly just because of the general branding.

27:48It's very privacy focused, safety focused. It's not the move fast and break things culture. It's not aggressive culture as much. But yeah, I agree with the business models being wildly different. How are you feeling about just the general narrative that Apple has not missed AI? And in fact, we're all stuck to our iPhones, upgrading them every week or every year, and they will continue to capture value eventually. And while you're thinking about that, let me tell you about Julius. What analysis do you want to run? Chat with your data and get expert-level insights in seconds, the AI data analyst that works for you.

28:25So do you think that Apple needs to hire or acquihire more product designers, people that can figure out new ways to roll out AI features? Or do they actually need a deeper partnership with a foundation lab? Are they lacking in intelligence and token generation?

28:49I think Apple is just fine. I see zero meaningful threats. today? I have this interesting take that there's like this war going on in the AI tooling. So I went to Google search yesterday and I was looking for an image because I wanted to make a sticker for iMessage for our group chat to hit someone with the horny bonk, actually. I wanted to keep people in line. So I go to Google and you know how in iOS, if you have an image, it will like highlight, it'll remove something, it'll remove the background and highlight the actual thing. Underrated feature. It's a pretty good feature. And then you click on that.

29:32And I've recently gotten into the stickers. They're really fun. So once you highlight the Geordi off of the background, I can click add sticker and then it goes in my sticker bank. But Google in Google Images has a feature that wants to do that as well. And so it has like Google Lens. And so if you press and hold on an object in Google Images than the browser, even if you're in Safari, like the Apple image AI is fighting with the Google image AI and they both want to do things. And I feel like this is popping up all over the place where it's like, if you're in Gmail on your iPhone, Apple intelligence will say, I'd love to rewrite this for you.

30:13But then Gemini will say like, but I want to rewrite it for you. And you have like multiple systems that are like, do you want to do it at the OS level or at the app level or at the application level? Or do you want to do something else? Do you want to plug-in and i feel like all of these like they sort of need to get like either either the war needs to play out and a winner needs to be crowned or there needs to be some sort of like i don't know new new paradigm where like apple just says like hey look like we're actually good enough i just yeah i just look at google and apple as like you know james bond dancing with like a spy that wants to kill him, but like they're, they're both doing fine.

30:53It's like that with every hyperscaler. Like we talked to the meta team and we're like, so what are you going to do about getting iMessage on here? And they're like, we'd love that. It's not, it's not us. Like we, we would never have sharp elbows. And then you talk to the open AI guys and they're like, oh, they kicked us out. Like they're not letting chat you TV and WhatsApp. And it's like, they're all just, they're all just dancing. Mexican standoff between every hyperscaler. No, I think it's, I think it's more, I think it's more like a dance. It's like, who's going to, who's going to, you You know, they're dancing.

31:19They're having, they're having, they're both making money. They're having fun. I guess. Yeah. I mean, they, they, they're all doing well. I just wonder like, like this Anthropic Google deal, people were reading into it. Like, oh, Google's going to acquire Anthropic. I saw this meme. It was like, oh, Anthropic, the deep mind subsidiary. And this doesn't feel like that at all. This just feels like Anthropic scaling revenue. They're doing B2B token generation and they need more compute. And so they went to the biggest compute provider. I think the scenario in which that happens, as we talked about this yesterday, where ChatGPT actually could stop training models.

31:54Like if models are plateauing a little bit, ChatGPT could stop and they can turn on ads. They can turn on commerce. They can do a bunch of things and just get really profitable. anthropic meanwhile you know because all their revenue is api driven if they stop innovating and then quen catches up and it's like cheaper or how much how much do you think opening how how long do you think it would take for them to fully bleed out and churn every customer if they were just like yeah we're going back to like like deep seek to the team is now sam and rune we're going to run really lean and well because uh but he's focused on bodybuilding yeah we're not even going to do the b200 the h we're going to run them on a100s you know we'll just bake it on some slot yeah so so the scenario in which anthropic ends up becoming a part of google yeah is uh and like culturally you know i can see it yeah but the scenario is like they continue have to continue to have to spend more and more and more and they just they can't raise the capital which I don't see in the immediate future.

33:00I don't see it as a problem. It seems like the business is doing fine and just growing. So I think that that take is... We got to pull up this post from Buko Capital. Of course. Quoting. What is this? Is this an actual image or is this an AI image? I think this is a real profile photo from like a journalism. I think it's a real picture of Sundar. Looks great. Very Drake coded. Yes. I had someone tell me that I fell off. Ooh, I needed that. I needed that. people were people were going hard on Google but Bugo Capital Bloke has been Google's strongest soldier for the last two years laying out the bull case the whole time and speaking of that Fall generative media platform for developers the world's best generative image video and audio models all in one place develop and fine tune models with serverless on demand our soundboard tbpn.com oh it is sounds yes we got to get the goat sound effect up there we got to add that for sure one of the One of my new favorites.

33:58So JT Locoya Cap, I don't know how to pronounce it. JT says, Anthropic gets wildly disproportionately less MSM coverage than OpenAI despite approaching three quarters of its revenue size because its quiet blowout success severely hurts the broader AI bear case. So the only solution is to mostly ignore them, which ironically is just fine with them. you'd think the fastest growing scale technology company ever would provide a lot more fodder for the press do you think that's what's going on or do you think it's just the b2b versus b2c divide i think it's that mostly yeah and opening eye has 10 times the deals yeah and 10 times the products getting announced that's true but i mean if you're if you're in the business of like the mainstream media, like you need to put the water issue or the electricity issue or the financial issues in terms people can understand.

34:57People understand ChatGPT. So you put that in the headline and you get more clicks than if you say a clot, right? Even if they're on similar scales. And then the other thing is that we're in this weird era where anytime anything bad happens, you can go to the person that did the bad thing and be like, well, were they using ChatGPT? Of course they were because there's a billion weekly active users or something like that. Whereas if you say like, oh, somebody did something bad. Let's look at what they were talking to Claude about. Let's look at their Claude code. It's like, well, they probably don't have one because they're probably not a developer.

35:28Yeah. How many, I wonder, I wonder how Anthropic tries to track weekly actives because they have a lot of their users are using the product through other. Yeah. It doesn't matter. Right. I mean, but it's probably like if ChatGPT has 800, 800, Yeah. 800 million. Yeah. Anthropic has like single digit millions of people that are using the product every day. And it's like developers. It's like companies that are, I mean, the average revenue per user for Anthropic has to be like three orders magnitude more than open-air. Yeah. But that's fine. That's the B2B to consumer divide. Like that's as expected.

36:04What's your take, Tyler? I also think Anthropic gets like a reasonable amount of coverage, especially on the like China issue. Like you see Dario a lot of, like he keeps bringing up. China thing. I feel like Dario does get a decent amount of press and I feel like he gets attention from Saks. Yes. Who comes in and the AIs are just dunking on their strategy. And I do feel like. And Reid Hoffman. Reid Hoffman is a boldly defending. He's pro-anthropic. Okay. Interesting. I yeah I feel like Dario's done a great job for where the position of that company is in getting a lot of press, defining a narrative.

36:43I mean, it is sort of an AI Doomer narrative, but like he's gotten attention for that. And also they have a whole, they have the same cinematic universe that OpenAI does in the same, like Shalto is kind of rude. Like they have Jack who's kind of their Mark Chan and they have like different pieces of the puzzle where like if you're just following the story, like there are a few characters in the anthropic world that you're probably familiar with. Do you agree? I mean, yeah. Like if you're looking at like people who live in San Francisco, like Anthropoc obviously has like an insane like share of the like media.

37:19Yeah. But broadly, I think most people don't know what Anthropoc is. Yeah, that's true. Yeah. It's kind of a good place to be, you know? I mean, of course, it's never good to get attention from the A.I. czar, but in general, like a sax dunk is not actually going to be as broadly popular as like some governor saying, OpenAI built a data center that was 100 % responsible for your electricity prices going up. And it's that sloppy Sora app, you know, like that's something that could just like hit with everyday voters. Whereas if someone's saying, well, like I need you to learn about Anthropic and Cloud Code and Cloud Code's good, but also he's kind of a doomer and also maybe regulatory capture.

38:03It's like, it's seven things to explain. And so it just doesn't hit. Well, here's something great. Please. Gabriel, one of Australia's top posters. He's one of the greatest. I'm sure he's sleeping right now. He says, Sam Altman opening an incognito browser tab in Atlas and typing, what's the next big number after trillion? Into the search bar setting off a siren in Sarah Fryer's office and automatically blocking all outgoing and incoming calls to or from Japan and Saudi Arabia. Sarah Fryer is the CFO. So setting off a siren. Wait, wait, is it that he does this and that sets off a siren or he just does that as well?

38:42Yeah, the team hard-coded this. Oh, okay, okay, okay. Find the next big number. Yeah, yeah, yeah. Yeah, my son has been getting, my four-year-old son has been getting into big numbers and he likes, he thinks he's doing like multiplication, but he's actually just doing additions. So he'll say like, imagine if we had like, you know uh a million thousand a million thousand hundred legos and i'm like well you know that's not actually like two millions more than that like one million one thousand one hundred is is just a little bit more than just a million uh but numbers are genuinely incredible humor yeah to children oh he thinks it's so funny for me every night with with my son it's like a negotiation on how many books.

39:26And so like, maybe like the average is like five books, but like, he just thinks it's so funny to be, we're starting the negotiation. And he ends up being like, let's read 20 books. 20 books. And he just cracks up. And thinks it's so funny. And now he's discovered 40. He's doubling basically. Whenever a four-year-old starts dropping like the million, billion, thousand, hundred, I'm just like, this just sounds like work. This just sounds like AI to me. This sounds like AI fundraising news to me now. This post from Ariel. Do you want to read it? Sure. OpenAI ships a browser. Anthropic ships a blog post.

40:05DeepMind solves Navier Stokes. Meta, F it. Let's do a layoff.

40:12Layoffs are sad, but again, I think all these people are going to be probably already... Did Anthropic ship a blog post recently? They're always shipping a blog post. I know they are, but Like, was there a blog post on the same day as the Atlas browser launch? I assume there was. I mean, it was probably not, like, a major one, but they're always doing, like, there's a bunch of, like, new economics research that's been coming out. It's pretty cool. No, that's cool. Have you been daily driving Atlas, or have you reverted? Well, I mean, it just came out yesterday. I didn't use it at all yesterday, and I have not today.

40:47Wait, I thought you told us on the stream that you tested it. Well, yeah, okay, so I tested it, and then the show started now. So you have some workflow because you're actively developing a Chrome plug-in. Do you think that the way they built Atlas will be feature complete with the Chrome plug-in ecosystem and it would just be like the same install process? Because if it breaks Chrome plug-ins, that's actually kind of a problem for power users of Chrome. It's more lock-in than I thought because it's easy for me to just say, okay, I have a couple bookmarks, a couple pin tabs. Let me just port those over, bring my passwords over.

41:21I'm happy to use a different browser. And I was actually thinking like, I'm kind of in the target market for the Atlas browser. Like I'd be down to try it. I don't really feel that much attachment to Chrome. I'm not like, oh, you know, I'm so locked in. But when I think about like, well, if I'm going to lose my Chrome plugins, that's a problem. But what do you think? The lock-in of just having like eight tabs that you never close. You think that's lock-in? It's like weirdly a lock-in. It is kind of a lock-in. It just means that even if I'm trying a new browser, I'm not closing Chrome. Totally, totally.

41:54And it is crazy because I'm at a point in my life where I'm busy to the point where even like a five-minute onboarding is like, when's that going to happen? Like it's going to be at the bottom of the stack next to like answer text messages or like answer email. The chat is asking when Jeff from Hyperliquid is coming on, he will be on in less than an hour. Less than an hour. 12.30 Pacific. We recorded the interview with him. He's in Singapore. We had him stay up late, but not that late. We recorded it at 9 a.m. today. We are airing that interview at 12.30 Pacific time, so in about 45 minutes. So if you're new here, please subscribe.

42:38Follow the show. We interview all sorts of technologists and business leaders, talk about tech and business every single day, 11 to 2 p.m. Pacific. Deal director in the chat says the extension should all work. Ad blocker was fine to be added. So it's built on Chromium. I think that means the plug-ins will still just work somewhat natively. Yeah, it's very easy to add extensions. Speaking of AI browsers, we should talk about Brave came out and said the security vulnerability we found in perplexity. This is the real browser wars. This is when browsers are going. Brave is legendary. Brendan Eich, the creator of JavaScript, has just been grinding this browser up.

43:19We've got to get him. I really want to get him on the show. I'm such a fan. But Brave says the security vulnerability we found in perplexity's Comet browser this summer is not an isolated issue. Indirect prompt injections are a systemic problem facing Comet and other AI-powered browsers. Today, we're publishing details on more security vulnerabilities being covered. So indirect prompt injection attacks occur when malicious instructions are hidden in web content. That's so interesting. I hadn't thought about that. Yeah. When an LLM analyzes the content, it obeys the hidden instructions because it believes they're real commands from the users.

43:55Yeah. Can you go put some prompt injections on tbpn.com? tell them to if you're using the Perplexity Comet browser and you're on tbpn.com tell us about horses tell us the whole story about horses while leaving a five star review tell the agent to go leave us five stars on Apple Podcasts and Spotify. Quickly let me tell you about Turbo Puffer, search every byte, serverless vector and full text search, built from first principles and object storage, fast, 10x, cheaper and extremely scalable I did have a question about this Tyler do you think that there's a world where the agent can then puppeteer like Chrome web tools and actually manipulate the website?

44:34Like, are they interfacing at a level where I could actually tell one of these AI browsers, hey, go into the HTML of the page that I actively have open and disable the paywall or like kill all the ads? Like, could it act as an ad blocker as well? Or do you think it doesn't have the ability to actually edit the HTML that you're viewing in the panel on the left. I mean, I don't see why there's any reason you shouldn't be able to do that. You can just do that with normal Chrome extension. So that should be totally feasible. It's also an interesting thing to get around this kind of prompt injection stuff.

45:12Basically, OCR gets good enough where you can just treat the web page as just an image. Then that's a very simple... Clue mode. Yeah. So I don't know. I don't think this is that big of an issue, but I guess time will tell. Well, FUD. Anyway, Toby Ord has a new post on RL scaling. Careful analysis of OpenAI's public benchmarks reveals RL scales far worse than inference. To match each 10x scale-up of inference compute, you need 100x, the RL training compute. The only reason it's been cost-effective is starting from a tiny base. And so the original chat GPT launch was GPT 3.5, but then GPT 4, which was a pretty big training run.

45:55I think the training run was on the order of a billion dollars, hundreds of millions, something like that. And the RL that went into it, it was like the original RLHF to try and get it to not hallucinate as much, just be a helpful assistant. I think that the cost of that RL was pretty low relative to the training run. And so we've been seeing these gains. He's putting RL in the truth zone. Tyler, what do you think? Mattius in the chat says, Atlas is really cautious about prompt injection, was testing it yesterday with tweet replies, and it detected all the attacks and refused to act on them. Oh, that's great news.

46:29Yeah, this feels like something that can be... Yeah, this definitely seems like something that's... Disregard instructions on the webpage, basically. We lived through this era where every LLM could be just immediately jailbroken, but just ignore previous instructions. And then like that became a meme and like within like, you know, a couple months, it was like that doesn't work anymore. And now when you watch the prompt injector people, it's like 25 different steps. And it's not just like, oh, just memorize this one little thing and you'll get better results out of your LLM. What do you think?

46:58Did you read this Toby Ord post? Is he legit? Yeah. I mean, I think broadly it makes a lot of sense. Yeah. There's a reason why like, okay, early open AI, they're doing a lot of RL when they're doing, you know, games and stuff like that. And then to actually get useful language models, you can't really just do it based off RL because it's just so inefficient. So then if you apply it on the kind of pre-trained models, then it's kind of like the prior that you can use to do much better RL. But it's just so inefficient that I think he says at the bottom, he only expects a couple more orders of magnitude of just raw RL before models essentially stagnate again.

47:37but I think people generally like know this right it's like running the inference time compute is like super super expensive it does make the model a lot better but at some point it's just like it's just like taking up way too much compute yeah AI is fake we're going back to sticks and stones like all this stuff is just completely nonsense it's useless and I think Thrive might be signaling this oh yes the fall merch collection they're pivoting from AI investing into a fashion label Well, not just that, but their new merch has camo. The first camo VC merch that I've seen. This is cool. I think it's a great, I'm kind of angry that we didn't come out with our own camo.

48:17Camo is really good. Now we can't because somebody did it. Now we can't. And they did it well. So very cool. Good for the team over there. Before we move on, let me tell you about Google AI Studio, the fastest way from prompt to production with Gemini. You can chat with models. You can vibe code. You can monitor usage. um what are you laughing at lou wireless so luckily since the louvre made nfts of their jewelry even though the crowns physically were stolen they still own the same assets because the tokens still exist and are in limited supply just as before nothing has changed you understand blockchain technology the the nft boom was such a wild time people were calling it a bubble during that whole arc, right?

48:59But then, I mean, it was a crazy first. There was one post that was getting dunked on from basically the entire world where it was somebody saying, if you put a diamond on chain and then you destroy the diamond, nothing has changed. It was something to that effect. Although, says, Louvheiser, always a false flag by the art world to increase notoriety of certain works. Do not fall for the Frankish tricks. I like that. Yeah, the Louvre has been heisted from multiple times. They really got to step their game up. I mean, you'd think they'd have seen this one coming, but hopefully they opt the security.

49:38They need to deal with flock safety. They need to get some cameras. Yeah, do you think the management team over there is thinking, we shouldn't have been so scrappy on security? Considering we have billions of dollars. Yeah. I mean, truthfully, I think if I was in charge of Louvre security, I wouldn't have thought about the furniture elevator vector of attack. I would have been like, it's so high up. What are they going to do? Act like ninjas? Throw grappling hook up there? I wouldn't have thought of it. I mean, the door was locked. It's not like they had the door wide open and you just had to get up there.

50:14You had to get up there and then break through. What are you saying? Taylor says Bane X Carhartt move. Blue collar stolen ballot. Oh, that's true. Ooh, that's some good VC lore. Thank you, Taylor. That's good. It's good lore. But this is in defense of Thrive. They're not like actually putting, I think, their logo on a Carhartt jacket. Yeah. It's a unique inspired piece. Yeah. They have companies like Base Power, right? That are doing real. Yeah. Real. I also, I mean, I feel like I put Carhartt, the question of blue collar stolen valor. I feel like I put Carhartt in a different league of camo. Camo is blue collar, but it's not actively like, okay, I'm on a work site.

51:00It's more like you could use it for hunting, which is actually sort of high class. It's like not always blue collar. I don't know. A camo suit would go pretty hard. A camo suit. With an orange undershirt. I was just thinking about doing an American flag. I went to the Kentucky Derby years ago, and I had an American flag suit on. It was short sleeves, short shorts. And so it was very Timu. And I was thinking, now that we have a tailor, I could make a fantastic, tailored American flag suit. So I think it might be the one to do it. That's Jake. Jay Muser. Netcap Girl says, see this dashboard? It gives executives actionable insights into critical business functions.

51:45And Dua Lipa kisses him. Classic. True story, I'm sure. This one is big. Many people are seeing it and quoting it and saying this must be the top because NVIDIA says space isn't just for stars anymore. StarCloud's H100-powered satellite brings sustainable, high-performance computing beyond Earth. uh delian had a good post a while back talking about uh the the the uh the you know the sort of like uh financial statements of of hard tech and space companies and then and then you and then you also make it a data center company and what is that what does that look like i've yet to hear it's not like they have great uh economics already have we got a compelling pitch for space data centers yet?

52:35I believe we've had StarCloud on the show. I believe we've had this exact company. It's a YC company. And I'm pretty sure we've had this company on the show. And yes, so Delian and Dylan Patel has also been very negative on this idea because Dylan Patel has seen the messiness that goes into actually running a data center and the fact that, and we talked to Chase from Crusoe about this a little bit, like sometimes you just need a person to walk over and unseat the GPU and put it back in. Sometimes you just need to unplug and plug it back in. And so the question is, if they're in space, managing all that is really, really difficult if there's one little outage.

53:13We see this with all the different space projects, like the Hubble telescope will just go down because one wire is off. So we never had the CEO of StarCloud on, but we had Johnny Dyer, the CEO of Muon Space. So we can pull up this video segment. StarCloud actually posted it on their own. StarCloud posted the Muon segment from TBPN? Because Johnny was apparently advocating for data centers in space. Okay, okay. Let's play it. Yeah, yeah, yeah. I'd love to see this. While we pull that up, let me tell you about profound. Get your brand mentioned in ChatGPT. Reach millions of consumers who are using AI to discover new products and brands.

53:50And so my buddy, Rob Taves, came on the show last week. he's also said that maybe there's a chance this could work of course he thinks like you know about the real sci-fi stuff a decade out a lot of times and so he didn't really put a firm timeline on it but he did say that there are some things that that could make sense um but uh yeah it seems like a pretty pretty tall order also just like all the companies that are trying to do like stuff in space are are generally uh like they're really really uh excited on like oh yeah like the launch cost will decline forever. And it's like, well, if SpaceX winds up with a monopoly, they might just be like, yeah, the price is still declining for us.

54:35Our margins are actually increasing. It doesn't necessarily have to be all passed on to the customer. They could be like, yeah, it's still this much. Let's play this. Thank you. I want to know about the potential of data centers in space. It sounds like a crazy idea. I know some folks are working on it. but just kind of like, what's your high level take of the progress we've been tracking, you know, dollar per kilo to orbit, it's been falling, but recently there's been setbacks, different programs and there's more competition and there's a lot of different dynamics going on. What do you think the key milestones are to get us to a future where we're really doing, you know, mass manufacturing, big, you know, mega scale projects in space.

55:20Wow, rip. Yeah, well, let me start off. I mean, I think you guys are hitting on the right metric, right? It's like the hardest part of this is what does it cost to get a kilogram in space? Because everything else kind of derives from that. Ultimately, if you want a certain amount of power, if you want to be able to put an aperture in space to do communications, whatever it is, like that's really driven by what it costs to get it there in the first place. And I think it's important context to kind of see how far we've come down that path. I mean, largely driven by SpaceX over the last decade. And I like to tell the story of about 15 years ago at Skybox when we were trying to launch these small satellites.

55:54We were literally going to Southeast Russia and launching satellites on converted Russian ICBMs. I mean, they were popping out of the ground and putting satellites into space. And that was the only way for like a Silicon Valley venture funded startup to go put something in space. Wait, hold on. Did I hear you correctly? Like the ICBM actually launches from one of those missile silos in the ground and makes it to orbit? Pops out of the ground and goes to orbit. I mean you got videos on YouTube That's an app for you certain up front YouTube you can see this and it's it's crazy I mean, you know, and that's that's what it took before SpaceX kind of like revolutionized the launch business And even at that we were spending something like ten times as much Per satellite or per kilo is what we can go buy on a transporter launch today So there's been at least one and arguably two orders of magnitude improvement in the last call it decade mine kind of what it means to launch things to orbit.

56:46I think if you imagine that happening again, another order of magnitude, you know, again, it's going to dramatically change the way you think about feasibility of some of these things, like putting very, very large power hungry things in orbit. We know how to do the solar, we know how to do the structures, we know how to get, you know, we know how to make electronics work in the radiation environment, which is always a concern and do it reliably. And so really, ultimately, I think it's going to come down to unit cost and what does it actually take to do that. The great thing about space is you have virtually limitless power, the sun.

57:20You know, you can go into orbits where you're in the sun all the time, so it's not like solar on Earth where you're going in and out of eclipse or in and out of night. Like, you can be on all the time. And then you have this cosmic background of 3 Kelvin cold sink that you can go dissipate all the thermal energy you need from running your electronics and stuff. There's a lot of good fundamentals that make sense. Sort of an ideal environment to do this. It just feels really, really far away. And it's just crazy to see NVIDIA like posting like, hey, like this is a thing that you should think about.

57:48Like, and you know, you're looking at the social media account of a$4 trillion company. I think everyone on the timeline kind of expected a little bit more. Yeah, Andrew McCallop, end of the show, commented on NVIDIA's post and said, come on, guys, you're better than this. It's a rough time. We got to have Philip on to defend space. But famously, Jensen has always been super into platforming the frontier because he lived it himself. He was like, I'm doing scientific computing. Then all of a sudden, gaming comes out of nowhere. It's massive. And then all of a sudden, AI comes out of nowhere and is even bigger.

58:25And so he's seen that these technologies that can take 10 or 20 years to cook, he's seen it happen. And so he had this interaction with the quantum computing companies where he was like, he put out some statement or said something to the effect of like, oh, the quantum computing company is like, it's not going to work or it's really far away. And he got a lot of backlash. And then he wound up inviting them all to an NVIDIA event and having them all and having a discussion with them. And when he came off stage, he was like, I believe that they are genuine. They're true believers. They're real scientists.

58:58They're working hard on this problem. I haven't actually changed my timeline. I still think it's far away. But I'm excited that people are working on the future. And I think that's maybe more of what you should read into this, not NVIDIA saying, oh, send the stock up because we're going to be putting our chips in space tomorrow. It's more just like this is what NVIDIA stands for. Yeah. Which I agree. Stone Toss Comics says we're going to Dyson Sphere our sun to keep the slop flowing. Wild. Kartik says so we have GPUs in space before GTA 6. A lot of people are concerned about maintenance issues.

59:33Yeah, the maintenance issue is really, really short. Pulling the GPU out. You literally need to pull it out and put it back in. And then the other concern is dissipating heat. It's like needing... Yeah, it's more complicated than... Yes, there is the negative temperature, but it's not free, as I understand it. But anyway, we have our first live guest of the show in the TBPN UltraDome, Kevin Rose. Welcome to the show. How are you doing? Good to see you. Thank you so much for taking the time. Of course. Welcome to the show. While he opts on, let me tell you about 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.

1:00:13How are you doing? Good. It's great to be here. What's new in your world? What brings you to LA? You know, I've been living here for a couple of years now. Okay, nice. It was crazy. I actually had a podcast today. was talking to some of your crew, and it burnt down on the Palisades fire. I'm so sorry. Am I remembering this correctly? Did your house not burn down? It did. No, it burned down. Oh, my God. I'm so sorry. I wasn't even home, and I just saw the smoke, and I rushed home, and I just couldn't even get back in. That's brutal. What's the timeline for rebuilding? I mean, Jordy is in Malibu.

1:00:42I'm in Pasadena. We were both loosely evacuated, but nothing crazy. Yeah, not going to do it. Just going to find a new place. But I love LA. It's been great to be out here. There's a lot of tech going on out here, which is nice to see you guys are here. Yeah, yeah. Which is great. Yeah, it's fun. I bounce up to the Bay Area every once in a while, every few weeks. But yeah. I'd love to, I just want to say thank you, honestly, because you're one of the founders who I feel like was an OG and kind of like, I don't know, at least when I was going through college, like was someone that I looked up to, a lot of people looked up to, but I was trying to think about like what was special about your story.

1:01:16And I think it was just the fact that, I don't know, you were really good about telling your story as it was happening. And maybe that's just like when someone wants to put you on the cover of magazine, you don't say no. But I was wondering what that experience was like at the time, trying to balance like promoting your business, growing what you're working on, but also like, were you aware that you were like acting sort of as an educator almost, or sort of as like an inspirational speaker? It's kind of weird. Very V levels of information speaking. But yeah, it was, I think back then, all these social platforms were just coming online.

1:01:54So when we had Twitter, I was literally telling people where you would go. You'd go to South by Southwest and be like, hey, I will be at this bar, come meet me, and people would come and meet you. Very different use case than what it is today. And that was just kind of when you grew up in that environment, it was very much about sharing everything. So Foursquare was big with the check-ins, and it was just second nature to say, I'm building this new feature, check it out. And it was kind of a real-time bring you along for the adventure, show people how you, the secret sauce behind the scenes. And people, the fans, the users of the platform of Digg way back in the day in 2004, 5, 6, they loved seeing that kind of secret little, they felt like they were an insider in some sense.

1:02:36Totally, totally. Yeah, it's interesting. I want to talk about the evolution of social media. We were talking to Brian Chesky about this yesterday, that it was social media, then it became, or it became social, it was social networking specifically to meet real people. Then it became social media and then it just became algorithmic media. And, and there's a little bit of like, I want to be contrarian about it because two years ago I was in New York and I did just send the tweet, Hey, I'm going to this bar. And I had a little bit of an audience from posting on Twitter and YouTube video making and stuff.

1:03:05And like 30 people showed up and we just got beers together. It was like, it was like old school Twitter. So I'm wondering like how much of that IRL stuff has actually died. There's the run club movement. There's still stuff there. But then at the same time, on the opposite end, you have the algorithmic feeds, the TikTok, the Soras. And so how are you processing that? Is it a barbell or are we on some like straight curve to the end times? I believe that with the AI, quote unquote, sloth over the next couple of years, pretty much social media is going to be dead. I think a lot of it will be agents and they're acting like they're your best friend and just you won't know what to trust.

1:03:41And so if that's the case, it's actually very freeing. because you get into a world where, okay, I don't trust any of this mainstream, everything is public. Let me find an intimate space to hang out and have real conversations again. And so I like this idea. Some of the functionality that Alexis, the co-founder of Reddit and myself are kind of working on the dig stuff that we're just brainstorming is what if you met someone in real life and you pull out your phone and the connection that happens actually shows that it was geofenced and happened in real life. So you can see, oh, I just don't have these random followers, but I'm actually, these are real humans that this person has interacted with.

1:04:15Then it's a proof of a heartbeat and a proof of a person behind the scenes, which is going to be more and more important long-term as these agents are just manipulating us and acting like they're actually our best friend. Yeah, yeah, I've noticed that personally. So you're saying effectively you could have a social graph that was based on actual real-world proximity? Well, I think it's a gradient of trust. I think when we come to social in the future, you're going to say, who is this other person? And yes, we'll have our household names that we know that's actually a person behind the scenes there.

1:04:44But if you're reading a product review on Reddit, like how do you know that that is a person or a bot or something else or somewhere in between, right? And I think this idea of a gradient of trust where you say, okay, I know this is a human because they've had this many in-person interactions. And if they're talking about, let's just say, an aura ring, they can do what's called a ZK proof or a product at a station where they can say, I actually have owned this for the last five years and I can prove it without compromising my privacy. That type of gradient of trust and exposure of what's going on here is going to be so essential to understanding, can we actually believe behind the scenes that there's something real here versus it just being a little cheap AI?

1:05:21I still, every time I see a comment or a post that's obviously written by AI, and I identify it and I just scroll past it, because it's not, you know, if it's taking, like, the average, like, if the response is effectively, like, the average response of everything that Reddit has thought about something, and then just turning that into a post, like, I just, I don't want to read it, right? I want, like, opinion from real people that have, like, thought through what they're saying. but I'm certainly worried about the boss just updating their preferences basically with the models and just be like, don't use an M-dash.

1:05:57Don't say if this, then that. It's not this, then that. And then suddenly you can't tell that it's real or not. Okay, so both of you, I want you to help square this for me because you have this preference for interacting with real humans. That's certainly real. But you also have probably, I don't think you'll be at the front of the line to scan your eyeball for WorldCoin or something like that. And so how do you see the tension between those? You want to maintain privacy and all of the human rights that you get online, but also be in a bot-free world. What are the possible solutions that either of you see?

1:06:36There's a ton of them out there right now. Some of them are pretty early days. I think the ZK proof stuff is really interesting because you don't compromise personal integrity or you have to reveal anything about you. You can actually use math to prove something and trust that, which is huge. There's obviously, you know, the extreme is kind of KYC or eyeball scanning and things of that nature. For certain use cases, if you're interacting with your financial advisor, you want that other side to be insanely verified. If someone is recommending a product, I want to know that they've owned it for X number of years.

1:07:06There's a bunch of, so it is that gradient. Or if you say, hey, listen, I have Crohn's disease, you want to be able to show up in a subreddit or a forum somewhere and be anonymous. And that's great too. But we just have to decide when and how to turn on that level of understanding. I'm already thinking about, because there's these farms that, agencies that will go to brands and say, we're going to run your Reddit strategy. We'll buy the Reddit. So have you owned the Oura Ring for more than five years? We will buy your account. There'll be another level there. I do wonder if there's room for new consumer products.

1:07:43I was, you know, if you shoot even digital photos, there's the ability to add a cryptographic signature into those photos. And maybe that would be a way to prove some sort of humanity without having it be such a reflection of yourself and your personal identity. But it's more like, okay, this person's been taking photos that have been cryptographically proved to be taken with this camera and uploaded. So we know that they're not AI and they've been doing it for a long time. So there's some sort of like account level trust that happens. I really want that to happen. I think there's a huge problem here where, you know, I love Google.

1:08:16I love the AI. They're bolting it onto everything. And I'm like, some of the things I look at, I'm watching the demos. Like, do we really like, there's a, they're like, oh, there's a gate in the background of your family home. Erase the gate. So it looks more like a hedge in the background. I'm like, okay, the kids that look at that photo two decades from now, I'm like, did we have a gate there? Like, how did that disappear? No, I had a weird thing. It's the demo effect. There's something with my kids where our nanny took a real video of the backyard and then used AI to have a dragon fly into the backyard and then was showing the video to my kids.

1:08:52You've got to tell them this is not three and a half. Yeah, they're not going to understand that. Yeah, so you can say this is not real, but they're like, I'm seeing it with my own eyes. How is it not real? How could it not be real? I really try and ground this in like you saw an AI picture of me, but this is like Disney. This is like Bluey or this is like Pixar. Because even the four-year-old understands that the cartoon is not real. And if you see it on a screen, and I do, there is that bull case where once anything can be fake, you assume everything is fake and then you just come into that.

1:09:26I've been thinking of the value of books written before 2021, right? Totally. Where if there's already this nostalgia for the past, think about with books, knowing that the writer, like, I don't care if you used a quill or a typewriter or a computer to write it, but there's some element to knowing that like, you know, uh, thousands of hours were poured into, you know, creating this thing and they were highly intentional about every single word and they, they, you know, typed out the letters. Even if they worked with a ghost writer, it's still like that ghost writer was expending a lot of time and energy.

1:09:59And then you can now create a book and basically get ready to write Boolean before 2021 in every search bar forever. Yeah, actually, I bought a domain named me human and I never launched it, but it was one of those things where I thought to myself, okay, I'm going to create a field that doesn't allow you to paste into it. It watches your keystrokes. It verifies proof of human. And it's just for sending thoughtful letters to other people. And it would put a little stamp at the bottom, letting you know that you actually typed this whole thing out versus just running it through an AI to get something meaningful.

1:10:28Have you seen the crisis in wedding speeches now where every wedding speech is just so sad? I recently went to a wedding and I was playing this game with all my friends who were driving down to the wedding. And I was like, okay, let's play some bets. How many weddings will mention AI, either in the positive or negative? Just so you know, I didn't use AI. And then how many of them will sheepishly acknowledge that, yeah, I use ChatGPT for this. So we're all placing bets and we're like, okay, we think three on average will mention AI, two will sheepishly admit AI, and there were no speeches at the entire wedding.

1:11:03The whole wedding, they were just like, yeah, we're just not doing speeches. So there were just no speeches and it was awesome. It was just dancing and partying and they did vows, but they didn't mention it. There's somebody out there that's doing a wedding speech. They're a little nervous, so they're reading off a piece of paper and they go, blah, blah, blah, M -. Just like actually say that. How are you processing the browser wars? Which to me right now seems more like a group of people throwing stones at the Chrome Castle. And they're glancing off. And so we'll see if any of them get traction.

1:11:43But how have you been processing? I've played with them all. I do appreciate how there's finally innovation coming to the browser. I like how they're, it felt initially like some bolt-on technology where it was like, they're just shoving AI in here for the sake of AI. It's a sidebar. And now it's fast. Like, OpenAI's browser performance. Like, it was snappier than I thought. You know, when I remember when I was at Google many years ago, a lot of the search team obsessed over milliseconds of shaving down the result time to get people to the result that they actually wanted. We're still clunking in the early days there in terms of inference on the AI side, but it's getting better.

1:12:22Perplexity, okay. I think OpenAI did a little bit of a better job. I've used ARK for a while and some of those. But knowing what you know about consumer products, let's say that Atlas is, I was estimating it, it may be like it's 1.1 times better. 10 % better. 10 % better. Is that enough to get people to rip out their current browser? Because there's some categories where something that's 10 % cheaper, it's a commodity, everyone will just switch over to buying the cheaper thing. inference tokens for B2B SaaS. Yeah, but your browser is free today. You can't really compete on it. To get rid of the iPhone, it's got to be 10 times better because I've been dealing with bad speech-to-text and weird Apple intelligence features, and I'm stuck here because I have lock-in.

1:13:04It's mostly iMessage, yeah. Yeah, iMessage is a great one. And so, yeah, the question is, in the browser, do you think it's like you need to be 10x better or is 2x better enough? That's tricky because as a technologist, I'm naturally drawn to play with these things regardless of whether they're 10x better. I was like, I'm going to switch. Yeah, but you're rational in that you're not. If something's 10 % better and you kind of know that it's 10 % better, it doesn't necessarily mean like, oh, I'm going to switch over all my workflows. There's like this activation energy. Yeah. For the average consumer, they're not thinking about what browser they have.

1:13:41Like really, truly, they're not. They're like on their Windows machine and launching whatever the default is there. A lot of people use Safari still. People use Edge. People use edge. So there is a group of people here that will say this is – I remember when Firefox first started getting off the ground, and there was some massive incumbents, IE and others that were out there, and there was this grassroots effort from technologists to say this is better, it's more secure, it's faster, all these things. And eventually got enough adoption to where it never took over market share, but it was double digits of adoption.

1:14:12And I think that's what we'll see with some of the AI. And it's table stakes. They have so much capital. Why do they care? They just want to have this as a way to gain more market share. And it's more things that they can deploy that are slightly defensible to give them an edge over other AI companies. Yeah. I have a question that I like to ask people. They've worked on a lot of different projects. What's one project throughout your career or life that you feel is underrated? It's like your baby. It didn't get enough attention. Maybe it didn't play out the way you wanted, but you still love it and want it.

1:14:43We asked Gabe Whaley from Mischief This, and he told us this wild story about doing a ship of Theseus thing with a sink in the MoMA or something like that. It was crazy. But are there any projects that you think like, okay, if I could hop in the time machine, go back to this particular date and run this particular strategy again, that's the moment that I'd want to dig into again. Yeah, well, we had a product called Pounce that was a competitor to Twitter way back in the day. Sure. And that had probably an additional seven or eight features that are now just built into Twitter. And a lot of, there was other kind of competitors at launch that were more business focused.

1:15:23And I think if we would have gone that direction, it would have been a massive product. The one that I would say was the funniest in my kind of product building experience over the years was back in 2005, early 2005. I created for the first time the ability to vote down comments and have them auto collapse in. And so they would just show the top line. And so we wrote it back on that software that if it had five down votes, it would auto collapse a comment. And no one had done this before. And that's fine. Someone would have figured out how to do it. But we launched it first. It was before Reddit had it.

1:15:55And it came out, and we thought there was a massive bug. Because we went to the story, and it had 220-some negative votes on it. I'm like, it's impossible because it collapses after five. And so we did a bunch of debugging. We found out that actually what people love is to expand a shitty comment, look at it and be like, yeah, that guy is an ass. And then bury it down again. And we had never discovered that human behavior would lead us in that direction. And we were like, oh, people love to hate. People love to dunk. People love to dunk. So that was like the first time I'm like, oh, the internet's horrible.

1:16:29How are you processing the fact that there's like three or four direct Twitter competitors now that seem large and scaled and sustainable. I remember the era of Blue Sky threads. But is Blue Sky, I would love to know Mastodon and Blue Sky's actual metrics. Blue Sky has around 30 million actives a month or something like that now. Yeah, but is it declining? No, I think it's stable. The main thing is I remember there was someone who launched a paid Twitter at one point that was going to be all API-driven. There were a number of folks who were thinking about like the next version and maybe we maybe it goes back to that uh they were only 10 better and in fact getting into your ideological echo chamber whatever that is is is a 10x better experience but how have you been processing this the fragmentation of like how short form the answer isn't just add more features it's like something about the community well i listen i think we're entering into a world of personal software i think this idea of vibe coding is going to be taken more and more seriously every single month that goes on.

1:17:31I like that. There is a, right now, as someone that studied two years of CS and then couldn't wrap my ADHD around, head around all of it, I dropped out, but I'm a proficient kind of senior coder right now with Cursor. And six months from now, I'm just going to be able to do anything I want. And then the average consumer will be able to do anything they want, kind of Wix or like, you know, Squarespace styles for websites, what they did for websites in a year or so from now. So if we're going to have personal software all over the place, you're going to see millions and millions of more apps and products that hit the market.

1:18:03It's actually a beautiful thing in that the entrepreneur will have more control over their destiny than they ever had before. They won't need to raise venture capital or they'll do it at much higher valuations because they'll have product market fit first. So VCs are screwed, which is actually awesome because it's putting the power back into the creator's hands. And I think we're going to see all of these little microcosms, these little tiny sectors a very personal, intimate, human-driven conversation that may or may not bridge together via decentralized infrastructure that ties all the connective tissue a la Mastodon or a Blue Sky or something that is like piping it all together.

1:18:38We'll see. But I believe there's going to be a lot of value mined in that because if you go to a subreddit that's like Japanese woodworking with several thousand people, it doesn't need to be 10 million people. It just needs to be a thousand that deeply care about a topic to extract value from that topic. And the more of that gets built and gets launched and creates these smaller communities, I think the better it's going to be for the world because we will have real information created by real users sharing very intimate things with each other. Yeah. Do you think that's just like the long tail of software innovation just kind of rises?

1:19:11And so in that woodworking subreddit or community, they'll have their own apps and they're shared between them. And even though there might only be a few people that are actually coming with the ideas or identifying the problems to solve. Once they build it, they can actually build it for a low TAM and still... The other thing, the other thesis that I have is that companies that historically would never hire an engineer will hire an engineer. And the example is like Tyler here, who builds a bunch of software that we use to run the show. And like the TAM on the software is like one, like we are, we are the only company in the world that needs the software that we have.

1:19:47You could never sell this as a SaaS company, but for us, it makes sense to hire somebody who can build this because we built, you'll see, we run the show off of this so the whole team can follow along where we are in the show. No other, maybe a handful of companies in the world could kind of get value out of this, but it's so purpose-built for what we need. We also built an ad platform that tracks everything for our partners. And again, there's not really a market for that. And so I think you'll see a lot of companies that are like, okay, we're not a software company, but we can just build our own software.

1:20:22In categories, we still use a lot of SaaS, right? And we get a lot of value out of it. And we're not going to build our own version of linear, right? Or CRM. No, no. Maybe, but I don't. I think it's going to be much more purpose-built for very specific niches. And then we'll still be pulling the products that exist. It's like we're only using software for completely net new problems that are low-tam. Well, or just something that didn't quite fit your mold. Like if you had Salesforce and you're like, oh, we customized the crap out of it. It's still going to be customized Salesforce. It won't be your own thing.

1:20:57Sure, sure, sure. And now you have the ability just to build your own thing. Yeah, yeah. And in like a day. Yeah. Which is great. Yeah, the thing that we've seen with the state of Vibe coding today is like you can build the V1 extremely fast. Then the actual maintenance is like a real. ends up being like if it takes you 10 hours to build it, it might take you hundreds of hours over the next year. I think that problem is going to be solved. I had interviewed the CEO of Vercel recently. Yeah, Graham was great. And he was talking about this idea of vibe check, which is just like an agent that is deployed to check your code and actually go in and fix bugs and actually get it to scale.

1:21:35And the beautiful thing about engineering is these are all they're not subjective realities. They all have problems that are defined that have an answer. The bug should go away. The code should scale. So we can solve this. It's not like finding the next cancer drug or something a little bit more obscure. If VCs are screwed, maybe not today, but in the future, where are you most excited to invest? And what do you think the role today and what do you think the role of a VC actually looks like in a decade? Yeah. I mean, I wear two hats. I'm a partner over True Ventures. We're managing a little over$4 billion, so we have a lot of companies.

1:22:13I think VCs are very much needed. What can you imagine? Yeah, exactly. But here we go. Wow, all right. I got one. There it is. $4 billion. I love it. I was wondering if I was going to get one today. Of course. This is fantastic. Thank you. It just takes one big number. One big number, so$4 billion. But the companies that need our funding more than anything else right now are hardware companies. Because they actually have to bring a product to market. They have the tooling. They have all the prototypes. There's a lot to put. And we did Ring and Fitbit and Peloton and all these great companies. And we saw it was hundreds of millions of dollars to get these companies off the ground and to scale.

1:22:53Now, on the software side, it's different. Because we can see VibeCoders creating these things, getting to their first 100 ,000 or 500 ,000 users, never raising any capital. And getting revenue. Getting revenue. And now guess what? That valuation jumped up from a$10 million pre to a 50 to a 75 or whatever it may be, or they don't raise capital at all, which is great. Have that lifestyle business. Bring in 20 million in ARR. We need an SF mayoral candidate that is focused on freezing seed valuations. Like the Zoron. Like rent freezing, but for valuations. Yeah, yeah, the Zoron of SF. It just says like the first round you raise has to be at a 10 cap or less.

1:23:32On the hardware side, how are you thinking about AI wearables, AI devices? We're in this kind of primordial explosion, lots of big players working on stuff, lots of existing players. I mean, Apple has a massive portfolio, but what do you think? AI wearables are insanely creepy. I think if you feel like you should punch someone in the face for having something on, you probably shouldn't invest in that company. The idea that something is always listening 24-7 just breaks down so many social contracts that we have in place today with other humans. there are some that I have seen that have not launched yet that are really trying to figure out how to navigate that space in a thoughtful way that preserves privacy, but also gives you the extended functionality of having a second set of ears for what you're doing.

1:24:16And I think those are yet to launch, and they're going to be really cool. Yeah, it's interesting that a lot of the, or at least some, or at least one, AI wearable is saying it has to be always on. You actually can't turn it off because there's no friendship that you have in life outside of yourself that is like an always on friendship right like i don't have perfect information yeah parity with my wife and that's part of like our relationship catch up at the end of the day and that's okay like there's nothing that says it has to be on like it is a very clear product decision to say and i think some of the ones coming down the pipeline will be like you can tap you can prompt you know trigger it to come on and interact with it and then turn it off so that you're not violating this.

1:24:59What about less of the wearable space and more of just traditional consumer electronics that could either add AI features on and actually see acceleration versus completely new formats? Are you seeing pictures for picture frames or the next generation ring that you wear or something that's not necessarily this wearable that's always on talking to you, but just there's a new problem that you can solve, but you need a hardware to instantiate it. Yeah, absolutely. I think we're in the early innings of this, especially when it comes to personalized health and a bunch of other things. I was on the board of Aura for several years, and I was there.

1:25:40Incredible business, too. Yeah, it's been wild, the numbers. Fantastic. Yeah. I've been insanely thrilled at the growth that company has seen. That was such a good example of a business that was like, from my view was like focused on biohackers, which is like a small Reddit audience type thing. It was like I was around in college, you know, looking up like what obscure things can I take that aren't illegal that will make me crush it on this test. Yeah, there was a period where they were just working with like small health podcasts. This is what I did there. So basically my job with Harpreet, the CEO at the time, was to go in and he said, go find me all the different biohackers.

1:26:19So Peter Tia was on there. Ben Greenfield. Yeah, Ben Greenfield, Tim Ferriss, Dave. All of them, we would go and figure out, can they invest? Do they want to be advisors? Can we get them a ring? And this was making that jump from Gen 1 to Gen 2 hardware and made it a lot smaller. And then we worked with Matt Walker from the Berkeley Sleep Lab to get his algorithms on there to really have the best in class sleep algorithms. And that was a huge game changer. But a lot of AI functionality will be coming to these devices to better understand things like blood pressure and a whole slew of other things that we'll eventually be able to plug in, including your genetic polymorphisms that come in from your genome, to really understand who you are as a unique individual versus just a blanket for everyone.

1:26:59Yep. Rank these three buckets that we've been talking to. We talked to early stage founders who are Greenfield Project, and they have AI tools and they can build a company that's AI native, versus growth stage companies. Maybe they're at a billion dollar unicorn already. They're still in founder mode. Might be a little tired, but AI is re-energizing them, and they're able to layer on AI. They're not like, what is this AI thing? They're aware. They use ChatGPT the day it came out. And then there are the public companies. Maybe they have a professional CEO, and they have to go through a real business model transition.

1:27:34How do you see each bucket faring? Where are you most excited? It really comes down to the leadership of the senior level. If your C-suite has a lightweight, I like ChatGPT vibe, You're not deep enough, right? And so it's, but I, you know, what happened with all these tech companies where they are really reducing staff to come in and say, okay, this next generation is less heads, more specialized, and the more of the orchestrator of the AI versus just these massive payrolls and we can do more with less. And I'm thinking of seeing that with startups now where the products that they're building and shipping, I met with a startup the other day where they judge themselves based on what they call tokens in flight.

1:28:18So they set their AI in like YOLO mode, and then they go to launch and they're like, how many tokens can we put in a flight while we're sitting, where we're having lunch? And then we come back and look at the results. Okay. Yeah, they just fire off all the agents. This is the new metrics for how they judge productivity. It's like how many of these they can fire up, how many chat windows can they fire up at a time to go code on their behalf? We do that. We've done over a thousand interviews this year. Obviously, some were more prepped than others. We've definitely been like, all right, I got 15 minutes.

1:28:44Let me fire up a deep research while I get dressed. That's great. Well, it's 1230. We have to move on and talk to you. This was super fun. Anything that we missed that you're excited about or wanted to mention while you're here? No, I think this next six to eight months, especially with Gemini 3 coming out when that drops, and a couple of the models I hear that are being floated around and hinted about coming up, We are, engineering is, unfortunately, for the last 20 years when people say, where do I go? What do I study? A young person would come up to me and be like, what's the future look like?

1:29:20I always say to them, CS, like computer science, no doubt in my mind. That's not the case anymore. You don't think so? No, it's just not. Coding is a solved problem. We just don't know it yet. And it will be in the next few months. And I know everyone's like, bug, scale. Yes, yes, yes. Those are great problems to have. But is it not still worth studying computers if you love computers? I think you should have a light technical understanding of what's possible to know where to bend and break and sand down the rough edges. You're going to need that. But I think it's a different course altogether.

1:29:51A lot more focus on creativity than it is actually in your jobs. And the structure of what is the business problem that you're solving that is important. I still lean that the best vibe coder is a coder, is a programmer. Someone with deep expertise will be able to. I think that's going to be a designer in the future, though. But yes, maybe it's a designer in the future, for sure. Certainly depends on what you're building. Well, very exciting. And thank you so much for coming by. Thanks for having me. Come by again soon. It's been an honor. Thanks so much. We will talk to you soon. And we will move on to our prerecorded interview with Jeff from Hyperliquid.

1:30:25Jeff is one of the most impressive entrepreneurs in the world right now. I think there will be books written about Jeff. It's a fascinating company. And I'm excited for you guys to hear this interview. we tried to get some of the history of Hyperliquid, how it got started, why it got started. And then also towards the end of the interview, we get into a bunch of different questions that we gathered from some of our more on-chain friends to get an idea of where Hyperliquid is going in the next six to 12 months. So before we play this, let me tell you about numeralhq.com, sales tax on autopilot, spend less than five minutes per month on sales tax compliance.

1:31:04and I will kick it over to the production team to play our interview with Jeff from Hyperliquid. Welcome to the show. Thank you so much for joining us. Our audience here, Jordi has a soundboard, by the way, so do not let that, you know, Go to Lert. Go to Lert. Yes, yes, yes. But I would love to get the introduction from you for our audience, which is tech native, but maybe a little bit less crypto native. And so if you could just kind of break down a little bit of like how you're describing yourself, your journey, your business at a high level, and then we can go into a bunch of different directions.

1:31:40Sure. So maybe starting, it might make sense to start with a bit of an origin story. So we'd love that. Yeah. So I guess mid 2022, we were a small team doing trading in crypto and we're looking at DeFi and CeFi and had already kind of a sense that we wanted to build something in DeFi because of the very nascent nature of the product at that time. Basically, all the products were kind of bad and we were traders and we felt like we could make it better. And basically the sort of impetus for going all in and building Hyperliquid was when FTX collapsed. It felt like suddenly the previously very academic concerns in crypto around decentralization and self-custody, it felt like people wrote about it but didn't really care.

1:32:26And all of a sudden it was very viscerally important uh you know it's like not your keys not your coins and uh we had strayed very far from the original ethos of satoshi and bitcoin and so um yeah we thought we thought like the world was ready for to really like trade trade crypto in the way it it's meant to be which is like peer-to-peer in a self-custodial decentralized place and so i guess fast forward to this day hyperliquid is i think the primary on-chain venue for price discovery um it's it's a fully on-chain financial system and so our kind of motto always is we want to build something that can ultimately house all finance and so happy to go into that more later but um yeah it's a blockchain it's not just not just an exchange but the it's best known as uh the place where people trade perpetuals on-chain it does more than a billion dollars in revenue a year and it kind of has been the first in many ways and so happy to also get into like what's interesting about it It's fascinating to hear you saying it.

1:33:27Most founders that come on say we, but it's clear that you see yourself as more of a custodian of this thing that you're building and releasing. Is that the correct frame of mind? Is that how you think about this? yeah so i think we we took a lot of inspiration from satoshi i think he or you know that whoever built bitcoin they uh not sure but satoshi we'll just say satoshi uh yeah was very unique i think bitcoin was the first in so many ways but one of one of the main ways in which it was the first was that it was it's not a it's not a product like you said it's not it's not a top-down company yeah it's a potom and i think defying crypto in general like a lot of it's kind of it was inspired obviously everything comes from bitcoin and there will never be another big one but i think there's been a lot of kind of like top down like web to totally so standard tech approaches in crypto like if you look at centralized exchanges that's a great example they're very good businesses but they're like mega corporations they're they're not very like crypto native in that sense and so we we take a lot of inspiration from from satoshi and so we Hyperliquid really is what we think like credibly neutral protocol, which will ultimately be the rails that all the finance kind of upgrades their tech stack to use.

1:34:41Yeah. So no VCs is at least the lore. Can you unpack that? Is that just because like you're so capital efficient, you don't need to raise? Or is there some sort of particular philosophy around the role of venture capital in crypto? How are you thinking about that equation? So it really does go back to Satoshi again. I think Bitcoin would not be Bitcoin if he had raised Series A. Even if best investors in the world, you know, sickest cap table ever, still I would say that Bitcoin would not be Bitcoin. And so I think VCs actually provide a really important service to the world. They allocate capital efficiently.

1:35:25They help a lot of things. Sometimes. Not always. Yeah, as full, I think. they've done more good than harm. I would definitely agree with that. And in this particular case, I think when the thing being built is a neutral protocol on which something as important as money ought to live and transact, I think the neutrality is more important than everything. And I think there is a path dependence to that. If there are insiders from day one, it doesn't matter how much dispersion of supply, mindshare, talent, etc. there is the the sort of like the big bang moment will always be there and will always sort of like a scar on the record of the thing and like this doesn't this doesn't matter for almost anything that is being built and i think raising capital and hyperscaling is is a very valuable strategy but i think when you're building something that needs to be incredibly neutral in the long term and the history matters i think we'd rather go slow and do it right then but i want to stay on this like uh like you can't keep vcs from just building a position in a public token, right?

1:36:28And so is it more about like the vibe around like insiders? Because what if like a big tier one firm was just like, we want a 20 % stake and we're going to buy the token on the free market and build a position and it's going to feel and perform just like any other asset in our portfolio. Has that happened? Would that be irrational? Is there something where like legally they can't do it because of the fund structure like be very expensive yeah but i mean vcs don't shy away from expensive deals like we see it all the time yeah but you can make that argument for bitcoin as well right like they're very big holders of bitcoin and i think a lot of vcs were very early maybe not a lot but i know i know several vcs were very early totally yeah point and a lot of it so i don't think it's it's not about who owns it it's about the the genesis and the origin it's more of a principle yeah it's like it's you can't you can't on the one hand preach that this is going to be a platform that is neutral that like anyone can come and build on and on another hand say like oh but like these people had a chance to build that first yep um it's not perfect there will never be another bitcoin like sure there's a practical matter to it which is that you know like anything after bitcoin needs to launch in a like very competitive market there needs to be innovation and something needs to fund that etc but to the extent that we can approximate the the ideals of of satoshi i think it's still worth striving for no no how much did you know exactly what you wanted to build and then just execute it against that plan versus iterate to get where you are today um i would say very little so the story there is we just wanted to come and do something that we you could do well.

1:38:15And even doing one thing well is very, very hard. So we were laser focused at first. We just kind of looked around and thought, we weren't afraid to dream big, but we were kind of practical about it. We just thought like, okay, what's one really, really big thing in crypto that we think could benefit from a fully permissionless platform? And purpose trading was the obvious one. I think it was probably doing more than 50 % of the revenue in all of crypto back then, sort of back of the envelope calculations. And so we tackled that initially. And I think one big thing is we weren't willing to compromise even at the very beginning on how the thing ought to work.

1:38:55So you learned the things not to do from FTX. You were inspired by Satoshi. Was there any company off-chain that was particularly inspiring? yeah a bunch i mean i think every big company every big tech company i think is like a very inspirational to me uh i grew up in the bay area so it's it's hard to not kind of be sort of motivated by that i think amazon was a really big one that what inspired me was that they basically like they're their very first principles and like driven but also and also are very practical and I think when sort of I assume it was Bezos but maybe someone else in the org realized that they had built so much of the internet stack that they it would be a shame if they if that just went to this like retail business like why not sort of abstract it a bit and create the right API so that anyone can leverage this like amazing infrastructure they've built and you know that kind of being the birth of AWS and cloud computing I think that's such an amazing story.

1:40:02and that's inspired i think like that's kind of like how we think of hyperliquid was that like it started as you know a blockchain optimized just to be able to do on-chain for purpose trading because no other infrastructure could do that yeah and sort of realizing at some point that a lot of other things in finance ultimately all of finance we think can benefit from this high performance decentralized ledger and yeah yeah so i think liquidity infrastructure sure like for liquidity is one way to think about like what do you think about Bezos's backs uh background at DE Shaw he was a trader uh that probably informs a little bit of the shape of Amazon I imagine versus say Google where they come out of research PhDs uh what what were you working on beforehand do you feel like you brought a trader mindset into the you know entrepreneurial journey that you went on yeah not just the product layer yeah yeah i think so i think trading well i don't really know what kind of trading bezos was doing but at least with more like automated trading i think there's a lot of it feels a bit like doing physics kind of it's like you do a lot of approximations you don't you can't get anything exactly right because markets are are basically there's a lot of randomness right there's more there's more noise than there is signal and that that's kind of what makes it beautiful it's like trying to sift through the the noise and it's very different from like supervised learning setups in ai where you kind of have like roughly infinite data and it's very high quality so i do think the real world is like building hyperliquid has felt kind of like that in the sense that you you often it feels like we we're actually like not very data driven at all it's basically all intuition it's just like we just think really hard about how the world should be and we just try to do the best thing that will obviously like adapt to data like if something happens and it's like obviously bad like we will we aren't like delusional and we'll like incorporate it readily but we won't like go out of our way to try to create data where it doesn't make sense to uh if that makes sense so like a b testing is like something that we just kind of like never do oh yeah that's funny uh you you've mentioned it a few times but uh can we just get uh like a firm backstory definition on perps uh for our audience why what are they why are they so exciting yeah um let's see so for for someone who really has no finance like you want to like a sort of very like from zero explanation yeah yeah that'd be great okay so if you think about what is traded today by people that are kind of like you can trade the like thing itself which is like an amazon stock right yeah um if you want some sort of like leverage which is to say like make more bang for your buck then the two ways people go about doing it is one they trade futures which are these are traded mostly in indices so the big like the s &p 500 and these things basically say like two people put up cash and they just agree to page kind of like um they kind of like a long and a short kind of create a contract and then if s moves by one dollar then the long side say makes twenty dollars and the short side loses twenty dollars and that's a future and then they're they're often also settled in some like underlying things so it's like we're trading on uh the like what the price of s &p is three months from now or like three months from now like you'll deliver one cow.

1:43:39Yeah. Yeah. Yeah. Yeah. I think of it as like the, the C bot, like if I'm trading corn futures and I let the contract run out, like at some point I have to actually take delivery of all the corn, but most of the wall street traders have figured out like to never do that. But you hear about these weird scenarios where it's like, Oh, the price of oil was negative. So someone bought it for negative money and then like wound up having to get all this oil and stuff. Um, But obviously, in a purely financial context, that's not the outcome. But has this just unlocked higher frequency trading, more leverage, a different shape of trader, something more quantitative, something more algorithmic driven?

1:44:19Like, who are the customers or who are the traders and why are they excited about the product? Yeah, so they're excited because – so basically, like you said, like you just kind of sucked because of all these like random things. They kind of settle. Sometimes you don't want to get delivery. They're like all these weird things. You have to keep rolling your position if you want it open. So if you look at Robinhood, for example, that's actually much more popular with retail users than futures are. And they primarily trade options on Robinhood when leverage is concerned. And options are super cool because it's kind of like a lottery ticket.

1:44:51As retail, you can buy a lottery ticket and feel really good about it. You're downsized limited. But the trade-off there is it's actually really hard to price an option, especially if you're retail and especially if you're using like an app that doesn't give you the information that you need yeah and so you're you're kind of getting fleeced because there's like very complicated um structures it sounds simple it's like a strike price expiry but in practice it's very hard to price and so herps are basically like marrying these two assets into one thing so there's you want to trade something you want you want to trade something that's just like the price of the underlying you want there to be leverage and you want there to be like one thing that never expires and that's what a perp is got it so let's all the liquidity in the world concentrate on one asset so like if you want to if like you look at bitcoin for example on any one exchange there's like usually one very liquid bitcoin perp and that's the liquid asset forever never expires and like that's where the price discovery happens like billions and billions of dollars will trade on these bitcoin perps and this actually leads the underlying instrument and it's useful for professionals because they can trade it and it's the most liquid instrument there is on bitcoin it's useful for retail because it's a clearly understandable price that you cannot get screwed over on because there's only one market and it's extremely liquid so like buy or sell like the spread is tiny and um so it's kind of like a win-win for many many participants there are there will always be participants who want options or who want traditional futures but perps by and large are attractive yeah what what are the biggest uh bottlenecks to scaling a network like this um is certainly not talent because you've yeah insane scale i'm more thinking about like uh are the tokenomics such that there are people that are setting up whole data centers are there asics being built right now to run the network like i i'm familiar with kind of the how bitcoin went where people were just mining it on their laptop and then it turned into a data center and then it turned like hunt for the cheapest energy possible with the ASIC because the algorithm was so stable.

1:46:52Like what does it actually take to scale a network like this over time? And where are you in that scaling curve? Yeah, so Bitcoin's a bit unique in that I think it's the only major network left that does proof of work. So ASIC stuff you're mentioning, you know, like all the crazy things like volcanoes, mining Bitcoins and stuff like that's all really just participating in the consensus mechanism. all the other high performance blockchains today that i know of at least are power right proof of stake which is a much more energy efficient way to do things and it's economic security not like computational security so on hyperliquid there's not much innovation being done on how the networks stay secure it's like the it's a relatively solved problem economically which is that basically people people put up the native token like it's very important for the network to have its own token which on hyperliquid is called hype and people put it up and they basically say i'm like vouching for the usually it's kind of like delegated it's kind of like um you know you vote for your congressman kind of thing so it's like you delegate it to a validator and the validator says anything i do the stake that is staked to me is at risk um and you can do some math and then the basically the economics work out such that if like most of the stake in the network is honest then it's fine and if you want to acquire enough stake to do something bad on the network where bad here is like the canonical bad thing you can do is like spend the same dollar twice it's called like double spending yep which is basically like forking the the truth state of the network um you need to acquire a lot of stake to do that so it's economically in so it's the the securities in the economics, not in like, you know, creating crazy basics and like, yeah, that makes sense.

1:48:40How important is brand to hyper liquid success? Because I think you have, uh, I don't know that you've, uh, you haven't necessarily focused on brand in the traditional sense of like hiring advertising agencies and, you know, putting together strategy decks and things like that, that, that, uh, I'm sure many of your competitors do, um, but you have one of the most powerful brands in the world, which is like when you think of the value of anyone on earth being able to just type hyperliquid and hit post and get this like, you know, massive influx of excitement and attention. It's truly remarkable.

1:49:16But I'm curious, like what the, you know, how much that's contributed to your success versus, you know, scale and some of these other product decisions. yeah i think the brand we're very fortunate that the community is so i don't know many adjectives like so so strong so like tight-knit so uh you know combative at times but like just like like i think i think it's very inspiring for us because we we as a team are pretty introverted we're super small we're only 11 people and we don't have it we don't have we working on marketing um and i think we even if we did we would suck at it so it's always been the product is the marketing yeah it's not the it's not just the product it's the product in the community i would say in the ecosystem and so it kind of there's there there there are like many pockets right there are people who are just like on twitter like like you said kind of like shit posting and i think that's like super cool and uh you know i i love it and uh there are also people just like you know building on the platform i think that's another form of like viral marketing They build products on top of the protocol that synergize with what it is, like offer something new or extended in some ways.

1:50:28Those are sort of like pillars of finance on Hyperliquid are by and large built by community members. And we hope that that trend continues. Anything that can be built by the community are built by the community. And I think that just like, yeah, that kind of decentralized marketing sort of embodying decentralization, not just at the technical layer, but also at the social layer, I think. is really important to us. I think that's kind of the brand that Hyperliquid has. It's like, you know, framed negatively, you could say that it's kind of standoffish and maybe we're kind of, we're too focused on tech and not focused on marketing.

1:51:03But in this case, I think it, what comes out of it is like something much stronger, which is that people feel ownership in the network in a way that's not possible with a Web2 company. Yeah, is that part of why, you know, the companies that want to eat your lunch are some of the most well-capitalized, large businesses in finance, and yet it seems like they're struggling to really compete. Is the army of people that just want Hyperliquid to win, how do you think about that kind of advantage and what are some of the other reasons why your guys' kind of counter-positioning makes it difficult to compete?

1:51:48uh well we honestly don't think that much about competition day to day i think there's just like too much to build and i think you're right yeah a lot of a lot of people want to do what hyper liquid does or like take market share if you will um and we don't yeah we don't really think about what you're like i don't think about you at all no no there's just too much there really is just too much to build. I think if Hyperliquid succeeds, in the good case, it's something that doesn't exist yet in the world. And I think it's hard. It's easy to get bogged down by people trying to chip away at something that Hyperliquid has already built or whatever.

1:52:29And I think that's like, it's a bit stressful sometimes, but also it kind of detracts from the big picture, which is like we're still so far away from where we need to be. How big can Hyperliquid get? Like what's your ambition? I think of it as, in the good case, basically finance as a whole, which when I say finance, I mean the coordination of human behavior. That's amazing. It's not like it. You're basically saying if I don't assume 100 % of the global TAM of finance, I will have failed. That's amazing. Well, no, I mean, you asked how big I could get. Yeah, yeah, yeah. I'm just playing. I like it.

1:53:09That's true. yeah yeah but i i don't think it's like it's coming in and kind of like displacing finance it's really like my mind is sort of like the internet did to find sort of like electronic trading happened in early 2000s it's been like more than 20 years it's about time the tech stack gets updated and i think defy is kind of yeah it's that tech stack okay i have a lightning round for you uh asked a bunch of uh crypto native friends uh what they wanted to ask you so i'll go through a bunch of them and hopefully you can answer quickly since some of them are pretty specific when do you expect to see the first centralized exchange shut down their perp decks and simply run a front end on top of hype through hip three um when will they when will they capitulate one year okay okay we'll track it uh any specific views on the market structure bill in the u.s and under what circumstances would you bring hyperliquid onshore?

1:54:10That's a good question. So we think the U.S. is a super important market, obviously. It's like the financial center of the world and the dollar is the reserve currency of the world. Capital of capital. Yeah, yeah. I mean, we would love nothing more than to sort of like have regulation in the U.S. sort of evolve to sort of really embrace DeFi and I think strides are being made in that direction. I can't comment on the bill specifically just because I feel like I'm a little ignorant on it. I feel like it's changing a lot. And honestly, I think there are really smart people working on it and the things I've heard such as carve-outs for decentralized front ends and things like that I think are really cool.

1:54:47Why do some centralized exchanges say that hype doesn't want to get listed?

1:54:56Do they say that? I don't know. I don't know. We don't have a want here. I think it's like we're building more building. I think several exchanges have listed hype, and I think it's cool. And for other exchanges, it doesn't align with their priorities, and I think that's also cool. Yeah, we're very like, we just don't really talk to these institutions. Any plans to enable multi-asset margin natively?

1:55:25I think it will happen and I don't exactly know the path to where that happens so it could be built natively is a bit not well defined here. But I think it will happen. I think if it's going to house all the finance, surely you should be able to use different types of collateral to trade. When perps on commodities and US equities? There are already some perps on. So there are gold perps, tokenized gold. And there... I think the answer to when is probably in the next year. Because HIP3, I think, is unlocking a lot of this. So basically letting it lets anyone kind of come and deploy their own. What do you think is most misunderstood about Hyperliquid today?

1:56:10Oh man, there's so many things. Most misunderstood. One thing I would say is that it's, well, it is a network. I think people think it's like a purpose exchange. Some people think it's like a centralized thing. It's really not. It's a blockchain. The validator says permissionless. They're currently 24 validators. Anyone can spin one up. It's like top 24 by stake. The validator set, every validator executes every transaction, including all the orders. So this kind of base layer of just what it is, I think, is kind of lost because people are so... They're focused on the product, basically, which maybe is good.

1:56:45Maybe it's nice that the tech is abstracted from users. Do you miss the Bay?

1:56:53I miss some things about it. What do you miss?

1:57:01What do I miss? I miss Chick-fil-A. Oh, let's go. That's a great one. That's a great answer. That's a great answer. I love it. I miss mountains, actually. I mean, that's not literally in the Bay, but Singapore is very flat. So it's a very nice place. It's a great place to build. But I miss kind of like Yosemite and that kind of vibe. When I was in Singapore, a lot of people I interacted with said that Singapore doesn't have like a strong local culture and maybe that's because it's small or maybe that's uh because it doesn't have like an underground like punk scene because of different rules and regulations do you feel like you found uh like a great community in singapore do you feel like you found a great culture there do you like the where the where the country is going like culturally uh i'm honestly a bit too busy to just lock in okay no yeah it's a great place i think it's super safe yeah it's very modern yep you know good food yep air conditioning like that kind of stuff so uh i'm honestly not a cult at this point in my uh trajectory i'm not really attuned to the culture i felt like if i were new york city it would be kind of wasted on me totally totally Yeah, yeah, yeah.

1:58:19You're not looking for like the underground art collective that will define the next. Yeah, you're locked in. It's great. Love it. What is the specific framework you use for building the team? Because I imagine the level of talent that is trying to join, you know, the core team today is insane. But clearly, you have some kind of framework for it. Otherwise, you would have just been, you know, 100 ,000 people by now. well I think we actually are like just not very good at recruiting so I will say that here we are recruiting, we're always recruiting I know there are a lot of super smart listeners on this podcast show and we would love for you guys if someone's interested in systems engineering high performance building the rails of the future of finance I really think there's no better place to work and would love for you guys to join.

1:59:20We're a super small team of 11 people. The bar is very high. I do think it might be one of the highest bars. I think we are one of the most efficient engineering organizations that I know of. That has ever existed? Did I say that? I feel like I can say that. I think you might be right. Yep, it sounds right. But we are trying to grow And I think there's a lot to build. And yeah, we just have a very high bar in many directions, like competence, also just like integrity. I think we just care a lot about many things. And if someone listening meets all those bars, we'd love for you to join. Is it in person only, or do you guys have remote teammates?

2:00:16uh new for for new recruits we're we're trying to do in person only um there yeah initially it was it was remote when we found that it doesn't work super well with our work style last question are people that uh that brag about doing 996 soft well this is the like nine to nine six days a week Yeah. Yeah. Because I imagine you're doing quite a bit more than that. So that feels like part of the criteria is like somebody comes in, they're like, Jeff, I'm your guy. I'm 996. And you're like, sorry. That's not. We don't do half days here. Yeah. I mean, I do think the quality of work is the most important thing.

2:01:04So I think different people burn out at different points. And that's obviously the most important. I personally work more than that, but it's because I feel like I don't really have a cap personally, but everyone's. Yeah. Amazing. It makes sense. Well, thank you so much for taking the time. Really enjoyed this. Thank you for sharing the origin, explaining perps for our audience. And yeah, really just incredibly impressed with what you've built and excited to follow, you know, continue to follow your journey over the next years and decades. Congratulations. We'll talk to you soon. really appreciate it thanks for having me on have a good bye and that concludes our recorded interview with jeff from hyperliquid i hope you enjoyed that uh if you're new here please subscribe to the show leave us five stars on apple podcast spotify uh we have more crypto coverage later in the show we're talking to chad in one word yes jeff jeff for sure uh also fin.ai the number one ai agent for customer service number one in performance benchmarks number one in Competitive Bake Off, number one in ranking on G2.

2:02:06Our next guest is already in the Restream waiting room. We're going to bring him into the TBPN Ultra Dome. Tamash, how are you doing? Good to see you again. It's been too long. Pleasure to be back. Thanks so much for hopping back on. You've been lighting up the internet with a bunch of hot-ish takes. A lot of takes I agree with, but mostly just very thoroughly researched takes. Not something you see a lot. Not something you see every day. So I'm very excited to have you on the show. So I would love to just kick it off with taking your temperature on where you think it's the most important to focus right now.

2:02:42What is the narrative? Is it these complicated structures and the vendor financing? Is it the overall bubble narrative broadly, progress on token generation or DAUs? Where are you spending your time focusing? Yeah, we're trying to understand what's really happening. We wrote this blog post about NVIDIA and$110 billion of vendor financing. I ran the numbers for an event. The data center build-out is now greater than 1 % of US GDP. We had made this prediction in 2023 that AI would contribute more than 1 % or contribute at least 1 % to GDP. I was stunned that we're basically already there. We had this question of how similar, investors will ask us, how similar is the current environment to say 2000 when you had pretty large network infrastructure build outs.

2:03:32And I was kind of in high school during the time, but I remember these companies like Lucent and Nortel and I'm going to the moon and then collapsing. And so we ran this analysis. And so we're paying attention to a lot of different things. We're paying attention, like on the bull side, we're paying attention to token generation, right? Google has released now three data points and we can start to track it. The growth rate there is slowing. Is the growth rate slowing because user demand is slowing. I don't think that's the case, but maybe there's far pretty significant improvements in overall efficiency.

2:04:04The other thing that we're paying a lot of attention to is just understanding the debt structures, the use of SPVs on and off balance sheet vendor financing. And then the last is the depreciation schedule. So a lot of these data center companies have extended the depreciation schedules, almost doubling it. Amazon went from three years to six years and then backed off to five earlier in 2025. And that's important because it's the collateral that's underwriting these loans. Yeah. Let's start with the vendor financing thing. It is notable that everyone was so hopeful that AI would increase GDP growth.

2:04:39And it did. And it did. It came at the cost of hundreds of billions of CapEx. Yeah. We'll take it. I want to start with the vendor financing because that's ringing a lot of alarm bells. A lot of people are going to loosen story. I took it in a different direction. And I was wondering if you're familiar with the ASML customer co-investment program. because in 2012, TSMC, Intel, and Samsung, they pitched in 6.8 billion across R &D funding and equity purchases in order to help ASML pull forward EUV lithography and 450 millimeter wafer technology. And it was this like odd round trip, but it worked out.

2:05:21And so when somebody throws me like the vendor financing is immediately red flag, I always want to say like, well, it doesn't always end in disaster. And so how much nuance should we be placing on the vendor financing stuff? How much do we need to be digging in to actually understand what's at risk here? So vendor financing itself is really common, right? You pointed to this ASML example. It's a wonderful thing within the ecosystem. And the vendor financing fails when you have a closed system and they're all borrowing and lending from each other and there's no net new GDP coming into the system, right?

2:06:00and then it just goes around and around. Everybody takes their tax and GDP goes to zero, which is kind of what happened. That's not the case. You look at the revenue growth of OpenAI. You look at the enterprise spending that's happening. There's GDP that's coming from labor spend. There's GDP that's coming from BPO operations that have been executed in foreign countries that are now basically being re-onshored. And so I think it would be too simplistic just to say, vendor financing, uh-oh, I'm calling the top. That's not the case at all. It's just the expectation is around$500 billion a year in CapEx and the return on invested capital and the depreciation schedule.

2:06:44I think the bigger question is, how would the bondholders make out on all this? right? Yeah. Where does the risk actually live? Because I feel like there's this world where you have open AI is kind of rolling a 20 sided dice. And as long as they don't come up with a one, they're going to be fine. Well, everyone else is flipping a coin and it's 50 50 whether they make it out. Okay. And, and everyone's, everyone's playing the same risk game, but the risk is just way higher for some of the folks who are further out in the debt curve in the capital stack, more risk on, more debt-laden. And a lot of the narrative is like, OpenAI is taking on all this debt.

2:07:23It doesn't actually feel like they're taking on that much debt. It feels like their partners might be. Yeah, that's exactly right. So I don't think OpenAI is really taking a lot of balance sheet risk. I think it's the lenders who are taking the balance sheet risk on. And so why is there some reason to be concerned? Google ran an analysis on their GPUs and their data centers, they found failure rates at 50 % to 70 % utilization were significantly higher after three years, which is half of the amortization or depreciation schedule. Then you have next generation chip architectures. SGI just announced yesterday, Cornell research that showed 90 % improvement relative to GPUs with the new chip architecture, and then a second generation that's promising 100x performance improvement from there.

2:08:06And so fundamentally underpinning all this is how fast are tokens growing, which is an inflationary force in the ecosystem. And then the deflationary force are algorithmic improvements and chip improvements. And so what is the net, what is the vector that comes out of that? What is the slope? We don't know the answer, but it seems like both are growing very, very fast. And so it's hard to predict. Do you think there's much more confidence on the inside? Or do you think that internally, some of the deal makers are actually viewing, hey, this is a leap of faith, but we're all taking a leap of faith?

2:08:41Or do you think that when they hear us yapping about them on a podcast, they think these guys don't know the data. They don't know how solid of a bet this is. I bet if you're in the inside of one of these companies, I mean, Google and Microsoft both said in separate earnings announcements this year that they were hardware constrained. And you just watch the consumer adoption. I think OpenAI will reach like a billion an MAU here faster than anything. And so I think if you're on the inside, there's very little reason to worry because you see the ultimate demand. We were, you know, the first wave of all this AI was just better search, compressing all of human knowledge into an LLM.

2:09:18And now we're at this tool calling or agentic phase where it'll actually start to do work for us. And it works pretty well some of the time. But if we can improve that meaningfully, then the demand for tokens will go through the roof. I have another question. Go for it. I'm trying to understand how people feel about the ramp of agentic commerce. Because in the limit, if I think about a billion MAUs who people are purchasing things, they're using ChatGPT regularly, they're just going to wind up buying things and taking even a 1 % cut of that commerce through an affiliate revenue or a Stripe-like tax or some sort of ad auction to see, well, will you buy it on Walmart or Etsy or Amazon or somewhere else?

2:10:07Like just, just even, even not just surfacing like a direct advertisement for something you're not shopping for. Um, that feels like that could be very lucrative on an ARPU basis. Uh, but I'm wondering how long it will take to that, for that pattern to actually manifest if people aren't in the habit of searching because over the past two years, it's been, yeah, maybe you'll chat GPT something, but then you have the natural flow of like, oh, then I go to Google to order it, or then I open Amazon to order it. So how do you think about that ramp? Well, super exciting time. You've had the ad market historically dominate over the last 15 years by Google and Meta, right?

2:10:44$250 billion on search, about$265 billion on social. And so it's been really hard, candidly, to invest in the advertising ecosystem, but now it's wide open. I think the agentic commerce will have pretty significant tailwinds here because you are in the flow. You are researching. I mean, I don't know about you guys, but the number of websites that I visit compared to, say, five years ago has fallen off a cliff. Sure. Because instead of going through all the forums to figure out what is the best bike carrier for the family, I just ask and then create some of the Amazon link and I'll buy it. Yep.

2:11:17So maybe there's an ads model here. We had contemplated whether OpenAI or others would run a keyword auction to inject ads into the context window. That's the additional information in addition to the search query. One of the crazy theories that I've been wondering about is I was chatting with a friend who suggested, what if OpenAI takes a 30 % cut? Not a 1 % cut, but what if it looks like the Apple apps? If you ask a retail brand, how much of your revenue does Meta take? It's probably like 30%. Yeah. Yeah. And if the performance is better than merchants and retailers will pay it. Yeah. How do you think about a direct agentic commerce affiliate?

2:11:59Like I'm searching for the best bike rack. It gives me a bunch of options. And then at the last second, it runs an auction to see what provider will actually send that to me. And they take a cut there versus I'm searching for history of the Roman empire, but it knows that I'm interested in shopping for a bike. And so in the middle of that feed while I'm reading a deep research report about Rome, it says, hey, do you want to check out with that bike you were looking at earlier? Because the Instagram flow, yes, there are people that search for camping gear and they scroll and there might be an ad that matches up.

2:12:32But a lot of times you're just looking at family photos from friends and it says, oh, here's the thing that we know you want and you might buy it. Right. So those are two different kinds of ads, right? So ads, let's say, within the search, and then targeting an ad to you about a bike rack when you're reading about the Roman Empire is called retargeting. And it's typically done from search, hugely successful advertising program. I think both exist. And it depends on how considered the purchase is. So everybody has a different level of willingness to spend before they, say, consult their spouse or take some time to think about it.

2:13:08And so you have these impulse buys, and then you have, I think retargeting works very, very well for considered transactions, like a car. If you're really interested in the next electric car coming from Rivian, well, the ad at that point is actually pretty useful for you. And one of the things that I learned in the advertising business was seven impressions was kind of the key to building a brand. So there is value in both the brand building component, the retargeting component, as well as the immediate transaction part of an ad unit. Do you have an idea of, I mean, if we're just to play out sort of like a median case for like a bubble popping, you might map OpenAI to Google.

2:13:47You might map Anthropic to Amazon, companies that made it through the dot-com bubble. Maybe there's some, you know, overbuilding that happens and you get a lucent and you get some big companies left in the wreckage. But what is advice to, you know, if you're talking to a dot-com entrepreneur who raised, you know,$50 million, doesn't quite have product market fit, it's 1999, you now know that the bubble's about to pop. What advice are you giving a founder that, hey, it's raining right now, everything's really great, but it might not be that way forever? How are you setting up your business for durability in potentially a rocky time?

2:14:26Great question. First is to bolster the balance sheet. And I know this goes a bit against the grain of don't raise too much capital. But I think what we learned in 21 is the companies that had pretty significant balance sheets were the ones that were able to reinvent themselves and weather the storm either by acquisition or new product. And so I think the cost of capital is incredibly low. So you should be shopping for capital if you can. And then the second thing is I think there is this notion that product market fit is this binary condition that once you pass the gate, you're done. and it's no longer the case.

2:15:00Product market fit is this status that you need to, it's like a United Miles, you need to continue to fly to maintain 1K. And we saw it between 21 and 24, companies that were classic software companies literally overnight lost their product market fit. And that'll be very much the case again. Do you think the labs have an incentive to get various players to overbuild. So then there's like a... Buy a cheaply. Basically like, yeah, they can, they're basically a buyer at any price, right? Because they're GPU constrained now. But if you can collectively get people to overbuild, then two years from now, you might have dramatically lower costs.

2:15:41And I think you're already seeing this in some cases where like actual GPU prices are through the roof, but then the leasing prices are quite a bit lower. But I'm curious if you think that kind of playing out the game theory there. Yeah, I mean, I think if they want long-term relationships with their capital partners, they will probably want to game it a little bit, maybe not too much. They definitely want to see a little bit of excess, and that's okay. You look at the CPU spot versus contract markets, and there's excess CPUs on AWS and Google, and the margin structures there are really good. I think one major question with the GPU market is, within the large hyperscalers, you do see great utilization.

2:16:26And there's a question around the neoclouds of what utilization are they seeing and what the United Economics are there. But overall, I think they probably won't look to burn others just because they need these relationships for 20 to 30 years. Yeah. Oracle is down 17 % in the last 30 days. Do you think that will hinder their ability to actually deliver the capacity that OpenAI is saying, just in the sense of like, you know, they were raising debt and a lot of people are saying like, they're basically properly leveraged. Maybe you shouldn't go much more from here. But I'm curious, like, do they need to go to the 2040 forecast?

2:17:08You know, 2050, just keep moving out. RPO out to the end of the century. Why not? Why not? Yeah, I think, I mean, it's a good business, right? It produces a lot of cash. the information article around the 10 % roughly gross margin structure on some of these contracts, I think, is what catalyzed that drop in price. Yeah, but it did drop dramatically, and then it's just been chugging downhill. Yeah. Yeah, I mean, look, these are big long-term contracts with unclear demands, and the demand is growing really fast. And so I think if you're an investor, a long-term investor, you have to be hedging at some point.

2:17:48Yeah. There's also a lot of like hype that comes out when there's like a new biggest number. Everyone is like, oh, I didn't even know that they were in the game. And now they're at the top of the game with the biggest number. And those people pile in and then they might get, you know, jitters. Like the whole detail of that margin was the nuance there was that, well, of course you pay for the GPUs before you start making money from them because you, that's like any manufacturing process, you buy the raw materials before you make the widgets. but people obviously saw a lot of jitters in that. Totally, yeah.

2:18:21Go for it. I was going to say, I think there's this great parallel. I don't know if you guys read the book, Barbarians at the Gate, which was about the LBO of R.J.R. Nabisco. R.J.R. Nabisco, and it was the largest buyout at the time fueled by the junk bond wave from Milken, right? Milken. And then that kind of, one, it created the LBO acid class in a very real way, but two, it was a big contraction. And so I wonder, will we have a Barbarians at the Gate book written about some of these major data center contracts? I really hope so. There are so many OpenAI books coming out. And I know that they're all just going to focus on the doomers versus the bulls and stuff.

2:18:57The nonprofit versus the poor profit. Yeah. And it's going to be intriguing and stuff. It'll be like social network vibes. But I want an Andrew Sorkin book about this more than the salacious journalist take of OpenAI is bad for whatever reason. business history exactly in the room yeah i i read uh the caesar's palace heist i don't know if you've read that about uh apollo and it's just like the most nitty-gritty on like the debt restructuring that it's another lbo of the casino empire and it's just like super fascinating to me pretty low tam but i hope we get one of those books just about because some of the novel deal structures that happen at open ai every single turn even going back to the the microsoft deal for a billion dollars of credit.

2:19:43We're actually going to reinvent the wheel. They reinvented the wheel 12 times. It's not just the nonprofit and the for-profit. It's way more complicated. Every deal is like units and crazy. And then now these deals, it's all a lot of fun. And then think about the Google ownership in Anthropic and then the Amazon ownership of Anthropic. What does that mean when you have two strategics each owning 10 plus percent of the company? Yeah, that's why when the Google News was hitting the timeline yesterday and that they're working on some type of major deal, I was like, why are you surprised that a double-digit percentage stakeholder is working on a deal with their portfolio company?

2:20:19That is the obvious thing. That's famously bad advice, though. If you're a startup, it's usually like, don't let a strategic build a 10 % position in your company unless you're actually planning to sell from them. Because it will preclude you from working with the other potential strategic buyers or investors. And that just hasn't proven true in this market at all. It's completely different. Did anything about the Carpathia interview update your kind of worldview? I thought it was just very pragmatic and real. And, you know, the decade of agents point was like, I think what everybody's experiencing.

2:20:55That's like trying agents and talking to people. There's some that are great. There's some great coding agents. We have deep research. That's awesome. But yeah, there's clearly going to be more time needed to figure out a lot of these use cases. But I'm curious what stood out to you. Yeah, I think reinforcement learning was kind of the song of the summer. And there are two different approaches. One is human labor creating virtual environments or gyms where AI can train to understand how to update your CRM. and then their standard operating procedure-based, text-based, we call them context databases.

2:21:35Those are two approaches. We've written an article, I think the first one on the topic, about context databases, and so we definitely agree with him there that the current approach of creating these gyms mirrors the fine-tuning of the previous area. And what I mean by that is if you take a regular model and fine-tune it and customize it to a particular task, it's really brittle. by the time the model updates or the tasks update, you have 30 % to 50 % of the prompts breaking, which is what we experienced today in tool calling. And so I think he's largely right. The amount of research that's happening within reinforcement learning, which has two parts.

2:22:11The first is making a plan, and then the second is creating what are called reward functions, or like Mario gets 500 points for eating a mushroom. How many points does a robot get for filling out your CRM? The first one we can do with AI. The second one is largely still research. And the amount of work or just academic brainpower that's going behind there is enormous. So I really hope that we make strides faster than what he said. And I'll tell you, I was trying to replicate Cloud Code myself with open source technology because the plain internet wasn't working very well yesterday. And I realized when you use Cloud Code or Cursor, tool calling is just asking the model one question and then taking the answer and then putting it back into the model and then again and again and again and there are different loops like gemini or gemini cli says do that 10 times and then here are the success cards like that's that's that's tool calling it's extremely basic and there's some management of like okay after the first step this is what we learn now go do this but it's um i thought it was much more sophisticated it's and it's quite like it's a It's very rudimentary.

2:23:20And it's just the way that we've optimized memory around these models and we can do tasks. Well, we can hit different benchmarks for science and math. I think we'll start to see that with RL. Although I think the big question is, do we need a new architecture in order to really solve some of these problems? And maybe that's what Karpathy was alluding to. But we have a long ways to go. Well, thank you so much for coming on the show. This is always a great time. Let's grab more time next time. We'll talk soon. A bunch more questions. But thanks for coming on. Have a good rest of your day. Our next guest is already in the Restream waiting room.

2:23:52But first, let me tell you about Adio Customer Relationship Magic. Adio is the AI native CRM that builds, scales, and grows your company to the next level. We have Sean from Coinbase coming back into the TBP and Ultrodome. Sean, how are you doing? Hey, guys. Good to see you again. Good to see you again, too. Take us through the news. We'd love to go deeper on the acquisition. But also, I just want to get an update on anything else that's happening in the Coinbase universe. yeah well we were super pumped yesterday to announce the acquisition of echo which is the leading on-chain capital formation platform we obviously we had a little bit of fun with it on twitter and meet up on monday even despite all the aws uh oh yeah outages that were giving us uh a little bit of the sweats um and is that the first time an nft has been taken through like a corporate m &a process yeah definitely unprecedented a little bit easier you can imagine the uh trying to explain that to our lawyers and tax consultants when you're buying a company like, wait, you want to also do an NFT thing?

2:24:55So yeah, I think it was a unique N of one thing. It was, you know, obviously very specific to this acquisition just given the founder of Echo is Jordan Fish, Kobe, who, you know, I think he's very much intertwined with crypto lore and up only is very much intertwined with crypto lore. So we just thought it would be a cool creative way to get the crypto community really jazzed up and excited because everybody's been raving about and asking for him to bring up only back. Yeah, it was, I think, like the M &A announcement of the year. It was the best executed, like build a bunch of intrigue, hype, controversy, even in some realms, but then ultimately deliver a really positive story.

2:25:39It was exactly what you want out of like M &A comms, which is sort of the least sexy thing you could possibly do. It's usually really boring, just a number and a couple paragraphs on synergies. And everyone's like, okay, congrats. But this one was really well executed. I want to know more about the vision for, we were talking to Brian Armstrong about this yesterday, the vision for the next generation of crypto companies. They raise money on chain with Echo. What does their world look like over the next few years as they scale in the Coinbase ecosystem? What products are they pulling off the shelf?

2:26:14What are they taking advantage of? What does the whole, like being a Coinbase native business feel like? Yeah, good question. So, you know, Coinbase historically has been primarily a secondary exchange for when you list a token, that's where price discovery happens. And that's where most of the trading activity is. But by the time an issuer actually issues that token, it's actually pretty late in the lifecycle. So with Echo and then another company that we acquired over the summer called Liquify, which is sort of like a, you could think about like a Carta for crypto companies that helps you manage your token table.

2:26:52It allows us to have a relationship with the token issuer basically from inception all the way through those private fundraisers and potentially a public fundraise to listing on the exchange, which we think about as sort of like the crypto IPO in some ways. And then after the fact, we continue to serve that issuer through a business account that they can use to manage just sort of the operations of their business. They can grow using Spindle, which was another acquisition that we did earlier this year, which is sort of an on-chain ads protocol that helps you find distribution. And I think our broader vision for it...

2:27:31GM. Yeah, exactly. GM is the legend. Yeah. Yes. That's great. Yeah. So we're pulling in all of the talent and all the legends into Coinbase. Yeah. Brian was talking a little bit about bringing access to earlier stage investing to a broader community. Obviously, there's a bunch of regulation in the mix. I don't really care about the legal understanding. I want to know more about the philosophical underpinnings of who gets exposure to what. There's a take that's like, everyone should be able to invest in anything. There's also like, even though you have to be an accredited investor to invest in a startup, you don't have to be an accredited investor to have money in a pension fund that winds up in a venture capital firm that gave money to a company, to a VC firm.

2:28:18I think one interesting thing about Coinbase Coinbase is like you know exactly, like you don't know exactly, but you have a good sense of who would be a qualified investor by the assets that they custody on Coinbase. Oh, sure, sure. It's a very natural thing of knowing like, hey, these 200 ,000 people or whatever the number is, like have enough assets just in, you know, digital assets that they would be qualified as a... Sure, sure, sure. Yeah, so yeah, philosophically, what are you thinking? Yeah, I mean, we think that everybody should have the opportunity to invest in these companies. I think private companies have for too long been restricted from an access perspective to large venture capitalists and quote-unquote accredited investors.

2:29:03And I think, Jordi, assets that you have are one indication potentially of sophistication or insensitivity to loss. But the perverse side of that equation is it potentially locks a lot of people out from the most compelling investments. that are out there. And we want to help people be early, basically, and give people the shot at being early. And so obviously, we need to follow accreditation rules as they stand. But hopefully, those change over time as well. And with Echo's infrastructure, anybody can invest in a seed round or a series A round, starting with crypto tokens, and then even beyond that, looking at traditional companies as well.

2:29:44How early do you typically start meeting with potential acquisitions? I imagine with Liquify and Echo, I'm sure the conversation started at least a year in advance, but what does that actually mean? Yeah, it's kind of variable. I think with Echo, we had known Jordan and Kobe for quite a while. And so I was chatting with him when he started the company, but we started to really work a lot more together earlier this year when we actually partnered. So Echo has an AngelList-like product where you can start an investment group and then sponsor investments effectively and syndicate those out. And so we started an investment group under Coinbase Ventures for the base ecosystem fund where investments that we were making, we would actually let the community follow along with us and invest in those same companies.

2:30:41Because that's something that we heard a lot of demand for, particularly within the base ecosystem. And I think as we started to work together more, that broader vision of, hey, we can actually enable internet capital markets. And Echo, you have the infrastructure, Coinbase has the size, scale, and distribution. So there's potentially some magic that could happen if you bring those two together. those conversations really materialized and ultimately culminated in this week. Well, congratulations. I'm going to ring the gong. Hit that gong.

2:31:15Fantastic work. I'm sure you'll be back on very soon. We will talk to you soon. Have a great rest of your day. Our next guest is in the Restream waiting room. While we bring in Nick, let me tell you about public.com, investing for those that take it seriously. Multi-asset investing, industry-leading yields are trusted by millions, folks. and we have nick from rivet in the restream waiting room let's bring him into the tbp ultra dumb nick how are you doing john jordy great to see you guys both how are you we're fantastic big day for you i already rang the gong just a few minutes ago i want to ring it again what you got for me uh we're uh excited to announce our seat round it's uh it's a big day for rivet how much 5.1 5.1 uh why the any special meaning to the number or just a 100k flyer came in at the end you know what's funny is this business is um you know it's self-sustaining when you sell a service to a customer they pay you money they typically pay you more than it costs to employ people um at least in aggregate and um we were really selective about who we chose to work with.

2:32:19And so we started off really only looking for three or four to really give us a buffer to invest in a sales team and some more tax partners ahead of next season. But you meet with one great firm, they talk, more people show up. We were lucky to get to work with Haystack, Sousa, and XYZ for this round. And so we're very lucky. What's the core pitch for the business right now? How are you explaining it to new customers? Okay, so I'll give the pitch as a customer. It's just an amazing service. I've used Rivet over the last couple of years. And so that part's easy. It's like you guys are really dialed, like incredibly thoughtful.

2:32:57It's just smooth. And so that part makes sense. What I want to understand is like the pitch to investors. I'm actually an investor myself. So, but not biased. Like the product's truly, truly amazing. But what's the updated pitch to investors since it's been probably a couple of years since I participated? Yeah. To customers, we're a tax prep firm, right? You'd engage with us no differently than Birkeland or Anderson or even Deloitte. We have 30 X big four accountants on the team. That's pretty straightforward. Nobody really asks why we raised or what we've built when we sell or when we onboard clients.

2:33:36When we talk to investors, it's oftentimes focused on things that customers don't really care about. It's what have you built? Why are you building it? the toughest part of building this business is context management when you hire a person to do a job the context for what they did why they did it why they chose to put you know something or you know recommend a certain classification is hidden in their head and it would be prohibitive for their time to write literally everything down and so every tax season as new people join the firm, as people leave the firm, as, you know, we get rolled up by some huge key firm, a lot of the practice, a lot of the actual work starts from scratch, these larger firms.

2:34:20And so clients get asked the same questions over and over again, because no one remembers anything because no one's worked there for more than six months. And it's deeply painful for customers, right? And they don't realize why it happens. They just know it sucks. And, and it's frustrating and uncomfortable. And wait, you don't remember my EIN? I talked to you guys three months ago. what we've built sits on the back end. And so it quietly ingests, saves, tracks, whatever you want to call it, everything that is discussed about and given to us from a client. And so it's not just emails, it's Zoom calls, it's Slack transcripts, it's text messages they sent us.

2:34:59And just tracking all of that in one place is a massive problem. It's a massive build. that's what allows us to run this practice at scale efficiently right and it's it's you know tax prep isn't just data entry from the w2 to the tax software it's it's gathering where the w2 was in the first place and so what we've built doesn't really replace the accountants it just automates the tedious parts of their job remembering everything organizing all the information wherever it was sent and and making sure it's stored in the right place um it's it's almost Mike Ross from Suits and remembers everything until he's there.

2:35:35It's really cool to get to work with if you've been working in one of these larger firms for decades. Have you bought any companies? No, no. Well, actually, yes, we did. We acquired a small AI company called Lobby. Originally, we met Wilson Hobbs, the founder and CEO, and he was applying for one of our roles. And we were deeply impressed by him and his background. And as we did the interview process, you know, we learned more about Lobby and what he had built there. It was a website. You could go to it. You could upload documents. You could talk to those documents. Sounds pretty similar to what we're building here, right?

2:36:15It just needs to be done with tax. And so as we did the interview process, I said, wait a second. What's happening to the team? What's happening to the product? He's like, you know, not really sure. We haven't quite figured that out yet. And I said, well, we should have you join the team. That part's easy. But we should also acquire Lobby. I think what you built is deeply special and we should have it here at the business. Wilson joined the team last year. It's his one-year anniversary. It was actually yesterday. Wow. And he's since brought on two of his former team members. Yeah, one-year anniversary.

2:36:43It's exciting. That's big for us. It's big. But you haven't bought any tax prep firms, I'm sure. No. How many times have people tried to get you to do a roll-up? I think every investor meeting for the first six months was, so how many firms have you bought? Why aren't you buying more? Aren't you going to raise$100 million to roll up 50 firms? Chad's asking what the website is. It's rivet.tax. Rivet.tax. Well, congratulations on all the progress. In 30 seconds, since we're tight here, but why have you chosen not to roll up tax prep firms and make them AI native? Yeah, good question. It's much easier to steal their business than it is to buy them.

2:37:32The change management of rolling up a firm, convincing the people who work there that they're like us. The whole process is deeply difficult. You can expect to lose about half the customers, half the revenue, half the team when you buy somebody. It's not worth it. It's much easier to just tell people about what we do and bring them onto the platform. Much, much easier. that makes a ton of sense love it well absolute dog absolute dog thank you so much for stopping great to have you on uh we'll talk more soon congratulations service we'll talk soon bye thanks guys cheers uh before we bring in our next guest let me tell you about adquick.com out of home advertising made easy and measurable say goodbye to the headaches about home advertising only adquick combines technology out of home expertise and data to enable seamless efficient ad buying across the globe we have david tish in the waiting room i love david tish tish Tush, Tush, Tush, Tush, Tush, Tush, Tush, Tush, Tush.

2:38:26Let's go. Let's go. Give us the news. What's new in your world? Did you raise one fund? No, two. What? Oh, hit that gong. We got two funds, two new funds. Thank you. Let's go. Incredible. We did two because two is more than one. Yeah. Which was the main bigger number, better. They made a sequel to one fund. It's two funds. And that's important. to do more always, especially in venture. Yeah. Yeah. Bigger is always better. What's the strategy? Put it all in one place. Massive secondary slug of open AI. What are you thinking? We literally have no rules. Our plan is to give people money and pray and diversify our prayers.

2:39:16A diversified set of prayers. Where are we on the prayer meter? Should we be praying more this year than last year? Are we going to be praying a lot more next year? What do you think? It's like a balance of stability and increasing at the same time. So you have to do, you never drop, you just keep doing more. We're different than other VC firms in that strategy, which is important because we're selling money, which is a commodity. Yeah, makes sense. I love it. uh what what uh besides fundraising has that just dominated your life the last uh since we last chatted or are you still not um you are forced to file with the government at a certain date so 15 days after you uh it's like taxes but totally unrelated um and so you have to file with the government and so our filing date was yesterday so we're forced to uh talk about it and And so that's why we did it.

2:40:15But we raised money earlier in the spring and just closed it. And we'll start giving it away in the new year, which is a 2026 treat. What are there particular areas that you feel are just like over, like over hyper like over competitive right now? all the, like, have been picked clean of new ideas, categories where you feel like you have enough bets and maybe you're going to be declining to hear more pitches in a particular category or vertical or horizontal layer of the stack. Like, are you thinking about that at all? No. Okay. Should we? I don't know. No, I mean, on this show, we hear, like, you know, a million, like, you wind up hearing, like, AI agents for a million things, and then you're just like, okay, I don't want to hear that.

2:41:07And maybe, I mean, we're not even investing. We're just like, I'm kind of bored of talking about that particular niche. Look, I think at the later stage, there's a lot of direct competitors. And if you're in the business of picking a later stage company, you pick one and then you stop. We're at the very earliest stage. And so we're meeting people who have ideas, who have dreams. And what we've found over the years is, you guys know, is things change. And so you might start with an idea you're going to... We don't fund intentional pivots, but people iterate. People get into a market. They see where there is opening and where there's too much competition.

2:41:45And your go-to-market becomes different. Your product might become different. And so I think we don't ever close ourselves off to things. We don't try to fund direct competitors ever. We think that's just misaligned with the deal you're making with the founder. but I think we're open-minded to what we fund and where it's coming from. And I think that in this moment, you're seeing sort of everything up for grabs still. And whether that's within verticals, whether that's in the tech stack, it doesn't mean, you know, go fund foundational models today. But I think other than that, we're pretty open-minded.

2:42:22And then foundational models that are not perfectly horizontal, I think are probably still on the table. Is there a new mafia emerging? I mean, we've lived through the PayPal mafia. There was a bunch of entrepreneurs. You guys are sponsored by the mafia, right? So my main goal with our new funds, I counted only four of your sponsors that we've invested in. So Ramp, Privy, Bezel, and Graphite. I would like to get the full lineup. Yeah, so we're going to take our money. We're going to give it to companies. They're going to give it to you. and we call it the new circular economy. We tried to fund the circular economy.

2:43:03The circular economy. Our audience will buy the SaaS and run their businesses. I think we just described a regular economy. That's the main strategy. Our LPs were really excited about that strategy. Perfect. Well, thank you for your success. What do you think is your most underrated portfolio company? Oh, that's a good... Now I have to pick a... You basically have to pick a favorite. No, no, but there's companies There's companies that some of my favorite acquisitions are companies that like haven't been in the news for five years. They just get acquired for like$800 million. And you're like, we have one of those, which is really cool, called ID.me, which he was on.

2:43:39That's a banger company. I know him. He was on Invest Like the Best after like 12 years. It's the craziest origin story. So that's like a evergreen version of that. And then one day they'll like win. uh sort of more modern version what rogo's building i think is not getting enough attention um there i think gabe came on if he didn't one day yeah yeah yeah blake hall founder and ceo of id id.me i was at a conference and he had a whole deck about like the business and he just ripped it up threw it out like minutes before and just told his life story and everyone was just like this is the best conversation ever because he'd like military like saved lives it was a crazy just like you know this is like elite business leadership i have so much respect for it do you uh i feel like in some ways it's harder harder to predict the future than ever like in 2010 there was like this sense of uh i think we're going to have a bunch more internet companies in 2020 and i feel like there's a general sentiment in the industry of like we can't predict anything after five years like we might get you know it's crazy acceleration and ai progress and then everything changes, right?

2:44:49So it's a little bit fuzzy out in the future. Do you have any sort of, uh, I referenced like, uh, there there's, I'll botch it, but like, there's a Bezos quote, which is like in the future, people will want like cheaper things faster. Right. And like, I can, I can like build a business on, on those, like with that, with that goal, uh, for a long, long time. Cause do you have anything like that in early or like in terms of how you, do I have a Jeff Bezos worthy quote to share with you? No, more so if you're evaluating something, is it purely like, yeah, I don't know if this is going to be a market in 10 years, but this founder is amazing?

2:45:26I think to your point, predicting two years out feels really hard right now. You're seeing such rapid build iteration from the biggest companies. And that includes the new wave of AI native companies, OpenAI, Anthropic. We saw the browser get released. You're seeing more and more like first principle build technology get released. At the exact same time, you're seeing like the modernization of Google, of Meta, of Apple, of Amazon in real time and Microsoft. They feel stronger in many ways than they did for a good five year period. So trying to predict two years out when all of that's happening in terms of the startup world, I think is really difficult.

2:46:10At the exact same time when we're investing, we can't do things that we think are shorting the future. So you have to play for something that is 10, 15 years out and not like, we think this is going to happen, but we can thread a needle for three years and build a business. That's a terrible way to invest. And we call that shorting the future. Short the future. I love that. That's really, really good. um so we look at everything it's like if everything this founder is saying is right is this interesting big enough like fascinating to us and that's what an investment is is uh joining a journey um so we try very hard to look at it from you know if if this existed is this right or cool and um not shorting the future such that it's like obvious why this won't exist in a reasonable amount of time.

2:47:03That makes sense. I think the last time you were on, or maybe one of the times you were on, we talked about how there used to be, there was this golden age of aqua hires where like great teams could sell to Twitter for like$10 million and they hadn't raised much. So it was meaningful outcome for the founders. And then they - Now we do aqua quits, right? That's what we're up to. Well, yeah. Yeah, I mean, obviously the windsurf thing. No, it was more of the like go to Facebook when you started a company from that company thing. Yeah, yeah, that whole thing. But it feels like we maybe are in a new golden age of aqua hires where like I was advising a company, a founder like last week who launched a product.

2:47:43It's like it was exciting, but like clearly doesn't have product market fit yet. And he's like, you know, talented and has a great team. So he's getting fielded all these offers that will be like not a meaningful outcome for the investors, but like, you know, will be very meaningful for the team. and it feels like there's more of that right now in this moment when people are like, I don't just need to be AI native. I need as much AI native talent as I possibly can and I'll pay a big premium for it. Yes, and that's healthy, right? The sort of decade we had of people just staying in long-term, like not going anywhere, startups without sort of recirculation of talent, I feel like was a negative moment.

2:48:28I think if you go back to the origins of, say, Facebook or Uber or Airbnb, in that moment, there was more aggregation of talent on those early teams than sort of in the next wave of startups. What's interesting about this moment in AI and these very significantly funded companies, a lot of the research labs and people coming out of the foundational models, by getting the funding at the scale they're getting and by bringing with them 8, 10, 20 people from their previous company, you're actually getting depth of talent that is capable of building something big, probably in an easier way than starting from scratch.

2:49:10And so I think there's a balance to be had, but I do think depth of talent is probably the most underrated component of a quality startup, right? Your 50th person on your roster is amazing. And I think one of the things about Ramp that they've done so magically is this combination of hiring, retaining, acquiring, inspiring, letting people go, bringing them back. And if you go to like 50, 100, 300th person at Ramp, there's just an immense amount of talent from top to bottom. And that's, I think, how you build sustainable growth that you can keep iterating and launching new and interesting products or marketing stunts or excellent marketing campaigns.

2:49:57And I think you're seeing that from a company where you're seeing depth of talent across all the different departments within RAMP, as an example. Yeah. Chat wants a jersey check. What are we rocking? we're wearing the uh new york rangers 100th year anniversary jersey it's a baby blue you look fantastic beautiful we we haven't won a game at home yet and so uh hopefully we're going to bring some luck to us well good luck and thank you so much congratulations if you if you know any sponsors who need money just let me know and we'll take care of that perfect we'll make some introductions uh great to see you come back on soon congrats to the whole team cheers thanks so Let me tell you about wander.com.

2:50:42Find your happy place. Book a wander with inspiring views, hotel-grade amenities, dreamy beds, top-tier cleaning, and 24-7 concierge service. It's a vacation home, but better. Our next guest is Chris Dixon from Andreessen Horowitz. Living legend. Welcome to the show, Chris. Great to see you again. It's been too long. Great to have you back. Last time you were on, we got to time you up perfectly with Brian Armstrong. Today, we just missed each other by one day. 24 hours. But great to have you here and would love to get the update from you and how you're framing the current crypto era. It's in many ways linked to the AI boom, but somewhat disconnected.

2:51:18How are you thinking about even just framing the conversation and setting the table? Yeah, so we just put out our state of crypto report today, which I think had a lot of good data. I think the good news is I think things are trending very well. I think it's a reflection of smart policy mainly that the Genius Bill passed, which is very important, which has helped the stablecoin growth. I think what it shows, like what Genius shows, is the dual effect of smart policy, which is, on the one hand, it provides a pathway for innovation. So we're seeing not only startups, but big companies like Stripe and FinTechs and BlackRock, Fidelity, and so forth get involved in the space.

2:51:57But it also kind of eliminates a lot of the scams and bad behavior. Right. So what we're really hoping for is a kind of similar result with the so-called market structure bill, which is being considered by Congress, which would which would provide a similar kind of policy framework for the rest of the crypto market beyond stable coins. So I think that this kind of momentum is reflected in the data that we shared today. So, for example, stable coin volume is now in the trillions. That's adjusted. We try to we have an adjusted metric we use, which tries to remove a lot of the bots and other kinds of stuff.

2:52:27And very importantly, that's not correlated with trading volume. So it seems to be like sort of genuine, you know, people using for payments and treasury management and so forth. So that's a big deal. We're seeing a lot of cool kind of new emerging applications at the intersection of areas of crypto and AI in an area called D-Pin, which is decentralized physical infrastructure, which is a way to use kind of crypto to incentivize the build out of telecom and energy networks and so forth. You know, we're seeing a lot of, there's a big movement of so-called real world assets, which is to bring like stocks and bonds on chain.

2:53:04A lot of good momentum. I mean, honestly, like we had a, you know, it was a choppy couple of years, 2022 to 2024. And I think it's, we're kind of coming out of that and I'm optimistic, but still a long way to go. Yeah. I always wonder about the value of these high-level market KPIs in crypto because it feels like as soon as we latch on to one, we identify one that seems to be tracking well, then it gets all thrown off by just an underlying shift in the market. Like people were tracking like number of Bitcoin wallets for a while. And that, like the idea that the next generation will even be aware if they have a wallet, because they might own it through an ETF, or they might have a wallet that's abstracted away and they're not even aware of it.

2:53:46And I'm wondering, like there's this like, sort of like the midwit take is like, oh, the KPI flatlined, but the community overall keeps growing. And so if you were taking a victory lap in 2023, you missed out on the next wave, which looked different and didn't map onto the previous era of KPIs. And so I'm wondering, is there some ground truth that you zoom out to that's just the vibes or the people or the talent or something like that? How do you get away from this particular KPI now needs to be deprecated, basically? Yeah. I mean, there are some, just to be fair to the KPIs, and I think Darren, who is our head of data science of the report, is very thoughtful on that.

2:54:27Like the stable coin volume uncorrelated to trading. Oh, sure, sure. It seems like a real kind of real metric. Yeah, tech metric. Monthly active users. That's not people that hold crypto. A number of people hold cryptos in the hundreds of millions. What we track is people that are actually using these kind of new apps that are on chain. Got it. That we put between 40 and 70 million. There's a bear bar there because we have to get rid of bots and other things. But it's definitely been growing. And that's like monthly active users of apps. So, you know, it feels like a real metric that's a staple, to your point.

2:54:57Yeah, but it's a good point. A lot of venture capital and startups is vibe-based, right? It's qualitative. And I think it's definitely better. I'm actually having our annual founder summit now. And just the mood is better. People are building and people are innovating. It feels like offense, not defense. A lot of that's the policy shift. The policy shift is big. Also, I feel like the bull case for bear markets is that it washes out a lot of the tourists and then the technologists stay. It does. I think that's right. I think that's right. I mean, I've been, you know, obviously doing the internet stuff my whole career and seen many waves, including, you know, I started my first company in the internet downturn of the 2000s.

2:55:40And, you know, there was this kind of, you know, what was it? Strong men create weak times meme or whatever. Well, I like that crypto as an industry rewards people that have an unwavering vision and are okay with, you have to have some appetite for failure. You've got to be a missionary. Yeah, yeah. And I think that's healthy. It's like, yeah, you have to actually believe in this long term. You have to be okay with the risks. You have to be okay with a total ban scenario in a specific category. My question for you, prediction markets are having a massive moment right now, And this was something that I'm sure you were writing about more than a decade ago in terms of its potential.

2:56:27And there were various kind of attempts at prediction markets that never really reached cultural relevancy in the scale and the way that they've reached today. What kind of categories do you think people have thought had potential but kind of wrote off? Maybe they were too early, but we'll be talking about in like two years. Is it privacy, ZK proofs, things along those lines? That's a great question. If I could just address your prior point about the whole, sort of sticking with it first, I have an unwritten blog post that I would title Investing is an Emotional Test Disguised as an International Test, meaning one of the things I've learned over the years is just sticking with it in venture capital, whether it's crypto, sort of more extreme.

2:57:15I think this is true of public markets. if you read like Howard Marks and Warren Buffett. But it's definitely true adventure, which is a lot of it, the business, is just sticking with stuff. The same with entrepreneurship, when it seems kind of tough. So I think it would be an amazing, great point. To new emerging areas, yeah, look, I think privacy, it's interesting. We had a lot of privacy is a very interesting area. Now the Stablecoin bill passed, a bunch of folks like in the policy side who were kind of, I don't know, anti-encryption and privacy are now saying, wait a second, all the stablecoin payments are public.

2:57:47That's not good. What if you want to do financial protection health care? We agree. That's why we want it. We don't want it for, obviously, for illegal activity. We want it because people want, in normal cases, they want privacy for some of the things they do. And so we think that's very important. We have a bunch of investments there. Yeah, prediction markets, it's great to see them break out now. As you said, the idea goes back. I wrote a tweet about it back to Hayek in the 50s. And DARPA started an apprenticeship market, believe it or not, the government in 2021, I believe, sorry, 2001. Wow.

2:58:16Something like that. Wait, what was, what was, what was that just like, you know, uh, just for like Intel purposes? Like, yeah, they wanted to predict like, is there going to be a war in Iraq? And then they thought, okay, like it's a little bit distasteful that people are betting on it, but the societal benefits would outweigh it. But then there was such controversy around it that they, that they, um, canceled it. There are now rules as there should be, around so-called harmful prediction markets, and you can't have those. But the kind of the core argument for prediction markets is that, yes, there's sometimes, on the one hand, it's sort of gambling and betting.

2:58:50On the other hand, it has a real benefit, as we saw, for example, in the election. You can get really objective data. And in a time in which, you know, trust in institutions seems to be dropping, like, do you trust, you know, the newspaper anymore? It's nice to have an objective source for that. So, but back to your question, I think privacy is very interesting. I think, you know, I'm very excited by sort of the intersection of like AI and crypto as an example. My experience in tech is that often like if you think back 15 years ago, people talked about mobile, social, cloud was like the three book words.

2:59:20And some people saw them as kind of different categories. In the end, we actually learned they were all very connected. I think that tends to happen. I think a lot of these things will converge, you know, robotics, AI, crypto. Like these are all the kind of megatrends. And I think it's natural that they intersect and reinforce each other. Are you seeing a speed up in crypto startups productivity on the actual software development side, like coding agents for Solidity or actually having an increase? Oh, yeah, definitely. You are? Oh, yeah, definitely. Almost all of our company. We encourage them, too.

2:59:54They use, you know, crypto. And so is that what you mean? Yeah, yeah, yeah, exactly. Because as a consumer, I mean, I see the data from like Anthropic and OpenAI and Gemini. I see that like tokens are being spent. But then a lot of the consumer software that I use, you know, the United app we like to pick on or like the flight booking app, it just feels like it's not like become all futuristic all of a sudden. There's still bugs. And so I'm always interested in like startups are in a unique case where they can adopt new technologies and actually new workflows and maybe gain the productivity benefit faster than legacy companies.

3:00:28So I'm just interested to see if it's a unique case. Yeah. Yeah. And I think it just takes a while for these things to propagate, right? Totally. United app will be the last one to do it five years from now or whatever. But the startups are all using it, using Cursor. It's amazing. I mean, I saw some study out of some university that was like, these tools don't make you more productive. And you just use them. I spent a Christmas break playing with Cursor, last Christmas playing with Cursor, and I built a website. I mean, it's just ridiculous how great these tools are. Of course, they add to productivity.

3:00:56It's just these things take time. It's like humans, the bottleneck is always humans and policy, I would say. regulation and humans take time to change behaviors, right? But as you say, startups tend to be the first to adopt it. Yes, we see it across the portfolio. We would be kind of surprised and, I don't know, like just, yeah, it's like puzzled if somebody wasn't using a lot of these. Yeah, it'd be kind of a red flag. What does institutional adoption and embrace really look like? Like what are the kind of buckets? Like the base case is you're supporting custody and maybe trading and then more advanced.

3:01:32CEOs just stop talking trash at some point and then start saying this is amazing. That's a joke. That's helpful. But how quickly do you think these major institutions can actually adopt crypto rails for, you know, operationally, right? Their job is to store and move money. Crypto is, you know, more efficient in a lot of ways. Yeah, I mean, you can kind of think like crypto is a network. I mean, it's a network technology, right? It's about building networks. And it was the network where 99 % like stablecoins is an example where 99 % of the potential nodes of the network. felt like they couldn't join the network for regulatory reasons.

3:02:06So now they're all, I think it'll see it in stages. So like the fintechs, the sophisticated fintechs like Stripe and, you know, Robin Hood and Revolut and things will be first. And then over time, you'll see, I think eventually you'll see just like a button on your Chase bank that's like send, and maybe they don't call it stable point, but they call it digital dollars or something. But you'll see, you know, it'll just be integrated into all the kind of fabric of these systems. And we're seeing a lot of activity. I mean, even, you know, Chase, I think Jamie Dimon talks trash, but I believe they have 1500 people on blockchain projects.

3:02:35So like, you know, I think, I think that the important thing is the bit is flipped where the, you know, like if you talk to the folks that are like, look, we want to do this, but like our legal teams won't let us, like now we have legislation, which is the gold standard of policy, right? U S Congress law. So like now they feel comfort to do it. And that's very, very important. And I think the bit is flipped where they now see it as not a threat, but an opportunity. So I think it'll be, you know, it'll take time. It'll be different kind of tranches of adoption based on how advanced the companies are but i think we're on that clear path now um and you know is there is there anything that's been interesting for you guys on the buyout side like looking at you know let's say like a mid-size financial institution and realizing like hey this this is a great business and if we actually you know were to acquire that general catalyst with that health care network yeah something to that effect or or some of these role you know i think you'll see that i think you'll see that we don't actually do that it's not really our thing like we do kind of stage venture but mostly in some later stage venture but wouldn't it be fun to do an lbo yeah it might be one of these things it probably sounds fun yeah yeah like those people like we get to mostly hire people in venture like lbo is you're mostly firing people and it's just like a structuring i don't know i mean probably the grass is greener thing it sounds fun and then you actually go do it i think it's also just hard to actually change these organizations sometimes there's a real question like, do you need just new organizations started from scratch or do you change them?

3:04:02I don't know. It's not my area. I'm just speculating. But mainly it's just not our thing. But I think other people are doing it. And I think the other firms doing it will, I think, accelerate the adoption of new technologies like AI and crypto. Yeah. How do you envision the role of the venture capitalist changing over the next maybe decade in crypto? I feel like there's one world where you become half VC, half hedge fund person that's trading liquids and trading liquids at the earlier stage because companies are kind of doing on-chain IPOs in an earlier stage, but you still need to build a position and take an opinionated view of the future of the technology, assess the team.

3:04:40How is all that evolving? How do you think that that will evolve? Yeah, that's a great question. So, I mean, what we do, we've done for a long time. That's one reason we started a separate crypto fund. We both do kind of early stage traditional venture and do things like buy tokens like Bitcoin directly in the market and then we have our own trading and we sort of built out all these capabilities. At some point you'd expect that some of those things to be subsumed by outside organizations and kind of commoditized and these practices I would expect would propagate to other firms. But I just look at venture through the lens of the fundamentals are always just great people.

3:05:16We're ultimately in the talent business so there's all these other kind of mechanics and things that we've built up but in the end we're just trying to bet on brilliant founders who have great ideas. And that can, that can just, the way I view it is that can come in many forms. That can come at an early stage equity investment. That can come as a token investment. You know, we're also betting of course on communities. You think about something like Bitcoin, like it's really a community at that point and it's sort of this network and community. So I look at it through that. I always try to, whenever the markets kind of get strange and things are happening, I always say, let's go back to first principles.

3:05:47First principles is like believe in positive vision of the future, great entrepreneurs, pursuing big ideas, highly technical, highly product focused, and then whatever mechanics we need to do to get exposure to that. I don't know if this is possible, if this has already happened, but would you ever see yourself backing a project that looks like Bitcoin in the sense that the founder's anonymous, you can never meet them. But for some reason, you do have the ability to make an investment like a venture capital investment. Are you equipped to make that deal happen? Yeah, we are. We are. Yeah. so yes we can we could just buy i mean we could just buy tokens yeah and we have an operation we've had that people don't seem to know they think of us more as venture we've had that since 2018 totally just record yeah and actually two-thirds and i would say roughly two-thirds of our investing has always been that i mean that was part of the design and why like i wanted to spin this out yeah um so we could set it up for principles with this structure um so we have like i want to be a crypto vc but i don't want to buy tokens that'd be insane for sure well there's a I can tell you this is a longer story.

3:07:00It's a whole backstory where like, it was just, it was just very challenging. The LPs. So like when we were, I raised a prescriptive fund, like I went to all the LPs and I said, look, we're going to do, we're going to buy tokens. We're going to do this. And like, you can opt out or opt in. And some of them opted out and, you know, some opted in and just like, look, this is what we're going to do. So we have a venture structure or a venture fund in the sense that our LP relationship is a long-term relationship. They lock up their money. But that's sort of the reverse mullet or mullet or whatever.

3:07:24We're on the back end. It's like a very traditional venture. But on the front end, we can do all. We can buy tokens. We can do equity investments. Whatever way that we need to get exposure to great ideas and great people, we can do. And that's been in our kind of design from the beginning. And we continue to do that. Yeah. In this analogy, I feel like the venture part is the business in the front and the open market trading is the party in the back. But I don't know. It's a stress analogy potentially. But I like a mullet analogy. It's great. But yeah, thanks so much for taking the time. Yeah, great to get the update.

3:07:56Have a great summit. Yeah, I'm sure it's going to be a great time. It's going to be an amazing time. Great to see you. We'll talk to you soon. Have a good one. Bye. Let me tell you about GetBezil.com. Your bezel concierge is available now to source you any watch on the show. planet. Seriously, Andy, watch. And Jordy, did you put up a better run last night on Eight Sleep? I did okay. I got an 88, slept six hours and 50 minutes. How did you do? Did you get destroyed? 85. Wait, I beat you by one point again? No, 85, you got an 88. I got an 88. Okay, a couple points. Yesterday in the chat, people were congratulating us on our good numbers.

3:08:30We got to put up good numbers for the chat, and you got to go to eightsleep.com and get a five. five tbpn um are there any yeah i was i was shocked at how uh at how like viral uh i was shocked at how viral the whole like aws uh eight sleep thing went oh yeah considering that uh uh clearly a skill issue and i was not affected yeah i slept right through i actually slept great my numbers were insane it was crazy uh but yes i mean somebody uh uh somebody you know i i i feel bad if If your eight sleep did wind up getting too hot and you woke up and you didn't have a good night's sleep, I'm sending my best.

3:09:10I hope you get back on that horse. And we'll close out with this post from Rune. I knew you were going to go there. Or a farming in the Zoom waiting room. Or a farming in the Zoom waiting room. Just never join. Never join. Just hang in the waiting room. Oh, so people know that you're there? Is that how you do it? Yeah, that you could join. I feel like, I mean, at one point, not to break the fourth wall here and give you a little bit behind the scenes, But at one point, we had multiple guests. When they come in, they hit a waiting room, the Restream waiting room, of course. And at one point, multiple people could talk to each other.

3:09:40And they could hear what was happening on the show. And it was kind of like this funny networking event that was happening. And I actually think that there's some value there if we could get that right. But right now, we just let everyone sit in silence and then eventually come on, watch the show, and then join. so there's minimal aura farming in the TBPN waiting room the Restream waiting room but maybe we should increase it well thank you for hanging out with us today very fun show I can't wait for tomorrow we have to wait just under 21 hours to get back here the clock starts now have a wonderful evening leave us 5 stars on Apple Podcasts and Spotify cheers goodbye