Elon vs OpenAI, Ads in ChatGPT, Leaked Tech Emails | Sean Frank, Bill Shufelt, Alex Mashrabov, Sonya Huang & Pat Grady

27:07To my knowledge, nothing has been shared publicly other than the sort of idea of him getting around 7%. Okay, okay. Yeah. That still feels low for a co-founder of a mega cap. No, but that was happening. That's like a lot of dilution that happened by that point. Makes sense. But, you know, it certainly doesn't feel like the Mark Zuckerberg or the Bill Gates or the Jeff Bezos scenario. I mean, yeah, if anything, these files make a lot of Elon's kind of antics more understandable. Like he's been the one saying that they're morally bankrupt. And here they just straight up say he would be pretty morally bankrupt to steal the nonprofit from him.

27:53Yeah. I don't think anyone ever expected that language came out that was this specific about their moment-to-moment thinking during that point in time. Yeah, what's the phrase that Elon uses? Scam defaultman? Scamual? Scamual defaultman? The timeline is really so stupid. And the portmanteaus. But I mean, one thing that is real is that it's very clear that at some point, the core OpenAI team was like, it's either like Elon as CEO, financial backer, like complete control. Like Greg Brockman said, this is the only chance we have to get out from Elon. Is he the glorious leader that I would pick? We truly have a chance to make this happen financially.

28:48what will take me to 1B, which is interesting because there are definitely people that have been taken to 1B alongside Elon with Elon in the glorious leader seat. There are people on the Tesla train. There are people on the SpaceX train where truly they are completely hands-off. Elon is the glorious leader of those companies. They don't have any control over Elon, but Elon has given them effectively billions and billions through either their direct work with the company or their investment in the company. There's a whole bunch of ways that those two things could actually be very compatible. But it would be weird, especially if you want more of an egalitarian organization, more of an equal board, more of a team dynamic.

29:31Well, you look at the run that Sam has gone on, and you could imagine a situation where they're sitting at a table and saying, this boardroom ain't big enough for the two of us. Yep. Right? Yep. And it really feels that way. Big egos, yeah. Yeah, yeah. Sam notoriously is able to draw in capital using many of the same sort of methods that Elon has. Yeah. I do think the financially what will take me to 1B, that's just a good question to ask. Everybody needs to be asking themselves this. Yes. This is something that should be taught in grade school, really. I see people on Instagram reels all the time asking this question.

30:13Alex Hermosi makes reels about this. He's asking himself this question. The only other question is, why aren't you asking yourself that question? You should. You should also get on 11 labs, build intelligent, real-time conversational agents, reimagine human technology interaction with 11 labs, baby. Continuing with the timeline, a federal judge denied OpenAI's motion for summary judgment. The case is going to trial, and Tyler's moving to... Oakland. East Bay. Should we read through some of Alex Heath's coverage in sources. Dot news. And talk about a scoop athlete that is really given. I call him scoop God.

30:54Really given Kylie over at Core Memory. I mean, these two are just going head to head. Yeah, yeah. It is a knockout, drag out fight for scoop athlete of the year. But they're both doing great stuff. and Alex read through thousands of pages in Musk versus Altman. So you don't have to. We'll go through some of the summary here. There's some news about Elia Sutskiver. Elia had early concerns about treating open source AI as a sideshow. In 2022, open AI's leaders seem quite concerned about the prominence of open source, the open source lab stability AI. We don't talk about stable diffusion that often anymore, but I believe Stability AI was the team behind that, right?

31:43Ahmad Moustak, I believe, runs it. Sutskiver voiced his worry over text with Maradi and others. Sutskiver, my trepidation around open source is that we're treating it as a sideshow, e.g. deaf not going far enough to really hurt stability. so they're not taking it seriously. And if open source takes off, everyone could standardize on that. Raghav says, bro, scooping harder than a Ben & Jerry's employee. Let's go. This is good. This is good, Alex. He's scooping harder than a Ben & Jerry's employee. We're going to use that for now on. This is fantastic. Maradi, Mira said, we're missing the opportunity.

32:25And let's just say, Mira's the real winner here. Yeah. She's winning today simply because the attention has shifted away from a good number of people leaving. Their own co-founder exodus drama, firings, who knows what's happening. But good to have the narrative shift and the vibes move on to OpenAI versus Elon Musk. So Miramorati said, we're missing the opportunity to set standards with this massive growing group of developers. People are hungry to build things and we should lean in and bring our tech to as many people as possible. long-term maximize our chance of maintaining lead, reducing competition.

33:02But if we do everything to get this in a couple of weeks at any cost out because we heard Stability is open sourcing a similar model, that's not in line at all with my motivation. So reducing competition, never something you want to see in discovery, always something you want to see in a pitch meeting with a VC, but you usually don't want to put it in writing. OpenAI leaders were divided over early investor Reid Hoffman's decision to start a rival AI lab inflection. And Reid Hoffman's interesting because he was one of the big early donors, I believe maybe 10 million, something like that in the millions.

33:39And he has not joined the lawsuit. He has not jumped in and said, hey, I need equity and open AI, I'm also wrong to like Elon. Now, in order to make the claim that Elon's making, you need a lot more correspondence, a lot more proof of misleading around the donation. And I don't think Reed necessarily has that, but also he doesn't seem to be saber rattling about it. So they were also already considering prohibiting investors from backing competing labs from an October 2022 exchange. Sutskiver, I guess I just felt betrayed by him founding a direct competitor while simultaneously telling me, quote, I could not possibly imagine you'd find it objectionable.

34:24And this is Inflection AI. Of course he'd find it objectionable. Inflection AI, which was ultimately sort of like acquired by Microsoft in that licensing deal. Was that Mustafa Suleiman's company, I think? Inflection? Altman, here's how I'd summarize my thoughts on this. Pros. He supported us in a moment when no one else would, and it was pretty existential. Okay, so we're learning more about the existentialness of certain donations when they came in during the OpenAI nonprofit era. There are times when if somebody didn't come in and write a check, it would have been very, very rough for them.

34:58And Sam says, Reed helped out at a key moment. I think OpenAI would have been pretty aft if he hadn't stepped up. Also, he was instrumental in getting the first Microsoft deal done. and has generally been quite helpful with Microsoft-related stuff, and he's generally a good board member. Cons, he's very motivated by collecting status, although I personally think he cares much more about OpenAI than inflection. He was blinded enough by the startup of being able to call himself the co-founder of a company. He made an uncareful decision. Also, at this point, I think at this point, OpenAI has the leverage to ask for a soft promise for new investors not to invest in competitors, but only a select few companies ever get to do that.

35:45And we heard stories of this during one of the bigger fundraisers, where even Glean was targeted as something that they didn't want their investors to also invest in. They wanted you to be OpenAI, ride or die. Of course, tons of funds said, I'd like a basket of labs, actually. I'm going to invest in all of them. I would like broad-based exposure to the category. I'm doing SSI and thinking machines, and I'm doing Anthropic too, and I'm also on the board of Meta. I'm also investing in OpenAI and XAI. You got to get a little taste of all of them. Brockman chimes in. He says, also an aside after talking to Sam Altman, I'm planning to meet Patrick Collison tomorrow in Demo DV3.

36:29will ask if he's interested in participating in the tender under the condition of not investing in AGEI slash big model competitors. So we talked about the Brockman diary. Well, yeah, the funny thing is he must, like, I'd find it hard to believe that he's not, didn't get to 1 billion just off of Stripe. Obviously, Stripe was much smaller back then, but Greg was a CTO of Stripe from 2010 to 2015. Stripe was founded in 2010. He was there for certainly longer than a four-year vest. Who knows? We don't know about his consumption habits. He might have been selling off Stripe equity as fast as possible.

37:11He might have been using shares to pay for coffee. I'll give you a forward contract. I'm going to use shares if you give me a free cup of joe. Satya Nadella was worried about Microsoft's position in AI when he started You're looking at open AI. Stripe was in 2017, Stripe was valued at 9.2 billion. Bunch of wee lads. Stripe. From Satya Nadella's deposition. The question to Satya Nadella, the CEO of Microsoft. Did you feel that your progress was moving more slowly than you had liked? And the answer, Satya Nadella says, I mean, always as a CEO of a company, I feel my job is to sort of be dissatisfied with the rate of progress at all times.

37:55And so yes would be the answer, which is both in the absolute sense, which is, can we build products that are more capable in any particular domain? And also, you know, vis-a-vis competition. There were others achieving things that we looked at and said, hey, that's great. And also, how can we make sure we're competitive with it? And so Satya Nadella was obviously motivated to invest. And now he has a huge stack of OpenAI shares ready to rock. Also, Satya Nadella almost wrote a book about AI called An Inflection Point. I think he definitely should write that book. That sounds amazing. According to an exhibit filed in the case, it was co-written with Marco Iancitti and was in development in 2023.

38:36From the first chapter, on Wednesday, August 24, 2022, with the Pacific Northwest summer showing all of its beauty, Bill Gates hosted a dinner at his home in Lake Washington, just a few miles from Microsoft's campus. No longer a Microsoft board member or even Microsoft's largest shareholder, Bill remained the iconic co-founder and trusted advisor of the company's senior technical leaders. Satya suggested the gathering, which included Chief Technology Officer Kevin Scott and a handful of top researchers. Food and drinks would be served, but the main entree was a hush-hush demo by OpenAI co-founder Sam Altman on a forthcoming release of ChatGPT powered by GPT-4, an AI built on large language models.

39:23Bill had long encouraged researchers to develop a truly accomplished AI assistant, but had voiced his skepticism about this particular approach. This sounds like I'm listening to an audible. Thank you. CrowdStrike. Your business is AI. Their business is securing it. CrowdStrike secures AI and stops breaches. baby. That's right. Crowd strike. Um, so that would be fun. I hope that book gets published. It'd be fun to have some of those little stories about, uh, we, we don't have, I don't feel like we don't have enough oral histories of AI who was in the room at what time. I feel like there's a whole series of books to be written about open AI.

40:02I know that they're writing, they've written a few, but most of them are sort of these like skeptical drama pieces. I'd be much more interested in, in two different books. One would be the technical histories, like exactly what researchers were doing what, at what times. And also, I want to know the feeling that was going through Ilya's head. And what was the vibe when he said, like, I'm hitting the button and I'm running the training run? What would you think? You're just describing Dworkesh's book. Yeah, but... The scaling era and oral history of AI 2019 to 2025. So, yes. And he interviews Hylia.

40:41Yeah, yeah. So I hear you on that being a great tour of AI progress across labs. Yeah. I think John's talking about business. I'm talking about within a single lab. Within a single lab, you're in the room account of, you know, narrow it just to like the GPT models. and so imagine the scaling era but just for the You basically want the acquired episode. Yeah, there's that. I mean, isn't there news that there's a book on Demas Hacibus that's coming out about DeepMind. That's going to be fantastic. It's written by Sebastian Maliby one of the goats of business writing. He wrote The Power Law, all about the history of Silicon Valley I know that book is going to be incredible but So I'd be very interested in the research side, how the research developed, who was doing what.

41:35And then I'd also be interested in the financial side. I want to know, okay, how did this SPV actually get done? What were the terms on this? Who was jumping in in this round at what terms? Who were they talking to? Who were their LPs? Were their LPs skeptical? We've talked to and seen a little bit of the other side of that. And then layering in all of this information from the deposition. Totally, totally. And some of this evidence that's come out that we read. And so there are a few writers that will take different cracks at the open AI story. I think the definitive book probably has yet to be written.

42:13But it will come, and hopefully Satya Nadella can play a role in that because he's such a key figure in the full story. Vanta. Automate compliance and security. Vanta is the leading AI trust management platform. Like a foghorn. Microsoft beat out Amazon when it initially started working with OpenAI. Elon Musk was opposed to working with Jeff Bezos and wrote the following in an early email to Sam Altman. He said, I think Jeff is a bit of a tool and Satya is not. So I slightly prefer Microsoft, but I hate the marketing department. And Altman responded that Amazon had started, quote, really dicking us around.

42:55Yeah. Such a crazy line. He's like, I don't like his taste in champagne. Have you seen what that guy drinks in St. Bart's? It's unacceptable. He should be taking it up another notch. And the sparklers with the champagne. The sparklers. It's just over the top. It's very vulgar. He knew he was going to get mocked. Yeah. Yeah. True. True. Yeah. I like a Tesla tequila over a St. Bart's champagne bathtub. I don't know. Credit to Jeff. He, of course, was a bit more locked in during that era. So the upside on Microsoft's initial$1 billion investment in OpenAI was capped at$500 billion. Hopefully they hit that cap from a filing written by Musk's lawyers.

43:36In November 2018, after dinner with Sam Altman, Scott told Nadella that OpenAI's new corporate structure offered both, quote, a commercial vehicle for monetizing OpenAI, AIP, and investment returns capped at$500 billion. That's not bad. A 500X bagger is going to move the needle for Microsoft, for sure. Altman claimed the nonprofit would eventually benefit because, though OpenAI has yet to make a single dollar in returns, if OpenAI ever does get to$500 billion in returns, the balance over that goes directly to the 501c3. That's exciting. Microsoft's board initially approved a capital investment of$2 billion, but ultimately decided to limit its initial investment to$1 billion in the hopes that a smaller investment would impress OpenAI to commercialize.

44:21Satya. Hey, we got it. We can't give them too much. Let's put a little fire under them. Let's make sure that they're thinking about dollars. Dollars and cents. Well, you've seen what happened with thinking machines. You give somebody$2 billion, maybe less pressure to commercialize. Maybe, maybe. More money on weight racks. Yes, yes. In exchange for its investment, Microsoft received a convertible limited partnership interest and rights to OpenAI's profits with returns capped at 2 ,000 % of its$1 billion investment. Microsoft CFO noted in an internal email that the cap is actually larger than 90 % of public companies and the limit on Microsoft's profits is not terribly constraining nor terribly altruistic.

45:03In fact, it was a good investment at Microsoft's request. True. True. OpenAI agreed to keep any mention of Microsoft's promise, 2 ,000 % of return on its investment out of its public announcement. Imagine going to someone and pitching them, oh, we'll give you a 2 ,000 % return and then actually delivering it. What a crazy, crazy story. Gusto, the united platform for payroll benefits, NHR built to evolve with modern small and medium-sized businesses. Get on it. Stop making excuses. This is the second update to Microsoft's partnership with OpenAI in 2021 included another$2 billion investment that wasn't reported and came with a lower upside.

45:42This is also a filing from Musk's lawyers. In March of 2021, Microsoft quietly invested another$2 billion in OpenAI. Neither OpenAI nor Microsoft publicly announced the investment, which was subject to a lower 6x return multiple. in place of its 2019 license to a single OpenAI model. Microsoft secured rights to commercialize any OpenAI model developed during the term of the agreement, except AGI, facilitating its commercial use of OpenAI's IP. Microsoft was permitted to embed up to 10 of its employees on-site at OpenAI. That's interesting. Anticipating increased product commercialization, Microsoft and OpenAI agreed to share any resulting revenue.

46:23Just three months later, in June of 2021, Microsoft released GitHub Copilot, its first product incorporating OpenAI's technology. Well, there are many, many more scoops in this piece by Alex Heath over at sources.news. So I encourage you to go subscribe, go sign up for his sub stack, and read the full thing for yourself because there's a lot of interesting stuff going on. But we should move on to the timeline. There's so much more news. let's find some post to run through. Everyone is having fun with financially. What will take me to 1B? Sitting on the edge of the bed. I love it. Feet up. You got to be asking yourself this.

47:04It's just so clear. This could have saved OpenAI from Elon and it's a diary with a lock on it. With a kitten. I think the diary framing is like woefully wrong. That's probably the worst part of this whole thing is because most of the thoughts that Greg's putting out are completely reasonable. It's like, yeah, we shouldn't screw this guy over. Do we really want him to have a complete authority? These are all reasonable things to be thinking out loud. It's just when you write them down, they get recontextualized in the courtroom. And then the fact that it's framed as a diary, very questionable. Yeah, I don't think it makes Greg look that bad.

47:41because he wasn't ultimately, it's not like he was secretly Sam's puppet master. No, no, no. And a lot of this went back and forth and there were lots of opportunities for both sides to yield. It was like a long ongoing negotiation. But everyone's picking their sides. Owen Sparks says, case closed. Elon has open AI dead to rights. Based on this quote here, he wrote, can't see us turning this into a for-profit without a very nasty fight. I'm just thinking about the office and we're in the office and his story will correctly be that we weren't honest with him in the end about still wanting to do the for-profit just without him.

48:17So bad blood. That's a bad one. That's a bad one. We will see. Anyway, we should move on from all this. Are there any other things in the timeline that we should run through? Here, Elon says they stole a charity, plain and simple. uh they're really on the on the uh for all the jurors they're really gonna have to i imagine you get if it's oakland 12 people on the jury i'd guess like four of them are driving teslas they're really gonna have to go check all the cars make sure they don't have the bought this before elon went crazy bumper sticker on it they're gonna have to throw those those candidates well they should also throw out any jurors that show up in koenigseggs because they're probably in the Kona Seg owner meet up and they're at Cars and Coffee and they're like, well, yeah, I mean, I don't really like, I don't like electric cars, okay?

49:13I do have a bias against electric cars. I need a V12 and anyone who drives a V12, I'm going to give them some leeway. I'm going to give them some leeway. The thing about a Kona Seg owner's meet up is that a lot of them are probably not driving the Kona Seg because they don't work that well. They're not reliable reliable uh frequently it's been frequently reported that they're not quite reliable the one day you get it to start cars and there's paparazzi there taking pictures of you just one of the chances seriously it's the worst one day the worst luck the worst luck anyway century good luck for your systems century shows developers what's broken and helps them fix it fast that's why 150 ,000 organizations use it to keep their apps working.

50:05Elon, a whole Mars catalog was asking chat GPT what chat GPT thinks. And it says, if I were grading this purely as an analyst, not as an open AI model, the documents are legitimately bad. Brockman notes, not ambiguous. I was asking, I was asking Claude about this. And it was very funny because it does, you can't not read into it like you're talking to someone at Anthropic because it's taking shots at both of them being like, oh, well, like, you know, Elon has XAI and that's a for-profit. So he's a hypocrite. And like, maybe that's just objectively true. And the model's just, you know, accurate, but it's funny just reading it in Claude's voice being like, Claude sitting there being like, I don't like either of these companies.

50:51Anyway. one more post Elon said yesterday we talked about this yesterday he was quoting the Kalshi odds on his case against OpenAI and he said I've lost a few battles over the years but I've never lost a war I think what this comes down to though is like what is the same thing as like what is winning the AI race actually look like with China what is Elon winning against OpenAI look like So the odds now, it's 68 % chance will Elon win his case against OpenAI. That's up from 34 % back in January 14th, just a few days ago. And so this new trove really did move the market, at least on Call Sheet. So we will see where this goes.

51:44If the U.S. District Court in Northern California sides with Elon Musk in Musk versus Altman, before January 1st, 2027. Then the market results to yes. Of course, this doesn't take into account appeals, which will obviously happen. And there's always the chance that there's a settlement before this. Although I think it feels like we're just past all that and we are fully in. It's going to trial. There's going to be some drama. There's going to be some courtroom sketches. So get ready because it's going to be beautiful. And I wonder who they're going to call to testify. A whole bunch of people will be on the stand, presumably.

52:18Joe Weisenthal What's he saying? He is highlighting Goldman raising$16 billion in record Wall Street Bank bond sale Joe Weisenthal says right here Toss to me

52:36Finally some good news Huge congrats to Goldman Sachs I've been following them a long time The whole team and culture is so impressive I can't wait to see where they're going and what they do next. Goldman's really just getting started. Today's really just day one. It's day one. It's day one for Goldman. Will Manitis, he thinks of Goldman as kind of the ideal firm, right? He has this idea of firm versus company, where the company is just the people, the firm is the core ethos. And Goldman is kind of the perfect firm. There are some absolute dogs over there. We talked about it a bunch, but one of the most legendary things going into the financial crisis.

53:14They know that real estate is going to sell off. They sell their corporate headquarters and lease it back for 10 years. They're not exposed to the financial risk of their building. You got to be careful if you're on the other side of a deal. If they're selling, why? Why are you buying? That's a good question. But a fantastic firm, fantastic results from them. And congrats. Joe. And speaking of the financial markets, You got to get in on them with public.com investing for those tickets seriously stocks options crypto bonds treasuries and more with incredible customer service. Joe has some more news.

53:50Yes. He's news maxing news out of AP Beijing breaking with the United States. Canada Canada has agreed to cut its 100 % tariff on Chinese electric cars in return for lower lower tariffs on Canadian farm products. Prime Minister Mark Carney said on Friday. Carney made the announcements after two days of meetings with Chinese leaders. He said there would be an initial annual cap of 49 ,000 vehicles on Chinese EV exports, growing to about 70 ,000 over five years. China will reduce its total tariffs on canola seeds. A major...

54:25Okay. Yeah. Canola oil. They're getting seed oils. They're like, we got to have them. Okay. Okay. Maybe this is part of a grander strategy. Oh. He's like, yeah, this is our version of fentanyl. So he actually did. Carney said that Canada's partnership with China sets us, Canada, up well for the new world order. So this is the CEDLs are the first step. Yeah, that was a crazy quote. That was a crazy quote. When I first saw that, I saw the video. I was like, OK, like, yeah, funny, deep fake. Yeah. Yeah. Prime minister saying like, new world, we're excited for this new world order. Yeah. But it was real.

55:09Yeah. Real. In other in other electric car news, Ford and BYD are in talks for car batteries. Let's give it up for some talks. U.S. car makers need, Ford, the U.S. car maker, needs more batteries for hybrid vehicles because it's shifting away from the full EVs. You know, they canceled the Ford Lightning, but they are going to do a lot of hybrids. And so they need a lot of batteries and they're calling up BYD to help with it. Ford and BYD are going to do a partnership, or they're in discussions for it, in which the American car maker would buy batteries from the Chinese auto company for some of Ford's hybrid vehicle models, according to people familiar with the matter.

55:51The two companies are still discussing how the arrangement would work. One idea is that Ford would import batteries from BYD to Ford's factories outside of the US. Some other people said talks continue. It's possible a deal won't materialize. The tie up, if completed, would pair Ford with the largest Chinese car company that has struck fear in much of the auto industry over its ability to produce affordable models that carry sophisticated technology. For Ford, it solves a problem. As the company pulls back from electric vehicles and ramps up its lineup of hybrids, it needs a battery supplier. And BYD is able to produce high-quality car batteries.

56:31We talk to lots of companies about many things. A Ford spokesman said, a BYD spokesman declined to comment. That's a good comment. We talk to a lot of people about a lot of things. Why are you focusing on the Chinese batteries today? Stop calling me Ryan Felton from the Wall Street Journal. Move on. I talk to a lot of people. Before becoming one of the world's biggest car makers in the world, BYD developed a robust battery manufacturing business, including batteries for hybrid models. It currently produces most batteries in China, but the company is building up capacity in its overseas plants as it expands to markets such as Southeast Asia, Europe, and Brazil.

57:13Bernstein Research estimated that BYD's battery shipments rose 47 % last year to 286 gigawatt hours. President Trump's trade advisor, Peter Navarro, criticized the idea on X. He said, so Ford wants to simultaneously prop up a Chinese competitor's supply chain and make it more vulnerable to the same supply chain extortion? What could go wrong here? That's Peter Navarro on X, the trade advisor. Last month, Ford said it would pivot away from making electric vehicles in the face of slumping demand and take an expected$19.5 billion in charges primarily tied to its EV business. So, very interesting news.

57:57Anyway, let's move on to the next big story of the day. But first, let me tell you about Shopify. Shopify is a platform, a commerce platform that grows with your business and lets you sell in seconds, online, in-store, on mobile, on social, on marketplaces, and now with AI agents. Agents. So. And relevant because the day has finally come. Not to see ads in ChatGPT. That's true. But they're coming. They're coming. They're coming. We're talking. We're no longer in talks. We're no longer in advanced talks. We're talking about in the coming weeks. Okay. OpenAI plans to start testing ads in ChatGPT free and go tiers.

58:38It's go time. Go tiers. ChatGPT go. They said, we're sharing our principles early on how we'll approach ads guided by putting user trust and transparency first as we work to make AI accessible to everyone. We got a check in with Mark Cuban. With Mark Cuban. But they say, what matters most? Responses in ChatGPT will not be influenced by ads. That is important. There's a firewall. There's a firewall. Editorial is over here. Ad sales is over here. They don't interface with each other at all. And so the models that generate the responses will not be aware of Who's advertising on what this seems extremely easy to do technically?

59:22extremely It's extremely good for product reliability It's what the consumer wants they want you want to know when you go to Google if you scroll down far enough You eventually get past the ads and you see the real results and you're gonna want that in your LLM even if there's an ad up at the top or in the middle, as long as it's clearly labeled, which they say they will be. So ads will always be separate and clearly labeled. Your conversations are private from advertisers. Plus pro business and enterprise tiers will not have ads. So I'm going to have to downgrade. Yeah. Because I'm on the pro$200 a month tier, but I want ads.

59:59I want to experience the ad product. Everybody's got to downgrade. But I think they're going to be making way more than$200 a month off me on ads. Oh, yeah? Oh, yeah. Oh, yeah? Oh, yeah. They're going to be advertising crazy stuff to me. They're going to be, every other result's going to be Koenigsegg. Here's a McLaren F1 in your area. You can go pick it up right now. It's going to be insane. Yeah, so people have been so concerned, specifically Mark Cuban. We obviously had him on the show to talk about this last year, about this idea of ads showing up in the results. And part of it, the reason I was never that concerned is, like if I just search best backpack for men, which is kind of a joke in itself.

1:00:39It's going to tell you Ridge. Well, a man shouldn't wear a backpack in the first place. So sorry to any backpack super fans out there. Tyler. Well, you're not a man yet. You're not 21. Oh, true. So you're good. Enjoy your backpack. Yeah, you should switch to a stainless steel briefcase that you handcuff to your hand. That's the correct thing for someone of your stature, someone of your importance. You've got important stuff carrying around. But anyways, when I search best backpack for men, I can scroll down and find a Reddit result. It's the second result after Nordstrom. But they also serve me a bunch of ads.

1:01:17I don't assume that the best backpack for men is the first ad. It's not like when it's clearly labeled and separated, I just assume this is an ad for somebody that sells backpack. Yeah, but now I'm aware of that particular backpack. Yeah. Maybe. So I think it's fairly aligned, right? Yeah, that's why I don't search like best backpack for men. I search best, highest margin, highest margin backpack, men's backpack. Because you appreciate from the perspective of the business. Yeah. Of the business. You want to support business. Yeah, and then, so I want to know what's the highest margin backpack for whoever I'm buying from.

1:02:06Then I go to shopping tab, and then it tells me, okay, this is the highest margin one. If you want to support that corporation, make sure the most number of dollars flow to their shareholders, this is the backpack you should buy. Back forward slash in the chat says he uses a Spider-Man lunch box. Let's go. See, I can appreciate a Spider-Man lunch box. And soon you'll be able to generate a Sora of Spider-Man in ChachyPT. It'll serve you an ad for that Spider-Man lunchbox. And the virtuous cycle of commerce will continue. MongoDB, choose a database built or flexibility and scale with best-in-class embedding models and re-rankers.

1:02:48MongoDB has what you need to build. What's next? Anyways, so Plus Pro Business and Enterprise tiers will not have ads. Yes. So they need to hammer, hammer, hammer the first bullet point here. Responses in ChatGPT will not be influenced by ads. We know that this is what they're going to do. Technically, it makes sense. It's completely the industry standard. Everyone in tech knows that this is how it works. But consumers will have all sorts of conspiracy theories. I mean, remember how big the whole Facebook listens to your audio and then targets ads off of that. And there's been so many viral conspiracy theories around how ads work, how monetization works on these platforms.

1:03:37There's going to be a huge amount of just noise out there basically saying that ChatGPT is polluted by ads, overly influenced by ads. And so they need to beat that drum really, really hard. It's good that they're doing it on the main account. It's good that they're doing it with Fiji Semos post. But they're going to have to emphasize this in comms on every podcast that they do, every announcement for a long time. Yeah, people seem to be really riled up about this. I'm seeing a bunch of comments on the post. They're upset. I don't get it at all. The whole point is that ads... Ads have made it so that wonderful services on the internet have been free for decades.

1:04:20they're generally aligned even target people report they like targeted ads it's annoying getting an ad that's not targeted like why are you wasting my time I agree this is funny, somebody in here says some poor 5th grade teacher grading the worst World War 2 paper ever turned in when it suddenly starts talking about World of Tanks and North for no reason just copy-pasting the ads honestly, maybe that's a feature though, you know, because you get two impressions. Two impressions. Yeah, there's going to be some hilarious hallucinations where someone's reading from their press release and they accidentally have an ad in there.

1:05:01Oh, that's going to be amazing. I love it. I love it. Well, the Wall Street Journal had a little bit of an overview on this. OpenAI to begin testing ads in ChetGPT and push for fresh revenue. The company will be showing ads in the free version of the chatbot as well as its cheapest subscription. Let's give it up for fresh revenue. Fresh revenue. So ads will appear at the bottom of ChatGPT's answers and be labeled. Interesting. The bottom of the answer? I want it at the top and the middle and the bottom. And I also definitely want to show the ads while it's thinking. Because you can run an ad auction and show an ad while you're waiting.

1:05:37I mean, Uber's done this while you're waiting for the car. There's no reason why that shouldn't be there, especially if you're doing a deep research report or anything. I do wonder how the ads will surface in the audio summary. So if you fire off a deep research report, it throws an ad at the end. If you're listening to that audio, will that play the ad at the end? Will that sound like a podcast ad or will it sound like something else? Like, will it be clunky? I'm assuming that they're not doing audio in these, right? Imagine if you do a Google search and it pops up a video. It's like trying to play sound.

1:06:11It's annoying. It could be. You'd be throwing your laptop. Yeah, yeah. But I mean, a lot of times people will go to ChatGPT, fire off a query, and then when the query comes back, they will just hit the play the audio button. And will it read the ad as well? I don't know. Maybe. But the company said ads wouldn't influence the chatbot's answers, of course, and that user conversations wouldn't be sold to advertisers. That also makes a lot of sense. Also, advertisers don't really care. Well, to be clear, my understanding is they will use what they know about you to offer better targeted ads to advertisers.

1:06:50They're just saying they won't explicitly be like, hey, we have here's this person's email and here's what they like and actually sell that specifically. Or even anonymous. There's not even an option to go to OpenAI and say, hey, I'd like to buy 10 ,000 conversations about backpacks. That's not an option. What you can say is you can say, I have a backpack. Find me some customers who want to buy backpacks. And they will say, sure, we'll do that. and we will go into the black box and sort and match you and give you the lowest cost per acquisition. You tell us that you're willing to pay$5 per customer,$2 per customer.

1:07:31We'll get you$2 customers. We'll do our best. This is the auction model. Yeah, and the thing here is if you are a business owner, you should be incredibly excited about this. Every time a major new ad platform has come out, Yes. It is an opportunity to drive a tremendous amount of growth really, really quickly because they typically under price, under price the ads in the short run, in the short run just to get volume, just to get a lot of customers in. So when Sean Frank comes on later, we can talk with him about this. I remember early days of Snapchat advertising. Connor over at the Ridge was getting tremendous results.

1:08:17Cisco. Cisco is critical infrastructure for the AI era. Thank you to Cisco for powering TDPN. Dean Ball is back on the timeline. We know Dean Ball. We know Dean Ball. He says, flowers, no way, reacting to the screenshot, that flowers hinted math before numerals. Pottery made by people of the Halifian culture who inhabited northern Mesopotamia between 6 ,200 and 5 ,500 BC painted flowers with 4 ,8, 16, 32 petals. Some of them have 64 petals. They were obsessed with exponential growth. They were obsessed with compounding, the power of compounding. It's the Claude logo. It is the Claude logo. That's hilarious.

1:09:08I love it. It's amazing. Anyway, Graphite. Graphite.dev. Graphite. Code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster. Where would you like to go next? Vic has a fantastic post. He says, had dinner with wife at a Mexican restaurant last night. looked at the menu. They were trying to raise prices from$18 to$24 for her favorite entree. Wife was like, I think we can have Claude make this. Told waitress, trying to gouge us, they done. One week sprint. Claude cloned and replaced. Cochinita Pibil and Carnitas. Restaurant manager freaks out. How do we solve this?

1:09:50This is going to happen so much in 2026. Just telling the restaurant owner that you're going to do it at home We cloned your entree with Claude. We cloned you. We cloned it. We cloned it. I just cloned it. This is pretty funny. Jeff Huber sharing this post from r slash teacher. Every year these kids come back with a new annoying quirk, dot, dot, dot. Claude boys are apparently the new thing. In my 10th year of teaching, mostly freshmen, ever since the pandemic, there's always a new thing students bring to school that they learned over the summer from the internet or wherever. The newest thing here is a flock of self-proclaimed Claude boys who carry AI on hand at all times and constantly ask it what to do.

1:10:34They have their entire personality revolve around Claude prompting an AI. When we went around doing an icebreaker, four of the five kids, some variation of I, lived by Claude and died by the Claude. That's their fact. Just about an hour ago when I assigned the first assignment of the school year, one of the Claude boys was bold enough to say, if Claude says I do it, otherwise I don't. I told him if he asked Claude, he would be getting a call home on the first week of high school. He asked it anyway, and it said to do the homework. That's amazing. This is a copypasta from the Coin Boys. Yeah, it's the same thing, except instead of asking Claude, you flip a coin.

1:11:09You flip a coin. The Coin Boys are apparently new things. This is very funny. I like the copypasta. This is amazing. Very, very funny. Well, you've been trying to make Jemmy Boys happen, which I think is underrated. We should get a copy of the pasta whipped up, of course, because we are sponsored by Gemini 3 Pro, Google's most intelligent model yet. They've got state-of-the-art reasoning, next-level vibe coding, and deep multimodal understanding. I have some advice for Jeff Huber. He's the founder of Chroma, and he's trying to change the world. And if he wants to change the world, he should raise capital at the New York Stock Exchange.

1:11:50Jeff, it's that simple. Head over to the New York Stock Exchange. When you want to change the world, you raise capital. Call Lynn Martin. Call Lynn Martin. Let's get into the mansion section. Let's get into the mansion section. I've been waiting all day. The first big piece is about the great property transfer. Gen Xers and millennials will inherit trillions of dollars of real estate over the next decade. And it's already reshaping how luxury homes are bought and sold. So over the next decade, roughly 1.2 million individuals with net worths of over$5 million are projected to pass down more than$38 trillion globally.

1:12:30That is a huge amount. Real estate is poised to play a significant role in the great wealth transfer. Gen Xers and millennials are set to inherit$4.6 trillion in global real estate over the next 10 years. And this is why it's so important to win the genetic lottery when you're born. We didn't. We didn't. I won't be inheriting anything. But stop making excuses and maybe get a time machine and go back in time and be born to someone with a luxurious state that you can inherit. Maybe even multiple. We already read about that. So M6. Let's kick it over to M6. What is that? Parents are buying earlier and bigger.

1:13:12It isn't a new phenomenon for well-off parents to give children a helping hand in securing a first home. But at the high end of the market, agents say more parents aren't waiting for kids to inherit their wealth. They are buying them luxury properties sooner. It's reshaping the definition of luxury real estate as more sellers cater to the tastes and preferences of a younger generation. The price points have gone wild, said Ian Slater, a Compass agent who works with ultra-wealthy families in New York. I used to commonly see people buy 3 million to 5 million apartments for their 25 to 30-year-old kids.

1:13:43Now I see them buying 15 million to 30 million dollar apartments for their kids. That is absolutely crazy. When you're buying for children, co-ops are a real no-no. Since many co-op boards want the occupants of the units to be financially independent. By contrast, condos offer flexibility, which is especially valuable for global areas. What makes you more financially independent than having a trust fund that just... Yeah, I don't know. I mean, I guess if you're a co-op, you want to just be around other people who aren't nepos or something. You want people who have built their own businesses or something.

1:14:26I don't know. I won't... Yeah. I won't stand for Nepo slander. Yes, yes. Seriously. I love when a Nepo utilizes every available resource and just absolutely dominates. Yeah, it's beautiful. You got to appreciate it. Well, let's look at Gene Hackman's Santa Fe compound, which just listed. Wait, how did Gene Hackman make his money? Actor. You didn't know that? No. I don't. You were joking. I don't. This guy was born in 1930. I guess. He's amazing. He's a Hollywood icon. He's not a hack. No, he's not a hack. French Connection, Unforgiven, Superman, the movie, Enemy of the State. That's where I know him from.

1:15:15Enemy of the State's a great movie. He's in The Firm. Guys, is this on the movie list? He's in The Firm. Are any Hackman movies in the TVP and chat movie list? Behind Enemy Lines, that's a great movie about a downed fighter pilot or spy, U2 spy plane. I think it's Owen Wilson who goes down and crashes and needs to get back to his team, to his side. It's a fun ride. I recommend it. So nearly a year after the shocking deaths of Hollywood icon Gene Hackman and his wife, Betsy Arakawa, Barcawa at their compound in Santa Fe, New Mexico. The couple's 53-acre estate is coming on the market for just$6.5 million.

1:15:59I had no idea that Santa Fe, New Mexico was so affordable. We were looking at a place in Nashville, which I would put at the same level as Santa Fe. John, I got to stop you there. Why, explain. I mean, saying that, is this a bit? Saying it's so at$6.5 million, it's so affordable? It is. I mean, we looked at a three-bedroom apartment that was like$25 million in Nashville last week. This is 13 ,000 square feet, 53-acre estate, and it's one-fifth of the price. I'm just saying, like, on a relative basis, do you go Santa Fe at six and you get 53 acres? Or you're living next to John Feo and you paid 25 mil?

1:16:43Which one would you go for? Pick it. Pick it. Which one would you go for? I'm going Santa Fe all day, personally. Hackman, the two-time Oscar winner. Let's see. There will be some buyers who are just adverse to purchasing property, blah, blah, blah. Let's move on to Washington, D.C., where the prices are way up, way over$6. $6 is low compared to this$28 million home. It's D.C.'s most expensive home, and it's now owned by the owner of the Washington Commanders, Joshua Harris. And this is one of those interesting things where you can build a massive private equity firm, a massive financial institution.

1:17:26Josh Harris, of course, built Apollo, the co-founder of Apollo. But he's known as the owner of the Washington Commanders, even though that's not like his life's work. It's an interesting thing, but when you buy a sports team, that's the thing that follows you around forever. And so the billionaire investor Josh Harris, the owner of Washington Commander's NFL team, and his wife Marjorie Harris have paid$28 million for a storied Washington, D.C. landmark with plans to return it to its original use as a single-family home. The sale is a record for the city. It's known as the Halcyon House. The federal-style building in Georgetown dates to the late 1700s.

1:18:07The roughly 30 ,000 square foot house, which has a large garden with a pool in the back, wasn't publicly listed at the time of the sale. And then the deal happened quietly off market. Would you live in a house like this, John? 100%. I knew it. Absolutely. Would you not? It's amazing. This style of architecture is extremely depressing to me. What do you mean? This is like the Lindiest thing. You know who lived here first? This was built for the first secretary of the name. No, I understand. Benjamin Stoddard. You are inhabiting. I understand the appeal for you. What is depressing about this? This is a beautiful house.

1:18:47Is it just not enough light? I don't know. Potentially, if they did a pretty hardcore remodel to it, really brought it. What's not to like? It looks beautiful. I don't understand. It just looks a little too vintage. Ridiculous. It does have a crazy, like, there's a roof inside. And it looks like a frat house. Do you see this? It does kind of look like a frat house. Look at this. Why do you have a roof under your roof? Do you see what happened here? They got roofs under the roof. It looks like they just built another house, an addition on top. And so there is some chaos here. It's a little too fratty for me.

1:19:26It looks like the floor would be sticky. The floor is sticky. Somehow I don't think this$28 million property has a sticky floor, but you never know. They have three TVs up there. That's sort of fratty. You watch three different football games. Well, if you own an NFL team. Yeah, you got to watch all the games. Yeah. The market. The previous record in D.C. was set in 2024 when Cantor Fitzgerald CEO Howard Lutnick purchased Fox News anchor Brett Byer's Fox Hall estate for$25 million. Mark and Hunter McFadden of Compass represented Dr. Kuno in the Georgetown sale. Harris was represented by Sotheby's International.

1:20:04Figma. Figma Make isn't your average vibe coding tool. It lives in Figma, so outputs look good, feel real, and stay connected to how teams build. Create code-backed prototypes and apps. I hate to interrupt the mansion section, but we have to pull up this video from the Houston Texans. Let's do it. They are, their new video is Call of Duty themed. No way. And I just got to wonder. Oh. We created COG. Yeah. Yes, 100%. We created the last thing. Well, let's switch over to reacting to the latest video from the Houston Tech. It is in the timeline. And while we pull that up. Here it is. Here it is. What is this?

1:20:44Can we get some? The match begins in 4-3-2-1. Eliminate the enemy team. And Kirk with a diving attempt. And he's got the catch. Oh, the hit marker. The hit marker sounds good. Long shot. It's so funny because we were just talking earlier this week. We did the hit marker sound effect. Hit marker sound. Air raid. Air raid. Kick. What sound effect? UAV online. Another hit marker. UAV online. That's us. That's not them, right? That was you? We got to figure out if they did this before. The graphics package. I didn't realize how iconic the text was in COD. That deep glow right there, the glowing letters, that's just COD to a T.

1:21:35They really nailed that. That's very fun. I wonder what inspired that. Anyway, let's move on to the last story in the mansion section. But first, let me tell you about Railway. Railway simplifies software deployment. Web apps, servers, and databases run in one place with scaling and monitoring and security built in. So this is a renovation. I want your take on this renovation. A Texas couple transformed a 945 square foot room into a space where they could watch sports and entertain friends. See, this looks fantastic. Thank you. This is what I want to see. This is what I want to see on the last place.

1:22:13I want to see this level of reno to the last property. But I feel like this does capture some of the opulence, some of the details, some of the texture. This is not flat white walls. There's lots of layers. There's trim. I like the blue, the way it goes, the brown, the gray couches. Everything sort of like harmonizes, but it's not bland. It's not flat. There's a lot of detail and layering there. Even behind the TV, you see a sort of like textured backsplash. I think they did a fantastic job on this. and they spent$465 ,000. So when financial consultants Damon Cronus and Julie Cronus decided last year that it was time to renovate their game room, they wanted to keep it fun for their two teenagers, but add a layer of sophistication.

1:23:02You always got to do that. They're in their early 50s and they drew inspiration for their new entertainment lounge during a visit to Manhattan and a stop by the UBS Arena Preview Club, a temporary showroom for the arena's premium space under construction. Styled in vintage hues of rich gold and deep blue with dark brown wood, the space had just the upscale feel they wanted for their home. They brought in pulp design studios and they transformed it. The owners wanted it to look like a lounge, like a private sports club. We took the inspiration and created our vision to suit their needs and lifestyle.

1:23:36The 945 square foot space was designed for versatility, allowing guests to enjoy different activities at once. It was tailored to host movie nights, lively cocktail parties, and Sunday football gatherings. The room had to have maximum flow for people to move about. And so you see they have a horseshoe couch, but there's a gap in the middle so you can flow around either way. So it's kind of like two L couches nestled next to each other. Very, very smart. Two L's. Almost making a W here. The custom two-piece. I like it. The custom two-piece sectional sofa set them back$26 ,000, and it allows guests to move in and out of the TV area.

1:24:14See, you can tell the floors are not sticky here. Yes, definitely. Which is key. The room is narrow, and we didn't want to have a sectional that you had to walk all the way around. And so for additional seating, Ottomans pull out from under the coffee table. Swivel stools accompany console tables beside the sofas. Tall stools are tucked under the nearby bar top, and custom banquet hugs a curve of windows. Performance fabrics and durable materials were chosen to endure active crowds. They're not letting the beer soak the walls. They got performance fabrics and durable materials in there. They're not worried about a sticky floor.

1:24:51A sticky floor here, a sticky floor there. It's going to be okay. They're going to mop it up. Those performance fabrics are built to handle the stickiest beer. maybe even the athletic brewing. Some athletic brews. You're chugging athletic brews. You're spraying them all over. You're cheersing, slamming them on your head. And they get on the performance fabrics. Honestly, so athletic brewing, which we have a few here. Athletic brewing really needs to do a hard sell to Cheeky Pint and be like, look, you guys are hardly drinking. We can tell. We'll have to ask him if he uses Stripe currently because maybe there's a one hand wash with the other deal there.

1:25:29Yeah, yeah, yeah. We can tell you don't have it in you to get hammered on the podcast. But we want you to be able to enjoy the taste of beer. Yeah, I like that. A little cheeky athletic brewing. I like that. Cheeky brew. Beyond watching sports across two TVs, the family enjoys competition of their own. An oversized scrabble board hangs on the wall to set the tone. They spent$25 ,000 on a table. Who knew tables can get so expensive? We've been looking at tables. I mean, I feel like we did. We're just trying to get a new table. I know. We want a bigger table. We want a slightly bigger table. We want some features, but very quickly you get up into these big numbers.

1:26:03So the round Cambria table was crafted from man-made quartz with a metal base. We bolted the base to the floor because Damon and Julie wanted to be sure it wouldn't topple around the teens if people are getting hammered down there. Standing on the table, they don't want it falling over, so they bolted it to the floor. The chandelier set them back$3 ,900. The Cavalino X-Large radial chandelier from Visual Comfort adds sculptural interest and focus lighting to the game table. Chairs,$2 ,400 each, covered in Everly fabric, having coasters. Casters allowed them to move freely, perfect for group activities so you can move them around.

1:26:41The window shades,$12 ,000. Motorized Roman shades create a blackout environment that's ideal for TV watching. That's a cool feature. And the banquet cost$27 ,000. The bottom is covered in ultra suede green and ochre from Crovat. And the back in Lulu Velvet is paprika from S. Harris. We had to do a custom banquet so that it fit the curve of the window to maximize seating. They're all about maximizing seating over here. Well, congrats. And also, fin.ai, be a number one AI agent for customer service. If you want AI to handle your customer support, go to fin.ai. What other news do we have, Jordy?

1:27:20Tell me. Harmonic. Harmonic. Vidya is investing in math. How much? They're investing in Vlad's AI startup, Harmonic. They're investing$120 million. We're going to hit. Emerson Collective is also joining as a new investor alongside existing backers, Ribbit. Ribbit. Ribbit. We need a frog sound effect. Mickey Malka. Mickey Malka is doubling down. Or Mickey. He's a big Robin Hood guy. Sequoia Index and Kleiner. Ooh, Sequoia. Yeah, we'll have to talk to the Sequoia folks about how Harmonic fits into the post-AGI age. Because we have Pat Grady and Sonia. What if it comes out that Harmonic is just working on a Chad Chibut-style app to help kids do math problems?

1:28:07That's it. That's the play. Yes. It's a trillion-dollar opportunity. Yes, I love it. Well, we have our next guest in the Restream waiting room. Let me tell you about Label Box while we bring him in. Get in the box. Real reinforcement learning environments, voice robotics, evals, and expert human data. Label Box is the data factory behind the world's leading AI teams. And we have nice and cold athletic brewing hazy IPAs. I will enjoy one of these while we bring in our guest, Bill Schufelt from the Restream Radio Room. Welcome to the TV Fan Ultron. How are you doing, Bill? Incredible to be here.

1:28:47Love the show. Thank you so much. even though I'm East Coast and not in the AI world, you know, I just sell things that hurt when you drop them on your feet over here, but I'm a huge, huge fan of the show. And so I'm excited to be here. Yeah. So, so great to have you. I was just trying to remember, uh, I listened to like about 90 minutes of you on a podcast. It must've been like four or five years. I listened to you on, uh, Logan Bartlett. That was a fun one. Yeah. He's great. Yeah. Yeah. He's awesome um what do you think about the pitch of getting are you familiar with the cheeky pint podcast from stripe have you seen this i i love it yeah okay so here's the actually thing about cold emailing them to get so the elephant the elephant in their pub is that they're not really crushing beers they're not getting actually drunk so i think you should make a special addition uh for the team over there so they can really start start uh indulging oh i i love that I'm a fan of them, Ireland.

1:29:46I've got a bunch of their books. So yeah, I would love to reach out to them. Fantastic. Let's take it back to the beginning. Since it's your first time on the show, give me the founding story. Give me what you were doing before. It's such a huge trend now, mostly thanks to you. Did you imagine that the trend would become so big? What were the inciting elements? Yeah, I think there's both the personal story and the category story and the health and wellness mega trend behind it. Even zooming out to the trend, I know so much of the mega trends in business these days and big numbers around AI. And it's rare to see something that is so solidified, very visible long term.

1:30:30And I can get more into how big I think the trend will be and why it'll be so big over time and why I've got line of sight to that. But my personal story really is I was working in the hedge fund world, had a great finance job. I came, I went to like, went to college, exactly. But I like went to college with that view, went right into it in my career, never thought I'd do anything different. But I was living this high performance lifestyle where, you know, 12 years ago I was about to get married. I was thinking about my future life and alcohol was kind of a ceiling on my work productivity, my relationships, my sleep, my workouts.

1:31:14You know, all the things that are like so commonplace in knowledge now that like, you know, all the fitness wearables, the podcasts like Peter Atiyah, Huberman, Rhonda, like all these things are like so evident now. But I was like living that and feeling it myself. I was, you know, I had a very intellectually intense job at Steve Cohen's hedge fund where you're graded on your output every day. I was at the desk from 6 a.m. to 6 p.m., 150 work dinners a year, then out with friends and family on weekends. 150 work dinners is insane. One for us, I mean, we have like a maybe not as intense as your schedule with Steve Cohen, but it's very, you know, we're here.

1:31:54We have to get to the studio at 9. We go live at 11. There's very little downtime. And so we do one work dinner and it throws everything off. It's like throws off the whole week. I have huge admiration for your always on. Like I think about that all the time. But yeah, we had 200 brokers and I had to do research dinners also. And you have to have good relationships. So when you like pick up the phone and need to convey something in 10 seconds that you know that person. So anyway, that's a long way of saying like I was very interested in health, longevity, my fitness, my performance. way before it became such a commonplace trend, and I was living that.

1:32:30And I really started saying to people, I really wish I could go out to all these dinners and be in the moment without alcohol. And I saw the trend happening over my finance career where people were just as excited to go to Barry's Boot Camp or something in the morning rather than go to a steak dinner at night. And so, but I just really wish there was another side of the road. Keith or boy, four Barry's Boot Camps in a day. So why not Shirley Temples? Why didn't you just become the Shirley Temple king drinking six of those? Like, what were you doing before you actually decided to build a new product in the category?

1:33:05Were you on just like soda, water, and lime? What were the options? Well, so I saw that huge impact on my life just by like moderating my drinking and ultimately stopping my drinking. And then thought, you know, if I could just bring moderation to the masses and just make it accessible. I have no view on bringing back prohibition, sobriety, or anything like that. But just making the other side of the menu delicious, match the culinary expectation and the experience. And as I started to look into the numbers behind it, 50 % of adults even at that point had 0.1 drinks or less per week. And it was really only the top 10 % of adults had more than seven drinks a week.

1:33:45And it kind of hit me that the adult beverage world is also just totally missing a majority of the population. and their needs. And so I'll skip like the years of business research and surveys and stuff like that to get conviction around it. But ultimately it was like right out of my lived experience and I wanted it to exist and quit my job to build that. How much, how much credit do you give just how good the product is with the success of the company? Because I think in CPG, you're not the first finance Chad to like wake up one day and say, I want to start it. Like, it's, it's kind of, it's kind of a common path.

1:34:24Like, you know, they're kind of like bored. They're, you know, they, they, they try some drink. Like I've met as an angel, I've met so many people over the years that like are trying to make the leap from finance to being an entrepreneur and CBG, because it's like, you're going from like spreadsheets to something that's like super tangible and you still have to use, you know, spend plenty of time in spreadsheets. and I think there's something really appealing about like going from like finance to like the tangibility of CPG but the issue is like so many people that are incredibly talented they just end up like they end up developing a product that's not amazing and then it just you can't even live up to your talent level or your potential because you're just like handicapped by the product and when I tried you know having this product even when I had it for the first time it's like it's just a fantastic product it tastes exactly like a beer it tastes like a sort of a cell you know celebratory moment or it tastes like you know winding down at the end of the day which is like it delivers something unique in the way that like a la croix or just some other like non-alc you know alternative actually works yeah that is exactly it and we have an incredible co-founder john walker and the two of us homebrewed on gatorade jugs and we took it down to the screws and reinvented the way non-alcoholic beer is made, rather than where historically non-alcoholic beer was a very industrialized, heavily processed thing.

1:35:45It was only a lager category as well. And so we took the time, and John said he wouldn't even sign on if I didn't agree that we wouldn't launch the beer if it wasn't indistinguishable from top craft beers from day one. And I think that's the huge differentiation. We made the very unattractive decision to investors to take a capital-intensive route. We've built all our own breweries. We've put over$130 million into our manufacturing. Just very unattractive to investors, if I can repeat that. Attractive now. Yeah, but it was so we could own our proprietary process, own the quality of what we were doing every step of the way.

1:36:27Where, yeah, as you were saying, if you were to nick surely a lot of CPG, or especially the competitors in our category, like you would find that they just totally outsource all the production there aren't brewers on the team and it truly had to be different quality and different experience and we really invested to change that perception of the category and that's still very rare in non-alcoholic beer there have been yeah 280 brands since we launched and a lot have had very disparate futures well yeah and what's happening with the category right now broadly because i've you know I've, again, as an angel in LA, there's so many CPG founders.

1:37:03I've seen so many of these companies try to come up, not just in kind of the non-alc beer category, but in other categories. And one of the things I always struggle with as somebody who, you know, dramatically reduced my drinking probably like six years ago from like drinking like a handful of drinks once a week to like drinking like one or two drinks, like once a month, really, like so really limited it. And so I'd get these pitches and people would be pitching like something that doesn't taste like alcohol at all. And I'm like, I, I, I'm the buyer here, but you're selling me a product that I can, I can drink a diet Coke.

1:37:44I can drink any number of other things. And so like your competition is just, is not like the non-alc category for some of these new non-alc brands, their competition is just like every single drink that doesn't have alcohol in it. And that just felt like a super tough physician. Whereas for this, it's like, if you want a beer, but you don't want alcohol, like your options are like much more limited. Yeah, for sure. And as, as I looked at the TAM of the market, the way I thought about it was a 50 % of the people really aren't drinking and 99 % of people are even people who do drink are not drinking like 99 % of the time they're awake.

1:38:21So there's a huge opportunity to meet new occasions. And those occasions you just described losing over the years. We're trying to bring those back, bring the social connection back, get people out multiple times a week. But I think that like the clearest way I think I could say this isn't like why I think it's such a big trend is, you know, the overall beer category per the Brewers Association is north of$100 billion in all retail sales. Craft beer itself is$28 billion, supposedly per the Crap Brewers Association. If you go back like 50 years, light beer was one of the biggest megatrends to ever hit beer, and that is an enormous part of the beer category today.

1:39:06But that was a huge trend up until about the turn of the millennia. And then we started to see like people, well, the light beer trend was driven by the nutritionals, where people wanted superior nutritionals to what existed previously. Then you had big trends in craft beer and flavored RTDs, which were, you know, light beer might not have that flavor. We want to like have the nutritionals, but the flavor and where non-alcoholic beer comes in is all the experience, superior nutritionals. It's about 20, 30 % of the calories plus all the flavor of great craft beer. And it's emerging in this new big mega trend that I think has really good line of sight to be the next huge thing in beer.

1:39:44And I'm a very delusionally optimistic person, but I do think non-alcoholic beer will be significantly bigger than craft beer and a huge part of the overall beer category in the future um yeah where where else in uh beer is there like meaningful growth well that's also the problem too is every other like most cpg overall but especially bev elk all the innovation is like a one-for-one substitute for the same consumer in the same occasion and and even some of them are like you're substituting one thing for many like a thc drink is a one one for like six substitute like um and so what we're doing is we're bringing in like a lot of people drink our beers within the same night as alcohol our 80 of our consumers drink so we're bringing a lot of people back into the category plus 25 of our drinkers are new to beer altogether.

1:40:39So this is really additive to the beer world for the first time. Yeah, and then as you think about those occasions, there's just a lot of occasions to layer in. I guess on the broader beer category too, the last few years have been the worst beer years of our generation in alcohol for a number of reasons. I think part of it is the category is not connecting to the next wave of consumers. If you look across the spokespeople in the major brand TV ads, most of them in their 70s, one of them's in their 80s. One brand's about to bring back an 87-year-old spokesperson from 10 years ago. This is not the cohort we need to meet.

1:41:24We are bringing beer for the modern, healthy, active adult. We have great aspirational athletes and chefs and people behind it. But more than anything, we're just trying to build a really timeless brand. And I could share more about our marketing and how we're thinking about that differently too. I want to get to that. But first I want to know about how is your business different or similar to other beer companies that are higher alcohol? Do you need to be 21 to buy this? Are you subject to the three-tiered system? Can you sell this online? What's different or similar to just higher alcohol beer distribution in sales?

1:42:02Yeah, I mean, as a finance chat coming into this category, I was like, wow, there is no tech in this sector and there is margin everywhere. And I very quickly got on the ground and realized that there's a lot of reasons why beer distributors are great at what they do. It's incredibly capital intensive. They're basically a big logistics business. And so we do largely plug into the three-tier system, but that's our opt-in. And then - So legally you don't have to necessarily because you're regulated slightly differently. Correct. And then we were the first beer brand to launch nationally on D2C also.

1:42:38And so that was like an absolutely enormous marketing advance where, you know, any high-flying brewery of the 2010s, you'd have to wait in line outside of their brewery for a limited release in the locale where, you know, we could, we could lose tons of money shipping beer all across the country. So is there any age restriction? Not in most states now, but yeah, typically it's, you know, we, we're trying to establish it for adults in typical beer occasions, celebrating life's special moments. I'm mostly trying to catch Tyler because, uh, I saw him grab one. And if, if there's some rule that he's violating, I'm going to call the cops right now.

1:43:18I mean, we are huge on the Michigan campus. Okay, there you go. I mean, some of the colleges across the country were enormous. This next generation is so much more moderate than prior generations. And I think the perception of alcohol has changed something like 14 percentage points in three years, which, you know, it's happening really fast. Talk about the formulation process a little bit, not to go too far back, but a lot of folks, when they're doing a CPG project, they go to a co-packer and they get a dog and pony show. And they come out and they're like, we formulated the perfect thing for you or flavor house or some supplier.

1:44:00They tell you this is the best thing possible. And then if you're not sharp, sort of the dog and pony show can kind of overpower you. Was there any point where you were working with a manufacturer and you had to sort of override them? Or what was the initial scale up like? It's really been all internal since day one. Yeah, I basically put my life savings into our first 10 ,000 square foot warehouse, Gatorade jugs, me and John's salaries a bit. And then raised an angel round to build that first brewery. Did 120 meetings, a huge rejection percentage. But in that, we wanted to approach it in a totally different way.

1:44:47And so we're confident we arrived at that. I feel like consumer is the one category where you just really don't know until you start selling the product. Like I can easily imagine getting this pitch and just being in the mindset of like, like if I'm if I want to drink a beer, I want to I want alcohol. And it just turns out it's like totally just totally false. Right. I wanted to ask, I know you did a ton of community IRL marketing early on, and I wanted to ask you about how you advise other consumer brands when they're early on and thinking about how to spend their time, what channels to really focus on.

1:45:33Because for certain consumer categories, if somebody's like, yeah, we're going to really focus on events and community, I'm just looking at them being like, you should just sell the product. Just sell the product. Spend all your time selling the product. Whereas you guys, I know, invested a lot in the running world and saw a ton of great results. Yeah, I mean, I invested a lot of my time in weekends. Yeah, like every weekend. Yeah, we had no more resources after all the stainless steel to put into marketing, so that was all on us. But just touching quickly on two things you said there, where a lot of people both, A, don't contemplate non-alcoholic beer.

1:46:15They're almost like non-considerers and then think they don't like the taste of beer are two things we combat over and over again because it really is hard to imagine non-alcoholic beer and the social experience until you actually experience the social experience without alcohol. And you're like, oh, I kind of really like this. I'm actually in it for my friends, the food, the moment, and not the alcohol that mostly hits you after the moment. That was like a big light bulb to me. And the thing we hear all the time is people say, oh, I don't like the taste of beer. Well, actually, the ingredient a lot of people don't like in beer is the ethanol, which at times is up to like 80, 85 % of the calories of beer.

1:46:53Like in a normal light beer, the ethanol calories is more than 80 of the calories of like the 95 to 100 calories. And so we actually do have a pretty big flavor advantage too. And like no comparison, large brand.

1:47:10in real life too. That was almost out of necessity. I was going around our early grocery stores like Whole Foods and nobody was even looking at the non-alcoholic beer shelf. And so I was like, okay, I've got to get out in the world and bring people in. And so I was running a lot of races and trail half marathons and Spartan races anyway. And so I just started pouring beer at those, giving out hundreds of beers a weekend and getting people to the actual table and trying the beer where they're happy, sweaty, thirsty, and actually considering it. How do you measure success with an initiative like that?

1:47:48You're pouring, you're sampling with people, you're seeing their reaction. Are you already at that time stocked in shelves in local stores and then you're trying to see, hey, if I do a bunch of weekends back-to-back, Am I going to see an uptick in sell through? Or is it more just like, this is like, I'm basically spending my time to get more people aware of the brand and I have more time than I have capital. Yeah, we have down to zip code and store level sales. And so I was able to see like ripples fairly quickly and could run regressions on that. Separately, like it has to be authentic to the brand proposition.

1:48:29Everything about this brand, the can, the mountains, the outdoor aesthetic, the taste, was all out of me and John's life, and primarily my life. And so where I was going to market the brand and talk to people, I didn't have to go to a focus group or Google where people like me hang out, because it was all built around me. And so I have seen a lot of people try to bring non-alcoholic spirits to races. It's an absolutely laughably terrible fit. But you can tell some VCs just pushing someone up to... Yeah, they heard you on a podcast and they're like, yeah, let's just run and do the same thing. It'll have the same effect.

1:49:07And a lot of brewing for Manhattans. We did also take a very intentionally long track with things too. So like that, like, yeah, talking to someone face-to-face is very unscalable and like doing things digitally is so much more scalable in awareness. but if you do it enough and if you're doing it all the time all of a sudden it is scalable on a long time horizon and even up until this year, last year we probably did 2 ,500 events as a company and handed out well over a million samples probably in all 50 states and three other countries and so those things do scale over time and then we've been really consistent with our marketing messages too I knew my core skill set isn't like me being a social media persona and just turning on the cameras all the time, we wanted to create a timeless brand that just compounds over time.

1:49:58We talk about the same things over and over again. I think of it somewhat as a Red Bull model of advertising consistency where most of CPG these days is chasing that next hit. It's like you've got to be on the next trend. The next video has got to be a home run. Or even shorter duration like celebrity trends. You know, we've had upwards of 10 celebrities enter our category. We have a whole range of superheroes at this point. And it's so funny because you probably know all of them, but you could list them all off and I've never heard of any of them. And that tells you everything about the level of focus and actually.

1:50:37It's probably harder to watch their movies now. I feel personally slighted. I can't do it. Yeah. Well, we've got one more, at least one more superhero coming into the category this year. This one's being launched at us from a French castle. It's going to be draped in Americana imagery, so we're excited for that one. But there's all different texts. There's the short-term celebrity spike and Hope You Can Ride It where we've kind of just done the thing and been authentic. Who we are inside the walls is who we are outside the walls over and over again. I think that really compounds over time and matches with our manufacturing strategy, our continuous investments in quality.

1:51:20Every one of our teammates is full-time teammates. And like all this capital investment is, it looks really good right now, but in 2018, this was super out of vogue for sure. Yeah. Well, yeah, the beautiful thing with CPG, it is incredibly competitive. You are going to deal, if you're successful at all, you'll deal with a million knockoffs and new entrants, but you can look backwards and say like, or even just look at your business today and realize like even with everything you know now, imagine trying to compete with yourself from zero. It would be an absolute nightmare. So it's just the benefits of great product, operational excellence, and then compounding.

1:52:02Talk more about the decline in alcohol consumption amongst young people. What are you seeing in the data? How real is the trend? And what are the key drivers of what's going on? Yeah, and how much, basically, how much does the alcohol industry hate Andrew Huberman? He basically needs to be in witness protection. Yeah, the new dietary guidelines were basically like, don't not drink alcohol. It was like as close to don't drink alcohol as it gets, and the industry was like, yeah. But at the end of the day, I think a lot of this is potentially skewed by COVID and generations that were maybe like, I mean, if you go back to the start of our careers, it was so easy.

1:52:51I lived above a bar in the West Village in New York and I had friends in that bar every afternoon. It was so easy to find where people were after work. And I feel for this generation that is kind of losing the work happy hours, the mentorship and stuff. And I don't think we're going to have generations that don't love social connection and blowing off steam and having fun. I think the industry and the world has to meet people where they're at. And so I do think they'll normalize, but I also don't think, you know, in a world of cell phones and videos always on and like high work performance expectations, it's going to be really tough to get back those like volume drinkers as well.

1:53:28So I think you've got to layer in more occasions throughout the week. You've got to find ways to keep people at bars longer with moderate beverages. So the industry on the alcohol side will have continued volume challenges. and then there's like new non-alcoholic functional drinks too of which THC is the most obvious and those are not like high volume drinks either so yeah they're low volume and they destroy your sleep so yeah not a fan talk to me about label claims uh the only real label claim that I see on here is non-alcoholic brew um it doesn't say it has creatine in it or protein or caffeine or anything else.

1:54:10Protein beer. Protein. But I mean, you've reduced the calories. Have you guys ever made like a MVP internally of the protein beer? The high protein beer, 50 grams. Why not? We've had some fun with like caffeine and protein and like funny videos and stuff. I do think there could be a world of functional beers in the future, but I also think like beer and function is just kind of different things for the most part. But that being said, one of the hardest things of starting a CPG business is the plaintiff's bar and how tight the FTC rules are from website to label claims to everything. And there are a thousand eyes waiting for people to misstep even the smallest tripwires.

1:54:57Yeah, that was why earlier last year there was a viral brand doing like nicotine drinks. Nicotine drinks, yeah. And that just looked like it was like, okay, you're just making stuff and selling it. Jordy, you don't think an athletic four loco could work? An athletic loco? Athletic loco has a nice ring to it. No, that one got cut immediately. What's the scale of the business today? Steadily compounding. I guess we've moved into the top five of overall craft breweries out of 10 ,000 craft brewers, top 20 beer producers overall. That's amazing. We passed$100 million in revenue in 2024 and have been growing very healthy double digits since then.

1:55:48Last year was the worst beer year of our lifetime probably and we had really strong growth into that. We're just coming back with more exciting partnerships, more marketing spend next year as the business grows and kind of steadily just chugging along And, you know, comparable, I think, could be the energy drink category, which Red Bull really put on the map in the 90s. Monster arrived in like 2002. And those brands didn't have a lot of 100 % years. It was just a like 5 % to 15 % compounder for 25 years. And that's really what we're looking for also is just to like make really high quality beer, timeless marketing, and just ride that health and wellness wave.

1:56:27Yeah, that's always the, I feel like founders can get a bit carried away of like, I'm making an energy drink. Look how big Monster is. It's like, or look how big Red Bull is. It's like, you don't, they didn't just like be like, oh yeah, we just executed well for six years and suddenly we're a$50 billion company. It's like decades and decades and decades. And I think I get excited about TPG founders that are like, you know, I think of like Peter with David, where he doesn't want to sell out to the last gen. He wants to just compound, launch new products, launch new brands to get to the point where he is the 800 pound gorilla in the industry.

1:57:09And so I'm sure you've had a bunch of opportunities to sell and turn them down to date. So I'm a fan of the compounding. Well, you guys know better than ever, too. The rules of private capital have totally changed since we launched the business. And we have an amazing cap table through all that rejection I went through. And just people who are super excited about what we're doing with a really long time horizon. And there's the whole public versus private debate. And do we want to go public and spend 80 % of our time on the regulatory of that versus how great it is for awareness? and then the public can also invest in the business who are like super fans in it.

1:57:48So there's a lot of push-pull on public versus private too that we're working out. Are you getting any pushback from other alcoholic beer producers? We saw this with the milk lobby pushing back against like the alternative milks and saying, hey, you can't use the word milk. Are there words you can't use? The alcohol industry definitely tries to put up walls around things like this. We can't use all the normal beer words like lager, stout, ale, you name it. We have to call our Mexican lager a copper, for example, or a stout, a dark. But I did look at those analogs and Beyond Meat and companies like that and how they made enemies with the incumbent industry.

1:58:38And, you know, we want to make it really clear from day one, like, we're not out to stay in a soapbox and kill alcohol. Like, we're here to add to it, bring a lot of new consumers into it, and actually create a really new, exciting, viable part of the industry that just helps build the whole thing. So, yeah, definitely learnings from that. Well, thank you so much for coming on the show. Thanks for all the drinks, too. We're working through it. We're working through it. I mean, it says under half a percent. I think if I drink 50 of these, I can get a buzz going. So hell or high water here. Yeah, there's a major placebo effect.

1:59:15So it's like having one with lunch or on stressful work days or like my commute beer is my favorite beer to online. Yeah, it's like all these new occasions. It feels a little riskier. I like it. Oh, it's delicious. Yeah. Thank you so much for taking the time to come chat with us. Great to meet you. Super, super impressed with everything that's built and come back on anytime you have news. Thank you so much. Love the show. Thanks so much. We'll talk to you soon. Great chatting, Bill. Bye-bye. Cheers. Let me tell you about Turbo Puffer, serverless vector and full-text search, built from first principles on object storage.

1:59:50It's fast, 10x cheaper, and extremely scalable. Our next guest from Sequoia Capital. Switching gears just a little bit. We're going from non-alcoholic beer to AGI. We have Pat Grady and Sonia Wong from Sequoia Capital in the Restream waiting room. Welcome. Welcome to the show. How are you doing? Cheers. Thanks for having us. It's Friday. It's been way too long, Pat. It's been way too long. I haven't seen you on the show. We missed you too. I missed you guys. Most shows, it's like, you know, maybe once a year we'll have someone on. Invest like the best. It's like once every few years. I like that we can just get you every week.

2:00:22We'll start scheduling next week's appearance too. Sonia, it's also been too long. We enjoyed the candles that you sent us after our last hangout. I'm so glad. I think you're the ideas for this year's swag. Just wait. Yes. Is the conference still called AI Ascent? AI Ascent. It's going to be on 420 this year. Oh, really? Whoa. Yeah, you were early to the whole Ascend trend. Everyone's talking about ascending this year. And so it's just a great brand for you guys. I love it. Anyway, take me through the article 2026, This is AGI. Why were you writing it now? What was the one thing that you saw? What was the inciting element?

2:01:00Do you want me to start? I think there were two. Let me start with there were two purposes to the article. And there was something that we saw that catalyzed the article. One purpose was to kind of put a stake in the ground because there's this thing that people call the AI effect, where once the technology is actually out there and working, we stop to believe that it's AI. And we start to just call it by the specific name of the technology. You know, that's not that's not AI. That's just our LHF. And so there's this mythical quest for AGI. And best we can tell, we're there. And so we just kind of want to put a stake in the ground and say, hey, look, we're there.

2:01:40And then you say, OK, well, what's the point of doing that? Why put a stake in the ground? Well, that leads to purpose number two of the article, which is it is time. This is meant to be a call to arms for founders. It is time to take stock of what capabilities are available in the world today and apply those to real world problems as vigorously as possible. And so there is enough great technology out there available to founders today to go solve so many real world problems. And particularly what we've seen over the last couple of months, and this is kind of the triggering thing for the article with Cloud Code and Opus 4.5 and these long horizon agents, which to us are kind of the tipping point that's gotten us into AGI.

2:02:22Like in particular, what is possible now with that as sort to the third big inflection point of the last five years is pretty remarkable. And so it's kind of meant to be a call to arms. Yeah, and we're not helping because we have physical goalposts in here. Yeah, exactly. We have to take the goalposts off. We've achieved it. No, no. Let's just mount it to the ground. Like we actually have to bring it. Your smell of AGI stuff was amazing. Yeah, see, it's there. We're there. Yeah, yeah, yeah. We'll bolt it to the ground. Congratulations. You're in the goalposts now. You're physically in the Gulfos because you have declared AGI.

2:02:59Take me through a little bit of the history of Sequoia, the history of previous technological revolutions. I don't remember a 1999, this is internet. But as you reflect on the work the firm has done, what can you tell us and what can you share about how the firm has processed these broad proclamations? about the impact of technologies that are coming and then arrival dates when they are ready and that maybe changes the stance of the firm. Yeah, I love that as a question. So thank you for asking that. We've been pretty quiet historically as a partnership. And so this idea of sharing our thoughts with the world is kind of new and hopefully it's a useful service to the founder community.

2:03:45If you go back to our founding, so in 1972 when Don Valentine started Sequoia Capital, he was the go-to-market guy at a company called Fairchild Semiconductor, which is a household name for people who've studied the history of Silicon Valley, but not really outside of that. It was kind of like the OpenAI of its day, or it was the NVIDIA of its day. It was the IT company of its day because it was the first company to really commercialize silicon-based transistors. It was kind of the company that gave Silicon Valley its name. And as the go-to-market guy of that company, Don's job was to take this magical new capability and figure out how to apply it to real-world problems.

2:04:22Yeah, make it useful. Make it useful. And that sort of led him into venture capital because a lot of his customers were these founders and he just kind of fell in love with them in their quest to build these businesses. And so I mentioned that because we grew out of this semiconductor revolution where all of a sudden computation is a thing that was available to the masses. We moved from that into systems in the next generation with the with the apples and the oracles, you know, taking these chips and turning them into more computing systems. We move from that into the networking and PC era, where all of a sudden you have endpoints on every desk and every office thanks to the PC, and you can connect them all up with networks thanks to Cisco and the router.

2:05:03We move from that into a more sophisticated form of networking, which is known as the internet. We move from that into applications, which are a more sophisticated way to make use of the internet. Those applications start to show up on mobile devices. And then eventually, now that everybody has one of these mobile devices in their pocket, there's an enormous amount of data. People started doing interesting stuff with data. And now we have AI. And so this lineage of Silicon Valley is kind of built up over the decades. And each successive wave has built on the one that's come before. and being in this very fortunate position to kind of be in the mix on a lot of these transitions and actually being the only partnership that's kind of made it through these different tectonic shifts and kind of has this institutional knowledge, you know, gives us this broader view on, okay, what's the importance, knowing what we know about past tectonic shifts, what's the importance of the current moment?

2:05:55And with that, I'll hand it off to Sonia because Sonia was the one who, you know, three, four years ago, prior to the release of ChatGPT, kind of called this for what it was and put out the original generative AI post back in the fall of 2022. The other thing I think that's interesting about this moment is I feel like people have had this sense of like AGI will be here when it's just like so obvious in the economic data. And like, that's like, that's when it'll really feel like, okay, that's when you can declare it of like, It's here. It's working the economy. And I think part of why I appreciated your guys' piece is it's like it actually is going to require a lot of human agency and human application of this.

2:06:35It may not just happen fully organically, running away by itself. Cold arms. Now's the time. Let's go. Yeah, I wanted to hear from you, Sonia, but I wanted to ask, like, has there been any type of shift in founders coming and pitching Sequoia with, like, an entirely, like, I felt like last year I kept seeing, like, agent businesses that were exciting and the teams were talented. But I felt like they were just, in many cases, just, like, kind of rebranded SaaS. It's like, okay, you're pitching me this as an agent, but if I log into the product, it's going to look like traditional software. And maybe SaaS is evergreen and we're just going to live in dashboards forever.

2:07:27But I kept wanting more, basically. And I think we have started to see some new paradigms. But has there been any type of shift in terms of the companies coming in where maybe somebody spent a few years at a lab and they're like, okay, we're actually going to kind of – we've seen the future and now we're going to really reinvent the, you know, the entire product experience from the ground up based on this new set of capabilities. I'll start, but Pat is our app layer guy. So you should, you should, you should talk as well. We, we framed this at the bottom of our posts, like from talkers to doers.

2:07:58And I think maybe that describes a little bit of what you're describing of like first iteration, um, gen AI companies, um, including some of the great ones. Like it did feel like, you know, it was, It was the next generation of staff was how it felt like. They weren't actually getting work done for you. Part of the reason for this call to arms is like we're going from talkers to doers, which is and I think it is both the long horizon agents is the thing, but it's both the underlying model capabilities and the harness around it. So like these models can now take action and iterate and persist in the world around us.

2:08:31and so what we're seeing now is that founders are selling companies that are not just doing the you know the you know newer brighter newer shinier version of sass but actually something that takes real agency and takes real time horizons to execute so for example a traversal in the troubleshooting space like when there's a bug that goes down datadog says like ding ding there's an issue typically you have a war room that goes back and triages what the hell happened goes back and scrolls through dashboards, goes through traces, et cetera. Eventually root causes it, fixes it. This is a, you know, several hour to day long process for companies.

2:09:07It's funny to think of the human analog there. It's like, if you have an employee that comes to you and they just tell you there's a problem and they're not doing anything about it, they're just like, we got a big problem here, boss. And you're like, okay, like, what are you doing about it? And it's like, so traversal is like, okay, like we're actually going to be what a good employee would do, which is like, we have a problem. Here's how we're going to fix it. And hopefully, that's just solved before it even gets to you, right? Exactly. And so in this case, there is no SaaS equivalent to that, right?

2:09:34It doesn't even feel like SaaS. It feels like talking to a coworker. And that's why the litmus test here is, can you actually sell work? I'd say the founders that came into our door a few years ago, they all saw that vision. It's not a new idea that you can build agents. I think what's new is that you can actually execute on that vision because these models are finally smart enough, capable enough to actually stay on the rails where you can just let them run. I would say founders have done a really nice job of making use of what is available technologically at the time. And I think over the last five years, we've kind of had three big inflection points.

2:10:07Inflection point one was ChatGPT when we saw the benefits of pre-training. And pre-training kind of gives you baseline world knowledge and some kind of instinctive judgment, so to speak. Inflection point number two is late 2024 when And 01 came out and we had reasoning. And so the ability to kind of take your time and really think through particular questions and come up with deeper answers and better conclusions. And then the flex point number three has sort of been these last couple of months where we saw the impact of these long horizon agents. And so you've got some baseline knowledge. You can do some reasoning over that knowledge.

2:10:40Now you can actually fail, recover, stay on task, persist all the way through to some sort of an outcome. And I think if you map those three capabilities to the applications that people have built, it's no surprise that the initial applications were kind of like good Q &A or good summarization or good basic content generation. Like it was the stuff that pre-training sort of gives you. The next generation of those agents could think a little bit harder. You know, like in the case of Notion, as an example, which I think has done a really good job of staying on top of all this stuff. You know, you could now tell Notion, OK, can you build me a database?

2:11:16And it could actually reason through, like, what kind of database do you need? How should I build? And it could actually build you a database. It's pretty good. And then now, with these long horizon agents, we're finally getting to the point where you can make a couple mistakes. You can test a couple hypotheses. And you can actually iterate your way toward the right outcome. And so, again, you can look at Notion. And you can effectively have your personal chief of staff doing stuff that you would otherwise be doing while you were in a meeting running in the background. you know or in the case of harvey you can effectively deploy an ai associate after a data room or a deposition or some other legal workflow and actually complete it end to end while you're doing something else and so i think this maybe one way to put it is with this third inflection point you don't have to babysit the model you can just like give it instructions and let it go and you can have a few instances of it running in parallel and you kind of go from this is a cool tool that allows me to be more productive to, holy cow, this is kind of like a team of coworkers that I have at my disposal.

2:12:21Yeah, totally. Help me reconcile these two things that I'm feeling simultaneously. I agree with your declaration, the call to action, just the idea of so much opportunity for founders to use AGI to create all sorts of different value in products and services. At the same time, towards the back half of 2025, I was seeing Ilya on Dorakesh talking about, hey, maybe we're in age of research, and Richard Sutton and Andre Karpathy talking about some of the more restrictive aspects of the research progress. And so is that something else that we're working on? How are you thinking about the progress that's happening in more purely research contexts?

2:13:08What even needs to happen there? Your role as venture capitalists in helping that happen? How do you blend those two things that both feel true, but feel somewhat at odds? I think it's an end. And we're putting our money in both camps. On the research side, there's absolute fundamental research that needs to be done right now. I'm personally, I don't know if you guys have read Dave Silver and Rich Sutton's Age of experience paper like that to me kind of outlines um the premise like we're going from you know learning from existing data internet data etc yeah um that was like the paradigm for pre-training to learning from experience which is like reinforcement learning the paradigm is you interact with environments you get reward signals um you improve um and so i think there's a lot that's going to happen in terms of learning from experience and and um have been following you know what rich sudden is doing with keen um and very excited about some of that direction i'll give an example also we backed a company called recursive intelligence they formerly help the chip at google and this is like recursive chip design so similar to um in the game of go you can have uh you can play your the agent can play itself and get superhuman at go you can actually get superhuman at chip design and very surprising things come out of that like you end up having chip design that looks you know not like the manhattan distance kind of like square grid stuff you actually have what looks kind of like alien like chip design uh layouts and so um we are absolutely backing research labs.

2:14:33But then if you freeze research progress where it is, I am convinced there is so much value to be built in the world. And it's just lacking founders, lacking agencies to actually go out and productize and build companies around those. One of my examples. I was going to say, how are you guys thinking about more passive product experiences? Because people talk about agents, and most of the time they're talking about like giving agents tasks like hey go do this thing and spend as much time as you need on it and uh you know we'll see the result right like selling work but when you think about actually working with people some of the best employees are just without direct guidance constantly doing work in the background and coming to you only when maybe they need some type of like input and so it's like more like you're just kind of reacting to the work that they're doing and giving guidance.

2:15:28And so an example I would think of is with Harvey. It's like if somebody sends me legal docs, I should get some type of notification from Harvey really when it's already done some type of like diligence on it. I already did the work. The usual. So I'm not even worried about, I'm getting a ping from Harvey, hey, you should pay attention to this thing. You're going to need to make a decision here. I've done all the work. And so, and, and even like with an email experience, like I would love to log into my email and not see just like the order at which the emails came in, but like, here's three decisions you need to make.

2:16:06And again, like when you work with great people, they're going to, if they have five minutes of your time, they don't just tell you, here's the last 20 things that's happened. They're like, here's the three things that you need to make a decision on. And so I'm really excited to see more of these like passive product experiences where the work is just being done and you only need to chime in when it's really necessary or if you're actually a bottleneck? Yes, I think this is a great question and great direction. And to me, this kind of gets at the software as a service versus services as a software.

2:16:39You know, the nature of software is really changing as we go from an era of apps to an era of agents. In the era of apps, you want a lot of surface area with your customers. You want them to spend a lot of time in your product. You want to put workflows around them that are going to be really sticky that they're going to get accustomed to. In an era of agents, to your point, things can just be running passively in the background. The actual amount of surface area you have with your customers might be de minimis. The amount of time they spend in your product might be de minimis. So the nature of the moat that you build is different.

2:17:11And the nature of the moat that you build, kind of what Sonia was saying with the age of experience, it's all about context. It's all about like the environmental context of the job that you are trying to achieve and the feedback that that agent gets as it runs off and does something and comes back and says, actually, you know, like you did this part a little bit, a little bit wrong. You need to do this part a little bit different. And I think in a perfect world, you have something like what Harvey does, which, you know, Harvey kind of bridges that gap where there are workflows today for people who want to do the work kind of how they've done it in the past, just a lot better, faster, cheaper with the benefit of the AI brain.

2:17:47And you can deploy agents to go off and do the work on your behalf and then come back to you when it is done. And I think we'll see a lot of that over the next handful of years where there are companies who can kind of bridge that gap, where they can live in the software world and have some of the workflows, better workflows because of AI, but some of the workflows people might be accustomed to. And then separately, they can deploy in the background with these passive agents that kind of function as co-workers who just come back to you when things are done. But it's two very different design paradigms.

2:18:18It's two very different business paradigms. The fact that they're so different is one of the reasons why I think it's going to be tricky for a lot of the incumbent software companies to make the leap. The same way it was tricky for a lot of on-prem companies to make the leap to cloud. Yeah. Can you help me understand where you sit on the level of software eating the world? We talked to Tyler Cowen about one of the problems with AGI showing up in economic statistics is that AGI needs to live in a digital realm. It has to affect the digital economy, the services economy, and that there's so much of the GDP is made up of the healthcare sector, nonprofit sector.

2:19:02He highlights a number of sectors that are sort of AGI resistant or AI resistant. And so that's why he has longer timelines for significant impact to the overall GDP figures. Do you agree with that? Or do you think that there's something different about AGI that should allow tech to finally move the needle on those stickier industries? Well, I do think as like I fully agree with the premise that, you know, this AGI stuff is only good. It's only embodied in the digital world right now. Right. And so like we we seem to be, you know, getting tantalizing close to solving the entire digital world. That is very exciting.

2:19:41At some point, the physical world does become the next bottleneck. I think the exciting thing is actually that like, hey, robotics is also going through renaissance right now as well. Like some of the smartest people are going out and working on robots. And so I do think like similar the pace of progress is so high right now. And I didn't think we were going to have digital embodiment that was reliable this early. And now you can just let the thing run on your machine. I think that increasingly we're going to get to physical embodiment and we're going to have a smooth curve there. I do agree with the premise of like if you want – if you believe in the super fast takeoff, recursive self-improving, like everything is just going to go vertical line.

2:20:18Like I just – yeah, we're not going to get there. We're not going to get there without the physical world. We're probably not going to get there even if we have the physical world. I just think the reality of life is really messy. So I don't think we're at ASI. We're not at that kind of like takeoff moment whatsoever. That's not the stake in the ground that we're putting. What we're almost saying here is, hey, guys, this quasi-religious concept of AGI is the thing generally intelligent. And instead of just waiting for it to arrive, it's here. Let's build that here. Yeah, there's so much. There's kind of this toxic idea that I feel like is ingrained in so many people's head, which is like AGI is just going to do that.

2:20:58Yeah. Like AGI is just going to do that. And then we're done. So it's like, AGI is going to cure cancer. And it's like, at least in the near term, from everything we're seeing, it's going to be a human, a group of talented people properly using the tools to solve all the most pressing problems that we have. And for young people, it's like, you don't want to be... It's extremely paralyzing to be like, in 2027, I'm either dead or a billionaire retired on a space yacht. And so I could probably just chill for the next two years because it's going to be one or the other. No, and I worry that we'll have this sort of like lost generation of people that have let be ingrained in their head that it's not worth trying.

2:21:38Now's the time to build enterprise SaaS. You heard it from Sequoia Capital. Get in the arena, folks. This is a call to action. I feel bad because a friend of mine was pitching me this idea of like an agent to help consumers switch between like insurance products. because like an insurance company will sell you in a policy. And then they'll just kind of jack up the price a little bit every year because they assume a lot of people aren't going to churn. And at the time, I think ChatGPT agent had come out, and I was just like, dude, like the labs are going to like one-shot this problem, like I think pretty quickly.

2:22:16I think you could work on this for like six months and get steamrolled. And like it's very possible that like if he had just focused on that one little problem. It'd be a decent business. It'd be like a solid business. You would have figured out something unique. Yeah, and it's a big category. And so I felt bad. There's lots of examples. How much do you guys, I wanted to ask, like processing signal, because obviously there's so much noise right now. And something that I've appreciated is kind of comparing like the X timeline to the App Store charts, because every single day on the X timeline, like vibes are fluctuating and some news will break or some product will be released.

2:22:50And the common response is like, it's so over for XYZ company. and then you look at the App Store chart and it's like the company that just launched the thing is like not actually even in the picture at all. Like they're not in the top 25 of any category. And so like I feel like there's this insane disconnect between the real world. The real world is using ChatGPT all the time, but they're clearly not paying attention to X, our little bubble. So it's like, where are you guys really looking for signal and noise, especially in the later stages where, you know, you're investing in companies that should be out in the real world at some point.

2:23:29Before I get to that point, I think it is so amazing that all of this AI discourse is happening on X right now. You could imagine a world where all this dialogue was happening behind closed doors or scattered across different channels. The fact that it's all on X is amazing. It's a combination of research and what it takes and what's happening with labs. I think it's awesome that it's all happening in one place. It's remarkable for young people. You can just jump in and learn by osmosis and understand who the players are. Yeah, it'd be bad if it was only happening in Valinor. Totally. Valinor.

2:24:01I still have to figure out what the elves comment was about. But anyways, putting that aside, yes, our job is to try to piece signal from the noise. We're obviously doing everything we can in the background, credit card panels, web signals, et cetera. And I'll say, like, sometimes we see a, you know, trending web signal take on X. And it's just wrong. Like, we look at our own source of the ground truth. The example, like, some of the stuff on, you know, ChatsGPT is losing market share. We actually cut it by, like, U.S. web traffic. It's a very different story. And so for us, it's like, it's really important that we're on X.

2:24:42I think all of the discourse happening here, very important. But then, like, let's get to ground truth. Most importantly, just know that as venture capitalists, if someone refers to you as smog, it's sort of a compliment, but also sort of a dig. So that's what you need to know in terms of Lord of the Rings references for where you sit in the ecosystem specifically. Is that short camera like Smeagol? Smog is the dragon that hoards all the gold and is an antagonist in The Hobbit, but does have the gold, which is kind of like what you do. You got a lot of capital over there in Sequoia Capital. Sorry, Jordy.

2:25:15No, what were you going to say? Oh, I was just going to say, you know, to the question on Signal. Yeah. Interestingly, one of the big themes of the last few years on X, thanks to Lulu, is the go direct idea. And I think for us, go direct is a little bit of try to, you know, let the world know what we're thinking. But probably even more so, just go direct to the founders. You know, the best source of Signal we get is the founders. We spend, you know, we spend many hours every day just meeting with founders out there doing interesting things. And we kind of have the ability to do primary research that not everybody gets a chance to do.

2:25:49And I think that's still the best source of signal. Yeah. What is changing in the AGI age in terms of adoption and just AI diffusing itself through the economy? This is something that obviously every company has to deal with because they can build a great product, but then they need to get it into businesses. If they're in enterprise SaaS, they need to get adoption, make sure that it's being used correctly. there's the forward deployed model, there's self-serve models. Is anything changing there structurally or is it all just, uh, you know, the same playbooks? And one thought is like, I think we're going from over the last second, we went from sales led growth to product led growth.

2:26:28Yeah. I think we're going from product led growth to agent led growth. Okay. And I think you see this most clearly, actually, if you're using, for example, cloud codec, uh, actively, like it says, Hey, you should use for database. You should use Supabase, uh, for hosting use for cell, etc and like it's choosing for you um the stuff you should be using and um i think increasingly like so it's it's most obvious to me with code like your agent is choosing your infrastructure um but i think it's going to happen across the entire economy like your chat is going to tell you like hey this is the um this is the place uh this is the travel you should be booking this is the you know and and so i think we are going that's like back to channel channel sales like old school yeah old school but i think that the difference is like you don't have competing incentives here.

2:27:13And I think the OpenAI guys tried to release principles that I think are, it's a very principled take on how do you want to tell users what to use. And I think product-led growth brought us closer to the vision of best product wins, but ultimately people are still lazy. They can't read all these reviews. They kind of default to what looks cool on the website. Whereas agent-led growth, remember, your agent has infinite time to go and make these for you. And so it can go and read all the documentation, read all the user comments, like figure out for your use case. Also has an experience like building on top of the infrastructure.

2:27:46It's like, like eventually it's like, well, you want to choose a product that's most reliable. Reputation management is going to be very important. The way you show up in these LLMs is important, both in the pre-training and in all the research that they do. And then also this whole idea of like, you know, SaaS companies that don't have customers, they have hostages. Well, if it's one click or one prompt to replatform from the database that was trying to keep you hostage to the database that's trying to keep you happy. Yeah, but it's also like you're saying, I think it's good alignment. Compare it to like if you're building a home and like the person like, you know, like running like the construction firm that you choose like uses really bad cement.

2:28:27Yeah. And then you have like a terrible foundation. Like you're going to be mad. You're actually going to be mad at like the firm. Sure, sure. Because you're like, why did you pick this? Yeah. You should have known. This is like, and so I do think there's a good alignment where it's like the best, like you were saying, the best product should win because the agent is incentivized to make sure that you're on the best footing and using the best product. Cost is a factor. Quality, reliability, all these things. I do think there's that good alignment. That makes sense. The agent kind of accelerate this best product wins notion.

2:28:57And best product for the human is maybe different from the best product for the agent sometimes. Fantastic. Well, thank you so much for taking the time to come and chat with us today. Have a great weekend. And we'll talk to you soon. Pat, we'll see you next week. What's the next week? I can't wait. I can't wait. Two for two. Have a great rest. Great to see you guys. Have a great Friday. Goodbye. Phantom Cash, fund your wallet without exchanges or middlemen, and spend with the Phantom Card. Meet Phantom Cash. An Andrew Reid special. A Sequoia Capital Company. We have some breaking news before we bring in Sean Frank.

2:29:31Sam Altman has taken to the timeline. line, he posts the melting smiley face emoji with a screenshot from Elon's court filing. In the court filing, here's what it said. Musk insisted that any new entity, quote, support the nonprofit mission and that OpenAI remain, quote, essentially a philanthropic endeavor. Now, Sam Altman's putting it in the truth zone. He says, here are the actual September 2017 call notes from Elon, coming weeks, top priority, got to figure out how we transition from nonprofit to something which is essentially philanthropic endeavor and is B Corp or C Corp or something must tell the story and not lose moral high ground.

2:30:22Absolutely vital. So we'll see if we'll see how Elon reacts to that. But if that is the case, Sam Altman is upset about it. He He says Elon is cherry picking things to make Greg look bad. But the full story is that Elon was pushing for a new structure. And Greg and Ilya spent a lot of time trying to figure out if they could meet his demands. I remembered a lot of this. But here's a part I had forgotten. Quote, Elon said he wanted to accumulate$80 billion for a self-sustaining city on Mars and that he needed and deserved majority equity. He said that he needed full control since he'd been burnt by not having it in the past.

2:31:03And when we discussed succession, he surprised us by talking about his children controlling AGI. I appreciate people saying what they want and think it enables people to resolve things or not. This is Sam speaking. He's no longer quoting. But Elon saying he wants the above is important context for Greg trying to figure out what he wants. and Hope's Revenge says, oh my God, girl, he's such a snake. The drama continues. The drama continues on the timeline. Well, let me tell you about AppLovin, profitable advertising made easy with Axon.ai. Get access to over 1 billion daily active users and grow your business today.

2:31:47And we have a happy AppLovin customer in the Restream waiting room. We have Sean Frank from Ridge. He's the CEO. Welcome to the show. How you doing? Good to see you. Oh, you got the polo bear on. What's this? Is that the Winter Olympics? Team USA? It is the newest drop, guys. You guys like it? I've put it on just for you. Insane. Fantastic. Fantastic. Insane. Brought the RL bear. Okay, so Emergency Podcast, we wanted to have you on because OpenAI introduced ads. How quickly are you thinking about ramping up, testing it out? Is Connor waiting in line outside of opening AI to try to be the first?

2:32:28Dude, no joke. We would be the first advertiser if they let us. We measure a lot of traffic service to Ridge. And the most revenue per user out of all traffic ever, I mean, Google branded, Meta, whatever, is ChatGPT. We get like$12 from every one of those sessions. For Meta, you'd be lucky to get$1 per user. The revenue procession from an AI user is just through the roof. We have some data. I don't know if you want to share it or not, but NorthBeam put out their 2025 report. They track probably 2 ,000 merchants, billions of dollars in revenue. AI search in January was 0.01 % of all traffic on all e-commerce websites.

2:33:14Now it's 0.7%. It's just shot through the roof. There's not ads yet. As soon as ads roll out, I am so excited. Okay. We would expect the dollar per user to fall based on ads, correct? But do you still expect them to be high intent? Because we were talking to Ben Thompson about this, and he was saying that they should have launched two years earlier because it's going to take a while for the ads product to work its way up. I imagine that the ChatGPT users that land on your website are super high intent because they've probably been asking ChatGPT about your product specifically or wallet specifically or product specifically.

2:33:53Now, if they're researching the Roman Empire and they just see a little ad at the end that you purchased and it's not targeted at all, they might click on it and check it out, but it's probably going to be lower intent. How optimistic are you about ChatGPT ads delivering at a level above Facebook quickly? Well, Google is rolling out their direct AI offer, right? And what that is, is if someone asks, hey, what are the best gifts for dads? They're going to have that specific question. I can now bid to show up with an offer for those people. It's like, hey, the Ridge Wallet's a really good gift for dads, and they're going to give you 20 % off right now if you click this link.

2:34:33So it's incredibly bottom of the funnel, like people who are ready to purchase. And that's Google's ad offering inside of AI, and that's coming out soon. So I think that was what pushed ChatGPT over the edge to rule this out. They probably do a carbon copy of that. So it's not going to be discovery ads like you get on an Apple Oven or a Netta, right? What it's going to be is, you know, like that last point of sale, almost like an affiliate, like push you over the door, like a honey competitor or whatever, which I think is. Would you be unhappy if the offering was just give us a dollar amount per customer and we'll go find you customers?

2:35:13something that's fully black box you're not doing any demographic targeting no keyword targeting or would you like those features well that is historically if you can tell somebody like here's 20 bucks find me customers and you're you just look at your average order size that's like the holy grail of advertising but it usually ends up being yeah a bit more but but is the is the initial product that they roll out does it matter to you or will you just take what they what they give you Yeah, it's what we've seen, the organic traffic from any AI search is already so valuable, right? Any way to get more of that, we would spend a lot of money for.

2:35:54But the beauty of the Facebook ad product is that they figured out a way to take half of everybody's revenue, right? Like affiliate is like people - Let's give it up for Mark. I'll take half. You do the work, I'll take half. TikTok shop affiliate, people don't want to give more than 15 % of a sale. They're like, oh, that is too much. But all performance marketing is, is just an affiliate in a different code. It's like they're able to, oh, you're going to generate your demand. And if you have a better offer, you'll get a better return. But really, they've just figured out exactly how much money they can take from everybody at the optimal level.

2:36:34And that's why they do$100 billion a year in revenue. So if the offer from ChatGPT and Sam Altman is, hey, give us 40 % of your revenue and we'll give you new customers you wouldn't get otherwise, I would take that deal because I'm already giving Mark Zuckerberg 50%, right? One thing that's exciting is I think once people start seeing more, once they start seeing ads for products, they will just naturally start using LLMs for more product research. Like I still find myself, if I'm doing product research, I'll still go to Google search quite a lot, even though in many ways the experience is inferior just because I'm so trained on that.

2:37:15And so I think as they roll out ads, there's going to be more organic activity. So hopefully as a brand, you should be getting more people just landing on the site organically because they're just finding out through their own research. Totally. Half of all product searches start on Amazon. So Amazon has been able to take over the product search box from Google, but they spent$500 billion in two decades doing it. So if ChatGPT can start stealing some of that glow that Amazon's been able to do, They'll have an amazing affiliate business and ads business and whatever you want to call it. You guys sell a ton on Amazon.

2:37:53What do you think Amazon's play here really is? Because they obviously don't want to give up the golden goose of their massive ad revenue business. They want to protect that. Alexa is not even in the picture as far as I'm concerned in terms of a shopping assistant. So what's the play? Well, Amazon has a beautiful ad arbitrage system where they spend – they're like the number one spender on Meta and on Google because they can spend$2 to get a user and then those users click on a bunch of ads on Amazon. It's a free money machine that they've invented. So, yeah. Double kill. Yeah, I think they would love to partner with everybody because once you're on Amazon, you're still going to click around.

2:38:42You're still going to do shopping and they have an ad product. The more interesting question is, what does Shopify do? Because Shopify is a product feed of all these merchants, right? But will they be able to win over any sort of these deals? They don't have an ad product. They make all of their money on me hosting my website and then a percentage of the sales I drive. And it's a very small percentage. It's a payment processing company. So it's like, how are they going to survive if more and more of those transactions are happening inside of ChatGPT or inside of Gemini? I think Amazon's fine. They have a big enough product, you know, they could feed it everywhere.

2:39:18And if you do come to Amazon, they get all their money back with all the clicks. But walk me through the risks of Shopify, because if I'm looking to, there's a lot of brands. like there's a ton of, I would say like the real like shopping that I've done over the last five years is like almost always on Shopify stores, right? Like Amazon for like essentials and things like batteries and things of that nature. But I feel like a lot of my, uh, like, like, um, light, like lifestyle consumption is happening on Shopify. And so you mentioned that not having that like ad flywheel but am i not still like discovering ridge and just popping over to your shopify store and like to complete the transaction and if they're effectively a payment processor like they're still making their money right well it's just if websites continue to be important like you know i i'm i'm going to operate on the assumption that yeah but all but calling calling shopify just like a website company like in my view it's like a catalog and it's a CRM and it's a payment processor.

2:40:35Oh, for sure. It does all those things incredibly well. Bro, I'm so bullish on Shopify. I love Shopify as a thing. But the challenge is the website will become less important in the future. Just like mobile apps had a moment, they're not important anymore. Where's the last time you guys opened a shopping mobile app? And before that, it was catalogs, right? So like the user experience has become easier and easier and easier. And the way you interact with brands has changed. And I think the website will kind of be replaced just by asking chat to buy you stuff. And maybe you'll still go to websites, but kind of like how you still kind of open catalogs.

2:41:12I don't open catalogs. Straight in the recycling. I'm like, I did not ask for this. Totally, right? And that's just like the worry of the website. So ShowVice is incredibly important for us, but they do have to plug into every single LLM, right? They have to be that product feed everywhere. Yeah, but I think they've taken the approach of just like being more quick to partner with the LLMs than an Amazon has. Because, again, they don't have that same like they have like less to lose. Right. They're not worried that because they don't have an ad business, they're not like like I feel like Amazon has to be a lot more.

2:41:49When you look at the when you look at the marketplaces that have partnered with ChatGPT, it's like Etsy. Right. Etsy doesn't really have anything to lose. Right. Like they're, uh, they just want more traffic, more, more sales. Like they'll take it however they can get it. Whereas like Amazon has something to lose. Right. You know, and it was big news when Shopify partnered with chat GPT, Intel, Amazon partnered with chat GPT and Amazon comes in with an equity round. But like that now all of those products are going to be feeding in chat GPT. And that's one of the worries of Shopify. It's like, how do you stand out right now?

2:42:28it's like beautiful brands with beautiful websites with shopping experiences. And that is different from what Amazon offers. But if all of those products are just going to be discoverable inside of ChatGPT and competing for the same space, I don't know the value of my website in three years. Like that's one thing that we're worried about at Ridge right now. Well, what about ad creative? There's this weird trend where DTC companies were effectively making TV commercials with TV production techniques, cinema cameras, high res photos, whatnot. But what wound up working on TikTok and Instagram Reels and YouTube Shorts was a lot of UGC, a lot of stuff shot on iPhones, a lot of stuff that feels more natural.

2:43:16Is there a world where maybe you want to get good at creating Sora videos or AI generated images, even if they don't look like a real photo, it looks more natural within that app, within the chat app. You're expecting to see AI images. And so you want to lean into AI image versions of your products, even if the background's a little crazy, hallucinogenic, it's what people will be used to. They'll be in that mode. Yeah. I think when ChatGPT launches their ad product, it's probably going to be thumbnails to websites with really aggressive offers. But right now, in our ad account, if you guys go to the Facebook ad library and pull up Bridge, we are running full-blown AI videos.

2:44:02And they're getting spend and they're winners in the ad account. So we're all probably underestimating how much it's going to affect the ad industry. By the end of the year, video is going to be... If you want to make AI video, it'll look just like human video. And the highest level productions are already being basically monopolized by AI. And something I've been thinking about that feels inevitable is two years from now, we'll look back and laugh that people would make an ad and then serve the same ad to 100 ,000 people or a million people. because it's like if you're going to spend a lot of money serving somebody an ad, you should serve them the most like hyper targeted like Sean Frank.

2:44:48I have an athletic brewing company beer for here for you. I got an offer for you, Sean. Just click the link and check out. Like eventually it will just get so, so, so, so targeted that it's basically like ads are being generated on the fly for individual customers. Oh, dude, that's for sure coming. Matt is working on it. And not only that, they'll put you in the ads. They'll be like, hey, look how cool you can look in this sweater. You should buy this sweater. And they want to do that because then the CPMs just go up. Right now, I get a$12 CPM on Meta. They would much prefer me get a$200 CPM and just show way less ads to everybody.

2:45:24So I think that the personalization of ads is coming. And we've already just seen it. In my ad account right now, today, maybe a third of all my spenders will be AI ads. And that's crazy. That is crazy. double click on the the ai ads that you're running are you running stuff that you could shoot that's indistinguishable from human photography human videography or is there unique value to saying look there's just no way that we could take a ridge wallet to the top of mount ever shoot it into space have it turn into cake have it turn into a cat like crazy stuff that's just impossible maybe with like a massive CGI budget, but truly like breaking the laws of physics?

2:46:06Or are you just trying to do like, Hey, it's just a Ridge wallet on a beautiful background. And you know, we were able to generate that. Yeah. So really what it is, is, you know, we have wallets with every college and let's say there's 200 colleges in America, right now I could get 200 jerseys and I could get 200 fans and I could have them talk about this particular wallet with this jersey, with this fan, like they're the very bespoke ad, and then serve it into a market like Ohio. I could do that, but that would cost, I don't know, $85 ,000 to run that campaign. Or I can have one guy do it in one day with a Higgs field or whatever, right?

2:46:44So that's what we're doing right now. Are you guys using Higgs field? We're having them with the show in just a couple of minutes. Yeah, I love Higgs field. So yeah, I don't think like, I think normies are using chat GPT now like it's fully broken through the mainstream but like like the all of the the tools around that I mean everyone knows lovable at bolt.new but like if you want to make videos at scale I think feels like the best easiest way to do that yeah so the pipeline is you specifically want to multiply your creative so you are filming one human video and then you're creating variations with AI is that the correct use case right now as that example No, no.

2:47:23So what we're doing is we have B-roll of actual wallets being used that we've shot, right? But we need an opening hook of a Buckeyes fan, right, in a jersey or whatever. So that is being done with AI. And then intros to the wallet. And you can pull up our attic right now. There's a woman, like Hunter, she's walking, talking about, like, hey, I bought this for my husband. He's like, you know, whatever. She doesn't exist. That is all AI. And it's because for us to find a woman hunter who's out in a field or whatever, we could do that. It would just take, you know, two days and it probably costs$2 ,000.

2:48:02And we can test that concept right now in real time inside the Facebook ad library. It's going to be interesting. I'm so curious to see what happens to the brands that like try to take like some type of like moral aesthetic or ethical high ground and just say like we never use AI. and what actually happens to those? It's one of those things. Yeah, it's like not using AI. It's like if somebody said, yeah, we don't use computers. I'm like, I don't even know what you mean. It's like everything's going to be AI. So look, incredibly bullish on ads coming to any surface of the internet. So if there's any surface of the internet that doesn't have an ad, I would love an ad to be there.

2:48:49you guys did the app loving ad read brand new to e-commerce and I spend millions of dollars there and I love it as an ad product so more of that is just really good for me so if it comes to Gemini or ChatGPT now I just don't think they'll do big beautiful display ads I don't think there'll be any video ads it's just going to be hey you're looking for a gift Ridge Wallet will give you the best price ever right now and then they'll take a 30 % commission. I really want to see more flashbangs come to the end of that. Watch out. Watch out. We got you. Do you have any flashbangs on your podcast? You're the one show that we would give you the effect if you want to flashbang your boys.

2:49:36Okay, cool. I'll get the ad words to that. We also have a horse that runs across the screen now. We'll save that. We'll save that. For right now. Wait, have you used any other of the AI ads?

2:49:55Basically, I heard Perplexity was rolling out ads in their product. Did you demo that? Have you thought about using that? I know it's smaller and earlier. Are they actually live? Do you have any experience there? No. And I have a list in front of me. I told you earlier, NorthBee did a report. and chat GPT is 99.5 % of all AI traffic to e-commerce websites right now. Wow. So it's a big one. So this is a big announcement today. So yeah, what does it take to actually get Ridge live? I don't know, man. You guys know Sam, dude. Hit him up. Say he's got his first paying customer right now. He's a little busy today.

2:50:36He's got to go back and forth between Oakland and traffic. no uh yeah we'll we'll put in a good word for you i would love to what what consumer what kind of broader consumer trends are you tracking this year we were talking about the sauna wars in new york city uh people setting up these like bath houses i was somewhat bearish on the trend like nationally i just think there's a lot of places in america where people will be like am i gonna pay a couple hundred bucks a month for this like sauna like bathhouse membership or do i I just want to buy one for my house. But what are you tracking? Well, on the sauna wars, I mean, Paws is a franchise-based sauna business.

2:51:19They have them in Venice. They have them in Southern California. I was at an e-commerce or like a private equity conference. I think those guys are going to expand aggressively. So I think they're selling territory rights right now. I used to live down the street from a Paws. I would go there. I enjoyed it. I was like, I like using a sauna and I like using an ice bath. And so I just bought one because it was like$100 every time I went. I just did the math and I was like, it's nice to go there, but it's nicer to just have it in my house. And so that's one of those things. I don't know how much actual spend they can capture long term or what the LTV is.

2:51:55If you get somebody hooked on using the sauna, eventually they just run the numbers. It's not like a gym where it's like, a gym, I'll pay that however much it is a month because it's nice to go somewhere out of my house to work out. And I don't want 300 different machines and all these different plates. So I'm not convinced that it's actually going to scale that well outside of, you know, major metros like LA and New York. Yeah. Also the cost per use of a gym is really low compared to a hundred bucks every pop of a pause. But they're going to expand like crazy and it'll be a great private equity transaction for somebody.

2:52:33But yeah, long-term, I don't think it's a good trend. The thing we're most interested in is fiber. I think fiber is going to have a protein-like moment. And protein's been on a mega trend for 20 years or whatever and has billions of dollars in spend behind it. Creatine had a little taste of that, but fiber will be the next thing that even reaches something like a protein. So we think fiber is like the obvious big trend right now that's going mainstream. Will we see you guys enter the category this year? We're secretly in the fiber space already. So we're not talking about it, but we have a brand bubbling up.

2:53:10We have real customers that are paying. They don't know it's behind a ridge, but we'll talk about that maybe offline. Amazing. So you're going to post your revenue every single month to invite thousands of competitors into the category, right? Yeah, seems to be the smartest thing to do, right? It's like the build in public movement brought to consumer is just an absolute nightmare. It's like if you brag to the world about how much revenue you're doing, there are going to be smart people that enter your category. They're going to do it fast, and they might out-execute you. I don't think they'd out-execute you guys, but I've yet to see a consumer founder build in public that didn't ultimately regret it.

2:53:52Oh, yeah. We're pretty public with rates just because I think you have to be stupid to try to sell wallets in 2026 or whatever. Yeah, but for you guys, it was never like you were just like, here's our investor update. I'm publishing it because it's so good. You guys don't have investors, but it was never like, I'm going to flex monthly on the timeline, which is the approach that I think a lot of people have had on the consumer side. Yeah. I mean, consumers and I fight, man, like there's no moats at all. It's, it's the digital equivalent of opening a restaurant, I think. So, you know, like you're, you're fighting for every single transaction, right?

2:54:33You guys were just having, you know, Sequoia on talking about it's a great time to sell enterprise SaaS because you guys have hostages and not customers. And it's like, yeah, I would love, I would love to have. Yeah. There's truly no lock in. Once you buy one, It's not that you have to come back the next year, whereas for a lot of enterprise SaaS, if you stop paying, your entire business shuts down. Yeah, the moat is scale, operational excellence, but those things can erode. Very team-based. Well, thank you so much for taking the time to hop on. It's great to see you. Next time you come on, come in person.

2:55:10We know it's not that long of a drive. Dude, I would love to, and guys keep crushing it. Once you actually have the OpenAI ad platform live, let's come here. Let's put it on the big screen. We'll get a bunch of monitors, and we'll just monitor the situation together. And we'll clap every time a sale comes in. Yes, yes, yes. And thank you for hooking the team up with Ridge Wallet. Oh, yeah. You guys are too kind. I was in the gym this morning with some of the folks. They had the suitcase. Look at the suitcase. There it is in the corner. There we go. Every single member of the team is rocking a Ridge suitcase.

2:55:49There are more Ridge wallets than you can possibly count in this Ultra Dome. It's amazing. We're huge fans. We love you. I carry a Ridge wallet every day. I love it. Talk soon. Thank you. Keep crushing. Goodbye. Goodbye. Goodbye. Restream. One live stream, 30 plus destinations. If you want to multi-stream, go to restream.com. I want to talk about ads more. I want to talk about Vibe.co's ads. They're a sponsor of us, but they've also been burning up the timeline with a number of billboard ads in San Francisco. Shiel Monat, friend of the show, says these ads are very clever. What else is in the collection?

2:56:24And I got a preview of some other ones. So this is Jensen on, Target Jensen on TV. That's a brilliant billboard. It only works in San Francisco, but Jensen Wong, the CEO of NVIDIA, his leather jacket is iconic. And so you see the shoulder of the leather jacket with Jensen. You know what they're talking about. If you want to get to him, you're going to go to Vibe.co. What else did they do? They did say target Mark on TV. They put just Mark Andreessen's head there. It's an iconic image, and it lets you know immediately how can you target someone like Mark Andreessen. It's hard. He's not going to be everywhere.

2:56:58Vibe.co has great targeting. So I got a little teaser of what else is available. We got Sam Altman, target Sama on TV with the McLaren app one. Is that real? And then we'll go to the next one. Target Zuck on TV, and it's a wrist with an F.P. Journe on there, or a Patek. No, you made these. And then the next one, they have a Target Chesky on TV, and it is a barbell. Yeah, so this is the type of marketing that you've got to be doing. You've got to be thinking, okay, how do I send a message? There's one more. We'll go to this. Target Augustus on TV. Of course, everyone knows Augustus Dorico. He's known for carrying a white monster.

2:57:36He's sort of like a human version of the white monster. If you put this billboard up, everyone's going to get it immediately. they're going to know that you're cool. They're going to know that they should go to Vibe.co and they should start targeting folks like Augustus DiRico, the CEO and founder of Rainmaker, who's known for having a white monster every once in a while. Not every once in a while. Pretty much constantly. Every hour. On the hour. He's powered by it. The way you drink Diet Cokes, I imagine he drinks white monster. Yes, yes. It's just like, oh, it's 2 o 'clock. No, no. The four of those that we just displayed, I did generate in Nano Banana with Gemini 3 Pro.

2:58:11Google's most intelligent model yet. Leah says, love trash day. Love hearing that truck. Humans getting stuff done. Cooperation, logistics, municipal services, civilization. Couldn't agree more. It's a reminder that we haven't collapsed yet. Dude, a four-year-old, the way their face lights up when the trash truck comes, and it's just pure wonder at the massive machinery moving up and down the street, the beeping when it backs up. That is childlike wonder that you should never lose. No matter what your age is, you should always take a moment to show some respect for the trash truck. I completely agree.

2:58:50I love it. Joe Weisenthal is crashing out. He says he never wants to use the web again. He said in today's Odd Lots newsletter, he wrote about how depressing the Internet's become. We'll read his piece. And we need to catch up with Joe. I think he's coming on the show soon. He says, I swear, I swear, this is not another piece about my personal forays into Vibe Coding. Okay, maybe it is a little bit, but not really. So last year, we did an episode with the COO of Perplexity, whose name is Dimitri. He says, which is kind of a hybrid LLM search engine. You'll be familiar with it. it's very tethered to the real world.

2:59:23So if you asked it about, hey, maybe the latest developments in Venezuela, it will obviously produce text, but also links to the latest reporting from reputable news outlets and other sources. It generally works well if you want to get up to speed on something that's happening right now. Anyway, one of the questions that Joe tried to get across with Dimitri, but he doesn't think he phrased it all that particularly well, is what happens to the future of the web as we know it if more and more of our news is coming to us from chatbot form. Let's just imagine people start regularly making perplexity their first destination each morning to capture up on the news.

2:59:57And let's just imagine for sake of a thought experiment that the traditional news publishers come out fine in this relationship. So the New York Times is and the CNNs and the Al Jazeera's of this world are getting adequately compensated for their reporting to be summarized and displayed on perplexity. It's hard for me to imagine the future, but whatever is beside the point. So he's just saying it's not an economic issue. Let's talk about the the vibes that would result from this. So he says, hey, this is his question. In the future, why do websites still exist if more and more people start consuming the New York Times' content through a chatbot?

3:00:26Why continue to invest in maintaining a well-functioning website called NYTimes.com? I suppose that to some extent this is a question that precedes the existence of chatbots. There was almost certainly a time when a majority, when a major priority for the New York Times was to have a well-functioning, elegant website. And while there are people still paid to work on it, this is probably not their top priority right now. A tech budget that might have once gone to web development might now be going to figure out the best strategy for Instagram reels or whatever. Nonetheless, do you have a comment, Jordan?

3:00:57Yeah, maybe this is just because I turned 30 and I'm unk now. But I like going into the world of the content. Like I like going to the Wall Street Journal app. I like going to the New York Times app. I like going to these places because you get it. I like the physical paper. We like the physical paper. This isn't going anywhere. For my cold, dead hands, me and you, Joe Weisenthal, we are keeping the newspaper alive. I guess you'll be on the web like some hot new young 'un. I will be reading the paper. Yeah, ultimately, I still care about, like it matters, I care a lot about who wrote it too. Totally.

3:01:33Like I don't want it just served. But that can be instantiated in perplexity or chat GPT. It can tell you who wrote it. You could, in theory, if the deals are structured properly. Now, you can't get, I don't think you can get the New York Times content in ChatGPT because they have a lawsuit. You can get Wall Street Journal content. So in theory, you could kick off your morning with a query to ChatGPT. Hey, tell me the most interesting pieces that are on the front page of the Wall Street Journal. And it would just do it for you. And it would bring that in. It would pay News Corp accordingly. But there's something fun about holding a nice piece of paper in your hands.

3:02:07I think it will die, but a very, very slow death. We're still reading newspapers. I imagine I'll still be reading newspapers in a while, and I imagine that we'll still be going to the websites. But he is correct to highlight that there is a change afoot. So he says, nonetheless, the web browser is, by and large, still a dominant piece of software for accessing content and services that exist on the Internet. And, in fact, if we step out of the news context and think about it, say, buying car insurance, then the website, as we know it, is still incredibly important. Anyway, I was thinking about this again because my vibe coding foray, easily the most annoying parts, were when I had to leave the terminal interface and do something on the web.

3:02:43So it's still an unbelievably annoying process to do something simple like redirect a domain name. I found GoDaddy's, where I've been registering various domain names for years, interface to be borderline unusable. Shots fired. Fact check, true. It is difficult sometimes. Every time I'm just trying to make sure that a domain is renewed, they're like, how about you buy 1 ,000 websites right now? Just buy 1 ,000 websites. Buy 1 ,000 AI websites. I'm a name.com guy. It's been a better experience. We have 1 billion websites for you. Yeah. Yeah, yeah. As I mentioned in the piece he wrote on Tuesday, one of the first things Joe did to make it so he could update his website directly from Cloud Code, rather than having to update the files via Cloudflare backend.

3:03:31This is the first very small thing, but the point is that every time I had to operate in the browser, it felt deflating typing in URLs, clicking buttons, checking boxes, pulling down dropdowns, entering fields. No thank you. He's genuinely curious what happens to the visual web that's designed for humans. I know there's plenty of work that's being done to allow agents such as CloudCode to go to a website, click around at what human does, just in the same way that chatbots were trained on billions of Reddit posts. and everything else on the internet. These agents are being trained by ingesting numerous sessions of people moving a mouse on the screen and so forth until they can mimic that process.

3:04:05So that's all well and good, but it still gets to this issue that we currently have much of the internet presented in the form of web pages that are designed to be logical and pleasing to the human eye. We still do so much on the web, but now every time I have to use the web, it feels kind of like a personal failure. It feels inefficient. And if the web is increasingly going to be navigated by bots working on a human user's behalf, then why even bother optimizing it for visuality at all? I don't know where any of this is going, but my guess is that it's pretty big structural changes are in store for what people think of as the Internet.

3:04:34Yes, it's interesting. I feel like the Internet is already living in apps so much. Like when I think about a world in the future where an LLM is just generating me an article that was in the New York Times that day. I could imagine it just generating it in a visual style that the New York Times decides themselves. You can already do this with Nano Banana. You can say, write me a New York Times style article about this topic. It will write it, instantiate it. It will search the web to find the facts. So it'll probably not hallucinate that much. And then you could say, turn it into an image that looks like a New York Times screenshot.

3:05:09And it'll do that too. So people will be met where their preferences are. I don't know. I think that, I mean, the flip side is like, he's making this argument that as more and more traffic shifts away from the web, it will get fewer and fewer resources within those organizations. So the New York Times will say, hey, we're not getting that much traffic to our website, so let's not spend that much effort making it better. But you also get 10 times as much leverage because the folks at the New York Times will be able to fire off an agent and say, hey, make it look better. And so you would imagine, I would hope that even as you go from, okay, maybe there's 10 people working on a website, and now five of them move over to Instagram and YouTube strategy, and four of them move over to generative optimization strategies and LLM and AI stuff.

3:05:59Well, there's one person left, the person on the Japanese soldier on the island who hasn't gotten the memo that no one's going to the website. Well, some people are going there, and they fire off a prompt, make it perfect, don't make mistakes, and boom, they're good. They're in business. So some cause for optimism. We'll see. What do you think, Tyler? I was just going to update on the Sam Altman. Please, give us an update. What's up? Okay, so first, just if you want to look at prediction markets, right? So will Elon win his case against OpenAI? Yes. Earlier today, basically at the peak of like the, oh my gosh, like Sam is so cooked.

3:06:32Yes. He was at 74.5 % chance. Whoa. It's now down at 54.1%. This is fascinating. We're watching them like duke it out in real time in prediction markets. You know, Elon, he doesn't have to worry too much because Nick, his strongest soldier, he's saying it reads as Elon's emails. It reads as obviously fabricated. Elon doesn't talk like that. Oh, well, you know, case closed. Case closed. I wonder if Nick will be called to testify. I'd love to see him on the stand. That would be very entertaining. Anyway, Jensen Wong doesn't care about gamers, allegedly. ASUS reportedly says the RTX 5070 Ti is no longer being produced.

3:07:18RTX 5060 Ti, 16 gig to follow. The gamers have been having a rough time. I don't want to play with, Jensen says, I don't want to play with gamers anymore. It's AI all the time, yes. Gamers are being squeezed left and right. We need to make everyone aware of the rights of gamers with our COD sounds and our flashbangs and our smoke grenades to let people know. I'll let Jensen know. Maybe we'll have to bring Jensen on the show, throw a smoke grenade, and when the smoke clears, he can tell us that he is still committed to gamers. Let's move on to World Labs. World Labs launched a box. They shipped a box.

3:07:57It's a video of they took World Labs. I think this is real. I have no idea how to process this, but it looks like they built a screen, put in a box that has an IMU, Inertial Measurement Unit, that can measure how the box is being rotated, and then it can re-render the world so you're looking into it in this box. This seems really, really cool. This seems like just a piece of hardware that people would actually buy and sell. We've seen this with the holographic display that you put next to your gaming rig made by Asus or Razor. I guess Razor is the one that makes it. This is really cool. I'm a big fan of World Labs, and I like that they're productizing this.

3:08:44I wonder if this is like a stunt, if this is just for fun. I really hope we can get our hands on one of these because this looks amazing. Well, there is more news in the AI talent war. The total number of employees that are leaving thinking machines. Alex Heath is hearing that it's up to six now. The true scoop god is on the trail of another scoop. He posted this at 2 a.m. He's up burning the midnight oil, getting the scoops on thinking machines. Maybe up to six people leaving. Naval says California-based unicorns. Do you think this is because of ads? Do you think all the thinking machines folks were like, wait a minute.

3:09:24We want to work on ads. They're going to do ads at OpenAI. I've got to go back there. I've got to go to OpenAI. I've got to get out of here. It's been my dream. Mira hasn't said anything about that. Or maybe they wanted to work on adult mode. Maybe. Maybe. There's lots of reasons. Or Sora. They could be trying to grind up the trough. The Disney IP. They want to get their hands on that IP. Yeah. They want to slough it up. And they say, we're not doing any of that at Thinking Machine. So we're going over to open it. Talk about this. Naval says, California-based unicorns being routed to the glue factory.

3:09:56The horse metaphor. Not mincing words. We will protect the unicorns. Unicorns are horses. We will protect them. They are. Do we read through David Friedberg's piece? Yeah, fantastic piece. California started with the gold rush and might end with the golden exit. We're coining. It has been underreported how much wealth has left California because of the asset seizure tax being proposed. It's important that we continue to call it the asset seizure tax. He's coining. A private poll was conducted amongst affected individuals a few days ago, and 80 % to 90 % surveyed they had already left California in 2025 or will leave in 2026 if the ballot measure looks likely to pass.

3:10:31Of course, the ballot measure currently is retroactive, so they would be subject to the 5 % one-time tax, and of course it would get litigated. Dave Freberg continues, two to two and a half trillion of assets gone, representing$20 billion of annual revenue for the state government and likely hundreds of thousands of jobs now at risk. Less reported is the bigger exodus underway from folks who are not directly affected but worry, as they should, that this law will quickly transition from billionaires to everyone else. The initiative actually gives California legislators the right to take anyone's post-tax assets anytime in the future based on a majority vote.

3:11:12This isn't about billionaires. It's a new tax system that simply destroys private property rights in America. All private property is now public property even after paying your taxes. It's not legally your property anymore. It's the government's. You're just borrowing it. Legislators will decide what you get to keep and temporarily use each year. Countless founders, CEOs, and other business leaders are actively looking to move their companies out of state, not just tech, not just AI, not just billionaires, but the core engine of California's prosperity since 1847 is unraveling. And here is how this initiative risks unraveling America.

3:11:4410 states have explicit or implicit prohibitions against asset seizure tax. Individuals affected in California and other states trying to do the same will move to these states that endow private property rights. California already has a 20 to 30 billion dollar budget deficit, an unfunded one trillion pension liability for public employees, unions, and 500 billion of debt outstanding. The state cannot afford to borrow more, much more, and will launch more asset seizures to meet its obligations. Asset seizures will transition to millionaires and eventually to the entire middle class as more asset seizures drive more people to leave the state.

3:12:23The deficit debt and job loss will spiral. The golden exit. No U.S. state has ever declared bankruptcy. In addition to California, dozens of other states face similar fiscal crises. Legislators promise future benefits that can't be paid where theft and waste have been allowed to run rampant and unabated for years. Struggling states will eventually request federal government assistance, as they always have in times of fiscal crisis, effectively federalizing state debt. States not in crisis will declare enough is enough. Individuals in those states will refuse to pay their federal taxes. Why pay for other people's mistakes?

3:12:59Some states may try to secede from the union, and a constitutional and civil crisis will erupt. I know this sounds crazy, but I think at some point, states will just say, like people, citizens will be like, Why am I sending 30 % of every dollar I make to bureaucrats in Washington that hate me? And so I think this sounds wild, but I don't think it's as far-fetched. Yeah, it's very dark. It's chaos. Nothing like a little Friday black pill. The answer is clearly just AGI pilling all of the California regulators. Just say, AGI will solve the deficit. We will just ask AGI to fix the debt, fix the fiscal crisis, fix the pension liabilities, fix the budget deficit.

3:13:53Don't make mistakes. And so much value will be created by AGI here. Just the income taxes will pay for that a thousand times over, a thousand times over. So you don't need to, you can respect private property. Don't worry. It's going to be fine. AGI is here to save it. That's the solution. Got to pitch them. Anyway, back to the AI world. We have Alex from Higgsfield AI in the Restream Waiting Room. Let's bring him in to the TVP and Ultradown. Alex, how are you doing? He's doing great. Great to be back. It sounds like you're doing great. We were just talking about the billionaire tax. You got something that, fortunately, your business is doing so well.

3:14:35Yeah. Focus on growth. Focus on the good things. Tell us the best news in your world this year. Give me that mallet. Oh, yeah. Give me that mallet. Yeah, what's the latest? What's the latest metric? Absolutely. So what's exciting about Higgs field is that we see emergence of AI native social media agencies. So really teams from maybe 10 to 30 people making commercial ads end-to-end with AI. Yeah. Yeah, we just talked to Sean Frank. He said he's a huge fan. He uses it for Ridge Wallet and has been able to generate custom ads for tons of different markets, local areas where it was just completely unfeasible to cast so many actors, to play so many different parts for every single different college team that he sells products for.

3:15:31Yeah, it's a fascinating time. So this is mostly driven by agents or brands going direct. What's the biggest driver of growth for you? Absolutely. We are excited to see Fortune 500 brands really hiring, primarily hiring, really hiring smaller agencies. Yeah. So that they then use Hipsville to make those ads. And I think, although the most exciting segment is definitely direct to consumer, especially when AI can deliver not just the efficiency, not the cost efficiency, but also higher level of personalization. So the example which you just brought up completely makes sense. Whenever someone needs to customize for 30, 60, 100 different user groups, let's say sport teams, Gen.AI delivers against this goal, and Pixel is at the forefront of that.

3:16:25What are the compute costs looking like these days? I mean, the big news is that you jumped from$100 million to$200 million top line in just two months. That's insane. Congratulations. Jordy, hit the apple. Congratulations. But on the flip side, I mean, you are raising money. Obviously, the capital markets are open. but if you were really spending on inference, that could be difficult. How are the margins developing? How are you optimizing for inference cost? What are you seeing on just inference cost trends broadly? Are you happy with the progress? Or what else can you tell us about the cost structure of this business?

3:17:18I have good news and I have bad news. Okay. Look, I think the good news is that video models get more capable so that over time marketers can just upload all the brand assets in one place and then the model just remembers all the contexts. And this helps to make not just one video but multiple videos and extend existing marketing campaigns. I think this is going to be a major unlock. Think about marketing agents who actually remember all the previous campaigns, all the performance of the previous campaigns and so on. Yeah. Although there is a bad thing. Like these models, they get bigger and bigger.

3:18:08There is no other path around. I think most of the video models this last year were maybe between 10 and 50 billion parameters and it is going to scale well beyond 100 billion parameters. So I think, although the good news is that in the same time we're seeing customers spending thousands of dollars a month on the platform. Which tells us that Slay likely make thousands of videos. Cost per one video is still quite below one dollar. Yeah. This is a very interesting time. So as we move from a concept of. Well, yeah, that's an insane. That's an insane reduction in cost because. Even the most efficient video shoot on your iPhone, you're going to spend a day, even if you're paying a market intern,$20 an hour.

3:19:01Yeah. When brands are doing trying to generate UGC content historically, you know, now they can use Higgs field and generate stuff on the fly. but historically you'd be like, okay, I'm setting a budget of like$200 an asset, and then you'd have a bunch of people like actually record on their iPhone to generate it. And so if you're saying it's, you know,$1 an asset or sub -$1, it's like a 100x reduction in cost. It's pretty significant. Absolutely. I think cost reduction has already, I mean, it has already happened compared to physical production. And I think this batch content generation, meaning that we make not just one video, but hundreds of videos simultaneously.

3:19:42I think this change happening this year. So the goal for Higgsfield is to lead this new wave of agentic behaviors and build cursor for video. What about ingesting the data on performance? Do you have API integrations with all the big platforms where you're ingesting ad data from meta and app loving or whoever else you're distributing video content through? Or is it PDF upload, CSV upload? Like, how are you thinking about actually creating that feedback loop? Because that seems extremely powerful. Yeah, absolutely. on that angle, I can tell that probably 2027 is going to be a year of major revolution in EdTech worlds as connecting generation with performance data is going to become a major point.

3:20:37This year, we collect a lot of these data from social media platforms. We build most of these integrations and establishing direct relationships with the platforms. We collect a lot of very valuable data points directly. as we see which generations results in an actual ad, so that we can actually run reinforcement learning. So streamline content creation process. I think this is the main priority for 26. And 27 is going to be a year of full revolution in advertising technology space. What's going on in the hardware side? What would be most valuable to you? We saw NVIDIA acquire Grok. with the queue.

3:21:24Grok's whole pitch has been very fast inference, usually for text-based LLM models. When I think about the work of a marketer generating a bunch of different video variations, if they fire that off and they come back an hour later, that doesn't seem like a terrible experience. So are you more focused on higher fidelity, larger parameter counts, over speed? Are you interested in what's happening in custom silicon in a particular area or do you just want more and more nvidia gpus at your disposal like what's what's important on the infrastructure side for your plans over the next few years this is this is a great question and i think we have seen that custom silicon can perform really well gemini 3 is the first multi-model llm which has these emerging capabilities, which is not a banana pro model.

3:22:19This model essentially made Adobe obsolete. Yeah, who needs to go to Photoshop? Yeah. Yeah, look, I mean, I'm not saying like, especially for social media marketing, if we're being serious. So, I mean, going and if I say, hey, Georgie, you assigned me to go and make five ads for you a day and try completely different stuff. I don't have time to go to Photoshop and push every pixel to make it perfect. I better use a model like Manabana Pro to really deliver against the vision. And it's semantic-based editing, not just like pixel pushing. I kind of miss when you used to see a really high-quality Photoshop on the timeline, and part of why you appreciated it is somebody really needed to put the effort in, or even if they're not great at Photoshop, they hired somebody, and nowadays it's like, oh, you just have.

3:23:22You know they got the lasso tool, and they cut that character out and slapped that text over there, threw a drop shadow on it. Now it's just a prompt. It's fascinating. Now it's just a prompt, and it's done with custom silicon. So I think we're going to see a renaissance of custom silicon for sure. I mean, Gemini and Nanabanana Pro are an amazing model with these emerging capabilities. And we really embrace those creators and social media marketers who create ads with AI end-to-end. So they don't have to go to Adobe Tooling at all. how are you tracking the problem of understanding what's AI generated what's real this is now a daily I feel like it used to be like even like 12 months ago it was like you know once a month you'd see something and you're like oh is that AI or is that real now it's like almost every single day I see stuff on Instagram all the time where I know it's AI but then I go to the comments and everyone's happy It used to be people would clock it, and then even if they couldn't tell, they would key off of the comments and then kind of pile on.

3:24:33Now people are just like, yeah, this is good content. What's your take on AI detection, how it's working its way through the internet? I think that we're going to see more standards emerging. I think in the European Union, it's a big thing. We're going to see something by the end of the year coming on that front, I guess. I can just tell you that direct-to-consumer brands, they primarily don't care if that looks like AI-generated content or non-AI-generated content. As long as it catches eyeballs and it sells, they just keep coming to Higgsville and making more of such videos. So the best-performing AI-generated ads, they look like AI ads, and they don't look like real ads.

3:25:17I think that's my main observation. Yeah, yeah, yeah. I mean, there's a certain element where, you know, the iPhone, the selfie video, it never actually wound up looking like an HBO show or a James Cameron film. It never wound up looking like, you know, it was shot with a$100 ,000 camera. But everyone just got used to it, and they liked the aesthetic, and we grew accustomed to that aesthetic, and it has its own aesthetic. And so if AI content, even if it retains its aesthetic for a very long time, even though I think it's going to be indistinguishable, it already sort of is, but even if it does, people could just become fans of that and be like, yeah, I like the way it looks.

3:25:57Yeah, interesting. Anything else? Very cool. Well, also, my buddy texted me. It's Alex Kasson. You guys just had lunch? No, no way. Absolutely. We're just chatting about that. small holes. Yeah. He, he's a great guy to know, uh, in this, in this town. So are you in LA right now? Yes, I'm in LA. You should have come, you should have come by the studio next time. Could have been sitting here. You could have been hitting the gong. Could have been moving the goalposts next time. Uh, absolutely. Anyway, thank you so much for taking the time. Well, yeah, congrats on the progress. The growth is absolutely insane.

3:26:31Uh, and look forward to talking more this year. Yeah. We'll talk to you soon. Thank you guys. Have a great weekend. Go bye. All right. We can close with one of the best ideas that I've heard recently. It's from Ryan Peterson. He says, we should have an AI transfer portal like in the NCAA. Couldn't agree more. Well, that's our show. The bomb has been planted. Thank you for tuning in today. Thank you for watching. Leave us five stars on Apple Podcasts and Spotify and sign up for our daily newsletter at tbpn.com. We send out our newsletter every weekday. Enjoy the weekend. Enjoy the weekend. We will be live Monday.

3:27:11We will be live Monday. Looking forward to it. And we'll be back at 11 a.m. Pacific. So tune in. We have a couple live in-person guests. Go touch some grass. It's going to be a lot of fun. Well, hope you have a great weekend. Thank you for being here with us. And goodbye. See you. Bye. Bye. Bye. Bye. Bye.