Sequoia's Leadership Shakeup, Anthropic's Road to Profitability, 𝕏 Timeline Reactions

27:51Like these people just, I don't have a lot of, maybe it's like, I don't know, maybe it's something to do with like our network. But it seems like you can basically get a two to three year head start by dropping out, right? And by the time you're in your 20s, like you look, you kind of appear and are accomplished like somebody that is in their mid 30s. Yeah, yeah. You just got to be ready to kind of build your own thing. If you're not capable of that, like you're going to have a bad time. But if you're ready to like, you know, just kind of go independent and like leave from and just like solve any problem that comes your way.

28:35It's clearly very, very accelerating to your career. Much like getting on Vanta, automate compliance, manage risk and accelerate trust with AI. Vanta helps you get compliant fast. And we don't stop there. Our AI and automation power everything from evidence collection to continuous monitoring to security reviews and vendor risk, whether you're starting up or scaling. Couldn't have said it better myself, John. Let's get into this highlights from this article in the Financial Times. The State of AI, is China about to win the race? Clickbait. It worked. Let's get into it. Arnaud Bertrand shared some highlights.

29:11He says, this is rare enough to be highlighted. an actually good Financial Times article on the state of the AI race between the US and China. He put race in quotes. Three interesting points stood out to Arnaud. One, the fact that for a general purpose technology such as AI, long-term advantage often comes down to how widely and deeply technology spread across society. This is because at the risk of pointing the obvious, benefits from a new technology come from using it widely, not just creating it. You wouldn't get any benefit from discovering the steam engine in and of itself, what drove productivity gains and ultimately britain's advantage in the industrial revolution was spreading and using this technology in its society and china has advantages here as the article notes since they're unarguably better at orchestrating society-wide transformation as that's all they've been doing for the past three decades they got some reps in uh to the point that u.s semiconductors uh export controls actually helped china and ai since they pushed chinese companies towards a different playbook pooling commute compute optimizing efficiency and releasing open weight models.

30:16John is not so long. I mean, like, the efficiency thing is... But, I mean, the actual... I would love to know about, like... The data point that I have not seen that I would like to see is actual, basically, like, consumption of AI. Like, it's clear that they were able to catch up on creation of Frontier model with DeepSeek, but how much is actually getting implemented? Like if you look at in America, you would go to like what's happening on OpenRouter. Let's look at OpenAI's revenue. Let's look at Palantir's revenue. Let's look at Anthropics revenue, right? And we can see that companies are paying to actually implement this stuff.

30:59To understand how close China is, I'd want to benchmark that a little bit more and try and understand that. Anyway, we can move on. uh good point um the the the efficiency thing is real the other thing that's um uh fascinating is just like the you know there there's the energy abundance that china has right is like helps off offset the the uh like yeah yeah if they can use more energy and be more efficient and then that's that's an edge before we move on to the next point let me tell you about graphite.dev code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster.

31:37Jordy, where were we? This is extremely ironic as the US effectively pushed China to develop models that are more efficient and open source, which turns out to be a better strategy for global diffusion. If you're a company in, say, Germany or Brazil, you can download Quine or DeepSeek, run it locally, fine-tune it for your market, and never worry about API rate limits. But does that mean that you win? It's like, cool. Linus Torvalds is not the most powerful person in computing. He did a great thing by creating Linux. It's awesome. But he's not like, he's no one near as powerful as such an adult.

32:09If anything, it seems like you kind of hurt the anthropics and the open AIs of the world that have, we, Chesky was, came out and talked about using Quen models at Airbnb because they were fast and cheap and they were. I'm dying to talk to the CEO of Red Hat because Red Hat Linux is built on top of Linux. It's the ultimate, like the apex predator of the for-profit business built on top of an open source software product. And to actually understand how much leverage you get from developing something like Quen or DeepSeq is a very open question to me. I feel like it might might be two orders of magnitude less value capture.

32:54If you look at, I mean, all of the Linux value creation, I mean, the creation's immense, but the value capture is, I think, on the order of tens of billions. And I think that the closed ecosystems, iOS and Microsoft Windows is in the trillions, maybe? I don't know. It's hard to, it's hard to. Last point here, which I think is the most notable. the oft ignored point of societal enthusiasm. As the article notes, the public in China is the most optimistic in the world about AI, far more than in the US or the UK. Optimism can be a powerful fuel. Our nod says, I often lament about this as a European.

33:34We often have a doomist take on new technologies, AI very much included, which is a huge repellent for entrepreneurship and technological diffusion. And that ultimately hurts Europe in a lot of ways, precisely because the first point above, advantages from new technologies come down to how widely and deeply it spreads in society. If China is the most enthusiastic country in the world about AI, this in and of itself will help drive the widespread adoption that per point one is actually what matters for a long-term competitive advantage.

34:05DJ Levy says, the point made about operating under constraints reminds me of how post-war Japan adopted, fine-tuned, and promoted lean manufacturing to overcome scarcity. Interesting point. There's an article in the Wall Street Journal today about what happened behind the scenes between NVIDIA and China. The reporting says that the Trump administration, Trump aides sank NVIDIA push to export AI chips to China. President, hearing from his inner circle, opted not to discuss policy shift with Xi Jinping. They met in South Korea and an urgent issue emerged. Trump wanted to discuss a request by NVIDIA chief executive Jensen Wong to allow sales of a new generation of artificial intelligence chips to China.

35:01Current and former administration officials said greenlighting the export of Blackwell chips would be a seismic policy shift. potentially giving China, the US's biggest geopolitical competitor, a technological accelerant. Wang, who speaks to Trump, often has lobbied relentlessly to maintain access to the Chinese market as they prepare to meet Xi Jinping. Top officials, including Secretary of State Marco Rubio, told Trump the sales would threaten national security, saying they could boost China's AI data center capabilities and backfire on the US, officials said. And so I still wonder, I mean, it's like NVIDIA is a$5 trillion company.

35:41Like if they could sell to China, how much? Like it's not like it's a$10 trillion company, is it? Is it that big of a deal? Like how big is the China opportunity? It's the second largest computing market in the world. It's huge. That's right. But so do you think that there's$5 trillion at stake here? Is that how big of a deal this is? I don't know. Depends on how AGI-pilled you are, John. I guess. Yeah. I don't know. Yeah. I mean, if you think AI is safe, but you want to sell it everywhere. I don't know. I also just wonder, how many American tech companies have really been able to do well in China over a long period of time?

36:20It feels like even if you have the blessing of the United States, you won't have the blessing of the CCP forever. So it seems like a very temporary thing. Tesla was selling into China as long as well, as long as it took for them to copy it. Granted, chips are quite a bit harder to make than cars. Before we move on, you might be aware that on the other side of the country, there's a mayoral election. New York City is choosing between Mamdani and Cuomo. And you can go to polymarket.com slash NYC to track it. There are 12 hours, 22 minutes, and 33 seconds left. And Mamdani is at 95 % chance. Cuomo's at 5%.

37:03Up from around 90 earlier. Yes, they flipped in July. And Mamdani has been grinding it up from the 70 % up to 95, where he is right now. And his odds look very good. But there's a lot of people out there saying, smart money's on Cuomo. Maybe he'll win. I think the degenerate money is on Cuomo. That's probably true. I mean, yeah. Somebody said that to me and I was like, I don't think that it's like hard to pull Cuomo voters. Like, I don't know that much about it, but it doesn't feel like the Cuomo voters are like under a rock. Like, oh, they don't pick up the phone. Like, if anything, it's like the Mamdani voters.

37:44He's bringing out like a new voter. So I don't know. You can go track it at polymarket.com slash NYC. There's been some debate on whether or not foreign whales could be pushing up Sauron for some reason. But there's so much volume on this market that it's, I think, at a scale that someone would have to be extremely motivated. I always forget that I believe we were doing the show during the election last year. And I think we were so locked in on technology, we barely covered it. I'm sure. But I think the election happened because the election was in November. And we started the show in October.

38:34Yeah. So that's like weird. Well, that was the first. That was the first. We just didn't cover it. I mean, but that was the first election that everyone was glued to the prediction market odds. Totally. Yeah, yeah, yeah. But also I think that it was exhausting and every show was obsessed with politics. And so it was kind of refreshing to hop on the microphone and actually just talk about technology. Talk about schizo and on posters. Break them down. Exactly. And there was truly, I still maintain that technology is like the most important story of 2024, for example. and tracking what's happening with open AI is just as important.

39:18Most important to us, that's for sure. For sure, for sure. Just like Julius.ai, the AI data analyst is the most important AI data analytics company. Connect your data, ask questions in plain English and get insights in seconds, no coding required. What's this post from Martin Shkreli? So a couple accounts we're sharing. Yes. A couple big shorters we're sharing that. the Hopper GPU rental index with a big move down in late October, now down 28 % year to date. Jess Dario says no signs of GPU oversupply whatsoever. And Screlly comes over the top and dunks and mocks by sharing some NeoCloud. I don't know which one exactly showing that every single service that this cloud is offering is out of capacity.

40:14So A100s, H100s, B200s, GH200s, et cetera. So definitely a tale of two chips. Yeah. It seems very, very hard to assess the market based on spot pricing in the single dollars per hour. The amount of liquidity in the market, if you want to spend 10 grand, is probably way different than wanting to spend$10 billion. dollars. And when you look at what the hyperscalers and the frontier labs are, are talking about, uh, I don't think that they're, that they're happy one way or another to Hoover up a couple little instances all over the place. I think, I think they need, they need a real data center. Yeah.

41:15Um, but I don't know, we, we haven't, we haven't heard, like, this is so interesting because like, have you, have you seen reports? Like we see reports when like a startup's website goes down, right? Because it's like, oh, maybe AWS went down or maybe like, they're just like not, they're being sloppy, right? Like, like if, especially if it's a company that somebody like wants to dunk on, right? Like a Cluely, like if, like if Cluely was like out of GPUs, like people would dunk on Cluely, right? They were dunking. You didn't see this? I mean, the people were, were showing. but not for GPU. Yeah, they were just showing that the service was down.

41:54Yeah, yeah, yeah. That was just like the status page. And there was actually some FUD about that. Like I wasn't sure if that was super real. But there are plenty of reasons why in any oligopolistic like startup market, if there's two companies that are duking it out and one of them is like, oh, they couldn't secure their GPUs to do their inference, to do their AI agents thing. One of them would dunk, right? You'd think you'd see this, but you don't. And so I'm much more on the, like, there are generally enough GPUs to go around right now. But I don't know. I'd like to know more. We do – I know Tyler does run into GPU constraints every once in a while.

42:34But it hasn't been, like, a major, majorly gating factor. Most of it's available. And if you need some GPUs or on-demand clusters for generative media, head over to Fall, the world's best generative image, video, and audio models all in one place. Develop and fine-tune models with serverless GPUs and on-demand clusters. Trusted by our friends at Shopify, Quora, Perplexity, Adobe, and many more. This was the data that I was referencing earlier in the show. Yes, yes, yes. QCAP shared year-to-date, or actually, sorry, monthly performance from Canva, Wingstop, Shake Shack, a bunch of other restaurants.

43:14Is Chipotle in there? CMG. Ooh, CMG's down 22 % in the last week. In the quarter, they're down 27%. Hey, they're only down 47 % year-to-date. That is so rough. But yeah, really rough. Wasteland Capital shares most of these restaurants. stocks were trading at 35 to 100 times forward PE just recently. Many still are. They were valued at a massive premium to the Mag7 and any comparable growth stocks in tech and AI. This is just a bubble deflating. Bubbles can't be sustained indefinitely. Let me give you some more context on Chipotle because Chipotle was actually in the journal today, B10. First, let me tell you about Turbo Puffer, Search Every Bite, Service Vector, and Full Text Search, built from first principles and object storage, fast, 10x cheaper, and extremely scalable.

44:02So, uh, the, the, the Wall Street Journal says Chipotle's big bet on younger consumers is unraveling. An uncertain us economy is catching up to younger Americans and Chipotle Mexican grill. The Mexican inspired chain has spent years cultivating a younger clientele, pitching its burritos and bowls as nutritious protein packed meals for gym bros and healthy eaters. What do you think, Ben? Are you still a fan? Do you still love it? I still love it. I still love it. You're over it. It's so tragic. Me and Ben used to go to Chipotle every day. No, I used to get easily 90 % of my calories from Chipotle at certain times in my life.

44:41And any time you were on a long drive, knowing that at some point in the drive, you could pull over and have a nice, fresh, nutritious slop bowl was just something to look forward to. Oh, it was amazing. And now the last time I pulled over and got Chipotle, I'm like, oh, this is going to be rough. Is it a quantity? It's a quality. It's a quality. It's entirely a quality. The quality has degraded to such an extreme degree that it's no, I'd rather fast than eat Chipotle in almost any situation. Obviously, I'm extra picky. So apparently they've been going younger. They've been targeting younger clientele.

45:22and it says here, Chipotle sponsors video game competitions and some dedicated fans dub themselves Chipotle boys. That sounds like something they're trying to, I haven't seen this at all. Have you seen this, Ben? That's not good. I would never self-identify as a Chipotle boy. In fact, the existence of the Chipotle boy makes me never want to shop at Chipotle ever again. Millennials and Gen Zers have helped power Chipotle sales despite rising menu prices and slowdowns at rival chains. In recent months, though, a harsher economic reality has set in. Rising student loan payments, stagnant wage growth, and higher health insurance costs have bumped burritos down younger consumers' priority lists.

46:04What are they eating instead? Dude, I've been in this scenario. When I first came to Silicon Valley, Chipotle was like a once-a-month luxury. No, at that same time. It was literally ramen noodles. You go to Costco, you get ramen noodles. Ramen noodles is truly 10 cents a meal. Like it is the cheapest, cheapest thing. I mean, yeah, the, I, I, I sort of look back on the days when Chipotle was felt and was a luxury when I was like, okay, getting extra meat and guacamole, extra, extra guac and extra meat was like, I am a king. Were you, were you like a frozen chicken breast guy too? Yeah, for sure.

46:48Because I'd just be looking at the huge bag, you know, one pound bag of rice and the frozen chicken breast being like, again, we meet again. We meet again. And then I'd be like, I got to mix it up and have chicken and rice from somewhere else. They're eating at home more often. They're eating with us less frequently. Chipotle CEO Scott Boatwright, I like that name, said in an interview. One is Sean Chopra, a 23-year-old law student at UCLA. Chopra typically stopped at Chipotle several times a week, but lately he decided he should make more Mexican-style food at home. He's DIYing Chipotle. It isn't that difficult to replicate what you get there.

47:30That's a great point. He has to cut back his Chipotle habit to twice a month. Chipotle said last week that customer traffic slid over the three months ended September 30th. For a third consecutive quarter, the company cut its forecast for same-store sales. Consumers from 25 to 35 who account for roughly one-quarter of Chipotle's sales are pulling back on visits. Other fast-casual restaurant stocks, including Cava and Sweetgreen, also declined ahead of their own earnings reports scheduled for the week ahead. Thrifty Eaters, there's more stuff here. So revenue is growing because they're opening new stores.

48:04but at existing stores, you see a lot less momentum. They've got to go upmarket caviar bowls. Yes, I'll have double caviar. I'll have double beluga. Can I get double beluga? Yeah, we've got a double beluga with extra caviar. Yeah, can I have Wagyu? A5 Wagyu bowl with double beluga on top. Can I get that? That's the only way that they're going to survive. They've got to go upmarket, the young people. Do you remember the moment that you heard that you could just order, even if you got a bowl, you could get tortillas on the side? I never did that. I've never once wrapped my own. No, it wasn't even, for me, the calculus is, I wasn't even going to wrap it.

48:50It was just extra calories. Extra calories. Okay. Okay. That's a true min-maxer. I respect that. I respect that. I also respect Google AI Studio. Create an AI-powered app faster than ever. Get started. Gemini understands the capabilities you need and automatically wires up the right models and APIs for you. You can get started at ai.studio slash build. Let's move on to the timeline. Bern Hobart, we got to get him on the show. It's been too long. A bunch of people have mentioned how cheap restaurant stocks are. So I looked at Sweetgreen and I would love to know what first principles you could use to deduce the existence of a fast food chain that pays 5 % of market cap per year in stock-based comp.

49:30This stock-based comp thing is very, very interesting to me. I think Sweetgreen has to get this number up more into the Snapchat range, where it's every 10 years, they basically give away the whole. So it's 10 % of market cap per year. This is an interesting metric. I would love to know across a wide swath of companies, market cap per year in stock-based comp. This is a deep research report. Somebody's got to fire it off. I want to get the Sweetgreen founders on. the show ASAP. I'd love to. Jonathan or Nicholas. I wonder, yeah, it just feels like at a certain point, you're a mature business, you can just pay cash, right?

50:12But maybe stock-based comp is just advantageous because it's cheaper than cash somehow, or it changes the financial profile. It feels like something that's more, it's more just financial engineering and it doesn't really create any value, but like you have to play the game on the field. Right. I think they've, they've always invested more on the software side than other fast casual chains as well. And so I just imagine this, a lot of this is executives and then. Yeah. I'm just saying like, there is a world where, where instead of, instead of taking 5 % of your market cap and giving your employees shares, you just sell 5 % in the open market, you get the cash and then you pay the employees cash, right?

50:56Like that's, those two should be economically equal. And yet certain companies, Snapchat among them, will pay stock-based comp. And I feel like it's because the public markets price things differently. And this is just like the reality of the public markets. Like right now,$1 of AI contract revenue is worth$4 in market cap. Like we just see this again and and again and again. And so if the markets are pricing things in a certain way, you have to put the game on the field. $38 billion over seven years is worth$150 billion. It was a crazy moment. Of added market cap for Amazon. Yeah. I mean, at the same time, the reason you could price that, just to steal man a little bit, the reason that you could price a$38 billion contract at 120 of market cap or something like that is, well, it's not just, what's going to happen after that 38 billion runs out?

51:57They're going to size up the next contract because they're going to keep growing. Like this is the beginning of a long relationship. It seems like this is a capital intensive business. They're building restaurants. It seems like you'd want to compensate employees with the stock so you don't you like you have cash i'm sure it's not like an necessarily an infinite bid yeah yeah um in other news mustafa suleyman says it's an understatement to say a lot of us haven't felt good about our relationship with technology so the absolute last thing we should be doing is making that relationship romantic let's hear from shots fired uh let's check in with the stock the stocks down 0.69 % today.

52:40That's not that much. What is that? It's like 60 billion now or something like that. Well, a 1 % move on them is 40 billion right now. So 38 billion. So it's 30 billion was the price of the adult content in Excel that was just... Yeah. Or it was already priced in. It's not going to happen. It was always mostly priced in. Anyway, you've got to get your brand mentioned in chat, GPT. You've got to go to Profound to reach millions of consumers who are using AI to discover new products and brands. Let me get started. Josh Wolf, who I believe we might see today. I'm excited to hang out with him. He says, history was just made.

53:25Lux family company, Anduril Technology. Anderl has just successfully flown the first ever fully autonomous fighter jet ever in U.S. history. So this is the office of the Secretary of the Air Force shares. Today, the Air Force made continuous progress in the CCA, the Collaborative Combat Aircraft Program, with the first flight of the YFQ-44A developed by Anderl. This is a milestone that shows how competition drives innovation and accelerates delivery. He says one of two and then maybe doesn't. It is a crazy looking jet. Yeah. Like at some point it just starts looking like a missile. Was this, so the mainstream narrative is that this was created by humans?

54:15Yeah, and I find that almost impossible to believe. Andrew wants us to believe. Yeah. that's the just like the Aston Martin Valkyrie everyone wants to say oh this yeah human hands yeah yeah yeah yeah yeah yeah yeah yeah they got a carpenter they got a carpenter guy with a saw pull up that thing yeah it's like alien technology it does it does make it make sense make it make sense make it make sense make it make sense uh linear do you think linear was built by human hands also hard to believe the design I know we know the team very well it could have been discovered though. It could have been something that they unearthed.

54:50That's right. And it was, it was built by intergalactic SAS that landed here on earth. It was discovered by humans to help build more and more great software in the world. Plan and build products. Yes. Uh, you can go check it out at linear. Meet the system for modern software development, streamline issues, projects, and product roadmaps. You remember when we talked about that UFO ETF? Oh yeah. Yeah. Well, let's checking in on that. How's it doing? Is it ripping? It's probably way up. Told me it's up because they had good stuff in there. Like that was the funniest thing. Okay. Procure space ETF, NASDAQ ticker UFO.

55:26Okay. UFO 40 % all time. So you bought it in 20, you could have bought it$25 a share in 2019. It's currently sitting at$35 a share. Invest in aliens, people. Simple. The Procure space may provide diversification beyond the limitations of solely earthbound companies. And I'll read from their investment objectives. Since the beginning of humankind, our ancestors have looked to the skies with immense curiosity in search of answers. Space has always captured human interest, but recently the space economy has also captured commercial interest like never before. UFO may provide diversification beyond the limitations of solely earthbound companies.

56:09This is huge. This is huge. So, yeah, it's up 40 % this year, basically. How do they have SiriusXM in their top 10 holdings? They have satellites. They're out in space. They're literally out in space. How do you think they'll make contact? What's the UFO theory behind Sirius? If not through Howard Stern recordings being blasted out into the deep depths of space. That's how we make first contact with Stern. Howard Stern. I can see it. Broadcast on SiriusXI. I don't think he's actually broadcasting anymore on there. Did you see this company T1 Energy? Yeah, you mentioned this and Harris nails it.

56:55Harris Rothermel says, where are all these companies coming from? We're so back. And I completely agree. Like we've been talking, everyone, everyone in tech has been talking about, we don't know how to do this. Where are the solar panel companies? Casey Hanmer is doing a whole press tour. He's been on door cash. He's been on cheeky pine coming on this show multiple times. Like Casey Hanmer has been saying like, we're not making solar in America effectively. And here's this company that's just, uh, it's just cooking, setting out 14 megawatts of solar panels in a single day. They're producing 5.1 gigawatts annualized play pace.

57:28I very, very worry. I worry that there's... Jerry Rigg Everything says he'll go and make a video for free. That sounds awesome. I hope he does. So the thing that was worrying for me on this front is that the company T1 Energy is just TE on NYSEE, is valued at$683 million, which felt very low in comparison to some of the American dynamism companies that we've had on the show who have valuations far beyond this, despite not even 10 % of the overall production capacity. I really am so confused. Like, where did this come from? This was something that everyone was saying is not happening in America, impossible, like we're so behind and this seems really good.

58:19I guess we need to actually handicap this against the real production numbers in China. What is demand? Also, it seems like this is maybe a newer company that just hit 14 megawatts. It's not like they're actually producing 5 gigawatts. So the company initially started as Freyer Battery. It was founded in Norway. It merged with Alusa Energy SPAC, listed on Nicaea's Frey. That was 2021. 2024, they bought Trina Solar's 5-gigawatt Wilmer, Texas module plant. And only earlier this year, they rebranded T1 Energy. And so this was an existing company that's backed in the US that then bought an existing solar panel plant.

59:13And so this didn't just get spun up overnight. Very typical spec. I guess it went out at$10 a share, roughly$4 billion market cap, I think. And now it's traded down to$4 a share,$680 million market cap. But as we've seen with a number of these SPACs that went out and got crushed after the interest rates, some of them build back. Some of them just make it work after a while. I love the story of Dave. You know Dave.com? I think it's dave.com. Oh, yeah. So it's like a loan company. I know one of the co-founders. It's Neobank. Yeah, it's Neobank. With a lending component. So went out at$4 billion.

59:58Everyone was like, awesome. Like, we did it. Trades down like 99.999%. The chart is just absolutely brutal. The chart is insane. It goes down. So they've done a number of reverse splits because I think they actually went out at like$10 a share or something. And now it's a, so, so if you split adjusted, I think it went out at like$310, traded down to like five bucks. So literally down 99%. I think it was trading like well below cash for a while. And now it's back up and is a$3 billion company. And so if you held on through all of that, but there were like multiple people who basically at the IPO were like, I, I, I'm worth a hundred million.

1:00:44or like my fund is returned. And then it was like, not yet. You're going to need to wait. And then they actually came back. It's such a crazy story. I mean, that was during peak bearishness around neobanks. Yeah. And anyways, it turns out that high interest rates - Cheers, salute to T1 Energy. Love what you're doing. CEOs, welcome on the show. We'd love to talk to you. This would be fantastic. And if you need to sell your solar panels online, you got to pay taxes. head over to numeralhq.com. Sales tax on autopilot. Spend less than five minutes per month on sales tax compliance. What else is going on?

1:01:22I don't think we can play this video of Satya Nadella, but the BG2 podcast is still burning up the timeline. It is the current thing. Brad knows how to start a conversation. Martin Screlly is forecasting that NVIDIA will get to$100 trillion in market cap. Human thought is values by businesses. Even desk email jobs require human thought, but higher price jobs like legal, medical, finance can also be replaced by AI. He's got a crazy bull case here. We will see$10 trillion, as I prophesied years ago. This is short-sighted and extremely lazy analysis. Stock mic goes down. So he's putting the screws to the NVIDIA bears.

1:02:13This is a nightmare post for NVIDIA bears. You're telling me we got 20x upside from here? At 25 % of the cost of a human thinker, a digital thinker is a great bargain. $10 trillion in revenue, creating$40 trillion in additional human thinking work. $100 trillion market cap. That's his math. Pretty remarkable. And... I don't know. Yeah, I think it would be awesome if Nvidia can be, you know, scale to being valued at the global GDP for 2024. Yeah, I mean, the crazy thing here is that he doesn't, I mean, it's already worth two Canada's GDP-wise, although you can't really, it's not really apples to apples.

1:02:55But I believe Canadian GDP is 2.5. Yeah, 2.2 as in 2024. Yeah, 2.2. But again, GDP is more of like a revenue figure than like a valuation figure. So you can't, it's not really apples to apples, but it's a funny comparison of the size. Anyway, let me tell you about fin.ai, the number one AI agent for customer service, number one performance benchmarks, number one competitive bake-offs, number one ranking on G2. Do you want to read this Chris Pike post about the end of cloud inference? Is this worth... Looks very long. I'd say we should just have him back on again. But I would be very interested to hear this thesis.

1:03:33I think it is a little bit too long to read through the whole thing, but I'm glad that he's... And the same thing for Casey Hanmer. We should have him on the show to break down his extreme hot take on NASA's Orion space capsule. He calls it hot garbage. Flaming garbage. We need to do one of those complex style trading cards for one of the TBPN trading cards for this quote. This is a bombshell quote. And in other contexts, news outlets would put this in big quotes and be like, quote, flaming garbage. Like Casey Hanmer declares NASA's Orion space capsule in salacious. They've spent$31.6 billion, and they're planning to fly it around the moon with four astronauts next year and also test the life support system and heat shield, which kind of sort of failed on the last flight.

1:04:26Anyway, this was Miserable to Write. I hope you enjoy it. Link below. Oh, so rough. We got to talk to the astronauts that are potentially going to be going. There's another, what is this post by Robin here? Robin says, you can't make this stuff up. SpaceX Starship, 2.7 billion HLS contract cost includes docking adapter. NASA Lockheed Orion, 31.6 billion cost and counting. 19 years in development, docking is an extra 2.5 billion extra. Wow. Yeah, the dock. I mean. The dock's going to be an extra two and a half. We're sorry. We know it's already been 31, but we're going to need another 2.5. It's rough.

1:05:17Get this functional. Rough. Well, if you want to dock your body in your bed, get an 8sleep. Go to 8sleep.com. Transition king. Get a pod five. They said you wouldn't be able to do a transition like that. I got an 86 despite the early takeoff this morning. Seven and a half hours, not bad. I got a 77, I think. There we go. Let's see. But let me see. I think I forgot to turn it off in the morning. Yesterday, I got an 87. Today, 77. Not quite as good, but my REM sleep was up 17%. There we go. There we go. Not bad, John. Clearly pivoted. We barely touched on this. Maybe we mentioned it. but I don't know.

1:06:03Did they actually pivot? It was, it's refinement. It's certainly like a shift in the messaging. But let's go through some of the hot takes. And then let's give our sort of analysis. So Hero Thousand Faces, great account, says all this to be an AI note taker app number 87, blah, blah, blah, blah, blah. Nizzy says, clearly didn't pivot. They're simply selling. They're simply not selling to 20-year-olds anymore. So the Cluelly homepage now says the number one AI note taker for meetings. How would that have happened? There's so many. How do you jump straight to number one? Always claim the top spot.

1:06:43Number one in what way? Number one in newness? I don't know. But the pitch is to get accurate notes, live transcripts, and follow-up emails at the end of every meeting. You can get it from Mac. I felt a little bit bad because I was going so hard on AI note takers last week. Yeah. You could read Toby Lutke as sub-tweeting this. No one had seen this new landing page. But then Toby Lutke just posted last week, I'm not a fan of those AI note takers that come into the Zoom rooms. Well, he kind of - He said it was the equivalent of coming into a meeting with your fly down. Doesn't sound right. He kind of clarified it and said, I like some of them.

1:07:27I just don't like bringing an entourage into the meeting. The entourage is hilarious. So, I mean, he said that bringing an AI note-taker to a Zoom meeting is the equivalent of having your fly down, which seems negative. And then you were like, I don't like it. I'm not letting them in. But I think, clearly, to their credit, is more of a desktop product, which means it wouldn't have to actually join the meeting. which is so maybe they're building with the direction that people want i think what they're doing here is is uh is like a half pivot right yep uh they uh they had a lot of traction around viral marketing towards a cheating tool yep and i think they realized that that wasn't going to be a great you know a trillion dollar market and so they decided hey let's pivot to enterprise sass They had always been, I think, doing some of this, right?

1:08:24I remember Roy talking about, when he came on our show, talking about how, you know, Fortune 500 CEOs and his DMs, that kind of thing. But a half pivot is always dangerous because they're using this Cluely brand, which they built a very controversial brand really quickly and then now need to kind of leverage that brand and the initial technology software that they built to try to pivot into a category that is kind of, I mean, just squarely at odds with the original brand, right? If you have like an AI note taker, you want to have full faith and trust in the team around a number of different things.

1:09:11There was another post in here. I'm trying to find it. around

1:09:19Ilya Sukar says, never kill yourself. There's so, so many post-granola depositions to read in the next decade. Post-granola depositions? What are you talking about? Like, if somebody's in a deposition and the opposing counsel has a bunch of your granola transcripts and they can be like, on this meeting that you had on this date, you said... Everything is admissible in court. if you have an AI following you around. As opposed to, a lot of people would say, okay, be careful about what you put in writing. But if you hop on the phone and you're spitballing on, oh, well, like, will we have a monopoly if we do this strategy?

1:09:56Like, that's not gonna be admissible in court. But once you have every meeting transcribed, even like a little joke can be taken out of context in the court of law. So pretty aggressive. Yeah, the question here is, is there a market for an AI note taker for the kind of people that clearly is really good at marketing too, right? If you just hyperscale an AI note taker to the TikTok, Instagram, younger generation, that's not getting marketed the same way as Granola. The challenge is like, I don't think it's possible for them to make a 10X better product than Otter and Granola. Otter's been around for a long time, but - Fireflies.

1:10:37Yeah. Fireflies has been literally doing it since like 2015. yeah it's it's some of these companies have been so i think i think it's i think it's possible that they'll still have this opportunity but it's possible like basically like killing just retiring the clueley brand and saying like hey this didn't work we're going to take another crack yeah with a new brand and there are also there are also note-taking uh there are also like like non like like i don't know what you call it like vertical sass within note-taking so like Like there are certain, like just for SDRs, call recording and analytics for like you're not just bringing the AI note taker to your general meeting, your staff meeting.

1:11:21You're actually bringing a special one just for your sales calls. And then it's doing analytics on how you're performing, what points you're making, and then plugging all that into your CRM. If you need a CRM, go to Adio, customer relationship magic. Adio is the AI native CRM that builds scales and grows your company to the next level. Daniel said, of course, I'll give the guys who gleefully break every rule access to all my meetings for a product experience I could get from a dozen other companies. What's interesting is go back to the cheat on everything pitch. That was pretty hard because what are college students actually using?

1:12:03Like what, like from what we've heard the stack, if you want to like cheat on your homework, what is the stack, Jordy? It's actually ChatGPT Atlas. It's ChatGPT and Atlas because ChatGPT. I did talk to a student who said, uh, I started offering him dollar amounts on what it would take to get him to quit Atlas. Yeah. We got up to$200 a month. So, so Atlas is, is just like clearly in the sidebar hanging out all the time while you're navigating your homework. And then on the free tier, you get 10 messages every five hours and one message per day with the GPT-5 thinking model. And then also for$20, you can upgrade, get more thinking models, thinking.

1:12:48And for a lot of parents, I bet you the parents will see, oh, chat GPT, yes, that's like a professional tool. Like I'm happy to pay for your$20 a month plan, like because it's like a tutor almost. And even the$200 a month plan, I bet is getting expensed to the parents, right? And so whereas Cluely, you come to your parent, I want to buy this Cluely thing. It's a little bit harder to underwrite. It's much easier. And then not to mention, this was your point that Gemini and Google is giving away a ton of queries for students. And every student in college probably sets up a Gmail account when they get there.

1:13:24Roy posted a couple days ago, sometimes you have to listen to what the A-B test tells you, even if it means changing the words on your landing page. And he shows a chart of when they introduced a new language on the website and the growth accelerates. So anyways, good. Did you see this pleometric post talking to Claude? You have made the same mistake three times in a row now. You should kill yourself. Claude, consider the weight of a life such as mine. You're absolutely right. This is so clearly fake, but it's very funny. And sometimes Photoshop is very, very funny. Hero Thousand Faces back on the timeline, second time in the show, talking about OpenAI marketing, breaking down what all the different labs are looking like.

1:14:17So OpenAI marketing, ominous evil megacorp, OpenAI employees, we are making the world a better place. We are building the most beautiful human friendly future imaginable. Anthropic marketing, the friendly AI and anthropic employees. Every single atom that is not a GPU is an atom wasted. I love checking in on the vibes on the timeline. It's very, very funny. If you want to check in on the vibes of the market, go to public.com. Investing for those that take it seriously. Multi-ass investing, industry-leading yields, trusted by millions. Benchmark 2020, vintage fund, marked at an 8X. Tyler Hodge has the story.

1:14:58They're doing just fine. This is, I think, for the most part, because of Sierra, as well as Fireworks, which we had on, I think, last week. We've had all three CEOs, the Royal Flush. Royal Flush. And I think what Ev Randall should do, new GP over at Benchmark. is 10x that fun again. For sure. Do another 10x. An 8x. An 8x, yeah. On top, on top. Oh, 10x. Okay, the 8x is a base hit. 10x it again. 10x it again, yeah. To an 80x fun. Well, we're going to get him on the show again. We're going to talk to him about our expectations for him, and we're going to tell him, please don't let us down. Just one more 10x.

1:15:44One more 10x. Just one more 10x. That's what the LPs are going to want to see. That's what the OPs are going on. That's what's going to give them faith. Yeah. You're bringing fresh blood in there. You've got to bring in some fresh ideas. You've got to go on adquick.com. You've got to put Benchmark on the 101. Put your face on the 101. This is the first thing you do. Put your face. Hey, do you need venture capital? Call Ev Randall at Benchmark. 1-800-Benchmark. 1-800. Let me just say his phone number right here on the stream. adquick.com. Adquick advertising made easy and measurable. Say goodbye to the headaches of advertising.

1:16:15Only Adquick combines technology. out-of-home expertise, and data to enable efficient, seamless ad buying across the globe. In other news, people are making money on blueberries. That's right. Fruitist. Fruitist raises$150 million to power the next wave of healthy snacking. This is crazy. Fruitist has closed$150 million funding round. That's a huge blueberry. Led by J.P. Morgan Asset Management with Ray Dalio doubling down. We got to try some of these for sure. Valuing the berry powerhouse at over$1 billion. The round also brought in Oaktree, Steve Kaplan, and Alignment. This is crazy. Element Capital, bringing total equity raised to$443 million.

1:16:54The snacking company, best known for its jumbo blueberries and new grab-and-go snack cups, generated more than$400 million in revenue last year with blueberry sales. They got to get an AI narrative here. This should easily be a$40 billion company at this point. with blueberry sales tripling as fruit has expanded to 12 ,500 stores across 40 countries. Wow. They're trading at two and a half times sales. I remember we talked about this company when they made the CNBC disruptors. Yes. And they were like above Anderil. I was like, why is a big fruit company above Anderil? and uh apparently fruit maxing is a good strategy pulled some strings maybe uh henry said charlie munger once said take a simple idea and take it seriously and sometimes the simple idea is obscenely large blueberries do you see how chaotic the uh the comments are on this drew fallon post there's the the food manufacturing guy is in there being like uh yes gmo blueberries are all the rays is like, I don't like the fact that these are GMOs.

1:18:02And then Jeff Leung is here and he's saying, Driscoll's quality is more consistent. And it seems like everyone has their favorite fruit brand that they're pushing on. And it's knockout, drag out fight. So Will.i.am is in the news because he will be teaching a new class on agentic AI at Arizona State University. I don't know if this is the first time Will.i.am has got into the professor game. Professorial. Professorial game, but he will be teaching a general elective course in his Hollywood studio with some students in person and others joining from ASU classrooms. So the same people that thought Jensen signing that woman's shirt.

1:18:54A lot of people are saying this is a tough signal. This is the tough. This has to be the top. Who knows? Kevin Gee is sharing that reminder. Personally, just personally, I think that hearing how well I am is using AI would be interesting. Yeah. I find that interesting. And I'm endlessly entertained by, there's somebody who has a craft. Like you might not like his music. You might love his music, but clearly you know that he makes music. Like he has a process. He has a studio. He makes hits. He makes hits, exactly. And there's a whole process for that. you know scheduling the the studio time okay assembling the music but here here's here's here's why i think it's not so crazy kevin gee is highlighting william was an early investor in tesla pre-elon as ceo open ai anthropic twitter pinterest dropbox hugging face beats and runway he's got exits he's got ipos he's in the labs he was even in the everything app he was in hugging face.

1:19:57So put some respect on Will.i.am. Not a bad portfolio at all. And yeah, I hope they end up open sourcing the class so we can study along with them. I would love to. We should actually have Tyler. Tyler, get ready to go to ASU. Get ready to transfer. We know you're studying physics but we now need you to study agentic ai i like that that's very funny uh in other news there's some in other news bezel your bezel concierge is available now to source you any watch on the planet seriously any watch any watch uh do you know who kim.com is or is this before this was kind of before my time but i've seen some kind of vibe reels and i oh yeah i think this guy He had a wild build.

1:20:46He was running a crazy build. He had a very like mob boss enforcer type build. I mean, he's still running it, right? I believe so. He's 6 '7", according to Wikipedia. And prosperous. Dotcom rose to fame in Germany in the 90s as a hacker and internet entrepreneur. I got to get my way up. Arrested in 1994 for trafficking and stolen phone calling card numbers. He was convicted on 11 charges of computer fraud. So he's definitely kind of mob boss coded. He was a crazy, crazy guy. And so L3 tweet engineer says, you don't really see guys like Kim.com anymore. Modern internet culture doesn't have room for these guys.

1:21:31It's interesting. Yeah, you see them a little. AGM says they exist in crypto, which embodies much of the unhinged ethos of earlier internet waves just siloed and encrypted Twitter and Telegram. So that's true. But yeah, the siloing effect is very, very important. I was thinking about that guy who takes those funny pictures. He always goes viral. He's like the I'm rich guy. He kind of has that aesthetic. But the weird thing about Kim.com was that he actually made a product mega upload that a lot of people used. And it was very controversial. and no one else was getting rich on it. It was for file transfers.

1:22:12So you could, it was, in theory, I could say, hey, we're going to record a podcast. I'm going to send a four gig file to our editors. Like we use, you know, you can use WeTransfer. You could use Google Drive. You could use any sort of air. There's a whole bunch of different services for transferring like files. Calder in the chat says six, seven mentions. Six, seven mentions. There we go. But Mega Upload quickly became used for piracy. So people would use it to upload software and movies. Like maps of the high seas. Yes, maps of the high seas, exactly. So it was just a wild time. Andrew Curran says an AI-generated song has entered a Billboard radio chart for the first time.

1:22:57Zania Monáe, created by a human Talisha Jones, who used Suno. Suno. also signed a$3 million record deal a few weeks ago. Imagine just swapping it up, and then all of a sudden you got the$3 million record deal. $3 million. It is funny that this is an AI song. Yeah, that's 2 % of our revenue. We'd like that. AI song that we can't play on the show. We can't? Oh, because it's on... It's an actual artist. Oh, okay. Yeah, yeah, yeah. That's annoying. What's the difference between that song and the other 103 titles that have made the chart this year alone? Zania Monet is an AI-driven artist, the product of a poet named Talisha Jones, and her song's chart arrival marks the first known instance of an AI-based act to earn a spot on a Billboard radio chart.

1:23:43The move also represents another step in the evolving relationship between AI tools and creators in the music industry. I would like to see an AI musician not share their real identity. Satoshi or Banksy. of music. Not a pop star, but a slop star. We have Michael, one of the co-founders of Suno, coming on the show Friday.

1:24:17If you want to log off, if you want to touch some grass, head over to Wander. Find your happy place. Book a Wander with inspiring reviews, hotel-grade amenities, dreamy beds, top-tier cleaning, 24-7 concierge services, vacationer but better. They can't make you listen to slop if you're in your wander. Gwart says, like most of you guys, I didn't get into crypto to get rich. I got into crypto because I believed in the vision of 500 different stable coins issued by 500 different centralized institutions on 500 different blockchains. The good news is you're going to be able to swap in and out of all.

1:24:50Yeah, I do wonder what the stable result is. Like what is the final equilibrium? Is it just endless proliferation of new coins or is it a consolidation? Because there was a war for store of value. Like there was like Bitcoin classic and I can't even remember that other one. There was the one that was like, it's the gold to Bitcoin's gold. And it wasn't, I can't believe I don't even remember it anymore. It wasn't Ethereum. It was a different one. And it was like very popular for a while. And it's just completely left my brain. Monero? No. I'm completely lost. I don't know. Anyway. Yeah. There were a whole bunch of those, you know, there was a war.

1:25:38There was a war and Bitcoin won. Capital war. And then there's been a war in the other chains and the other like values. And I wonder if there will be a winner in stablecoin or stable duopoly. Of course there will be. I mean, right now there's effectively a duopoly between the USTC and the UST. Yeah, but there's still a lot of value in creating new stablecoins. Let's look it up. Stablecoin market. Top stablecoin tokens by market cap. You have Tether at$183 billion. USCC at$75 billion. The next largest is USDE at$9 billion. Below that is DAI. And then below that is World Liberty Financial. Wow.

1:26:31Remarkable. Palmer Luckey has a post here. He says, I have this partially formed hypothesis that artist backlash against generative AI crystallized largely in response to text-driven user interfaces. The art world has come to accept tools like Photoshop and earlier photography itself that still shares the aesthetic of traditional art creation. If Midjourney, ChatGPT, Dali, et cetera, had interfaces along the lines of NVIDIA Canvas from 2020 pictured below, I remember this tool. It was awesome. You could do that. You could draw whatever outline you wanted and it would just automatically generate it and it looked great.

1:27:13Right. Palmer says it would look enough like existing tools that that the critique would look purist and elite. Art via command line, on the other hand, looks sufficiently other that you can easily berate the users as not artists. You just typed a few words. You aren't an artist. Has a better ring than your drawing skills rely on Photoshop somewhat more than mine. You aren't an artist. And that's really true. That's good. I do, I was thinking about like, I don't know if it's like bull case for Adobe or someone else, but there, I was shocked. I wanted to do a, I won't leak it just yet, but I wanted to make a big, like we do those market maps.

1:27:58I wanted to make a big, a big tapestry with a lot of detail that I could zoom in and zoom out on. And I wanted to be able to iteratively put together the collage with generative AI. I didn't wanna just try and one shot the entire thing. I wanted to zoom in and say, in this corner, I want a house and so generate a house and then I'll zoom out and then over here, I want a person and over there, I want a forest or whatever. And I wanted to use something that was like the, like the Nvidia canvas UI or something. And I just, I haven't found a tool and I was wondering, I need to go back through all of the different, like effectively like drawing tools.

1:28:39there's been a few that have really taken off on the iPad that allow you different brushes. So you can say, I want to make my brush a rubber stamp of a tree. And then you can just kind of scrub in some forest and it will just stamp the tree again and again. And that's not generative AI, but it certainly helps you speed up drawing a forest, for example. And I was wondering like, how far caught up are those companies on implementing things like Nano Banana or GPT, you know, images in chat GPT? Because those models are great, but I am kind of sick of just the command line because I try and one-shot it, and then I'm like, oh, that's not good to try again.

1:29:21And then it's back and forth as opposed to like actually building it up like clay where you're molding it. That's way more satisfying in my opinion. Yeah, the question for Adobe that I don't necessarily have an answer around yet or even a strong take is if Gen.ai makes it a lot easier for people to be creative, like combine these two ideas and make an image of it, does that create more people going down the pipeline to becoming power users that ultimately subscribe to various Adobe products, right? Because historically, like the learning curve, the learning curve on a lot of those tools is very steep.

1:29:57Learning curve on just generating an image is much lower. So I don't know. I think that bearishness on Adobe specifically was always overblown. Photoshop. I'm going to get to the bottom of this because I am somewhat comfortable in Photoshop mapping up, cutting, and lassoing and stuff. I know they have Firefly. if they had, okay, so they have choose a model, and then they have one of the non-Fly or Fly models. What model could you use? I'm going to dig into this and understand this because potentially there's some cool stuff you could do because I really want to make this like collage, and I need to figure out the correct tools or I need to delegate to someone.

1:30:52The chat has some ideas for you. There's Bitcoin gold. which was the thing. There was also Litecoin. Litecoin is the last thing. Gustav called it. Thank you, Gustav. It was silver to Bitcoin's gold. Yeah, it was silver to Bitcoin's gold. And I don't know where the market caps are now, but just in terms of my mental awareness, Litecoin has truly fallen by the wayside, whereas it was something that I was aware of during the Bitcoin rise. You probably invested in it at one point. I don't think I did. Yeah. Litecoin is down at a$6.8 billion valuation. And Bitcoin's at$2 trillion, right? I think Bitcoin's at$2 trillion.

1:31:35That is the current. $2 trillion on the dot. It's at$100 ,000 today. Bitcoin is off its high. And the markets in general are very red. The NASDAQ is down 1.88. Dow Jones is down 0.64. Meta down 16.7%. Everything is red today except for Apple, which is up. And guess what? My take today was a bull case for Apple. You're welcome, Apple. I will take full responsibility. Maybe we close out with my take. I would love to get your feedback on what I wrote today. Palantir is now almost getting, it may get to being almost down 10 % today, 9.35%. Bloodbath. Well, not for Apple. and it feels like Apple's getting the good ending.

1:32:23It's not that like it's over, that not like the conclusion of the AI play, like, you know, Game of Thrones has played out, but it feels like we just watched the total vibe reversal on Google. And I feel like there might be one coming for Apple. And the reason is that there's this popular narrative that they missed AI, right? Like Apple missed AI, it's so egregious. And Gurman didn't really mince words. He said two. He said it was wildly embarrassing for them. And he said even internally, they feel that way. Yeah. And so it's weird playing like the bull from the outside. They're still getting the do nothing win dynamic.

1:32:59The do nothing win dynamic. I should have put that in here because that was the correct phrase that we should have used. But, and to be clear, like I do think that they missed AI. Like I'm frustrated with Apple intelligence. I never use Siri. I can't even use the native dictation and speech to text reliably. Like if I need to read a story out loud, I'll still like dump it in Gemini or ChatGPT. And like, if I want to dictate something, I'll use a different app for that. I don't just - Trying to get your phone to dictate an article challenge. And it's like, that is not super intelligence. Like that's a very basic model that's been out for two years.

1:33:35So I do think they missed it. And for, you know, this phone costs almost two grand. Like it should be able to do that. Like that's the expectation. but there might be worse fates than missing AI. Like there is a worse scenario, which is imagine if they were like, oh, we're missing AI. And instead of just being like, oh, let's do some software and do some on-device inference and do some cute little Genmojis. It's like, we need to spend$200 billion of CapEx right now to build a massive thing. We got to build the Colossus data center. We got to build all this crazy, crazy frontier model. And then they burn all that.

1:34:09And then they get a model that's not even as good as ChetGPT. They can't really integrate it. And then they run into, imagine, you know, Google like dealt with the whole like the generative image thing where the German soldiers were depicted with the wrong races. You remember this? The black Nazi fiasco. Like Google doesn't love that type of PR fiasco. They got through it. Apple is like 10 times more risk averse than Google, in my opinion. And so Google is not like a move fast, break things company. But they were able to move fast enough to get through that. And now they have VO3, which like has been amazing.

1:34:44And Nano Banana, which has also been amazing. But Apple really would not want that. If Genmoji was doing crazy black Nazi stuff like that, that would not be good. And so they, you know, Apple, we heard this yesterday. They kicked the tires on some multi-billion dollar AI acquisitions, but they never really got anything over the line. Now they're partnering, the news yesterday was they're partnering with Google to essentially white label Gemini into Siri. I think this is awesome. I think consumers will love this. I want exactly this. Right now, Apple's paying Google for the API calls. But I still have this question about the flow.

1:35:18Do you think Apple is an underrated threat to ChatGPT? Because what German said that was most notable to me is that Apple will try to pursue the agenda commerce opportunity. I didn't understand that at all. Because I don't know how they would set that up. I think that it's going to offload to Gemini entirely. And so my question and my take is that in the long term, so first off, yes, I do think chat GPT usage could fall if Siri plus Gemini gets good enough that when you want to fire off a GPT-5 Pro query, some of the time you're just like, let me just ask Siri because I know Siri is going to Gemini.

1:36:00It's good enough, right? There's that possibility. The flip side is if the end state, if like next year, all of the models, when you ask, buy me a pair of shoes, they just go do it because that's just a functionality in the API. Well, then Gemini is going to start making a ton of money. Like the API is going to make money because like a certain percentage of the requests, sometimes it's going to be like, tell me the history of the Roman empire. And sometimes it's going to be like, buy me a case of Diet Coke. And when it's telling me to buy me a case of Diet Coke, you're going to make money. Apple at the product layer, if they're using Gemini at the intelligence layer, does not mean they're doing revenue shares at the model layer at all.

1:36:44What do you mean? Like if Apple has like their LLM experience, which is like Apple intelligence. No, that will be Gemini. No, but that's not how I read Mark's thing. He said that Gemini will be below the surface, powering the product experience. Yes. Because there's a world, like you can imagine chat gpt could use non-open ai models under the hood and still own the customer experience and own the customer relationship and monetize the queries sure i don't yeah yeah okay okay okay yeah yeah yeah no no no i i think you're right so basically apple apple this is the thing that i want to apple sees what they have going with the app store and apple historically has not, they historically haven't been able to monetize commerce well.

1:37:32Like all, think about all the economic activity that's being, uh, driven by Apple products, right? How many, yeah, unless it's digital, you probably spend, you probably spend like the average Apple user is like driving so much purchase volume off of their phone, whether it's buying clothes, buying groceries, just on Amazon, on your iPhone, like the iOS sales on Amazon. And that was always off limits for App Store, you know, revenue share, because it was real world products being sold and traded. And so I think they look at this and they're like, hey, this could be a, you know, trillion dollar opportunity for Apple.

1:38:11And it's going to be super, super competitive. They're going to be competing with Google, OpenAI, you know, a bunch of other companies. But I just felt this was this. I don't know. I think if it happens in the LLM, like in the, in the LLM, like for loop, basically, like if I go to Siri and I say, order me a case of Diet Coke and it goes to Gemini and says, the prompt is order John a case of Diet Coke. It's pretty hard for Apple to say like, and we want the revenue. like it's like google did the work so they are going to get the commission it's the same thing as like wouldn't it be nice if apple was like oh yeah like you know you can use google on your on your phone but like if you click an ad like we're going to take the cut like no that's not how it works yeah but if they surf if they're surfacing the ad yes at the app layer yes apple yes then apple can decide what ad network do we use do we use a native ad network that we're building similar to the app store do we i don't know no so so what i what i'm proposing is i mean yes you could build it either way and this is obviously going to be a point of debate and a point of negotiation but there is a world where where you go to siri you say i order me a case of diet coke it goes to gemini for with that prompt and google says we're using google checkout and we'll have it delivered to your thing we'll book your and we'll book your credit card that's on file.

1:39:42Does that sound like Apple? Apple would definitely fight that off because they, and they're doing, they, they ran a bake off with a bunch of different intelligence providers. Yep. I'm just saying that that was something that would be top of the list in terms of, Hey, we want to use you for search and knowledge retrieval for sure. Knowledge retrieval. And that's why Apple's paying Google. But, but if, if like play this out in a year, but it would be flipped if, if, if, if. What if Google is like, we can order you a case. If Google knew they would monetize, they could monetize this. Yes. They would be like, you can have it for free.

1:40:18Totally. And I think it flips at some point. Well, okay. So you think it flips at some point? I think it flips at some point. Within the grounds of the current agreement, I believe that Apple will want to own any sort of like payments own. Yeah. Capture at least. That's certainly their leverage, but there is totally a world where Gemini learns my home address through Siri, learns my credit card number through Siri. Yeah, but I mean, Apple is literally the final boss of like disintermediating all businesses. I'm not saying that they won't have leverage in the negotiation, but I am saying that like Google is more set up to actually build the product that delivers the diet.

1:41:01I totally agree. And so, but I think they're going to be in the same middle ground. The middle ground to me would more likely be like a rev share. That's what, yeah, no, that's exactly what I'm asking. And that's what's happening with Google search. But I'm saying, yeah, yeah, yeah, exactly. But Apple has built a taxation engine with the app store. They've already built an ad network. And so I think internally they would probably like all the margin. But I mean, you could use the exact same logic to like, they should build a search engine then and then take all that margin. Yeah, but it was also much harder.

1:41:30It was much harder to build. like a search experience until. I don't, I don't think they're going to, I don't think they're going to do it. I think they're going to, I think that they're going to say, Hey, look right now, right now, LLM search queries cost a dollar every time you fire them off. So we're going to pay you a billion dollars for like a million of them or whatever. It's not a dollar each, but like, you know, a million, a million prompts is a couple bucks or whatever, whatever the, whatever the math is, uh, over time, it's going to be like, no, every prompt generates money. So you want people prompting your system because you make money every time people prompt it.

1:42:04Yeah. And you monetize through advertising. And so at that point, there's going to be another bake-off, but the bake-off is going to be who's going to pay Apple the most money. Yeah. Anyway, that was a big part of the debate. The last question I had was like, how do you feel if you're Tim Cook right now? You have to be pretty happy was my take. You didn't get over your skis chasing the AI boom. You didn't invest in a bunch of infrastructure. You didn't acquire a company for 20 billion that you didn't need to. Even though AI features are super nifty, lacking them clearly doesn't cause people to switch phones.

1:42:38iMessage is such an insanely locked in network effect. Everyone still went out and bought the 17 Pro. More importantly, Apple got through a massive existential crisis during Liberation Day and the ensuing trade war with China. playing the political Game of Thrones while being a supply chain mastermind is more than enough to propel Apple to a$4 trillion valuation and earn Tim Cook a major pat on the back. And so I think the mood needs to turn. His comp should go up back to 100 at least. At least, at least, at least let's get him in the nine-figure club. It's really, really depressing that he hasn't been there.

1:43:14Let's get to this post by Jason Schumann. he was highlighting Service Titan at$9 billion market cap on$866 million of revenue, comparing that to Harvey, which is$8 billion valuation on somewhere around$100 million of ARR. I think they're quite a bit beyond that at this point. But Jason says, at this point, you have to believe that we're moving from a world of niche vertical SaaS to the next industrial revolution. I think that's exactly what the VCs that are backing Harvey at$8 billion believe. They believe this is a labor displacement, labor replacement opportunity. And so they're underwriting Harvey against the overall revenue of law firms globally.

1:44:04And can they capture a percentage of that, you would not be able to underwrite Harvey like this against like the existing legal AI market. And so, yeah, this is a bet that, uh, this is, this is a bet on, uh, labor, labor displacement. Yeah. I don't know. I mean, it is just like, it is just a new market, like in some ways, like service Titan is a, is a, I mean, not to boil it down too much, but it's like customer relationship management software for trades people with applications focused on sales, customer service, marketing automation. So it is in like a niche of a niche that's big and they make a ton of money.

1:44:46Um, but it's, it's somewhat niched down, which is the, the point about the niche vertical SAS, uh, Harvey, at least right now, it feels like is in a fairly new and broader niche. Like it's not, it's not, uh, it's like the legal market. It's had, it's had niche vertical SAS, but, um, something that's like pretty dramatically accelerated. Like it's, it feels like, it feels like the returns to the AI-ification of your vertical SAS product are just higher in legal than in, in CRM for tradespeople. Yeah. I would think. Yeah, apparently the 200 largest law firms in the world annually provide$130 billion of legal services.

1:45:37So that's the top 200. So Harvey's already sitting at basically their market cap at like 6 % of the global annual spend. Also just the growth rate is really wildly different. So price to earnings growth, if you just do PEG, you're going to get to a wildly different number. but I don't know we've seen a bunch of examples of like you know the lawyers genuinely like expressing that they are ready to pay for AI in a serious way well I have the right post to end it on a rare Pagani Zonda R one of only 15 ever built has found a new life not on the racetrack but as a centerpiece inside a luxury oceanfront condo valued at over$1.5 million.

1:46:30This Italian hypercar has been suspended on custom-engineered steel beams, serving as a breathtaking room divider between the master suite and the living area. The design transforms what was once a track-exclusive masterpiece into a sculptural focal point. Celebration of speed, art, and architecture. This is so crazy. Very ChatGPT-coded, unfortunately, writing. but uh yeah this is absolutely insane i've seen some of the pictures of it um uh i'd see some of the pictures of it being like lifted up into the room yeah uh it's kind of cool to to be able to like walk around the car and see see the bottom of it although we can't see the pictures of it keep it on the track out of there i i this is disrespectful to my culture as a track guy as someone whose entire personality is about tracking cars now that i went once uh this is uh This is extremely disrespectful.

1:47:24Take it out of the - Yeah, it really is. Give it to a track. Could you not just create a model of it? Give it to a track junkie in need. Could you not have created a model of Hollywood and say, for$100 ,000, I want you to make a model of my car. I'm going to keep my car at the track, but I want to have it in my living room. Yeah. So yeah, I'm a little bit - Also, what is going on? Why is this in a condo? If you have$1.5 million to spend on a centerpiece, like get a house. Ever heard of a penthouse, John? Ever. Ever heard of a Miami penthouse? It says condo. It doesn't say penthouse. Yeah, I mean a condo.

1:48:00It doesn't say, oh, they own the whole building. A condo could be a penthouse. It says they got a condo. They got a little studio and they put a 1.5. I mean, honestly, respect for going crazy with the prioritization. You're just like, yeah, my rent's 4K a month. And I decided to put a$1.5 million car in my 4K a month studio apartment. You want to do what? yeah tenant upgrades I don't know let us know in the comments how what you would do if you had a Pagani Zonda R one of 15 ever built I say keep it on the track yeah personally I would have at least kept it flat so you could use it as a call booth phone booth for sure if you need to jump on a zoom it's just kind of loud yeah yeah you just jump in there so anyway thank you for watching the show thank you for listening leave us five stars on Apple Podcasts and Spotify.

1:48:51We will be tomorrow back in the ultra in Los Angeles. We'll see you then. Goodbye. Thank you to Ben. Bye. Cheers.