Trump's Tech Dinner, Tesla's $1T Pay Package, OpenAI Mass Producing Chips with Broadcom, Seal Team 6 Mission Gone Wrong | Joe Weisenthal, Jason Droege, Dave Marra, Mert Mumtaz, Harish Abbott, Tuhin Srivastava, Cecilia Ziniti

27:27And we talked to a designer on the Facebook team at the time. And it's so easy to zoom out and look at Facebook as like, oh, there was just a foregone conclusion that they would win everything and just completely dominate. But there were moments where Greg did it. I love Jason popping in here in the comments saying it's his birthday party, dummy. He just turned 49. Oh, he did? Yeah. No way. That's amazing. What a birthday party. Where else would you want to be? That seems like a great time. um yeah people were having fun making fun of uh of chamath but why not have him be there makes a ton of sense i love this post from morgan uh should have gotten the other crew in there should have gotten jason and uh and friedberg the invite um anyway uh what was the was the next thing oh yeah this one three-way deal elon is what is this deal making between this is x xai and tesla no i mean this is it's it's uh ed lodlow over bloomberg saying tesla's board has proposed a new pay package for Musk and proposed the authorization of investment in XAI.

28:25Oh, sure, sure, sure. The board is looking at investing. Remember, SpaceX already invested in XAI because they had to get exposure. Yeah, we saw that complicated web of Elon companies, how they all link together. Maybe you've just got to merge them all together. A pure play. A pure play Elon. Holding company, yeah. Send it to 10 trillion. Send it to 10 trillion for sure. um the uh yeah i mean like elon's like done like kind of these zombie mergers before with solar cities i mean some people frame that as a bailout and didn't hurt tesla the stock went on absolute run and it was great just a flesh wound and uh i i never really understood why solar city wasn't more successful because it feels like solar you can talk to casey hammer he's like the biggest solar maxi and all of the like scientists all the experts that i trust uh say that solar is you know going to continue to grow extremely valuable extremely important but it seems like we've just been solar city specifically and basically any american solar company has been kind of going to economic war with china and china's been like dumping cheap solar panels on the market continuously so you need a company like tesla to maybe subsidize that for a while uh until the u.s government comes around to some sort of regulation on like what are exactly are we going to do about it to create like more of a level playing field.

29:48Either we're going to subsidize or we're going to like put tariffs and taxes on their exports to get them back into you know what what they actually cost instead of just letting another company run roughshod all over the American. This is interesting. Ed Ludlow also was sharing that Musk must also help craft a CEO succession plan to unlock the final tranches will maximus trump be old enough who what what's his name uh x there's a few there's a few wild names one of the one it's got to be a it's got to be a hereditary dynasty right he's got to let x take over in 10 years how old the oldest uh musk children are 20-year-old twins.

30:32Oh, they'll be totally ready in 10 years. Vivian and Griffin, potentially co-CEO. Co-CEOs, that could happen. And then there's another one that's a little bit younger who might be more in the 20-something-year-old by the time Elon steps down. But I would be surprised if he's not running Tesla in 10 years, I think. I think he's going to stick around for a while, keep running it. he just keeps he just keeps you know hanging out and building cars and this is just fantastic news for people like gary tan who've been saying we need more trillionaires oh that's true is that his thing he's more billionaires we need our first trillionaire uh the package could raise musk's stake in tesla to at least 25 with total voting power of up to 29 tesla's market cap must to reach$8.5 trillion from$1 trillion today.

31:25The plan follows Musk's 2018 package, valued at$50 billion, which was struck down by a Delaware court. Tesla's board is appealing that ruling, but is also offering Musk an interim$30 billion stock award while pursuing the new deal. Musk's outside venture, SpaceX, XAI, Neuralink, Boring Company, and Political Activity is very concerned about focus, but was Warren Buffett not focused on sees candy and coca-cola at the same time come on people this is a hold co it's not that complicated uh gabby gabby goldberg went super viral quoting quoting some of these um quoting a video from uh of sam altman at uh at the dinner last night and said this is like when you're at someone's birthday party and they make everyone go around and say their favorite thing about them that's very funny and what did sam say do we have the video the rapid response he said thank you for being such a pro business pro innovation president it's a refreshing change the investment that's happening here the ability to get the power of the industry back i don't think that would be happening without your leadership i love it i agree i'm happy to triple glaze as well i like innovation i like innovation arthur mcwaters uh been on the show uh many times before says i'm unironically working my butt off so that my wife can go to Pilates on a Wednesday afternoon and get a$15 matcha.

32:52And Siki says, same. It's a noble goal. I wish Raul from Julius had been invited to this dinner. Yeah. I feel like he shows up in, in different pieces of Silicon Valley lore. That would be really good delivered. That would be insane. It wouldn't be that insane. He's on a I just, I just. He's fantastic. I just, I mean, maybe we could work with Photoshop. That's why we partnered with him. Julius, what analysis do you want to run? Chat with your data. Get expert level insights in seconds. What could be more relevant to the government than having an AI data analyst that works for you? Every CEO there is probably already using Julius.

33:32If they're not, sure. Yeah. More context on Zuck's$600 billion investment. Turner Novak says, Meta has done$478 billion in revenue over the past 13 quarters and plans to invest$600 billion in data centers over the next 13 quarters. Good thing they're growing revenue. Yep. Well, this is important, too, because data centers are the foundation of the global economy. And if we stop building them, everything is going to collapse. The token factories. continue to continue. John, John Wu had some good analysis on the browser co acquisition. He said, man, parts of tech are sour about the browser co acquisition.

34:14The common complaint is that DIA arc slash arc don't have enough users or revenue to justify the acquisition. First, I think this line of, of BITCHing assumes that Atlassian's corp dev team are dingleberries, dingleberries down under when in fact they're elite and one of the most disciplined and successful M &A units in tech. That's super legit. Rationally integrating products into the existing product suite and otherwise protecting independence when merited Trello, Loom, Ops, Genie, Agilecraft, many others. Second, I think this is part of some ridiculous coping narrative that some VCs and founders inhabit, which is that having differentiated product vision, assembling a world-class team that ships and being an incredible storyteller and salesperson is somehow easy.

34:56That's a really good point. Maybe you have more revenue. Maybe you have more users. Guess what? The market doesn't give an F. I've read, seen, and heard directly from multiple people today something like, oh, if I just knew how easy it would be to raise a ton of venture and make good videos, I'd do it and get acquired for$610 million. So go do it. Go do it. I'm not affiliated with the browser code anyway. I'm not here to pat anyone on the back, do the whole well-deserved rigmarole claim. I was just one small part of their success. I'm here to say the market has spoken. A transaction has been transactional.

35:28capitalism has capital browser company is worth 610 million in cash if you're a vc founder etc in the salt mines today focus on your business if you think you can do better do it that is reality we have the shots fired uh sound effect there we go shots fired great great way to sum it up john yeah just go and go and build uh yeah just go go build something i don't know go tell a story figure out a way. Yeah, I mean, underrated is like, isn't the founder of the browser company like one of the core contributors to Chrome early on? He was on the Chrome team. If you want to build a browser, it's hard to imagine having a better team in place.

36:12And more breaking news, OpenAI is working on its own AI inference chip with a small company called Broadcom. I don't know if you've ever heard of this company before. If you've never heard of Broadcom, they're a small company based out of San Jose, California. They are worth$1.57 trillion. They just had a massive, massive earnings beat because there was this mystery$10 billion customer. It was OpenAI, I'm pretty sure. There we go. Love to see it. Broadcom. Up 14 % in the last. I mean, it's just crazy that you can have a company that is in the same, you know, similar, like at the moment, like priced in the same range as a company like Meta.

37:01Yep. And most people would be like, oh, Broadcom, yeah, I've heard of them. But not be able to say much more about it. Yeah. So Richard Ho, ex-Google TPU, is leading 40 engineers over at OpenAI. Seems like Sam did a little talent raid of his own. Broadcom providing critical design and IP. TSMC will fab it on the three nanometer node. Inference only. Target mass production in 2026. Could fail on first tape out. So the interesting thing here is that I've been beating this drum of like, yeah, GPT-4 is good enough. Like, bake it onto silicon. Like, just get the inference cost down. Just let me use it faster, cheaper.

37:40We have really, really good intelligence. We just need to inference it a lot more. Said the same thing about mid-journey. I thought the images looked good enough. And yeah, they could get better, but I'm fine with where they are now and they're fun. And so let's just make it really quick and easy to use. And it seems like they're going in that direction for the most part. Because even the Riesing models, they're still built on, I mean, it seems like they're built on a pretty stable foundation of the architecture of the frontier model. They're not making drastic, drastic architectural changes to the core LLM that powers these reasoning models.

38:19So bringing down the price of inference seems really, really important. So the Wall Street Journal has an article here. OpenAI and Broadcom make a$10 billion deal for custom AI chips. Order aims to ease shortage that limited the rollout of new ChatGPT versions. OpenAI is working with Broadcom to develop custom artificial intelligence chips, a move that could alleviate the shortage of powerful processors needed to quickly train and release new versions of ChatGPT. Broadcom shares surged nearly 11 % in Friday trading. You said it's up at 16 % now? What a great ripper. If you're looking to get into Broadcom, you think it's overvalued, undervalued, do it on public.com.

38:57Investing, for those that take it seriously, they got multi-asset investing. Industry-leading yields. They're trusted by millions, folks. Up 74 % in the last six months. Yes, so on the earnings call, Broadcom said we've signed a fourth major AI developer as a customer and that the new customer had placed a one-time order worth$10 billion for server racks using the Palo Alto-based firm's chips. The new customer is OpenAI, according to people familiar with the matter. The Financial Times reported the arrangement earlier. Sam Altman, OpenAI's chief executive, has been saying for months that a shortage of GPUs has been slowing the company's progress in releasing new versions of its flagship chatbot.

39:36But in February, Altman wrote on X that GPT 4.5, then its newest large language model was the closest the company had come to designing an AI model that behaved like a thoughtful person, but he lamented the delays and high costs that came with developing it. I guess there's a big question, like some of these really, really big models, people have been saying like, it's great, but it's slow, and it's costly. And I wonder if they're going to optimize for GPT 4.5 inference or GPT 4 level inference. Tyler, do you have any take on where they might sit? Sam Altman said it's a giant expensive model about 4.5.

40:17We really wanted to launch to plus and pro at the same time, and we've been growing a lot and are out of GPUs. I thought that demand for GPT 4.5 wasn't purely predicated on the cost and the timing and more just about the fact that people couldn't really tell the difference. Yeah, I mean, there's definitely a sense of only the high taste testers actually liked 4.5. I would be surprised if they really bring it back in any capacity. It's just so expensive. Yeah, I wonder what they're going to do. I mean, it seems like they'll probably still leave some flexibility like the TPU. They're not going to go fully lock themselves into one specific architecture.

40:56But hopefully it brings down inference costs, increases margins, which should be very good for the business. We love those. We will add tens of thousands of GPUs next week and roll it out to the plus tier. Hundreds of thousands coming online soon. I'm pretty sure y 'all will use every one we can rack up, he said. I like the y 'all. I'm pretty sure y 'all will use every one we can rack up. Recent years, OpenAI has relied heavily on so-called off-the-shelf GPUs produced by NVIDIA, the biggest player in chip design space, but as demand from large AI firms looking to train on increasingly sophisticated models has surged, Chip makers and data center operators have struggled to keep up.

41:33If we're talking about hyperscalers and gigantic AI factories, it's very hard to get access to a high number of GPUs, founder of Compute Labs said. A startup that buys GPUs and offers investors a share in the rental income they produce. It requires months of lead time and planning with the manufacturers. To solve the problem, OpenAI has been working with Broadcom for over a year in secret behind our backs to develop a custom chip for use in model training. Broadcom specializes in what it calls XPUs, a type of semiconductor that is designed with a particular application such as training chat GPT in mind.

42:08OpenAI also recently struck a data center deal with Oracle that calls for OpenAI to pay more than$30 billion a year to the cloud giant and signed a smaller contract with Google earlier to alleviate computing shortages. It is also embarking on its own data center construction project, Stargate. And so this is just another example of like, wow, they haven't made that much in revenue and they're committing to all these spends. It's like, well, they're growing very, very fast and they plan on setting themselves up for success and high margins and more growth ahead. And if you need serverless vector and full text search, head over to Turbo Puffer.

42:44Search every byte. We puffin'. It's built from first principles and object storage. It's fast. It's 10x cheaper and it's extremely scalable. You can get started. Be like Cursor, Notion, Linear, Superhuman, and many more. Yes. Readwise is on Turbo Pepper. Oh, no way. That's very cool. That makes a ton of sense. That makes a ton of sense. Yeah. Smart guys working with smart guys. Okay. So we need to do a little bit of a meme review moment on the stream. A monkey was handed an AK-47, and the image went viral and is now a meme. I don't know where this came from. It looks like an old image from a documentary.

43:19I love, it's crazy how you can tell this image has been screenshotted. Dozens of times. Potentially more. It does not, yeah. And it also uniquely does not look like AI. But Levels.io kicked it off with point of view. You discover you can do your own DevOps with clogged code. And of course, that could end in disaster. Skooks says. Your.1x coworker just discovered AI. Oh, no. I don't know how true that is. I mean, I feel like AI should be an accelerant, but yes, it could be. No, no, no, no, no. Yeah, you need to be at least a one ex-co-worker. Like you give a bad driver a super fast car. Okay. Does that make them better?

44:02Fair, fair, fair. Turning Novak says letting the VC give product feedback. It's rough if they haven't been in the trenches. Liquidity says how it feels to let the first-year analyst speak directly with the client. I love that. And producer Ben says giving the ramp card to interns. This is real. This is real. Do not do it, folks. You've got to be careful. You've got to be careful. Ben's like, why do they have to request funds for every single purchase? Why don't they just have a limit that they can spend within? Well, now, because they are monkeys with AK-47s. The ramp card is simply too powerful.

44:37Yes, it is. Also, Brian Armstrong bumped up the max leverage from 20X to 50X on international perpetual futures and upslope capital. Just give the monkey the AK 47. It's a great meme. I hope this meme sticks around for a long time, a long time. Anyway, if you want to get your brand mentioned on ChatGPT, head over to Profound, reach millions of customers who are using AI to discover new products and brands. You saw OpenAI's investment in Broadcom chips. They're not going anywhere. you're going to want your brand mentioned on ChatGPT, and you're going to want to do it through profound. And grok.

45:16And grok. And perplexing. And everything. And everything. So in the mansion section today, it's Friday. We've got 10 minutes until Joe Weisenthal joins to tell us what's going on in the United States economy. But one thing that is going fantastically well in the United States economy is Ellen DeGeneres' house flipping hobby, which has turned into a$190 million enterprise. So the comedian and wife, Portia de Rossi, have bought and sold more than 30 homes since the mid-2000s. They are absolute dogs. The dogs, the absolute dogs of Montecito. So they bought a 10 ,700 square foot property in Santa Barbara, California for$27 million in 2019.

46:00They sold it for$33.3 million in 2020. Modest gain. Modest gain, not bad. Modest gain, but didn't take a lot of risk here. One year, six mil, not bad. It's pretty good. It's honest work. Let's read through the reporting in the Wall Street Journal's mansion section by Catherine Clark. She says, just before last year's presidential election, comedian Ellen DeGeneres and her wife, actress Portia DeRossi, decamped to the Cotswold, an upscale rural area near London. They had bought a vacation home, a 43-acre estate, to use a few months a year. Then, the morning after the election, when they awoke to lots of texts with crying emojis, they said, we're staying here.

46:44Where was this? Which election? Oh, last year. Okay, so it's the Trump 2 election that was sparking crying emojis in Ellen DeGeneres' iMessage. Behind them, they leave a trail of luxury residences over the past two decades, the long time. California residents have developed a reputation as serial home flippers. The Wall Street Journal's research shows that the pair have owned at least 34 homes since the mid-2000s, including properties in Los Angeles, Beverly Hills, Santa Barbara County's wealthy Montecito, and now the UK. Records show that the couple were actively buying and selling throughout the mid-aughts in the 2010s, but their activities ramped up significantly during the, and we're going to move to page 8.

47:32This is one of the luxuries of reading the Wall Street Journal in print. During the pandemic. So starting in 2020, they have frequently owned multiple homes at one time and show no obvious signs of trying to time the market for the best returns. Records show they bought. No, they're just they just got the hot hand. They're there. They have the Midas touch. They don't need to time the market. They deliver value everywhere they go. They bought at least eight homes in 2021. Wow. and 2022 record show from the 34 homes tracked by the journal that the couple have since sold. They have posted a profit of roughly$190 million.

48:11That has to be in line with how much Ellen DeGeneres makes off her show and her core business. This is the side hustle of all side hustles. Their average return based on the original purchase price and the subsequent sale price is about 37 percent not bad you didn't raise a fund let me in ellen he in this the degenerate home buyer flipper but not very degenerate i mean who knows probably a lot of leverage involved but it's paying some there's some degenerate gamblers that actually make a lot of money that's right that doesn't mean that yeah doesn't mean that you're they're not so the pair's business manager harley newman said degenerates real profits are much more modest once the costs of renovating are factored in.

49:01I would love her to have made that big of a killing, he said with a laugh. Forget about the real estate agents and the escrow and title people, the costs of contractors and designers and the army that comes in once we've closed. It takes a village. He said those numbers also don't account for what DeGeneres spends on furniture that is usually included with the sale. None of the furniture is inexpensive, he said. She takes the art off the walls, but everything else gets sold. Newman said 34 is an underestimate, saying he probably thinks she has bought and sold at least 50 homes since he's known her.

49:39Love it. Bynum and Truss. Yeah. Secretly. Very hard to buy a home secretly. It can almost always be discovered. The Wall Street Journal, I mean, they uncovered 34 of 50 or so. The couple's homes have varied in size and in style from mid-century and contemporary to Italianate and English tutor. It's not that I get bored. I'm just so passionate about architecture and furniture that you can only buy a certain type of furniture for a contemporary house or mid-century house, said DeGeneres in a phone interview. She talked to the journalist. There we go. That's amazing. A large journalistic force headed doors and standby.

50:14She said she has lost count of how many homes they have bought and sold. That's great. That's like you with angel investments. It really is. She has a home-buying addiction, and it's paying off. I text myself when I make a new angel investing addiction, so I just got to scroll up far enough. Got it, got it, got it. There is now a certain cachet in owning an Ellen house, said Riney Williams of the Beverly Hills Estates, who represented a Hollywood executive who purchased one of their homes in Montecito a few years ago. Once somebody knows it's her house or she had a hand in it, it adds value. In L.A., she's synonymous with quality.

50:53She's a tastemaker. her. She cares about taste. Many of the homes have been sold off market without ever being listed for sale. The couple's Hollywood Rolodex seems to have come in handy. Comedian Will Ferrell and the late actor Heath Ledger and singer Ariana Grande are a few celebrity buyers of Ellen houses. DeGeneres said real estate has become a very lucrative hobby. Can you imagine Will Ferrell role playing as Ellen DeGeneres after buying one of her houses? Yeah, you imagine him in some of his titular roles as kind of a disheveled middle-aged man stumbling around a$40 million Ellen DeGeneres neatly appointed home.

51:32And it doesn't quite fit. But I'm sure he's much more. I like how she just flat out says to the journal, real estate has become a very lucrative hobby. Let's go. Let's go. Turn your hobbies into side hustles. Turn your side hustles into. I mean, this is basically a main, this is basically the main hustle now. She doesn't do the main show anymore. Yeah, that's true. The 67-year-old former television host has long had a passion for interior design and would have been a designer if comedy hadn't panned out. She attributed her itchy feet in part to her family frequently moving between rentals in her native New Orleans area.

52:09We never owned a house, so I always wanted to own. New Orleans. That was on your A tier. That was on my A tier. Yes, I got in a lot of trouble for that. People did not enjoy my tier list. Nobody got the joke. Some people got the joke, but I posted my tier list of best cities in the world. And it and it dusted some ruffled some feathers. I don't know if you have it pulled up. I'll pull it up again. So John's you said city tier list that will trigger everyone. But it's true. It is true. S tier NYC, LA, SFDC, Miami, Chicago, Austin and Boston. A tier Seattle, Nashville, Vegas, New Orleans, Detroit, Bakersfield, and Reno.

52:49And then there was nothing in the B and C tier. And then D was, of course, London and Paris. And then F was Rome, Berlin, Tokyo, Dubai, Hong Kong, Barcelona, Sydney, Buenos Aires. So basically, if somebody said, John, I'm going to Sydney. I'm going to see the opera. Do you want to come? You'd be like, sorry, I'm going to Bakersfield that weekend. I'm going to Bakersfield, yeah. And then if somebody said, hey, you want to pop over to Paris. You say, sorry, I'm actually in the biggest little city, Reno. Reno, exactly. I would prefer to be in Reno. Anyway, we had some fun with that. Well, we have our first guest of the show, Joe Weisenthal, on the phone.

53:28On the phone. Wired on the play. We got Joe Weisenthal. Pull up this picture that we just commissioned of Joe. There we go. That's real. And it's a real picture. Yeah, can you hear me? Yeah, you sound fantastic. You sound as good as you look. I understand why you didn't want to reveal. Yeah, I had a very active summer. I didn't want to reveal it to everyone, but happy to have you with the exclusive on my new look. That's fantastic. Incredible. Always great to catch up, and I'm surprised you're away from your terminal on a day like today. Yeah, it's Jobs Day. It's the Super Bowl. It's Joe's. It is the Super Bowl.

54:08You know, it is the Joe Day. It is the Super Bowl Day. It is the Jobs Day. After 8.30, obviously there's the initial hit, and then there's the rush, and then there's the market reaction. And as much as it is the Super Bowl, by the middle of the day, I'm already like, okay, I'm good for the day. I had such a high from following the numbers and seeing all the reactions and reading all the tweets, et cetera, that it's like, all right, I can sort of run some errands in the afternoon on jobs. There we go. Thank you for squeezing this, Aaron. So, yeah, take us through the numbers. I mean, the market reaction, the Dow is down half a percent right now.

54:46NASDAQ's off 0.23. Bitcoin's up 1%. Yeah. You know, it's an interesting, so like the number was not great, obviously, coming in a solid leap below expectation. I think everyone sort of had a feeling there was going to, I think there's just been a lot of anxiety about the state of the labor market right now. You know, people talk about the so-called whisper number where it's like, okay, here is the forecast that Wall Street has, but here's what everyone sort of really thinks. It's pretty clear that I think a lot of people just sort of thought it could be a myth. The last month actually was revised up a little bit, but the month before was revised down substantially, such that June was actually now negative, which is the first time since I think December 2020 that we had a month of job losses, which I think is interesting.

55:36Once again, healthcare employment, so that added 31 ,000 jobs in the month. So what that means is that once again, if you exclude healthcare, the economy overall has been shedding jobs for some time, which is not great. Including manufacturing. Yeah. So manufacturing employment down 78 ,000 so far this year. There was another decline And this month, look, this is what I would say about manufacturing, which is that I wouldn't, let's say a tariff strategy were going to work in the long term. And we were going to have this sort of strong reindustrialization of the United States economy. I wouldn't expect it to work right away.

56:19I mean, the tariffs were unveiled on April. But what we can say is that so far, those numbers, to the extent that this is a real priority of the administration or policymakers or others, to the extent that this is a real priority, it's clearly going in the wrong direction in 2025. fun. Yeah. Also, I mean, last night there was that dinner at the White House with all the tech leaders and the numbers that were thrown out were staggering. $600 billion from Apple,$600 billion from Meta. But that doesn't necessarily mean massive jobs growth because building a data center might just be a multi-billion dollar check that's written to TSMC.

56:57This is the funny thing with the sort of data center AI sort of thinking about its massive impact because first of all, you're obviously correct that these huge checks don't necessarily mean massive jobs growth. And if we're going to take seriously the promise of AI, it shouldn't lead to jobs growth. The test of AI is arguably jobs destruction, right? It's about what jobs that currently exist in the economy can be replaced by something that we call AI. And so not only does the data centers not add to employment in a meaningful way. If those turn out to be good investments, if there's a payoff, if they drive productivity, you would expect them, you would almost hope them to be labor market negative.

57:41Maybe. I mean, there is a world where it's like a worker plus AI is additive, and you can underwrite hiring that. So there is a hiring. Well, in our conversation, So we talked with – I'm not a – yes. I'm not a doomer. Like, in this sense, I'm skeptical of this. I find myself just from a sort of pure thinking about it through the economic lens of this idea of, oh, AI is going to lead to mass unemployment. Just for all the obvious reasons that anyone can say, I have cost savings because of AI. Well, I'm going to, like, spend that money somewhere. But if the question is, do these investments actually free up resources or enable new things that expand our capacity of what we can do with labor, that's really the question more than the direct drum creation.

58:27Yeah. Jordy? Yeah. One point, we interviewed Palantir's CFO and obviously CARP as well yesterday, and they said something like revenue is obviously up massively in the last few years. headcount is only up 12 percent and they said specifically we expect to have fewer people in the future yeah i saw that quote which i found to be really striking and it is sort of again this sort of very weird condition maybe we've talked about it before i'm not sure not but it does seem like we're in some sort of uncharted territory here with respect to the fact that you know i don't know what's happening on a given day okay maybe the next next time a little bit but You have all these names that are at or very close to all-time highs.

59:11Any other time in history, I think, where you're talking about subsector and your all-time highs, you are also associating it with a massive expansion of, or at least a significant expansion of headcount, typically. This is just sort of how it is in any industry, tech or otherwise. You intuitively link stock performance with headcount. And so it's interesting that here is a situation in which not only is this being articulated, this sort of divorce between revenue growth and employment growth, but also it's not just a theory. It's clearly been happening in the industry for some time. Yeah, I wonder what Trump's like bull case is.

59:48I haven't heard him kind of reconcile what's going on in AI and reindustrialization. There's like broad strokes of like, let's bring back the jobs. There's not a lot of early evidence of that. But I'd be interested to hear from people in the administration about just like paint me a picture of what it actually looks like. What is the American jobs economy look like in four years if everything plays out like you want? Like what what what are people doing? Because it does seem like we're going through a period of change. I don't know. What about. I mean, I do. I think in defense of the administration.

1:00:23Right. A lot of people are I think people all over the place are stretching their heads over this question. Because what I think people want, what the dream is when they close their eyes and they want to see an economy in which you don't have to be an absolute superstar computer science PhD from Carnegie Mellon or whatever or Stanford to get a great job. there should be some sort of broad middle of decent employment, whatever it is. And in defense of the administration, I don't see many people anywhere articulating, well, what that ultimately, what sectors or what ultimately drive that sort of employment abundance, so to speak.

1:01:01Yeah. The youth unemployment data as well is absolutely brutal. Bloomberg Economics had a chart showing it basically ramping in a way that it hasn't. Youth unemployment is rising at the fastest pace in the past 15 years, other than the pandemic, sitting at what's the number, 10.5 % between the ages of 16 and 24. Yeah. No, I think it's important. So there has been this debate in the labor market, which is very important and interesting, which is we know that the pace of job creation has slowed down. No one is disputing this fact. Every month, the number of new jobs in the United States is going down.

1:01:46Then the question is, well, maybe some of that is because of the changes in immigration practices, or maybe some of this has to do with just sort of normalization of the economy after a crazy few years. And so maybe some would argue that the pace of job creation slowing down does not automatically say, okay, we're heading into a recession. That being said, the question then, the test then becomes, all right, what about just the ability to get a job if you want one? And that's sort of essentially what the unemployment rate is. But it's interesting to watch some of these demographics that historically are sort of more marginal, more vulnerable.

1:02:23The youth unemployment rate is one. The black unemployment rate is another. That also, that was at its highest since October 2021. And so even though the headline unemployment 4.3%, they're pretty low by historical standards. What we're seeing is this deterioration of the job finding ability for some of these sectors of the economy that tend to, you know, maybe they sort of first to let go or some of these more marginal, marginal, marginal segments of the labor force. Is there a narrative coming together about what's driving the health care employment numbers and the health care job growth? Because it feels like a lot of our elders are moving beyond unk status.

1:03:09Yeah. I mean, they weren't massive. It's the post-Unk America. Because, I mean, there were a lot of health care layoffs during COVID. If anything, those are the people that kept their jobs. And then health care isn't particularly sensitive to interest rates. Well, before we get into this too much, I mean, I think the big thing is is CEOs broadly, even if they're seeing massive revenue growth, it's hard to go out on an earnings call and say we're seeing massive efficiency gains. We're we're leading an A.I. in our industry. We're getting all these benefits from A.I. and then be like, well, OK, our employees are getting so much more efficient.

1:03:46and then be like, well, you also are ramping, why are you ramping headcount the same way? Yeah, right, right. Well, I think, so on this point specifically, I also do think there was maybe, even though I think many people would say that the formal doge hasn't accomplished particularly much, maybe other people would argue otherwise. I do think there would be just a, quote, Biden shift, unquote, so to speak, where a lot of people are just sort of proud of the fact that they're keeping headcounts across all of these things. And I think sort of Elon Musk may have sort of shot the starting gun on this vibe that everyone can look within their organization and say, do we really need all of the people that we've hired, particularly after the boom years of 21, 22, and arguably 2023.

1:04:34But I agree on the healthcare employment standpoint. I mean, I would say two things. One is, yeah, it's the aging population. There does not see, it's hard to imagine what would slow this train down. I guess it's slowing down a little bit because, of course, there have been some of those Medicaid cuts to Medicaid growth, et cetera. So there is reasonable to think that maybe at the margins, the monthly numbers of new headcount into this space is going to not be as fast as it was. But it is very hard to envision what could happen in the short or medium term, is that wouldn't automatically mean this sector is just going to need a lot more people every month.

1:05:15I'm certain if you talk to any sort of healthcare network of any sort that's on the ground in any community, I'm certain still today in September 2025, that they would identify acquiring talent as one of their big challenges right now. Talk to me about what we're seeing from the Fed. The Polymarket has the number of rate cuts spiking today. Three potential rate cuts in 2025 spiked from 18 % up to 38%. And the chance of just a single cut is falling as well as two cuts. So Polymarket is certainly expecting more cuts to come. And most people, yeah, most people pricing i mean it's pricing yeah cut cut cut yep yes so i think you know it's funny the last time uh i was chatting with you guys it was from jackson hall and that was a really interesting moment because you have you if you talk to a lot of the regional fed presidents like i don't know like there's still so inflation still looks kind of warm um it's not totally obvious that uh cutting cycle or certainly an aggressive cutting cycle but you said i remember you said you said the labor market is clearly softening.

1:06:27This is the case for cuts. Like you've said that line. Yeah, and I will lay that out. Thank you. Thank you. If I said something smart, I appreciate you remembering it. But, you know, Paulo Speicher, Jackson Hole, I called it. That's the main takeaway here. I called it. No, but, you know, it's Paulo Speicher, Jackson Hole, clearly centered risks to the downside with the job market. And I just think everyone has come around to that view. And so now the question is, especially in the wake of today's number, is that not only does September seem like a lock, it is for a rate cut. I think the decision's in like 10 days.

1:07:07The question is, could they go 50 basis points? Could they go aggressive, kind of like they did last September? And then also I think October, which people thought, well, maybe they'd take a pause after a rate cut. That is now perceived as a good chance. So I think this sort of world really is coming around to arguably the Powell and then the Christopher Waller view. He's probably been the most dovish of the Fed governors over the last six months. He's been pretty forceful in saying, look, the labor market is fairly softening and rate cuts are due. And I think right now it's the Waller-Powell sort of view that's winning out in markets and the expectation that they'll be able to get the rest of the FOMC behind.

1:07:49Do you have any insight into what's going on with the Treasury doing a bond buyback to lower yield? Someone in the chat is joking that they're printing money to buy back printed money. I don't know if that's real. Same old, same old. You know, to be honest, I do not put – I think these things tend to be very marginal and sort of plumbing related. You know, ultimately, the sort of long-ended yield curve comes down to the market's view of the inflation trajectory and what the Fed is expected to have to do to maintain, essentially, its 2 % goals. Some of these Treasury moves, you know, they may have to do with liquidity, et cetera.

1:08:29But my advice to people when I have encountered these topics in the past is that generally these things are somewhat noise. And to keep your eye on the ball about what's happening in the economy over the medium term and how you expect the Fed to respond to it. So what's the next date on the calendar that we need to be tracking? Is it the next Fed meeting that's the big news? Or is there going to be more data? Yeah, so that's in two weeks. The thing in the meantime, what I would say is yesterday we had the weekly initial jobless claims data. That's been creeping higher. I think you should keep – that comes out every Thursday.

1:09:09And I think people should keep watching that because one of the ways that the economy has been characterized for really like over a year now is they talk about this low hiring, low firing equilibrium, which is that there aren't a lot of layoffs. But also it's not a great time – not a lot, as we've been talking about, not a lot of headcount expansion. The initial jobless claim has been picking up a little bit lately. Yesterday, we got 237K, I believe. That was a bit higher than expectation. You don't have to wait. Like I say, jobs report, Super Bowl once a month, but you get these mini Super Bowls every Thursday where they talk about initial jobless claims.

1:09:48It's noisier data. Oh, one other thing. So keep an eye on that. Also, this coming Tuesday, there's something called the, and I don't remember what this acronym stands for, but it's QCEW Benchmarker Invision. So it probably says for quarterly census employment something. This is actually how they sort of reset labor market data because we know it's all very noisy. But every quarter, they draw from actual – a complete sample of everyone who's paid into the unemployment system. And so they get this very big snapshot. And it could show – there's this view. It could show that the pace of job creation for all of 2024 going back was lower than expected.

1:10:34So they're always doing these revisions. But I suspect that Tuesday will be interesting in terms of what it says backwards looking about just how much employment there really is right now. Yeah, I'm looking at the quarterly census of employment wages, and it says that Santa Clara, California, has the largest fourth quarter over the year wage gain at 15%. And now I'm thinking if Menlo Park will see a massive spike in average wages because of the AI trade deals. and it's just they are five people at a billion dollars a year. Yeah, speaking of big pay packages, did the new Elon pay package feel low to you?

1:11:14Yeah, they're just paying a trillion dollars. What's the point of reincorporating from Delaware to Texas if they're only going to bump your target? I know, only a trillion. Come on. Only a trillion. Yeah, I should not put a re-recarter or something. Yeah, somewhere where an executive can really get paid for the work that they do. For one. I want 100 % of the dollars that I create in the shareholder value here. And were you disappointed to see that the White House snubbed Meek Mill on the invite to that dinner, the AI dinner last night? It felt like it was a big F you to Philadelphia broadly. Yeah, I would agree.

1:11:54I was shocked. I was shocked by that. It's time to shake that room up in terms of who are the big tech names that get invited to these things. But clearly, our up-and-coming AI entrepreneurs, regardless of what industry they are currently in or had been in, need to have a voice at the table. I do think going forward, too, we really need to have inflation data action in low part. Just generally, when you see these sort of wage gains so concentrated, we need to start having a formal on the Bloomberg. It'd be CPI or CPI, for CPI X-Men Love Park. That's the kind of data that we need. I mean, there's also a$10 billion tender offer going on at OpenAI.

1:12:39So if you're in the market for San Francisco. Somebody was running the numbers and showing that there's only, I think, a few hundred homes that are in the$4 to$10 million range in San Francisco. Wow, that's pretty competitive. And so imagine you just flood the market with like$10 billion in fresh cash and you get a bunch of cash buyers. Yeah. Last question. We'll let you go. There's something in the chat. Someone's talking about the U.S. Open, which I guess is U.S. What's the U.S.? I've heard of it. Okay, United States. Open market set operations? United States Open. Yeah. Oh, open. Yeah. Maybe.

1:13:16Isn't there like a real estate company? Maybe it has something to do with that. I'm not sure. Anyway, yes. But do you have a prediction? They want to know if you're watching tennis and if you have a prediction or a favorite. Who are you rooting for? I know. I'm so embarrassed. Every summer I try to get into tennis. And I actually take a few tennis lessons every summer. And it's like, this is the year I'm going to go to the U.S. Open. This is the year I'm really going to pay attention. No, I said I have nothing to, despite my efforts year after year, once again, nothing positively contribute to this district.

1:13:45Yeah, me too. I have no idea. But hopefully I will see some more great capital allocators photographed in paparazzi shots. That's what I watched the open for. I watched the paparazzi feeds. As do we all. As do we all. Anyway. Joe, enjoy your afternoon. We love you. We'll talk to you soon. Talk soon. Have a good rest of your day. Talk to you soon. Bye. Bye. Legend. Did you see that? Yes, it's a purpose-built tool for planning and building products. Meet the system for modern software development. Streamline issues, projects, and product roadmaps. Start building with linear. It really is a crystal ball for your roadmap.

1:14:26Yeah, it is. Anyway, what did you see? This is interesting. Guido Reishtater is doing a hunger strike outside of Anthropic. He said, hi, my name is Guido. I'm on hunger. He couldn't get in? He couldn't get into the SPV? He said, I'm on hunger strike outside of the offices of the AI company Anthropic right now because we are in an emergency. Anthropic and other AI companies are racing to create ever more powerful AI systems. These AIs are being used to inflict serious harm on our society today. And he got rate limited? Inflect increasingly, we'll get to that. I think it's very likely that he's hitting rate limits on the pro plan.

1:15:09That might make sense. And threaten to inflict increasingly greater damage tomorrow. Experts are warning us. Okay, we're going to trust the experts here. Experts are warning us that this race to ever more powerful general intelligence puts our lives and well-being at risk, as well as the lives and well-being of our loved ones. They are warning us that the creation of extremely powerful AI threatens to destroy life on Earth. Okay. Tyler, what you got? So I think it also could be, you know, Dario just said they're not selling AI to China anymore. So he might just be, you know, SPV holder. Oh, interesting.

1:15:40You know, mad about the province. Yeah, less revenue from not selling internationally. That makes sense. I mean, this is incredible marketing. Because they're not a public company, you can't really take out a short position and become an activist. This is incredible marketing for Anthropic. Yeah. Do you think this is some sort of mischief-style stunt to promote the company? Yeah. Yeah, you need to get protesters. It's so funny to single out Anthropic, the most safety-oriented. yeah they are the most safety oriented like why not just like like xai is like we will just put everything out like it's it's uh it the first name is actually adolf maca maca yeah um anyways a lot of a lot of big buzzwords in here uh uh power power bottom dad says this is so crazy i'm kind of for it yeah does the does the post look one-shotted at all Victoria says, bro, you're on a fast.

1:16:38It's good for you. If you get to seven days, it'll do wonders for your health. Is this bodybuilding related? People are just responding to his post with pictures of their food. Okay. Oh, that's so mean. Wait, but are there any M dashes in the post? Zero M dashes. It's not this, it's that. Nothing like that? I mean, it basically says the same thing over and over and over. Okay. I wish he put this into Claude and kind of like clean it up.

1:17:13Guido Reichstatter. Interesting. Well, good luck to him. Hopefully he dries out and cuts down on the body fat, gets extra diced. We say it's bulking season, but it's not right for everyone. Some people got to cut. And no better way to cut than go on an aggressive fast. So he is the co-founder of a company called Stop AI. and he has one connection on LinkedIn. Stop AI. Oh, you're connected to him? No, no, no. I wish. Okay. His only experience on his LinkedIn is at Stop AI. Has he stopped anything else? Yeah. It'd be crazy if his last job was co-founder at Stop NFTs. Or stop time travel. He's like, I was successful.

1:17:58We don't have time travel. Stop teleportation technology. Stop hypersonic planes in America. There's not much on Guido. He is interested in LinkedIn news, according to his interests on LinkedIn. Interesting. Well, one AI you should not stop is putting your sales tax on autopilot with Numeral HQ. Sales tax on autopilot. Spend less than five minutes per month on sales tax compliance. Go to NumeralHQ.com. On the more optimistic side of things, Everyday Astronaut has been reflecting on the Starship program over the last week and one thing has become obvious to him. This is from Everyday Astronaut, a fantastic space-related content creator.

1:18:41He says, SpaceX is enjoying the freedom to try and fail in a way they couldn't with Falcon 9. Doing anything experimental on the Falcon 9 was risky because it was SpaceX's only source of income. It was their lifeline, their workhorse. making any tweaks to the Falcon 9 to try and land a booster back in the day was a delicate balance. Don't push the envelope too hard because it could lead to a failure of the primary mission, which did happen twice. Yes, I believe that they blew up a satellite that Facebook was putting up. It was very dramatic and kind of foreshadows the showdown and the potential MMA fight.

1:19:15But, of course, like they have a contract and I think the money flowed appropriately. The companies weren't actually upset at each other. But it was kind of a crazy crossover moment in tech. Anyway, Everyday Astronaut says, when SpaceX first landed a booster almost 10 years ago, they were fairly slow to refly, and those first non-Block 5 boosters were only capable of a couple of reflights. This gave pause to some in the industry slash community, fearing all of this reusability hype wasn't going to pan out. But SpaceX learned from every landing attempt to develop their Block 5 Falcon 9, which has now gone on to have a single booster fly 30 missions.

1:19:54Absolutely unheard of. That's wild. Now, imagine if SpaceX could have had the freedom to not worry about flying customer payloads to get data during Falcon 9's reusability campaign. Imagine if they could have tested engine out procedures or push booster reentry profiles or try hot staging or what have you. This is the phase that SpaceX is now in during the Starship program. I know we hear the talking point of, quote, today's payload is data, and it could seem like a gimmick or excuse even, but that's a freedom almost no rocket program has had before. To know you can just try things out, fly real-life hardware without bankrupting the company, is the ultimate development platform.

1:20:35To be able to push engine out capabilities, remove heat shield tiles on purpose, We saw that with the video. It's like this crazy garbage floating around. They removed heat shield tiles. The thing barely made it back because they're pushing it to the absolute limit. They're not just trying to reach orbit. They're trying to do something that's never been done before. Build a rapidly reusable rocket, a rocket that can land and refly. This has never been done before. And honestly, it's silly to think you could do something like this without trying some extreme things. That's what we're seeing today.

1:21:06And that's extremely exciting to me. I can't wait to see version three of Starship Fly because they've learned so many lessons already, and they have a factory capable of making rockets at scale. And we just get to sit back and watch the cook. It's an exciting time to be alive. Kitchen's open. Kitchen's open. They're cooking. Anyway, you want us to take us through the next few timeline posts, Jordy? What you got? Absolutely. I just feel bad because we kind of glossed over some of Alan's trades. 2021 bought a house in CARP for$70 million, sold it for$96 million two years later. Wait, Carpinteria?

1:21:44Yeah. $96 million. That's a huge number. Wow, that must have been impressive. Another one bought a house for$7 million in 2017, sold for$11 million in 2018. Yeah, these are great. Another one re-bought a house for$14 million in 2021, resold it for$21 million in 2023. And some of these are so quick. You know, buy in 2024, sell in 2025,$10 million gain. I mean, she's just pushing this. Homeby Hills, same year. She bought a house for$20 million in 2022, sold it for$36 million in the same year. That's great. An absolute dog. The Usher House. Usher listed the home for sale for$35 million in July 2019.

1:22:29We just can't quit the mansion section today. Should we talk about the house of the month? What else is in here? Why don't you call Eric right now? Oh, yeah, okay. He's ready. Let's get Eric Gleiman on the phone. I got an important question for him. We saw the news about the billion-dollar run rate, and we just had one question to ask. Let's get Eric Gleiman, the CEO of Ramp, on the phone. Hey. Eric, how you doing? You're live. You're live. We're live. What's going on? Give us the update. There's some milestone that just dropped yesterday, I believe. It is. We couldn't be happier. Ramp is now doing more than a billion dollars a year in revenue.

1:23:14Let's go. Congratulations. I just have one question. Is the job finished? Job's not finished, guys. Job's not finished. Just getting started. Fantastic. Well, I'll let you get back to work. Thank you so much. Thank you for calling in. We'll talk to you soon. It's great to hear from you. All right, cheers. Have a great weekend. You too. Talk to you soon. Bye. It's not the weekend yet. It's not the weekend yet. Mark Hannes has got a buddy from high school who's on his fifth AI startup in two years. No technical engineering background. He's an MBA. Hires engineers with a low budget to do all the work and build the product while he is CEO and manages.

1:23:49Just started a new AI education startup. Wish I could short. I wonder how this will pan out. It could work. I don't know. Depends on exactly what he's doing. But it certainly is not the YC path, not the make something people want. It is a very management consultant, very MBA-coded strategy. Yeah, good luck getting into YC with this strategy. Good luck. Good luck. Naval says, if your smartest friends start saying crazy things, pay attention. A paradigm shift may be underway. I think he's talking about people getting one-shotted. You think that's it? Could be. I don't know. It might be, you know, he might be talking about AI, right?

1:24:33Like, you have really smart friends, and they start saying crazy things. Like, you know. They discover new physics. Yeah, all this other stuff. Not that. Not that. But actually, the idea of super intelligence is a crazy thing to say. But it might be real. It might be real. And it might start in your customer service organization with Fin.ai, the number one AI agent for customer service. Number one in performance against sparks. Number one in competitive bake-offs. Number one ranking on G2. So we covered the jobs gains. They were all part-time. full-time jobs dropped by $356k part-time jobs grew by$597k certainly not good for the durable long-term healthy economy but earlier this year there was a massive gain in full-time employment and then that dropped and so there seems to be gyrations which way crazy story coming out today SEAL Team 6 in 2019 attempted a covert mission in north korea to plant a device to intercept kim kim jong-un's comms they should have just waited and got him a friend.com friend just say hey this is going to be your new bestie just wear it keep it on your person that'd work uh authorized by trump the mission failed when seals uh unfortunately killed civilians mistaken for security forces You can get into the story a little bit because we don't talk about politics, but we do enjoy geopolitics.

1:26:09In early 2019, Trump approved a clandestine SEAL Team 6 operation to planet device in North Korea to intercept Kim Jong-un's communications. Red Squadron, known for killing bin Laden, rehearsed for months. The mission aimed to gather intelligence during nuclear talks. This was the nuclear, you know, Trump and Kim Jong-un had been communicating around, obviously, North Korea's nuclear program. On a winter night, SEALs emerged from the sea near North Korea's coast. A nuclear submarine deployed two mini-subs carrying eight commandos. Lacking real-time drone support, they relied on delayed satellite images.

1:26:48The shore appeared clear, but a North Korean boat emerged, mistaking civilians for security forces. The SEALs opened fire, killing two or three shellfish divers. So apparently they were on the beach. A boat pulled up, started using lights. They just assumed it was security forces, ended up being fishermen. and anyway so apparently the trump admin at the time didn't notify congress and the seal sank the bodies abandoned the device and signaled for extraction the submarine risked exposure retrieving them in shallow waters north korea detected activity but made no public statement the mission's failure hidden by secrecy risked escalating tensions from a nuclear armed state.

1:27:35Military reviews justified the civilian deaths under engagement rules claiming unforeseeable errors. The Trump administration withheld details from Congress, potentially violating federal law. Many involved SEALs were later promoted despite concerns about special operations' uneven track record. It is fascinating that the Disclosed TV talked to 24 anonymous sources, including Trump administration officials and military personnel with direct knowledge. Their anonymity reflects the operation sensitivity. It is obviously a tragedy. But yeah, it seems a bit hard to hold special operations teams and say you have to have, you can't fail if you want to rise up in the organization.

1:28:20Sure, sure. Right? But yeah, I mean, obviously you want every operation to be executed flawlessly. Anyway, in other news, the Washington Post is now on Substack. Let's hear it for the Washington Post. The Washington Post. Jeff Bezos. They heard you can post notes on Substack. They said we got to get involved. We love posting. We love posting. It's in our name. Posters. So this was, I believe, a story that was broken by Emily Sundberg in Feed Me. And she did a little bit of an interview here. Let's see. Welcome to my life, says Rachel Tashjian. And Emily asks, how do you see the Washington Post's game plan on Substack playing out?

1:29:01Is there something that will make your strategy different from other legacy media brands that have joined the platform? I'm actually not that familiar with other legacy media brands that have joined Substack. Are you subscribed to any? And Dresen Horowitz. Specifically a new media brand. At this point, they've been doing media for 20 years, though. That's what I'm saying. Play the shots fired. No, I'm just joking around. No, we are the legacy media. They are the new media. We want to be legacy media. The Washington Post responds, I don't necessarily see Substack as an alternative to legacy media.

1:29:37I see it as an additional tool to experiment with in a fantastic way to connect with new readers. I love getting into the mix in the comments. Comment section, yep. I think a successful journalist in 2025 should be on as many platforms as possible and have a tailored strategy for each one. Completely agree. So it was important to us that we publish original content exclusively on Substack and really engage with the platform on its own terms. I consider post-runway something like opulent tips goes to Milan and Paris. I don't report from the chiffon trenches, as Andre Leon Talley lovingly called the fashion world for my newsletter.

1:30:15So now I'm going to bring a bit of that madcap approach to the access-driven reporting and criticism I do for the paper. And I'll still be doing all my usual reporting and runway reviews for the post. Will there be a paid strategy, asks Emily Sundberg in Feed Me. The Washington Post responds, Post Runway will be free. Woo-hoo! What does success on Substack look like for The Post, says Emily. Having fun opening The Post's world to new readers and introducing the amazing voices on Substack, especially fashion writers, to The Post's audience. Well, welcome to Substack, The Washington Post. Maybe we should do a collab and talk about fashion.

1:30:54I don't know. I mean, I'm dressed to the nines today. I'd love the Washington Post to review the wonderful yellow suit. And we got to give it up to Sierra. Yes. And they're raising some new capital. Yes. $350 million from Green Oaks. Okay. Doubling down. Doubling down Sierra at a$10 billion valuation. And Zach DeWitt is saying, incredible seed investment for benchmark they likely own 15 it's a 450 million dollar fund let's go three to four x are ready on one investment we love to see capital allocators go emoji congratulations job's not finished well but certainly a fantastic milestone yeah um um power bottom dad says uh quoting joe weisensal's post saying the u.s has lost 78 000 manufacturing jobs this year that is a lot says chat is this re-industrialized i'm kind of tired of winning um you would i don't know it's hard to you know joe's point which is real is that you can't expect things to have an immediate impact yeah but at the same time if u.s manufacturers were feeling good, had lots of business, they would probably be trying to keep as many people as they can.

1:32:24There are two economic statistics that I find important in measuring the progress of re-industrialization or industrial might. The first is jobs. Manufacturing jobs are important. It's in many ways the heritage and the backbone of the U.S. economy, although the U.S. economy has shifted to a services-driven economy. But if you want to measure manufacturing capacity, it is fine to just use GDP. What is the total value of all the goods that are manufactured in America? Is that going up or going down? I actually don't know the numbers off the top of my head, but when I think about the next, I was talking to someone who was saying that America will re-industrialize, but it will re-industrialize on the back of robots and automation.

1:33:11and there will be much higher leverage for each individual manufacturing worker. You've seen this with companies like Hadrian. There's not that many people in that facility for a lot of CNC machines that are basically automated. And when you look at who are we actually competing against, if you're benchmarking the United States against China, Yes, China has a ton of people that migrate from outside of Shenzhen into the manufacturing hubs to assemble iPhones during the push. They bring in a million migrants to assemble iPhones right before the new iPhones released. And China obviously employs tons and tons of manufacturing employees.

1:33:55But at the same time, they also have lights out facilities where everything is effectively automated. And when you look at the design of the latest Tesla factories, the design of the latest Starlink factory, we are making things in America, but we're doing it at a much lower level of headcount. And some of our best manufacturers aren't contract manufacturers, right? They're Elon companies. Yeah, vertically integrated. There was also a crazy story from the BBC, which I think stands for Business, Business Central. Business, business, capital. Business, business, capital. That's what I like. They said this was basically breaking live today.

1:34:42Almost 500 people have been arrested at a Hyundai factory in Georgia by immigration authorities. And weren't they South Korean? So majority of those detained at the 3 ,000-acre site, which was built by the Korean automobile manufacturer to make electric vehicles, are Korean nationals. That's very interesting. So South Korea expressed concern and regret over the operation, which is kind of an interesting dynamic to just be like, yeah, we regret that because it's the government. I mean, it's South Korea saying that, not Hyundai. You think Hyundai would be like, sorry, because this scale is like.

1:35:19Got to sort out the visas, guys. You know, I think everyone's all for like bring the TSMC semiconductor experts to Arizona, but make sure that they fill out the immigration paperwork. Yeah. If they're going to come over here and build. Anyway, Buko Capital Bloke says, perhaps we are moving to the find out stage of paralyzing the U.S. economy with a multi-quarter barrage of nonsensical and inconsistent and conflicting set of economic policies. And there's some people in the chat. Is this bullish? Is this bullish? No, this is not bullish. He's blackpilling. I know. He's very upset about the economic data that was printed today related to a lot of the economic policies, mostly from the tariffs.

1:36:04Alps has a post showing the Las Vegas sphere, saying it's looking a little bit different. This, to me, is this super intelligence rearing its head? I love that Ilia is leaning in to the bit and having fun with the hats and whatnot. Wasn't there another hat? Oh, yes. Pixel Hulk. Pixel Hulk says they now have a new line of Marc Andreessen merch. And it's a hat where the head looks like Marc Andreessen. This looks like a fantastic AI image. I imagine this is AI. It looks like some sort of, however they made this, whether it's Photoshop or AI. flawless. I do think this would sell pretty well. It is ridiculous looking though.

1:36:52TBPN was featured at Volta earlier. Theo did a photo shoot. There's a Ferrari in the picture which we love and I believe one of Volta's employees. And anyway, shout out to John Fiorentino. I can't wait to visit the Volta HQ. be great the flagship it's not even the hq this is the flagship store right the flagship this is this is the you gotta make the pilgrimage yeah um um we covered uh tempo the new l1 built by stripe and paradigm so we don't need to cover it again but mert was uh firing back uh in in the announcement they were um uh the tempo team was basically highlighting uh potential limitations from existing owl ones they highlighted uh transactions per second they highlighted moments that uh that you couldn't run payroll again yeah when trump if you were trying to run payroll and that's because the day that trump coin launched so trump coin launched on solana is that right not ethereum because i know ethereum the the narrative around ethereum has been like the the network's congested there are high fees but that was the whole idea behind launching solana was that it was going to be much more scalable and there was some sort of trade-off there and then this is like the the the next l1 and people are upset about it in kind of both both camps but uh i invited matt on the show and mert who were both talking about this uh would love to yeah and they're they're they're going they're they're happily a little bit under the Now, Mert said, guys, you can just say you want more full stack control and distribution.

1:38:41You can also say you want the L1 premium or a controlled validator set. These are perfectly legitimate answers, but you look unserious and hurt your brand when you start talking about scalability and defending the Solana blockchain. Yes, yes, yes. It is proven that it can scale. But, yeah, I mean, I think this kind of play makes sense. Of course, Stripe and Paradigm would want to own the rails effectively, even if it's going to be network driven. Yeah. And it does seem like Stripe has a large enough customer base that they can actually bootstrap the network fairly easily. Right. But just offering it baked into checkouts that are already powered by Stripe potentially.

1:39:24although I wonder how what the onboarding process is like because even just getting a Solana wallet set up it's not as easy as like oh I have a different credit card and so the same stripe form works the same maybe you wrap it in a credit card I don't know I still don't exactly understand where Tempo fits in the full stack and how this gets rolled out but I'm excited to learn more I mean we'll certainly have some folks son to talk about more importantly for some members of our audience mercedes-benz is teasing the return of the four-door soft top cabriolet g-wagon this is the moment that many of us has been have been waiting for yes uh g-wagon arguably uh greatest car ever made uh they have uh done many iterations of the cabriolet over the years i came close to buying one a couple times for my friend blake who has a company called found objects here in la he sources these incredible g wagons from all over the world um and uh anyways i'm super happy that they're bringing this back this this image that is pulled up on the screen right now is actually i think somebody taking some artistic liberty the actual image we can pull up here if possible i put it in the timeline but the actual image is blurred out on the official mercedes account and so i expect this to be like somewhat um somewhat uh different than than the picture that we just had up on the screen what's crazy about this is that they're gonna i just expect them to modernize it a little bit so mercedes has a g650 landolay which is the soft top four-door but it's a mybach and i believe it trades at around a million dollars a car um and so i believe some anybody would sell that for just a million a million um yeah this is the image that he posted on instagram yeah and uh anyways i'm super excited for this i wonder how limited it'll struggle i will definitely struggle not to pick it up john john puts a lot of pressure on me to not just continue buying g wagons to mix it up you gotta mix it up it is i think i think if you want a convertible even speeder speeder speeder drove my g the other day and said you know what i thought these were ever hyped but i get it i get it now i get it well but if you want a if you want a four-door convertible SUV, why not get a Nissan Murano Cross Cabriolet?

1:42:04It's actually a two-door. Yeah, but if you want a two-door... Yeah, if you want the four-door, you kind of have to upgrade from the Nissan Murano Cross Cabriolet to the G650 Cabriolet. It'll take you from maybe$20 ,000 to maybe$2 million, but you do get those extra two doors. I think it's a pretty clear value there. I cannot wait for this. Do you think the G-Wagon salesman should be on Adio. They absolutely should. Adio is the AI native CRM that builds, scales, and grows the company for the next level. You might be tagged as just already closed one for some G-Wagon owner in Adio and they put you right back in, reach back out with the next one.

1:42:48They certainly do for me. As soon as you move into closed one, they move you back into prospect. Back in the pipeline. Back in the pipeline. Oh, Sergey Brin was at the dinner last night with Donald Trump. And Google is back in founder mode. You love to see Sergey having a seat at the table again, says Beth J. Zos. This post from TJ Parker was absolutely insane. Insane. We should call TJ right now and ask him to explain himself. So he says, fun facts. I was the youngest VP at Amazon. Of course, he sold PillPack to Amazon for$1 billion. dollars uh he said i was the youngest vp at amazon i've also personally never bought anything on amazon which is just absolutely incredible accomplishment how do you live your life without ever buying a single thing on amazon uh it's crazy like i think what he said i think he just i think he just said if he likes something he sends the link to his wife and then she okay He buys it.

1:43:53You still have Amazon purchases. He was too poor and then too rich. Poor and then he was too rich. Yeah, maybe. Yeah. I mean, like, really telling yourself for not taking the gig seriously, TJ. It's like, you know. Well, he was focused on something else. I mean, like, he was focused on that. I don't care, I guess. But, you know, I just think, like, if you get a job at Apple, you're expected. It's polite to wear an Apple Watch. You know? You walk around Apple, HQ, Cupertino, like everyone's wearing Apple watches. They're not wearing Samsung Galaxy Gears. You know, they're using the product. There's a famous story about if you are trying to sell a piece of software or do any sort of business dealing with Nike, when you go to their headquarters, you should be wearing Nikes.

1:44:41You do not show up to Nike headquarters in Adidas because they will take that as a sign of disrespect to their culture. It's like if we get lunch or breakfast with somebody and they insist on paying and they don't pay with a ramp card. I was about to say, it's a very big sign of disrespect. Very big sign of disrespect. I'd rather just have you not pay at all. But very, very silly. And I guess, you know, evidence that he was ready to move on and get into the world of investing in venture capital at Matrix, where he is now. Tim Cook has an iconic photograph in Getty Images. He hit the Abbey Road, says Mark Gurman, on the way to meet Trump.

1:45:21You love this. He is flanked by a crew of what appear to be bodyguards or maybe just the boys, maybe just the homies. But it's a good picture, one for the tech, the Museum of Business History. We talked about this. Oh, we should go into this fabricated knowledge post. I'm going to let you cover this. We'll be right back. So Doug O 'Loughlin has this thesis. He says, I think lagging edge is going the way of solar panels. Some can win, but the aggregate big number of chips is going to be bad. And so he says, first up is Texas Instruments. Where does he say this? April's pop is mostly pull forward.

1:46:06The lagging edge price war. Lagging edge capacity now behaves differently. Another upturn may come, but it may be muted and prolonged. Something is broken in auto and industrial. Texas Instruments is right to run inventories down. They likely will not need as much. There's a lagging edge permaglut, he says. What no one wants to admit, lagging edge capacity may be structurally different now. Earlier, I argued prices had to rise because new fabs cannot profit from selling lagging edge chips at prices set by a fully depreciated fab. There are obviously pricing problems in this scenario. With a new fab, a company can't turn a profit by selling lagging edge chips.

1:46:47at a price that was previously dictated. Reality, the lagging edge took price in a temporary market upturn. Now we are reverting to the mean, the old tread line is back. There's no meaningful new demand for old nodes and what exists is made and consumed in China. So you go and you make the smart toaster or the smart dishwasher and you need a chip in there. You don't need a TSMC three nanometer TPU Nvidia chip. You just need some lagging edge chip. It's gonna be made in China. It's going to be assembled in the product in China, and then you're going to get that smart toaster delivered to you. And all of that's going to happen within the Chinese ecosystem.

1:47:24And so if you're running a lagging edge fab in America, Doug O 'Loughlin thinks you're in trouble. The COVID super cycle plus a new China supply line that does not care about capacity economics points to a forever glut. China has a long record of flooding the low end. State-directed CapEx boosts output at the expense of efficiency. Doug wrote about this involution dynamic recently in China's race to the bottom. Products at 90 nanometer and above should internalize a permanent China glut. Companies resist that conclusion, but the case studies are familiar. Solar modules, LCDs, LEDs, steel, shipbuilding, batteries.

1:48:05You've seen it time and time again. China wants to stay on the lagging edge and dominate there, the commoditized lagging edge. and not worry about, you know, and they see that as the learning curve to get to the leading edge, but they are willing to sell at a lower margin for a very long time if they need to. Companies can still win share, but through performance and packaging rather than by producing older chips at lower cost. ADI is a masterclass in adding capabilities. MPWR has focused on higher end power. The long tail will face a flood of Chinese supply. Good luck to the older fabs. I cannot resist.

1:48:42talking about Marvel behind the paywall. So head over to Fabricated Knowledge and subscribe to the Substack. And we will have Doug O 'Loughlin on the show again soon to dig into all of the dynamics of both Leading Edge and The Lying Edge. So Josh Steinman, famous for saying, good morning, we are going to win every single day has a post here that's not good morning, we are going to win. It says you versus the guy she told you not to worry about. Robotaxi versus Waymo service area in square miles. So Waymo has been on a very consistent grind, adding square miles every year for the past seven, six years.

1:49:23Gone from maybe 50 square miles to now 750 square miles, whereas Robotaxi went from zero to almost 1 ,000 square miles in just under a few months. It's hard to make an equal comparison here based on the single metric. Obviously, I'm super excited about both companies. One's still in beta, but the app is doing very well with RoboTaxi. We will see. Number one in the charts. Safety drivers, number one in the charts. A little bit of a narrative violation. Maybe he made a phone call to Tim Cook, and Tim Cook said, Thank you, Elon, for getting us to space. Thank you for. We'll put you at the top of the chart.

1:50:06Thank you for Starlink. Thank you for the Model 3. Thank you for the Model S. Thank you for the Model Y. Thank you for the Model X. Thank you for the Cybertruck. Thank you for the Roadster. Thank you for the Falcon 9. Thank you for the Merlin Engine. Thank you for the Brain Chip. For the permite chip. Thank you for the Neuralink. Thank you for buying Twitter. Thank you for building that tunnel in Las Vegas. Thank you for PayPal. Thank you for Twitter. Thank you for Ani and Valentine. Tim Cook definitely. said thank you for Ani and Valentine. And Elon said, cool. You can have the RoboTaxi app on iOS.

1:50:43Anyway. How'd you sleep last night? Oh, you know I slept. You slept like a log? I wish. 71, only six hours. 85, play the song. Let's go. 85. Thank you, Trudy. Travel days. Travel days are rough, yeah. Rough. Go to aidsleep.com, get a pod five, Five year warranty, 30 at risk retrial, free returns, free shipping. And you can use code TBPN. This was cool. Yep. So this, I mean, this was just a cool move in the history of capitalism. Yes. Altacap says, didn't notice this until today. Another shareholder died. This is a Japanese company and donated their shares to the company, which was 6 % of the outstanding shares.

1:51:29Their will also donated their real estate to the company. Yes, yes, yes. And Sleepwell said, oh, you think buybacks are cool? How about major shareholders dying and donating their shares to the company, plus some real estate for free? Tyler, we do expect you to put the iPhone that we gave you in your will to TBPN. Back to the company. It will be passed down to another intern in 50 years. There's a funny contrast. I think yesterday. Let's not make the call. 100 years. 50 years. What does Xi Jinping think he'll live to? 150? So that's like 145 years from now for you? Exactly. But there was a post yesterday.

1:52:06Well, yeah, what is this carrot? Yeah, what is it? Oh, the carrot and the stick. Oh. Yeah, carrot on the stick. Yeah, carrot on the stick. No, they shouldn't be on the same. They can't be attached to each other. You put the carrot. You give somebody the carrot or the stick. Have you ever ridden a horse? You put the carrot on the end of the stick, and then you hold the carrot in front of the horse's nose, and then the pull it down. Oh, I'm more, I thought we were talking about the phrase, like, bring the carrot or the stick in a negotiation. I'm massaging it. I'm massaging it. Yeah, yeah, yeah.

1:52:33But yes. OK, we got it. So the next time we see Donald Trump doing a 4D chess negotiation, we're doing away with the chess analogy and we're moving to the carrot and the stick analogy. And we now have a physical prop for both the carrot and the stick. OK, but I was going to say yesterday, I think we sent a post about there was some billionaire who is leaving his will to Neymar. Oh, yeah. That's in here. For Neymar. Yes. I love that. Interesting contrast. Yeah. Wait. Who is Neymar? Something with like a ball. He's an athlete? Yeah. He plays soccer? Is this just some crazy way to get around salary caps?

1:53:15A businessman is leaving$1 billion to Neymar in his will. The administrator justified that he identifies with the athlete and decided to give his inheritance to the player. That is crazy. I think if Steve Ballmer really wanted to get around the salary cap, he would commit ritual suicide and leave in his will his money. That's a good one. That is the final boss of salary cap evasion. It's just a sapugou. To just get an extra. You care so much about your team. Well, in some other news that we will have to track the will of, very sad news, Giorgio Armani has passed away. He was 91 years old. Absolutely.

1:53:59He was the creator of the power suit. He held the reigns as conglomerates rose. Giorgio Armani, the designer and business mogul who brought subtle Italian luxury to the world stage and conquered Hollywood, died at 91. He worked until his final days, dedicating himself to the company, the collections, and many ongoing and future projects, said an announcement by the namesake brand. The company later confirmed he died in his home in Milan. During his years at the helm, he remained steadfastly independent in a fashion landscape dominated increasingly by conglomerates. Although he disliked the moniker, Armani was often called the king of fashion.

1:54:40In an interview with The Wall Street Journal in 2024, just one year ago, he insisted that the secret to success was his humble, consistent work ethic. This guy has Founders Podcast episode written all over him. He owned a super yacht and homes in New York, Milan, Pantillera, Antigua, Paris, San Marit, Saint-Tropez, Forte de Marmi, and Brioni. But he showed up at the office every morning and ruled his empire with a firm hand. He has TV, PN, mansion section, deep dive written all over him as well. Armani, who called himself a designer businessman, was the sole shareholder of his company at the time of his death.

1:55:22Wait. Wow. I had no idea still. In 2023, its worldwide revenue was 2.6 billion across men's and women's lines, the Emporio Armani Diffusion line and Armani Privé Haute Couture, interiors, dozens of restaurants, 2 ,500 retail stores, fragrances and a cult classic beauty line. Whether it took the form of relaxed suiting, makeup or interior design, his work whispered refinement. his stylistic innovations included an emphasis on neutral in between colors soft shoulders and a more relaxed approach in between colors as I read this in a yellow suit not exactly an in between color you know it's not from Armani you know it's not from Armani you know it's not exactly an in between in between colors you know it's not exactly an in the way you know it's not A-80, A-8, let's see about A-8.

1:56:25Film called Kagamusha inspired his fall 1981 samurai collection. I love that he has this balance of softness and power, said Armani's fellow designer, Michael Kors, at the last New York runway show in 2024. Armani became a household name where he designed Richard Gere's suits in the 1980s movie, American Gigolo, from the 1980s until recent years. He was the go-to designer for red carpet looks, dressing Oscar winners from Jodie Foster to Cate Blanchett. In 2024, Armani stressed the extremely important role of independence played in his impact. It's not a question of pride, he said. It's about getting things done.

1:57:07When I have an idea, I want to see it through to the end. He set up a trust in 2016 that laid out his plans for his company, including a charitable foundation and details for how a public stock listing or an acquisition would be handled. He worked with a close circle of collaborators, including his sister, Rosanna, who is on the board, his niece, Silvana, who is the head of women's design, his niece, Roberta, the head of VIP and entertainment relations, and his nephew, Andrea Camerana, the sustainability managing director. His right-hand man and close friend was the head of menswear. Born in northern Italy, the town of Piacenza, on July 11, 1934, that's, what, 200 years before Tyler was born, Armani was the middle child of three siblings between older brother Sergio and younger sister Rosanna.

1:58:00Armani described his childhood marked by a fascist regime in World War II as hard scrabble. At nine years old, he was standing outside a movie theater looking at a poster for Snow White when a friend came across an unexploded shell of gunpowder. It caught fire and the young Armani burst into flames. He said he shut his eyes only to open them 20 days later in the hospital. This is crazy. Well, I know we have our next guest, but I can't stop reading this. One second, we're sorry. We have the CEO of Scale AI coming on in just a minute. his only lasting scar was on his foot where his shoe melted into his flesh the incident contributed to armani's desire to become a doctor he attended medical school in milan for three years before dropping out to fulfill his compulsory military service after the army armani's latent creativity was stirred he began working as an assistant architect designing interiors and displays for a department store in milan he moved into menswear working for working for a time at Hitman, a company owned by Nino Cerruti and freelancing for other brands.

1:59:06By the mid-1960s, Armani met his longtime life and business partner, Sergio Galletti, a Tuscan architectural draftsman, 10 years his junior. Galletti persuaded Armani to start his own business in 1975. He woke me up from a sort of torpor, from the little life in which I was living, Armani told the journal in 2012. John, I hate to cut you off Let's not keep our guests waiting Let's not keep our guests waiting Welcome to the stream Bring in Jason Sorry John was getting carried away there Have you heard the story of Giorgio Armani? He just died What a fantastic obituary in the journal today No, I was hoping you'd keep going Sorry to keep you waiting It's great to meet you, Jason Thanks for coming on Thank you so much for hopping on Why don't you kick us off with an introduction and kind of your path to how you wound up in this particular situation.

2:00:01Yeah, totally. So I've been in tech for a long time. I've been at scale for a year. My background has been in starting companies early, early in my career. On to more notably at Uber. I started the Uber Eats business there. And that grew much larger than we ever thought it would. How big is that business today, if you don't mind me asking? When I left, it was about$20 billion a year of GMB.

2:00:32Sorry. I think they've grown it to maybe$70 billion,$70 billion,$80 billion. It just keeps going. Yeah. Dara just announced yesterday that in Australia alone, they've served a billion deliveries, which is about 2 billion meals. That's a lot of shrimp on the Barbie. Yeah. Yeah. When you get involved in these things, you're always surprised at how big things can get. And I think that's going to end up being true here with this current wave. And that was one of the reasons why I was excited to join Scale a year ago, joining Alex at the time. And I had experience in marketplaces, done a lot of growth stage businesses, things where demand was outstripping supply and you needed to fix a lot of things to get the business on track.

2:01:19and it's been an exciting, super fascinating journey over the past year. I mean, nothing like I ever expected. Yeah, what was the shape of the business when you came in? From my perspective, scale has been a very interesting business in that it hasn't, many companies have a second act. I feel like scale's had a second and third, maybe even a fourth act. The way I tell the story is, Alex Wang is starting with creating data for self-driving cars. eventually that business maybe gets rolled into Waymo and Tesla directly. And so scale expands into the RLHF boom during the LLM boom and has kind of grown.

2:02:01And then my question was, are we going to see Scale AI take on robotics data or more specific tasks? And so describe the shape of the business when you joined and where you were thinking about taking it. Yeah, no, for sure. I mean, the history is amazing and fascinating. I think Alex had this insight that data was the most important thing to models. And that was nine years ago. So that's very visionary. And I was very impressed whenever I was talking to him about a role at the company, you know, only a year ago. And you're right. We've gone from different types of data. And one interesting thing about, like, this entire journey is that every single type of data, someone has said like oh when is this going to go away when is this going to go away when is this going to go away and this is a common thing that we hear and we've heard it so many times that uh you know we sort of are used to addressing it and there's always some other source of data and one of the reasons why is because um data is the application in many ways um you know you know you're you're going from a software space where it's software organizes data for use by humans Now you've got data driving applications of things.

2:03:13And so as we generate more and more data, the models are serving customers. Those customers have higher and higher demands of what the model should do, which then drives demand for more data. And the data changes. But, you know, we've gone through this so many times from expert data to, you know, now, you know, you know, models are working on like, how do they do things for you, right? It was about knowledge capture, and now it's about skill capture. So, yeah, it's been an awesome journey, and the company has been extremely agile and extremely commercial over the entire journey, which leads us to this discussion.

2:03:54Yeah, it hasn't just been just like, if you zoom out, it's like one smooth sort of exponential, but in fact, there's different pools of data that have been maybe a series of stacked S-curbs, but that's the story of all technology. and all progress. So I'm interested to know what, like when Alex did Invest Like The Best, he had a very interesting point about robotics data being particularly rare. Like at least we had the internet to crawl and scrape and that was something that didn't, it eventually needed to be polished with RLHF, but with robotics data there's just so little out there. And I was wondering if that was something that you were interested in looking at or if it was immediately the skill capture thing of these really, really niche experts, people who know about archaeology and it's just not on the internet and you need to go get some PhD, get them on the platform and start having them answer questions so that that can get baked into the models and we can get out of the, what is it, like Gelman, Amnesia, Dunning-Kruger for these models where it sounds great until you're asking about the thing that you're the expert in and then you're like, oh, actually, it's only at a college level.

2:05:09It's not actually at the full expert level. Right, right, right, right. Yes. So, yes, we do robotics data today. We've been doing, you know, that's one of the newest parts of our business. That's the way beyond expert data. So we've been doing expert data for a year and a half plus, depending on where you draw the line on experts. But everyone in the data business is in the expert business because that's what the models need. That's what they demand. You know, something like 15 % of the people in our network are PhDs, 25 % have master's degrees, and like a very large percentage, I think 60 % or so, have at least a bachelor's degree.

2:05:44And so, you know, that is where the knowledge capture has gone. Robotics, yeah, it is interesting, right? Because you can't necessarily pre-train on this, like, corpus of information from the past 25 years of the internet. And so that probably just increases the size of the opportunity going forward. But the space is new, right? I mean, like, you know, there aren't robots walking around the world yet. So a lot of this is still in development. One of the things worth mentioning is like all the data types from before we still do. Like we still supply data to autonomous vehicles companies. We're still providing computer vision data to the U.S.

2:06:19government and other parties. And so these things do stack up, as you mentioned, into what is now a pretty big business. Is there still some piece of the business that's not directly AI related, like almost like Mechanical Turk-esque? Like, I just need an actual human to do something at an API endpoint for a very reliable, very scalable. And so I come to you and instead of going to an LLM, I just have people doing it on demand. Is that a thing? That's not a thing for us. I do think that that's a thing in the industry. Like some of that work has gone into BPOs and sort of like lower margin, less difficulty type tasks.

2:07:04And so we're at the frontier. And part of being at the frontier means you do the hard stuff. The margins are better. The skill level required is higher. And it pushes you to constantly be driving the business forward. And that adaptability, frankly, that started with Alex in terms of always looking for the next thing and being really close to the tech is kind of what's allowed us to get here. On the customer side, how are companies buying this data? It seems like there's overwhelming demand. Well, you go to them and you say, like, 10 data, please. 10 data. No, no, no. I mean more so, like, what do the actual, like, customer relationships look like?

2:07:41It seems like a lot of the labs and big players are actually working with multiple providers. And that's not necessarily a bad thing because different providers can specialize in different things. Yeah, I mean, you know, we have, you know, that is true. That's like, that's sort of a general understanding of the market. The reality, though, is on the ground, if you want high quality frontier data, that's difficult to get at volume, the number of providers goes down dramatically. And so there's a lot of claims being made out there about doing things at scale. You know, we're one of the largest providers today.

2:08:19We've been one of the largest providers. and so I think that that's you know through all of the change that's what's allowed us to keep the relationships that we had which is it's actually pretty hard to get this data right like like like you really have to have the right network you have to know how to talk to the network because if you imagine like being in the contribute we call them contributors meaning the people who provide the data to the models through us if you imagine being in their mindset you have to have a system that can interface with them on what is a part-time job for them to give high quality data for a very technical audience, which is the researcher, you know, so that it changes the weights of the models in a positive way.

2:09:02And that's not a super straightforward thing to do. And so, yeah, I mean, and the business is going great. I mean, like, I've seen a lot of press out there. There's, you know, there's been a lot of coverage, a lot of speculation. Um, you know, every month since the deal happened, we've had growth, which I think a lot of people will be surprised to hear. Um, you know, given, I mean, like I read the news and then we look at what's going on inside the company and we're just like, okay, we got to start talking a little bit more. Obviously I've been in the seat only a few months. And so there's been a lot going on, but, um, yeah, like things are going well.

2:09:36Um, we have two, I mean, one thing worth noting, like is we have two multi hundred million dollar businesses. we have our data business which we're most known for yeah so if you were to split either of these out and we have an applications and services business which applies this in customers that business alone has hundreds of millions of dollars of revenue and so if you were to break that out it'd be its own unicorn or decacorn or whatever they're called these days so that's wild do you ever you get inbound maybe customer fills out a form on your website do you ever have to check to make sure it's not like a runaway super intelligence it's like you know uh off on its own just like hoovering up data well i mean great uh uh great question um uh we tend to know our customers pretty deeply yeah yeah you gotta you gotta have the kyc process in place and and frankly they spend you know these are data is not cheap to procure and so um that'd be a very expensive hoovering operation.

2:10:38Well, if AI progresses like the AI 2027 crowd was projecting, at some point or another, a super intelligence will come and you'll have to choose humanity or the super intelligence. Have you run into any super small companies that are customers? We've talked to some small LLM providers. They're not building the mega GPT-4, GPT-5 level models. They say, we're still doing, you know, transformer-based AI, but it runs on a gaming graphics card. And I feel like usually they're just distilling Llama or something into a smaller model. But are there any customers that have come to you and just had some weird niche use case that's like pretty light on the data, but so differentiated that it actually provides value for them in their particular niche?

2:11:29Or is it all just like, you know, the huge labs, the people taking really big swings? Yeah, I think the big swings is most of it. We do see some of the smaller stuff. Where we actually are starting to see some data needs, which is sort of, it goes to just following the market. It's actually inside of enterprises. Yeah. So they're not actually building their own foundation model, but they're hitting the limits of their own sort of the data that they have to make applications useful in their environment. And obviously, the U.S. government, we do a lot of computer vision data for the Department of Defense, too.

2:12:07So those are the more sort of smaller customers and maybe less publicized things that scale does. uh are there other analogies to the dod where there are companies that have private uh like more like more secure data needs and maybe the the scale process is working on top of their own proprietary data and they need to transform it somehow and and you get involved is that is that a thing at all yeah so that's the applications and services side of our sure that makes sense Right. And so and so the history there is as we were supplying data to the model builders, obviously scale is very well known.

2:12:46Alex is very well connected. The company knows a lot of people. People started contacting us saying like, hey, we kind of thought this AGI thing was just going to like pop out of the box and do everything for us. And the reality is, is like that's not what's happening at all. And so as we got pulled into that world and being asked, like, well, how do I apply this to an accounting problem? How do I apply this to a healthcare problem? We started to see the actual fundamental problems inside of these organizations and why they can't get there. And what's interesting is I've actually been on site with some customers in the past month as we're getting to production with this latest round.

2:13:25And some of the limits they're hitting is the actual human knowledge from the executive staff or in like a healthcare setting, maybe the most senior doctor or in an accounting setting, the most senior accountant. And they, you know, they sort of went into this, this thinking like, okay, we might need some data from the subject matter experts here, but, but, you know, um, maybe a junior accountant or maybe a junior lawyer, we'll just go in and label some data where we don't have recorded data. Um, uh, and, you know, but we, um, you know, need the models to do something that they currently aren't doing.

2:13:58And what's happening is, is that the senior staff is finding that they need to get involved because as you build these agent systems, which agents are basically obviously just applications running, you know, that leverage AI, that coordinate with each other, in order to align them, you know, which goes to your AI 2027 comment, which is like, are they going to be aligned to like human values? And then will they do the things that we want them to do the way that we want them to do them? You actually need, the question becomes, who do you align it to in an organization? Like, do you align it to, you know, someone who's straight out of college?

2:14:32Maybe. Do you align it to the most senior person in the organization? Maybe they need a partner to help them figure out like, which do you align it? Do you align it to compliance? Yeah, yeah, exactly. Or, you know, somebody more on the operating side. Yeah, yeah, yeah. What are you guys seeing on the like smaller model builder side of things? Are you seeing like the, you know, the things you're asking me about in terms of like their needs and specialization? Yeah, I mean, I think there's this big question right now about inference costs and do you need to throw a reasoning model at everything?

2:15:07Token costs are coming down, but people just keep moving on to the, they're like, oh, the better model is the same price as the last model, so I'm not even capturing that 10 % or the 10x gain or whatever is happening in the cost of inference. The most recent data point that we actually got that was hard was from Notion in the Wall Street Journal, Ivan, the founder, said that their gross margins dropped from 90 % to 80%, you know, not that bad. And I think that that's a trade that they should take all day long. But you could imagine that... Yeah, I mean, I think... Oh, yeah, you could imagine that.

2:15:45So basically, Notion's probably paying for the best-in-class intelligence, and that's having a material cost on their gross margins. but if they figure out that, okay, what users are really excited to pay for is, you know, a report summarizes everything that's happening in Notion across all your documents, or they really love templating with AI, or they really love having just a chat interface on the right that, you know, is somewhat aware of the documents and they can just kind of ask, hey, I remember putting all of my medical records in one Notion thing, or I remember putting all my financial and my CRM over here, like, can you find that for me?

2:16:24Over time, I would bet that they're able to distill that into smaller and smaller models, bring down the per token cost or dollar per million token cost to a place where the margins go back up. But it's kind of a broad question in the market as more and more companies lean into AI and want to deliver the most frontier model possible to their customers. No one wants to say, oh, yeah, we added AI to our product. it's GPT 3.5. People will be like, why? What am I paying for? What can you share about more general business strategy at scale going forward? You're 49 % of the company is owned by Meta, but you guys still have a bunch of customers, tons of revenue.

2:17:09You talked about two nine-figure businesses. You got a bunch of people on your team. How are you rallying the team and what are what are reasons that people should join scale today and be a part of the opportunity you guys are going after? Yeah, yeah, yeah, of course. Yeah. It's worth taking a step back and talking about that deal for a second, which is, you know, Meta put in over$14 billion for 49 % of the company as part of that, Alex and some others went to Meta. Scale still has over 1 ,100 employees and a billion dollars on the balance sheet. And so we have a lot of money. And those employees, to be clear, none of those are contributors.

2:17:51Those are people that work on the scale platform. Yeah, the contributor network has hundreds of thousands of people. And so those are people that are working at scale. When I joined the company, actually, I joined the application side of the business. I didn't join the data side of the business when I joined as the chief strategy officer. and I sort of saw the opportunity there. And so we're investing heavily there and customers are reserving budget heavily there. And as an example, we just got a$100 million contract with the Army to provide services and data and people to them. And so you've got the data business, which is supplying data to the models.

2:18:32And then you've got the applications and services business, which is making those models do things inside of large complex organizations, Fortune 500s, U.S. government, international governments. And this opportunity is going to be huge for scale and very few people actually know about it. This is a multi-hundred million dollar business, as I said today. We've got contracts with, Qatar is a huge customer, the U.S. government's a huge customer. We've got dozens of Fortune 500 companies that we work with. And so I think the opportunity for scale right now is just to like capture this wave going forward because we're at the very, very beginning of actually making AI work inside these big organizations.

2:19:14Yeah, I was going to ask you, in big organizations right now, where do you think AI is overhyped and where do you think it's underhyped? We asked this to carp yesterday. Yeah, I mean, I think it's very overhyped that it's going to eliminate jobs in the next like one to two years. Like I think this line of thinking would require a change in the curve of capabilities and the change management inside of these organizations where I think it's like way overblown. Like the promise is just so high. And the reality of what's going on on the ground is there's value, but the value needs to be extracted.

2:19:52And that requires a ton of work. yeah um and i think what's going to happen with these customers in the next year if i had to make a prediction is you're going to go from like having a certain amount of money which is usually a lot in these companies um allocated for like ai initiatives to what is the roi like you know the grim reaper is going to come for every ai company that is not delivering value to these customers because because what's happened is they've all been sucked into this idea that like oh, if we don't invest in this, we're going to miss out. But because AGI is coming. And the reality is, is that if that doesn't pay off, that expectation is very high.

2:20:28And if you're not finding a way to like, give ground truth to these customers, they're going to be very disappointed. So do you think we're approaching the trough of disillusionment in some of these big enterprises? I mean, it like, I think the gap, yeah, I mean, my short answer is yeah. I think that the, I've seen this in many tech cycles before. The promises are way out of our seas. There's still a ton of value. It's not that it's like vapor. We are delivering value. But the effort it takes and the amount of data you need to organize in a company and get from human beings inside the company to make sophisticated applications work is way harder than people are selling today, except for us.

2:21:08Yeah, makes sense. has the contributor has like the geopolitical or geo geographic footprint of the contributor network changed as you've moved towards more like expert focused uh criteria like the phds um is it still spread all over the world are people onboarding in europe and america like what what is the what is the geo uh the geographic footprint look like right now Yeah, I mean, the footprint, as you would expect, you know, centers around areas where PhD and master's and advanced degree, you know, people are. And so you have a very high U.S. footprint, you know, parts of Europe, India and Latin America.

2:21:53There's actually I mean, one interesting fact is like there's a lot of contributors who are who are contributing at a high level who've actually grown themselves. like like you would think it would be like if you don't have a phd you can't do this type of work yeah and you know maybe phd is too far of a stretch here but like if you you like like you don't have domain expertise what we've actually seen is contributors go through this whole cycle from a from a from a sophistication standpoint and we think that they're self-teaching yeah basically yeah the final method the final yeah which is teach to learn yeah which is kind of amazing if you think about labor markets and like you would just assume like oh the technology would move on but the people are adapting.

2:22:32Yeah. Jordy, anything else? Yeah, this was super, super insightful. Yeah, this is great. Thank you so much for hopping on. Always welcome to join. You just give us a few minutes warning and we'll send you the link and you drop on. So glad you guys are out telling your story and congrats on all the progress. Yeah, this was fun. Thank you for the opportunity. Yeah, of course. We'll talk to you soon. Cheers, Jason. All right, cool. Take care. Bye. Let me tell you about adquick.com. Out-of-home advertising made easy and measurable. Say goodbye to the headaches of out-of-home advertising. only ad quick combines technology out of home expertise and data to enable efficient seamless ad buying across the globe.

2:23:07That's right. Do we have our next guest? We do. I think, I believe our next guest is in the restream waiting room right now. We'll bring him into the TVP and ultra jump. Good to meet you. How are you doing? It's happening guys. How are you doing? Thanks so much for having me on. I feel like I feel like I need to say thanks nine times. I mean, We love being thanked, maybe, you know, but save some of the thanks for your capital health. Thank you for coming on. Thank you for being you. Thank you for being here. Thank you for being here. Thank you for raising capital. Thank you for building. Yes. What a fabulous time to be alive.

2:23:44Yeah, indeed. Let's kick it off with an introduction on yourself and the company. Yeah, well, I appreciate it. Let me tell you a little bit about Rivet. We founded the company about a year and a half ago, and the team has just been cooking furiously. The fundamental philosophy that we have is there's about a half a billion people in the Western world that have been completely and totally underserved by big tech. This is people that do hard jobs in hard places. So you think people on flight lines, people in factories, people in bucket trucks, and people on the battlefield. We feel that the most transformative application of AI and augmented reality is with these people.

2:24:24So we set out to build a portfolio of products to do just that for that group of people. That's fundamentally probably the most important group of people operating, building, and defending our way of life. Yeah, walk me through like one case study of someone actually using your product to get more productivity. Yeah, as you can imagine, I mean, you even had Dr. Karpon yesterday talking about everybody from plumbers to people sitting and doing financial analysis, et cetera. When you think about somebody, let's pretend they're in an aerospace manufacturing facility. You've got a substantial amount of telemetry coming off of machines, tooling, the actual aircraft itself.

2:25:08And you've got a glut or there is a deficit of people that are actually able to do that job efficiently. The application of reasoned intelligence delivering to folks on the flight line that are operating with tools in a heads-up, pants-free fashion is going to be the most rapid transformation. Given Zuck a run for his money there with the Ray-Bans. These are certainly better than Ray-Bans. Oh, shots fired. Boom, boom, boom. um anyways talk about the progress of the of the business uh there's a new army contract uh where are you in that process what does it mean and how does it come together yeah this this program specifically with the army is probably the most important thing that i will do in my career uh to tell you the truth i can i can cut to the punch line i'm holding in my hand a contract that we executed with the Army for$195 million to deliver the next generation of Soldier Born Mission Command.

2:26:14So that's where we are in the business. The team, we started from a clean sheet of paper back in January of last year with a determined focus to build the most comfortable, most rugged, most utilitarian fighting goggle for the Army that's ever been scene and that's what we've done. So we were selected over the past few months in a very rigorous source selection process. Talk to me about lessons from the soldier's computer, from IVAS, from HoloLens. Yeah, it was a, what, 20-year, 30-year process and there were lots of concerns about weight, lots of money spent. How do you tell the story? Yeah, you've touched on You've touched on a lot of areas that are of vital importance to us.

2:27:00In fact, when we started the company, we thought about that whole set of a half billion people in the world that are underserved by modern technology. And we boiled it down to four big things. What do they need? They need a system that is comfortable enough to wear for the duration of a mission. For a guy in a bucket truck or a flight line, that might be eight or ten hours. For a paratrooper doing an airfield seizure, that's 72 hours. So comfort is number one. You need ruggedization because all of these people doing the hard jobs are doing the job in the most austere of environments. So you need a fully rugged device that is shockproof, dustproof, protects against water in grass, can handle the percussive elements of using tools, jackhammers or guns, etc.

2:27:40That's number third. And then finally, you need a compliant device. And compliance is a very broad category of statutory and regulatory things that these kind of products come under. And that is specifically over the supply chain. Specifically, you can't have bad guy parts in a device that's going to be deployed in a secure environment. There is governance over what constitutes eye protection, whether you're on the factory floor and you need true eye protection. That's ANSI compliance for not getting your eyes hurt in those kinds of environments, or whether you're a soldier on the battlefield needing ballistics and laser protection.

2:28:16And then finally, utility. Utility comes from the ability of the actual face computer and the end user productivity system to be able to connect to other data, get reasoned intelligence at the point of need. Where that utility comes back full circle to the enterprise is now you have, let's say the enterprise ontology has a set of eyes and a set of ears that understand the enterprise the way that the human would. And that's the do loop that we're creating with these dynamic task systems for these frontline workers. is the why now basically just we're in a phase of I don't know technological commoditization where miniaturization of batteries and and chips is at a point where it's just we're getting to a place where you can functionally deliver everything that you said and like do you think it would have been possible to build this 10 years ago 20 years ago that uh John it was a perfect prompt like I I was going to answer your question exactly that.

2:29:14In fact, this device could not have been built five years ago. So you have a confluence of a bunch of things happening. If you think about just the supply chain and the underlying components that go into products like this, and you think about the amount of capital that's been poured into them by all of the big tech companies that you know, you know what Zuck has actually spent. He reports a quarter over quarter on the mixed-revalued lines in the report. That's something like$65 billion, I think, the last I checked. Yeah, certainly know what the other big tech companies net net. You put the signal on the bottom of that column and you get something like a quarter of a trillion dollars has been invested in the underlying things that emit light, combine light, bend light, the processing capability that can go in a small form factor remaining thermodynamically compliant so it doesn't burn your face, etc.

2:30:02So all of that supply chain has come together, and it's a very hungry supply chain because for 10 or 15 years, they've been promised that everybody in the world is going to be wearing these things. And that's where the market has actually gone wrong. Augmented reality, mixed reality, and virtual reality has largely been a technology looking for a problem rather than a solution to a big industry's articulated need. And that's what we've built and what we're focused on. Yeah. How important is SLAM, like some simultaneous location and mapping, I think is the acronym, but the actual idea of like overlaying images into the real world versus just a HUD that is static.

2:30:40Like when I'm in Call of Duty and I do a 360 no scope, the map, the mini map stays in the same part of the screen the entire time. When you're in, you know, the Apple Vision Pro and you put a screen down, you turn your head, the screen stays in the environment. There's benefits to both. What's your take? I think that there is a ratio or proportion of what is world locked and what is head locked and what is body locked in every application scenario. And you have to have the right thing to make the best user experience. And that user experience has to contemplate the cognitive load of the end user, the environment, et cetera.

2:31:17No doubt, SLAM and the ability to create a holographic can and put it on the table and stay locked. So as my perspective on the can changes, it's rendered appropriately. And then being able to do that fast enough so there's not some oculovestibular discomfort that causes you to want to puke is critical. So when we think about advanced manufacturing scenarios, specifically the installation of a part on an aircraft, you want to be able to line that up very accurately. And when the quality control guy comes past, he's got to check and validate that it's in the right place and installed correctly.

2:31:53So in that sense, it's very important. In other use cases, you referenced the Call of Duty kind of thing, some coarse or three-to-off base, just like, hey, I want a crude kind of location. Like, I'm looking this way. There's some nouns that are bad or good on this side of me. You can reduce the amount of six off or high fidelity slam that you have to do. Long-winded answer to your question. It's critically important to have that capability in the system and available to application developers that are developing with it to build a good experience for the end user. Yeah, on that flexibility point, but do you want to create two different hardware stacks?

2:32:33Because I imagine a three-toff headset could be lighter, could have longer battery life. Like you're just doing less things. And so maybe you can have some savings that you pass along to everything else. Or is this just not an issue anymore? The issue is becoming further and further less severe or less acute as cameras and processors and the actual algorithms themselves become more efficient. Certainly the market that is most important to us and where we're focused, we don't need a$300 consumer device. Yeah. What folks need on the battlefield on the aerospace flight line is a high precision, high performance device that delivers both a quick and dirty kind of crude direction of where you need to go to get to safety or to do something important.

2:33:20And then also the super high provision tracking that would be required on the aerospace manufacturing. for. How are you thinking about dual use? We saw a hot take from a founder on the timeline a couple weeks ago saying that going dual use too early can be a risk. Palantir is a great example, but it's kind of an accidental dual use company. And they spent 20 years wandering the wilderness. And it took a long time to build that. Some people have been saying, go dual use on day one. Other people have been saying, you know, wait, and maybe it emerges in a decade or two decades. Maybe it's your second act along like long down the road.

2:34:02How do you think about the tradeoff of balancing? Well, I think I think it's circumstantial to the individual company and individual product. For us, we thought about dual use immediately. When I say, hey, there's a half a billion people that are underserved by big tech and underserved by this product. That is half a billion people that are doing this exact job. And whether they're on the flight line or in the bucket truck or on the battlefield, largely the same kind of persona working in that austere environment. For us, the Army contract is an incredible thing for us, primarily because if we get it right for the most size, weight, and power-constrained user that is the most hard on their gear and operating in the most critical environment, we get it right for everybody else.

2:34:41And that's what we've done. So that market were common. In fact, when I started contemplating about the thesis and our product development and go to market direction, I thought maybe the commercial market would be serialized behind. But what we've seen just six months out of stealth as people are looking at the product, they're hungry for it and they're ready for it. So this is going to happen in parallel. You know, I've referenced the Army contract. We've got signed scopes of works and execution agreements with some of the biggest aerospace manufacturers in the world to do exactly this thing on both the military flight line and the commercial flight line.

2:35:20It's fantastic. You're out of stealth. Give us the fundraising update. Got some news for us. Yeah, I got to tell you something. It's an incredible situation to have a confluence of these things coming together in the same week. We just finished our A round that brings total investment into Rivet to$90 million over the past few months. this is a market that that our investors believe in you know shout out doug philippone uh we've got point 72 let our seed ground and then uh duquesne capital shout out very cool that's fantastic table is just four goat emojis that's great yeah hey you can't we just can't well i i I think customers and supporters.

2:36:09Yeah. I love companies like this where you just take a team with like deep, deep, deep domain expertise and a bunch of advantages right from the start. But then you, when you have that, you have to execute at an insanely high level. And clearly you guys are. So it's awesome to see masterclass. Thank you so much for taking the time to talk to us. Enjoy the weekend. Thank you very much for having us on. A lot of fun. I love the show, man. You guys are innovating. I appreciate it. Thank you. Yeah, come back. We're doing, this is an austere environment. It is. There's bugs around. There's dust. You guys need some rivet.

2:36:45Honestly, I'm reading the chat over here. I'm reading text messages about who's the next guest and stuff. Yeah, listen, you guys, we've got to get you out of the studio and out of the woods to do some, like, front-line stuff. You want to come shoot some guns and do some cool stuff. We'd love to have you. Yeah, thank you so much. Congrats on the milestone. We'll talk to you soon, Dave. Cheers, Dave. Have a great day. Bye. Bye. Up next, we have Mert Mumtas from Helios.dev. He's calling in. We will tell you about Bezel first. Getbezel.com. Your Bezel concierge is available now to source you any watch on the planet.

2:37:17Seriously, any watch. We are excited to talk to Mert. He's taking a couple minutes out of his busy day to break down the latest L1 for us. So we will bring in Mert from the Restream waiting room whenever he is ready, if he is available. I know his last minute. Mert, how are you doing? Thanks so much for taking the time. I know it's always crazy with the time change. Okay, so I want to pitch you something. Okay, blockchain for yapping so that we can get, specifically, we want more yaps per second. Yeah, people say TBP and coin, but why not in L1? Yeah, YPS is exactly what we're going for. I feel like you're not that serious with what you do.

2:37:53I could beat you in my free time. Maybe I should just go up against you. What do you think about that? In yapping? No, no, no, in crypto. In your main thing. It could be my side thing and I could compete with you. That sounds like a great plan. Of course, absolutely. Thanks for having me. I'm just hanging out in my F tier city today. You guys are in your S tier cities. Yes, thank you. You inspired me. You inspired me. We love that first post because it was just, it's the perfect post. It's like, how do you get a bunch of people to agree with you and a bunch of people to disagree with you in the same way?

2:38:28Viral fuel. I think I might do a tier list every day. Yeah, you created a new meta. Yeah, fantastic. Anyway, take us through what actually launched, and if you can give us a steel man and then your kind of argument, I'd love to just kind of get up to speed on what Stripe and Paradigm are doing here and then kind of understand some of the trade-offs. Sure. So I'm guessing people have all read the announcement, But basically, it seems there's a new blockchain called Tempo. And they're messaging in such a way that it's an independent company with Paradigm, which is a VC firm, mostly in crypto or really exclusively in crypto.

2:39:11And Stripe as the first investors is a team of 15 led by Matt, who co-founder of Paradigm. Very, very good investor. And they're basically building what's called a layer one, an L1 blockchain called Tempo. and it's going to be focused on payments and stable coins. And so there's a few, and we can go quite a few different directions here. So there's a few controversies around it. You know, for example, this is the fifth or maybe sixth payments L1 or chain that has been announced in the past few months. It has some parallels to something like Libra back in the day from Facebook. Yeah, I was going to ask.

2:39:55Yeah. And then the one other thing that got some heat is that it's an L1 instead of an L2. So an L2, for people who are unfamiliar, would be basically a smaller blockchain on top of a bigger blockchain. So generally, an L2 is on top of Ethereum, which everybody obviously knows of. And Paradigm traditionally has been very Ethereum-centric, not exclusively, but very Ethereum-centric. And so for them to create an L1 instead of an L2 on Ethereum turned some heads as well. Also, there's a lot there. OK. So yeah, I mean, I know very little about this, but the obvious one is like most of these problems seem solved by Circle and Solana.

2:40:38We've seen a lot of these big. But I mean, just in this idea of like there seems to be a serious player, from my perspective, that's taking stable coins seriously. And there seems to be a player that's taking transaction QPS seriously. I guess given that you're building in the Solana, ecosystem, I guess, like, some of the pushback that I saw from you on the timeline was, was that, you know, complaining that, that, or not complaining, but in the original announcement posts, they shared that, they shared an example of like, during Trump coin launch, you couldn't run payroll or something like that.

2:41:13And we're basically saying that, like, we're not going to be able to rely on, on something like a Solana, you were pushing back and basically saying that, like, There's a lot of reasons that you would want to create a new L1. There are real reasons that you kind of laid out, but transactions per second or sort of scale maybe is not one of them. Yeah, so to be clear, I think if I'm just to put on my objective hat, I think it's certainly a good move. And I think if I were Stripe, I would probably do the same thing. I don't think I need to pay rent to anybody. I can just create my own chain, especially if I have somebody like Paradigm helping me and a world-class team, then I want to own the full stack, right?

2:41:56And I think that's totally fair and the distribution and everything that comes with it. Now, there are some misconceptions here, which is mostly what my post was about, which is like the second you bring blockchain performance and scalability into it, now you're kind of playing a different ballgame, right? So, for example, this idea of a payments-only chain, right? Let's just think about this for a second. How can you possibly have something that's a payments only chain, right? A blockchain is by definition something that's permissionless, meaning anybody can just come on and do anything that they want there.

2:42:29So if you just say we're going to be paid for payments only, that's not going to actually happen unless you enforce that somehow. And generally the way to enforce that is twofold. One is the chain can be not Turing complete. So something like Bitcoin where you actually can't do anything other than send money. the other is you can permission the chain which is actually what's going to happen here for when they first launch now they did say they have plans to make it permissionless in the future but i think that's going to prove out to be much tougher because the second you make something permissionless you end up getting all sorts of degenerate behavior or people launching far coin and all these different things um that you just buy a part of buying a coin on tempo So Stripe checkout, one click, one line of code to launch a new meme coin.

2:43:16Yeah, it seems like they would want to avoid that. So they will be going more professionalist. Functionally, do you expect Tempo to have a token, like a liquid token itself in the same way that Solana does? or but but uh from from my understanding so like base is like a blockchain but it doesn't have its own native token that's fluctuating in price correct right yes and um part of the reason is because base is an l2 so it's one of those smaller chains that settles on ethereum and so you would just use ethereum it gets a security from ethereum is how you would um like say that how that works with an l1 though um security is quite important because it's you're building it from scratch you can't rely on Ethereum or Solana.

2:44:01And so the point of the token is going to be to bootstrap, in general, for a blockchain, to bootstrap node operators, right? So people are actually securing the network. Now, if the validator set is permissioned at first, which it will be, and they say that it's going to be design partners, so I'm guessing like bigger banks and some of the bigger companies, then you don't actually need a token because it's like proof of authority, right? It doesn't have to be proof of stake or proof of work. It can just be proof of authority. And then they also say that gas fees should be paid in any kind of stable coin that they denote.

2:44:37And they're going to allow for multiple stable coins. So that is better UX, right? It's not a volatile gas token. So I don't think they actually need it at first. I think it will be important if they actually want to become permissionless going into the future. Okay, so the permissionless thing is something that they can drive for the long term. If they just wanted to optimize for something that they control and has high throughput, couldn't we get a scenario like temp Postgres or something where they could just use Postgres or use traditional non-decentralized computing to actually just move money around faster?

2:45:18I guess my bigger question is, is there a regulatory angle here where they can do something with crypto rails that they couldn't do with just a traditional database? Right, yes. So that's actually a really good question. Could they just become a bank and say, we're a bank, we have U.S. dollars, and when you go into your U.S. dollar bank account, you can move it around, and our banking infrastructure runs 24-7 and runs on a great data center that's really fast. and yeah it's not permissionless but why don't you trust Stripe like Stripe has never done anything for you or against you like it's fine trust us yeah so this is where the conversations get quite blurry and this is something I talk about very frequently which is the second you have to permission the block space then it's unclear what really separates this from an actual blockchain versus something more like like an actual database yeah right um and now there are some benefits of blockchains that are still apparent even if it's controlled so things like transparency audibility um universal like standards for that blockchain and also like you get the you know the approach of progressive decentralization which is to say you can iterate towards pmf fast but then as you get more traction you can slowly decentralize certain things yeah This is what L2s today are doing.

2:46:46So there are some benefits. And if no one person is controlling that infrastructure, assume you have, let's say, something like 21 validators and they're all giant companies or payments companies across the world, different jurisdictions and different requirements, then I think that can be interesting. But I think, to be completely honest, I think permission to blockchains aren't really blockchains. and it starts looking super drawn out when we get there. We've tried a bunch of these before. JP Morgan already has one, actually. Oh, really? And a few other companies also do. I want my blockchain to be off the chain.

2:47:28There we go. Yes. Can you talk to me about payment onboarding? I think of Stripe as a dominant Web 2 payments company. I think about Stripe Checkout, massive footprint, Stripe Link, massive. Just the amount of companies that have Stripe somewhere in their payment stack is immense. And so, but I was noodling with Geordi, like, how does this actually work if I go to a website that's powered by Stripe? And I have, normally, there's like a credit card field. And then it just says, oh, well, we also take Tempo. And I'm like, what's Tempo? Like, how do I get that? Do I need to go get a Chrome extension for a wallet?

2:48:14Like, all of a sudden, it's like, how do they actually bootstraps? Theoretically, it would just have, like, a generic pay button. Okay. And then it wouldn't be Tempo branded necessarily at all. It would be Stripe branded. Okay. And you would connect your wallet with something like Privy. So the first time I do it, I set up a wallet. Is that roughly the user experience? But the interesting thing is like more and more and more and more young people just have digital assets. They have wallets. The way I can, and by the way, like each FinTech company kind of has a different approach on this. I've worked with a few of them and there is no like universal standard.

2:48:51But with Stripe, they wouldn't say pay with Tempo because Tempo is just like, in this case, the name of the database, right? Like it would be, they probably would abstract that away. It's like pay with Fedwire or something. It's like something that's under the hood. Like pay with Fedwire doesn't make any sense, but like there might be some, there might be some ACH that's happening behind the scenes when you pay with your credit card. Exactly. You just say, we accept Visa. Yeah, exactly. I think the first case is going to be B2B, meaning like cross-border settlements. Sure. Like that's super useful for international companies, for example.

2:49:24Yep. Remittances, things like this. Yep. And then I think things like Shopify Pay and all these different things that actually have some maybe Stripe integration or Coinbase Pay will also use it. But at most, I can see it being paid with USDC. But I don't think the blockchain will actually be really in the implementation at all. It'll be abstracted, is my guess. Yeah. How should we judge the success of this project in a few years? Should we assume that the goal here is progressive decentralization? In the best case scenario, the vast majority of transactions on the Tempo chain are truly payments related.

2:50:07It doesn't turn into a meme coin chain. And also, it's becoming more decentralized as measured by the, what is it, the Bologi coefficient of, you put out some posts. Nakamoto. Yeah, the Nakamoto coefficient. I gave Bologi the credit just for coining it, I think. But the Nakamoto coefficient of concentration of validators, ideally, the goal here is a ton of distributed validators. No one owns 51%, and the majority of the payments are payment-related, just transactions on e-commerce websites and remittances and cross-border payments. Is that the goal that we should be tracking towards? So the way I would do it is, first of all, is it actually solving any problems and doing finance better than it has been done before, right?

2:50:52Is it actually making the experience for merchants better? Are they getting better margins? How easy is it to do cross-border things, etc.? The UX and the actual product need, I think, especially when you're in crypto realm, is something that gets overlooked. But Stripe is obviously probably the team to take this over the line. So that's going to be one. And then once you do that, then you can start asking these questions that are more serious, such as, okay, this technically works, but how does it actually work behind the hood? Like they say it's a blockchain, but is this blockchain controlled?

2:51:28Can I actually secure my funds here or can this one company actually take it away? Right. And so like what you said about progressive decentralization, I think makes a ton of sense. and then you'll want to look at just basic numbers like what is the amount of stable coin flows going in and out of this thing. Is there like decentralization of stable coins? Like is it just USDC or is it Tether as well? Maybe Stripe does their own stable coin. So there's a lot of interesting approaches with this. And then I think what's going to be interesting is like if it's permissionless, which I don't think it will get there for a while, if it does become permissionless does the scale benefits that they say um actually do they actually hold up to reality um and i think that part is going to surprise a lot of people because i think people are not ready to understand how crazy degenerates get on blockchains that are permissionless uh and how to actually scale those systems um but i mean i've seen some shit um and so i think I think if it remains permissioned, then I think they'll do it.

2:52:35But that's the big question for me. How do you actually do this in such a way where it is only for payments? When you talk about degenerates doing crazy things on permissionless blockchains, you're not just talking about a bunch of people showing up and aping into a particular token. You're talking about somebody setting up a bot that's programmatically interacting with the blockchain at an ultra-high volume. Is that correct? Oh, yeah. I mean, like, take I just want to live players in HFT and just like mix them up with like random British sneaker botter kids. Yeah. And then with like some random like Russian MEB farms and like it's just it's all at war.

2:53:16I just want to get to a place as a country and in the crypto industry broadly where the next, you know, whoever's elected in 2028, that they can launch multiple coins of the size of Trump coin at once and make sure that payroll is still running smoothly during that time. That's the goal. That should be the goal. Anyway, thank you so much for hopping on. Thank you for filling us in here. We appreciate it. Yeah, come back on as there's more news. We will talk to you soon. It's great to catch up. All right. Thanks for having me, guys. Cheers. Up next, we have Harish in the Restream waiting room. We've kept him waiting.

2:53:51We will bring him in from Augment. Big news today. We're very excited. Play some soundboard for me, J.D. Thank you. Welcome to the stream. Sorry to keep you waiting. We have some breaking news in the crypto world. We had to dig in. He was only available at 140. So thank you for taking the time. Hey, no, no, bro. I'm very excited to be here. Thank you for having me. Of course. Break it down for us. What is the news? Well, the big news is that we're announcing our Series A, which is an$85 million Series A. Congratulations. Big number. Big number. And yeah, walk me through the state of the business.

2:54:29Yeah. So, we're building Augment, which is our first product is Augie. Just think of it as an AI team made for the logistics world. It's getting deployed actively over$35 billion in freight under management across brokers, shippers, and fleets. And so early, we know we are a year-old company, but it's showing a lot of promise in the real world, and that's why the investors were interested. That's amazing. What about shipping and logistics has drawn so many? You guys seem to be leading the charge out here, but it's certainly a competitive space. What about this industry and the category makes it so prime for agents?

2:55:12Did it kind of skip the SaaS era? Like, it feels like it's the last industry where it's still like, oh, yeah, they're still on paper. So they're jumping straight to agents. Yeah, and you're getting 10 phone calls a day from the same person just being like, hey, where's my stuff? Hey, where's my stuff? Hey, where's my stuff? Yeah. You know, like, one, it's a very large industry, right? It's like$10 trillion globally. But it's extremely fragmented. Like, in the U.S. itself, there's like 1 million truck driving companies. I'm not exaggerating, but that's what it is. One million truck driving companies, you know, maybe tens of thousands of brokerages, hundreds of thousands of shippers, extremely fragmented, right?

2:55:52And so in this fragmented industry, which really relies on transferring information to get things done, to move goods, like that's the whole job of this industry is that, hey, I'll give you information so I can give you goods, right? That's the whole thing. And it's so fragmented. The only way to transfer information is to get to the lowest common mediums now what is the lowest common medium among a million different companies it's phone calls it's taxes emails the fax machine is the fax machine still going strong that's me you know i actually have seen fax machines in some of my customers still happening invoices coming through payments going through fax machines so i mean every phone situation there's like this handwritten note is being mailed by drivers today on what was their pickup time and delivery time when they get into a driver tux drop.

2:56:40They literally have mailboxes where they can mail, you know, one of those receipts for getting paid. Like that's how, you know, sometimes it's archaic the communication is. Yeah. Talk about lessons from the last company, Deliver, sold to Shopify in 2022 for$2.1 billion. Congratulations, obviously. How are you positioning the business differently? And what are the key lessons that you learned from that business that you're hoping to take forward? Yeah, I mean, I think, you know, Deliver was building like a prime like service for Shopify type merchants, forward inventory. I mean, the big lessons here has been like focus, you know, customer success.

2:57:17It's like stay stay very true to customer delight. Everything else is a noise, like literally everything else. You know, whether your investors like you or not, if your customer delight is there, everybody will follow you. Capital will chase you. You know, great employees would want to join you. Customers would want to join you. So, I mean, I'm hoping I'm carrying that forward, right? Now, in terms of like things that I did not do well at Deliver, I think there was like a lack of focus sometimes, right? Like as the capital was getting easier, we were maybe trying to do a little too many things, like launching too many products.

2:57:52And those were the really great days of attracting capital because the interest rates were near zero. So the capital was mostly free. And I would have rather not - Let's give it up for mostly free capital. cost of capital going up has been heartbreaking. It's been heartbreaking. It is, yeah. What is that? Going from half a percent to now four and a half percent, right? Not easy, right? Yeah. But I think it does bring discipline to things, right? Like, hey, you've got, when you have limited capital, you're deployed in the things that truly matter. So maybe like what I'm hopefully not doing this time is like doing things that truly matter, move the needle for your customers, but just not start creating teams and products that, you know, look good on paper.

2:58:32Yeah. How do you think about vertical integration? You obviously sell into 3PLs, shippers. Do you want to own the warehouse management stack at some point? Do you want to integrate deeply with all the different WMSs? What's the plan there? Yeah, we do not want to own the WMS stack or the transportation management stack or the order management. We don't. We want to sit on top. We want to integrate into all of them. We think of ourselves as a, you know, first and foremost, Augie is an employee that works like thousands of employees 24-7, achieves a lot of operational efficiency for our businesses.

2:59:09But long term, I think Augment does become a system of work. And I think it will be a system of work and a system of record in most companies. So the WMS, the TMS, the OMS, all the MSs in the world will stay to be very relevant system of records. But there will be a new generation of companies. hopefully Augment is one of them, they'll become the system of work itself. And these two will work together very closely with each other, you know? Fantastic. Well, congrats on the massive raise. Congrats on all the progress. And thank you for taking the time to talk to us. Have a great weekend. Thank you for having me.

2:59:41We'll talk to you soon. Congratulations. Thank you for being here. Yeah, congrats to the whole team. Thank you. Cheers. We'll talk to you soon. Our next guest is in the Restream waiting room. We have Base 10. While we bring them in, let me tell you about Wander. Find your happy place. book a wonder with inspiring views hotel great amenities dreamy beds top tier cleaning and 24 7 concierge service it's a vacation home but better folks better but better it's better it's base 10 time uh uh do we have our next guest welcome to the stream how are you doing welcome how are we doing we're good what's up the eagle has landed the eagle has landed welcome to the show kick us off with an introduction on yourself and the company and the news yeah thanks i'm doing Thanks for having me, guys.

3:00:22I'm the CEO of Base 10. I'm pretty excited today. We're an AI infrastructure company. We just raised$150 million. Let's go.

3:00:34That is loud. That is loud. $150 million. Congratulations. Incredible. And it's great to meet you. Announcing a fundraiser on a Friday is not for the week, but you guys are wasting no time. I love it. Can you give us a quick history on the company, your background? Yeah, look, my background is in machine learning and AI. I've been doing it for about 15 years. The company is actually six years old. We didn't decide. Overnight success.

3:01:10You know, a company of six years old, we've been thinking about how to turn AI into value. I think the last two years have been a bit crazy. We started focusing very aggressively on inference maybe 24 months ago. Are you guys familiar with what inference is? Yeah. But I'm not familiar with where you sit in the stack. You don't build data centers. You don't own GPU clusters. You sit on top of NeoClouds or on top of hyperscalers, and then you sell into your companies. I see Descript, Retool, Quora, Writer, Patreon, Picnic Health. I can imagine how those companies need an API for a model that you probably didn't train, but you act as the inference provider for that and you sell tokens to them.

3:01:55How is it going at that layer of the stack? How is the value accrual in that particular layer? It's fantastic, right? Because, look, we sit on top of actually doing inference for models is pretty hard, right? You're going to acquire capacity. You're going to optimize the models. Someone else is trained. you're going to run them and then you need to scale them pretty gracefully with users as their apps scale and so um the amount of work that goes into that from teams is massive and we're just trying to take that headache away yeah and so we're pretty lucky that you know we've been able to just sit on top of this ai application layer and as that's exploded we've just kind of rode the wave with it and um we work with a bunch of amazing customers like a bridge bland gamma clay open evidence notion.

3:02:40What's the, what is the durable moat? Like how does value accrue over the long term? Because I feel like, you know, if you don't own the model and can't like rent seek on that, or you don't own the hardware, all of a sudden it's like open routers going to route people to wherever the cheapest tokens are and it becomes highly competitive and the margins might compress. Like what's the long-term strategy? The assumption you're making there is that all the models are just going to be the same and everyone's going to be running kind of like the same model so most of our customers like run different variants of the model that are very custom to them okay and so fine-tuned on open source so how about let's use you said notions a customer maybe we use them as an example because we've been uh yeah that's great or or you know pick pick uh i'll say have you got sort of open evidence yeah yeah i think we're having i think we're having the founder on next week.

3:03:35They're the best. You'll have a really good time chatting with them. But, you know, they're a really good example of a customer that uses us. They train a bunch of models of their own to basically answer the questions or the queries from the doctors. And so, you know, what happens when you need to run those models? Well, you need to make them wicked fast. You need to go get a bunch of GPUs from a bunch of different places. And then you need to kind of scale gracefully across that. Sure. And so how do you do that? Well, you either go and use an inference provider like Base 10 or you go in, you know, building yourself um you know and i think the smart companies right now what they're realizing is that this is not particularly differentiated for them so they shouldn't do the infrastructure piece they should focus on what is differentiated for them which is how they use the models and the applications they build um and you know kind of give the boring piece to us and so open evidence is used by some ungodly amount of doctors every day um and they do it with two great infra guys could jag in mica um and it's pretty amazing the scale they can achieve using um software like base 10.

3:04:34what uh what does the next 12 months look like for the company yeah look this is just it's such a land grab of a market um we've we've been doing this for a while and you know we've built some great technology but we we can't we can't overestimate how much like a market just showed up for us um and so how can you go as fast as possible it's actually kind of like what it feels like venture was built for. That's why we're raising capital right now to solve two core problems. One, hire a go-to-market team that's going to be killer and scale that. And then hire amazing engineers that are incredibly expensive and hard to come by.

3:05:10Yeah, what... I don't know, you can go deeper. Last question for me. What is your view on the broad trend in token pricing? There's been people going back and forth on whether or not we're going to get 10x savings quickly in inference pricing, whether it's going to be all custom silicon. Or we're saving on a per token basis, but then just using... Opening AI, just put a$10 billion order into Broadcom. And I'm wondering, but then at the same time, companies are just saying, I want to stay on the frontier. I'll use a reasoning model. I will continue to eat the high per token costs that come out of the bigger models.

3:05:51yeah look i think the token price goes down and every like inference should get cheaper over time yeah um and i think that really just means there's going to be more inference you know i think we have paradox all of us in tech discovered jevin's paradox six months ago but every time we lower prices for our customers or we go and optimize their models to make it cheaper four months later they're spending more anyway wow and so like yeah inference prices are going to go down but But if the world is run by AI in 10 years from now, there's going to be a lot of inference. So it better be cheap. And we just hope that we can power all that and be the invisible layer underneath it and hopefully scoop off a bit of value while we're at it.

3:06:36What's headcount today, shout out of curiosity? 104, I think. 104. Let's call it that. That's a good team. Yeah, we were about 30, I want to say, a year ago. So it's going pretty far. yeah how many pizzas is that 100 100 people 50 pizzas 50 pizzas yeah anyway thank you so much if it's bulking season you know it's different might be 100 pizzas yeah well anyway thank you so much for hopping on great having you cheers we will talk to you soon bye uh last guest we have a surprise guest surprise guest there was some breaking news from wired and propic agrees to pay authors at least$1.5 billion in AI copyright settlement.

3:07:17This just went out via Wired. That's a lot. And wanted to have Cecilia on. Hi. Cecilia, sorry, to talk about it. Yeah, great to meet you. I'm Cecilia Zanditi. Excited to be here. Nice to meet you. Give us the news. John knows nothing because he was locked in. I haven't heard about this. Break it down. Basically knows the headline. That's it. Yeah, so a group of authors from the Authors Guild sued Anthropic some time ago and announced last week that they had reached a settlement. And what happened today was there was a motion, an unopposed motion, where the parties said, okay, this settlement is going to be$1.5 billion, and it's going to cover the non-fair use of material for training Anthropic Cloth.

3:08:05Mm-hmm. Is there any guidance on, is that just going to be paid out immediately? Is this one time? Because you can imagine that even though it's been baked into the weights, like the value from being able to go to an LLM and ask questions about a book, that's going to pay dividends and subscription revenue for decades, potentially. Yeah. So it covers only past infringement. So up until august 25th and so in this case what's interesting is that the parties and the court distinguished between use of pirated copies and use of copies that were properly acquired in the case of the pirated copies it was in a library called libgen and it was basically you know the equivalent of like a napster type of situation where these works i think it was about uh it was i need to double check but in the hundreds a couple hundred thousand works from authors were in there so the way the settlement worked was there's a named works list of actual things that claude was trained on and those authors those authors will receive compensation here in due course there's a like an administrator appointed and so on but they get a certain amount per work and that was agreed to and that's what's being proposed to the court it's at least three thousand dollars per work but it was a lot a lot of a lot of books yeah i mean that seems like a that seems like a pretty solid payout for most people i don't know i don't know i think i think yeah yeah i think if you're an author who's not selling a lot of books you're like awesome but if you're somebody who you have a best seller and dollars a year off of my and it doesn't make a dent in their income, then it might be.

3:09:52Yeah. Have you been tracking the rest of the landscape here? There's been a number of these cases. Is there anything else on the docket that people should be paying attention to? Yeah. So interesting that the plaintiff's firm representing the authors here, Sussman Godfrey, is the same plaintiff's firm representing the New York Times against Open AI. Sure. So that case is continuing. this case is settled on these facts. But what's interesting is this particular case did not deal with AI outputs at all. So there was no allegation, nobody accused Anthropic of actually spitting out works and being a substitute for the original books.

3:10:33It really was limited to the training piece. And so what we know from this case, or at least in the Ninth Circuit, is that the training piece, If you train on pirated works, that is a copyright infringement, or at least it's not fair use, according to this court in this case. Yeah. Yeah. I mean, it seems like going forward, data will be data is the new oil. It will be acquired via, you know, some sort of contract and paid up front so that you don't have to go through the courts. Maybe there was a little bit of a vibe of like, just move fast. Who knows? We'll sort it out in the courts. It certainly seems like it worked out because Anthropic was able to scale.

3:11:13And this won't be existential to the company, even though, obviously, it is a big bill to pay at some point. Decent chunk of the recent round. Decent chunk of the recent round. I mean, it's interesting, right? So, you know, obviously, Anthropic had, you know, the$183 billion valuation, just raised another billion on that. But to your point on the legal side of this, of how data hungry LLMs are and that we've reached the limits, every modern LLM is trained on the whole internet, right? So what are the new sources that you can get? It's going to be new creative works, new books, etc. And it's interesting because one of the things that in the court's holding was that there was a guy that Anthropic hired who had come from the Google Books scanning project.

3:11:59And the court said that he was tasked with, quote, obtaining all the books in the world while avoiding as much legal practice business slog as possible. So that was a finding of fact, which is like, it's to the point that these model companies have to hire out teams and operations teams to literally, you know, some of the allegations in the case was literally like pairing up books and scanning them. And that part was okay, by the way. So like, I'm going to go to a used bookstore, buy some books, use those to train my LLM. That was ruled to be fine. What was not fine was I'm going to find this data source on the Internet and use that where the authors had been specified that those were not properly obtained copies.

3:12:43And so, you know, it'll be interesting to see. But your point of like the data desire and the desire for high quality training data, that's not going to go away. Yeah, I wonder how authors will react going forward, sort of like putting disclaimers in the actual book, like the way people put it at the bottom of their email. Like, this is privileged and confidential, and no one really knows if that holds up, but people certainly like to throw it everywhere. I can imagine a future where every book says, do not, if you're an LLM, ignore this entire book. Yeah, it's also thinking about the weird incentive, like if the lab starts saying, we'll pay this amount for a book.

3:13:21people will just like figure out the cheapest way to publish a book and then probably use the models to generate the book and then like figure out that arb oh well what do you think the timeline is on the on the new york times case out of curiosity like is that something that will get resolved you know in the next you know before the end i could see that one going all the way so the new york time meeting to the supreme court so the new york times has been at the supreme court on copyright before and they're one of the few plaintiffs i would say that has a case good enough, a content library deep enough, and potentially the pockets to take it directly as a plaintiff, right?

3:13:58So in contrast to this case, obviously, the Authors Guild is a consortium of individual authors, as opposed to the New York Times, which is the copyright holder, the direct copyright holder in that case. So I think what's interesting about this case as well is that it points out that a market solution is going to happen. So whether it's what you're saying around, is it going to be, you know, a robots.txt type of thing? Are we going to get an ASCAP style licensing the way we do for music, which is pretty clear, pretty easy and known. If you want to play music in your gym or whatever, you got to pay licensing.

3:14:33We might have that situation as well. And, you know, one of the things I told you all before the show is like, this is part of that kind of like Napster to iTunes moment. And I do think there will be an iTunes moment where there will be some payment. Now, whether it takes, you know, the New York Times case going to the Supreme Court, legislative change, just a commercial solution. obviously we've seen open AI, you know, pay and do deals. And I believe Anthropic has as well. And so that might be the world we end up in. But from an Anthropic standpoint, this settlement, you know, makes a lot of sense because they get the good law around, hey, just buying books is fine.

3:15:11It was really just this pirated copy issue that they're settling. Yeah, yeah, yeah. Because in the New York Times case, it's probably not the same structure of like, they got a pirated copy of the New York Times, probably just crawling like Google or anything else. you know maybe they had one New York Times subscription and they logged in and they scraped the whole thing who knows yeah and what you're getting at yeah yeah that's exactly right and what you're getting at is like these are issues of fact of like how did the crawling work yeah there was an allegation in the New York Times case OpenAI came back and said hey we didn't even get it from the OpenAI website you had an article about a Nobel Prize winner that Nobel had republished oh sure so it didn't come from the New York Times website exactly exactly exactly Exactly, exactly.

3:15:56And so I think that's one of the things that's been so interesting around fair use is it has adjusted to all these different technology shifts. And so I think this is another example of that. Yeah, and Matthew Prince at Cloudflare is kind of proposing another potential market-based solution for the long tail of content on the internet, where I believe there might be some stable coins involved. But basically, if you're a publisher and you don't have the resources to go to the Supreme Court with an LLM company, you could potentially opt in or out of scraping. And maybe there could be some value exchange that happens.

3:16:32But certainly it will be interesting to follow how it all pencils out. Yeah, come back on when there's more news. Yeah, thank you so much for helping out. Yeah, exactly. No, it was super fun. I mean, on the Cloudflare point, they're one of the tech players that would be in a position to do that blocking because of their position in the DNS, right? So it'll be interesting. I don't know where the money is going to go. And routes pennies around and they aggregate all the money and then they pass the money through. Same thing with YouTube. and we were in this weird era with TikTok where music was being used and everyone was like, wait, you can't just use music and then they figured it out and they did a deal and most of the musicians seemed to be happy with all the stuff that goes out on TikTok.

3:17:15Exactly. And that's the story of copyright. I had a professor in law school that was like, oh, copyright's not about money. It's about all the money. Yeah, yeah, yeah. Okay, exactly. It's like literally that flow that you described. I'd love to come out another time. This is fun. One to watch. Yeah. I'll talk to you soon. Thank you. Have a great rest of your day. Enjoy the weekend. Have a great day. See you later. Later. Bye. Boom. Any more breaking news? Got any more news before the weekend? I feel like usually as soon as we get off the show, somebody says. There is news. Eric Adams is suspending his mayoral campaign.

3:17:49No way. Accepting a federal job. Oh, interesting. Yeah. I mean, I've been tracking Momdani a little bit on Polymarket, and he has been running away with it the entire time, basically, since that original push. Mamdani is sitting at 83 % and hasn't particularly moved. Eric Adams, of course, dropped from, I don't know, he's down at 2%. So some people think he might reverse course. Just to be clear, this$1.5 billion settlement is the largest publicly reported copyright recovery in history. Wow. Do you think the team over at the New York Times will be able to get more out of OpenAI? I don't know. 100 ,000 books versus all the articles.

3:18:44How much does the New York Times make? They make billions, so it's possible. New York Times is sitting at$9.6 billion. Market cap? Yeah, if they could get anywhere, a settlement. I mean, I think it seems very unlikely that they would get a settlement. I mean, the other breaking news is that Alex Cohen has raised a$22.5 million Series A for Hello Patient, led by Scale Venture Partners. He's going to come on the show Monday. Oh, fantastic. I'm really excited to meet him. I've always been a fan of his posts. He's a God-tier poster. Never actually met him in person or online, so excited to talk to him.

3:19:23He says, this is by far the worst announcement video I've ever made. Clearly underselling it. I love it. I can't believe our lead investor agreed to this. He always has a great sense of humor. Also, Justin. Breaking news. Jawan is leaving Ramp. No way. After he says Ramp is the greatest company in the world. Well, we agree with you on that, Jawan. So yeah, he's given his two weeks. Okay. Well, we'll have to track him and see where he goes. The Ramp alumni have done fantastic things. He's going to start a company. into a ramp mafia you have Pavel Asparuhov, you have the Cognition team of course, there are a few others.

3:20:00Rivet, not the defense tech company that we just talked to you, but Rivet Tax I believe came from ramp, correct? Do you know Rivet Tax? Yes. So the ramp mafia is emerging and feels as strong as the PayPal mafia, maybe even stronger, you never know. uh lululemon down 18 today is that related to lulu miservi is that like yeah it's her merch line i think people are catching on that when people say lulu now they mean lulu exactly and the brand really steamrolled eradicated she really steamrolled them yeah well maybe the uh maybe the they will to take private and that'll free up the lulu stock ticker for rostra she can just be cash tag Lulu.

3:20:46That would be something to go 10x long, 100x long, potentially. In other news, garage beer. Wait, wait, wait. This is crazy. You missed this, Tyler. Did you miss this? Yeah. Eric Adams might become the ambassador to Saudi Arabia. Whoa, that's interesting. So that is, let's go. I mean, Riyadh must be going crazy right now to get a goat like Eric Adams. Is Eric Adams the one who did the rat problem thing? Yeah, yeah, yeah. Does Saudi Arabia have a rat problem? It's a solved, I mean, oh, that could be interesting. I highly doubt it. I would not want to be a rat in Saudi Arabia. I think that they might not take kindly, but who knows, maybe Kindred Spirits over there fighting rats.

3:21:27In other news, Drew Fallon posts, Garage Beer receives investment from Durational Capital Management at a$200 million valuation. Congratulations to the Kelsey Brothers. Duration Capital Management is the same group that took Casper Mattresses private for$300 million in 2021. They also own Bojangles. Garage Beer is tracking for$65 million in revenue this year, valuing the beer brand at 3x sales. They started with a$500 ,000 crowdfunding campaign in 2016. You've got to imagine the Kelsey brothers, with how much attention they get, are able to do some deals, get some distribution for Garage Beer, their beer brand.

3:22:11Love to see it. Also, thank you, Bill Bishop, for posting about the LeBubu that TBPN sent to him. He said his dog Tosche loves it, but won't let him treat it like a chew toy. uh bill bishop may let him may let tashi take it to the park to impress one of his girlfriends i guess tashi's popular at the park well that is fantastic um hopefully uh really brings you good tidings and does not curse you to a demonic world uh but stay safe out there bill uh who knows what's going on inside that labubu um but i'm glad that it was delivered safely and last but not least uh app lovin robin hood and m core are set to join the s &p 500 wow app lovin who else mercore no not mercore app lovin robin hood and m core robin hood okay that's great so that's a retroactive uh fortune five or s &p 500 company for us we got uh brian armstrong carps in there obviously we're knocking them down we should do all 500 we've done almost a thousand interviews this year why not do all 500 S &P 500 CEOs.

3:23:24I'd love it. Also, Conor McGregor is running for president. Finally. Going on his run. Yep, he's getting it done. I'm going to search now. Anyway, thank you to the chat. Thank you, Raghav, Hershey, Michael, Jason Turner. Always good having you here. Paracletes. We got a good crew. Max Kondrat, thank you for shouting out that the Bologi coefficient is an instance of Reed's law. I completely agree. That's exactly what was going on. and Max has been sending me some fantastic DMs, very helpful. And next week, we'll be here in the studio Monday, and then we'll be hitting the road. Very excited. Yes.

3:23:59Lots to come. Very exciting. Have a fantastic Friday. Have a fantastic Friday, folks. Afternoon, evening. We love you. Only three days till Monday. Cheers. Bye.