Mark Cuban, Vlad Tenev & MORE | Larry Ellison's Cloud Scores Big, Silicon Valley's Family Life, Qasar Younis, Minna Song

27:59Mag-8. Put Oracle in there. I mean, founder-led company. Good luck. He's been on an absolute run. And he's, I mean, it's crazy where Oracle started and how they're still conquering in essentially the same market. Obviously, very different technology, but have done quite well. So Oracle, a software company that had long struggled to find its niche in this market, gained a foothold through scoring some major customers. Today, it's helping power Elon Musk's XAI from a data center in Utah, assembling a cluster of tens of thousands of AI chips for NVIDIA near Singapore and recently inked what is likely the largest single cloud deal ever with OpenAI.

28:40We talked about this a few weeks ago. This is OpenAI effectively committing to these sort of massive inference bills years in advance. We're talking tens of billions of dollars. Is this inference or training? I thought this was training news. Models will be trained there is what Brody Ford says. I was talking about the OpenAI cloud deal specifically. Yeah, it's a move away from Microsoft, which is the underlying, the undercurrent here. Oracle is on the hook for tens of billions of dollars to develop unprecedentedly large data centers amid energy and material shortages. shortages. That includes a plan to spend more than$1 billion a year just to power one new megasite in West Texas with gas generators rather than wait for utility connection.

29:28With all of this spending, Oracle recorded a negative annual cash flow for the first time since 1990. That's absolutely wild. And questions remain about the longevity and margins of offering infrastructure to train AI models. Oracle, founded in the 1970s, had long faded from tech prominence, certainly hadn't faded from prominence in things that matter. Exactly. I was about to say. Racing. Racing. So I won't accept this. Having a founder that looks 40 despite being closer to 80. Yeah. On the road to its return in the AI era were internal battles to convince chief executive officer Safra Katz to invest in the cloud, recruitment of key customers like TikTok, and the rise of a brash executive named Clay Magorick.

30:16Magorick. McGork. Bloomberg spoke with more than 30 current or former employees and customers to understand how Oracle can deliver on its towering promises. They detailed spiraling costs, accelerated timelines to satisfy open AI, regulatory roadblocks, and the hiring of hundreds of former Amazon Cloud staffers, speaking anonymously to discuss the inner workings of a famously private and litigious company. Oracle declined to comment. So as cloud computing picked up steam in the late 2000s, Ellison had little reason to be excited. Oracle's namesake database, which is purchased outright and installed directly on customer servers, is one of the world's great cash businesses.

30:53Adopting a cloud sales model where users rent access to software would almost certainly come with lower profits and a need to collaborate with rivals. After years of writing off the cloud as marketing, Oracle's tone changed as its business came under fire. Amazon pushed deeper, cheaper database options on its cloud customers. Application rivals like Salesforce promise easier installation and updates. Ellison saw the need to get Oracle's Fusion apps, commonly used for finance and resource planning, on the cloud, according to people familiar with his thinking at the time. There were competing efforts through the 2010s within Oracle to build a cloud for internal and customer use, an initial attempt led by Thomas Kurian.

31:32Kurian, no way. He's the head of Google Cloud now. Okay. Yeah. I didn't realize he was at Oracle before. Yeah. So a separate group made up of largely Amazon alumni led by Don Johnson, pushed an approach called Bare Metal, in which clients wouldn't have to share their servers with other renters, allowing Oracle to target privacy-focused customers. The group also focused on building smaller data centers, which let them expand in countries that weren't yet viable for industry leaders. The Bare Metal approach won the blessing of Ellison, who remains a key decision maker despite handing off the CEO title more than a decade ago.

32:05I absolutely love that. It's like, hey, somebody else can handle all the stuff I don't want to do. I'll still handle the big ticket decisions. What was that famous Larry Ellison quote during COVID? Larry Ellison. I'll keep reading while you find it. The new product came to be known as Oracle Cloud Infrastructure, or OCI. Kurian departed the company in 2018 after his approach was sidelined. He then joined Google to lead its competing cloud service and remains in that position. And there's a chart here, if we can pull it up. Oracle Cloud is expected to be majority of business by 2029. And Wall Street is projecting breakneck growth.

32:49You find the quote? I don't remember exactly what it was, but basically he was just hanging out in Hawaii during all of COVID. But he paid all of his workers like the full amount when they were in benefits through May or something. They did some shutdowns of the office. But it was just very funny where he was just like, I will be in Hawaii while this chaos is unfurling. And he was hanging out in Lanai. Why have a compound in Hawaii if you're not going to go there during a pandemic? Yeah, it's like the spot. I mean, everyone was out there during that time. When offering the new product, Oracle had to overcome a long-held reputation for aggressive sales strategies.

33:24Major pitches for OCI include cheaper pricing. Thanks to a simplified set of products and good integration with other clouds, Oracle salespeople often take jabs at AWS's sprawling set of tools, telling customers that paying for Amazon's cloud means subsidizing expensive moonshots like a satellite project. It's Coupier. Coupier. Coupier. Interesting. Yeah. I mean, there are a lot of tools in AWS. You don't necessarily need all of them. Some of them probably higher margin than others. It's just funny to hear the internal strategies of the Oracle salespeople putting the screws You really wanna subsidize this satellite project with your money?

34:05Yeah. Do you even care about this? Yeah. Have you even thought about it once or something? On the - Did you know you were paying for it? On the battle card. It's just like, do you care about satellites? Yeah. Okay. So during the pandemic, they landed Zoom, who needed the capacity. Interesting. They also got a$2 billion commitment from Uber in 2023. And then during that era as well, they got a TikTok course owned by ByteDance. And that was the whole narrative of like, if TikTok sells, it'll be the Oracle, and Oracle will just host all the data, and TikTok will be kind of its own company, and there's no antitrust considerations because Oracle doesn't have any consumer-facing social media products.

34:45It always seemed like a weird fit, but like a decent fit. Like not huge synergy, but certainly solves all the security concerns because Oracle's like the most American forever company. It's like just built American data centers. So they landed TikTok. That account quickly surpassed a billion dollars in run rate. And then it eclipsed the rest of the cloud business. I remember seeing the cloud bills from, what was it? Snapchat, I think, was paying maybe Google Cloud like a billion dollars a year. And that's on like, what, 15 billion of market cap or something. So like a pretty significant investment.

35:24I mean, storing videos is a lot, a lot heavier than storing text. And so when they started working with TikTok, they actually, that's when they started getting familiar with building out these more AI-focused data centers. That makes sense, because they need to do the inference for the recommendation algorithm. That's when they started buying a bunch of GPUs. Yeah, that makes sense. Well, they should also get on graphite.dev, code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster. You can get started for free at graphite.dev. internal accolades have flown have flowed to mcgork we're probably just butchering this butchering that name or where are you the 39 year old has rapidly climbed company ranks and reports directly to ellison let's go one of oci's first employees after a stint at amazon and recently relocated to nashville where oracle has pledged to move its headquarters i didn't know that it's nashville interesting june he was promoted to president and is seen by company leaders as a potential successor to his 81-year-old boss.

36:22Do you think he went there because John Fio's out there? Probably. Nashville? Yeah. Isn't, who else is out there in Nashville? Didn't Ben, who's the guy from the Daily Wire? I forget. Anyway. Nicole Wischoff was there. She just moved back to San Francisco, though. But yeah, bit of an odd choice, but you know. Yeah. No sailing in Nashville, right? So McGork says, sometimes you look back and say, I'm not sure that I was entirely qualified at each step of the way. That's the best, though. Seems to have worked out so far. That's great. Good for him. Those who have worked with him describe him as highly effective and conflict prone.

37:01He once told an executive his actions were effing stupid in a room full of leadership. Talk about managing up, according to allegations in a 2021 lawsuit. Blunt bosses are typical in a company run by Ellison, who once described his leadership style as management by ridicule. We've been going through all the different Larry Ellison books in the group chat, talking about what's the difference between God and Larry Ellison. Like, God doesn't think he's Larry Ellison or something like that, right? Isn't that one of the quotes? Legendary, legendary. We're having David Sender on the show soon, and we'll have to get him to do a deep dive on Larry Ellison for us, since there's so much more to the character Larry Ellison in the history.

37:45there's that famous book where he gave a writer, you know the story, right? Larry Ellison gave like biography rights to an author and said, you can write whatever you want, but I get a rebuttal in the footnotes. So the book is the writer like documenting, like shadowing Larry and following him around and the internal workings of Oracle. And then there'll just be a footnote and Larry Ellison will be like, I wasn't being too aggressive in that meeting. They were being too aggressive. Like this is wrong. It's like a debate. It's amazing. Anyway. Jeremy shared something in our group chat earlier. He said, with his top guys, this is about Ellison, from the difference between God and Larry Ellison.

38:26With his top guys running the day-to-day business, he was free to do the things he enjoyed. Brainstorm with his engineers. Yep. Reread Pose, Annabelle Lee, visit Japan, entertain women, buy things, give interviews, and dream. Ah, the dream. That's amazing. Yeah, it was Ben Shapiro, Stefan Silesi says. And Nate says, move to Nashville for Volta Cafe. That, of course, is John Fio's latest caffeine cafe. We'll eventually do a live show from the Volta Cafe. We need to. The aesthetics are fantastic. You should go check it out. We should do that show on 1 ,000 milligrams of caffeine. I'm down. The ultimate challenge.

39:06Pump me full of it. Straight into the veins. OCI's importance has grown as the cloud has become a bigger part of the company's Business today, about 23 ,000 employees report up through McGork. Over the last two years, more than 600 workers have joined from Amazon, according to analysis run by Workforce AI. Amazon's recent five days a week return to office has made poaching easier as OCI remains largely a hybrid or remote workplace. Interesting. Today, the company's biggest driver of OCI, AI is the biggest driver of the growth. Majority of Oracle's backlog or booked deals are tied to customers training or deploying AI models with GPU-based servers.

39:43We've talked to some founders who have said OCI is deeply underrated. Everyone talks about Azure, GCP, AWS, but OCI, there's some good deals there. OpenAI is slated to become Oracle's largest customer and has marketed the work as Stargate, an effort announced in January at the White House. The companies have struck deals for more than five gigawatts of computing power. That's a lot of power. An unprecedented sum and enough energy for millions of American homes. Work for OpenAI is happening on an aggressive timeline. The data centers are targeted for completion by early 2027, with many of the servers running as soon as next summer.

40:20According to people familiar with the plans, facilities are being designed for training AI models with the assumption that they could also power the applications. So they will do training and inference. They need to be designed for hybrid workloads because they don't think that they will be necessarily training on an ongoing basis forever. and that at a certain point the models get good enough and then you just gotta get them diffused into the economy. You gotta get GDP growing and GDP's not gonna grow on training alone. Providing infrastructure for AI. Training is generally thought to be a lower margin business than inference because it requires more advanced semiconductors and cooling.

40:52Exact timing and design details for the open AI work could still change. Finding data center developers and power providers to support this project to palatable prices has been a challenge. costs for many goods and services have gone up owing to tariffs and vendors capitalizing on intense demands if you have some power transformers or tents for Alisa is a buyer the price of tents is going up though yeah you're gonna be gouging them and so Oracle cash flow has gone negative on the cloud build out but just barely and they and they and they assume not like you're turning some other folks. The site Oracle plans to power entirely with gas generation is located in Shackleford County near another Oracle server farm in Abilene, so they're expanding outside of Abilene in Texas.

41:41It's being developed by Digital Bridge Group's Vantage Data Centers and will have a compute capacity of 1.4 gigawatts, making it among the largest known sites. We've seen a couple groups pop up with one gig clusters. This is 40 % larger, so significance. But semi-analysis often, there's different ways to measure power. Like it's, and people get kind of funky with it. It's like, are you pulling 1.4 from the grid? Is it 1.4 to the chips? And there's like the concept of like compute capacity versus IT power, all these things kind of, you lose a little bit at every single stage. So we'll see where this cluster actually nets out.

42:25And we will let Dylan Patel be the final arbiter of how insane this particular buildout is. Part of OCI's popularity with AI firms is its longstanding focus on bare metal servers. This has become the standard arrangement for AI work. And Oracle was the... Missed cloud, but made a key decision. Going bare metal. Early on to go bare metal. Instead of platform as a service. It ended up working out. Yeah, yeah. For old Larry. Users and employees will also point to Oracle's networking quality and willingness to do high levels of customized work. Semi-analysis, shout out semi-analysis in Bloomberg today.

43:02Popular industry analyst listed Oracle Cloud near the top of its rankings for AI infrastructure vendors in March, citing its cost effectiveness, networking quality, and strong customer service. This would add a category for how aggressive the sales team is. Sure. Battle card ranking. How much trash they talk about the competitors. Another of Oracle's largest customers is NVIDIA, which uses OCI for internal development and to power its own cloud infrastructure service. The chip maker, codenamed Pathfinder, internally rents capacity from an Oracle cluster of H100 chips in Japan and from a data center being built on Indonesia's Batam Island.

43:42That's crazy. I'm getting into islands over at Oracle. Oracle is also in talks with meta platforms and XAI for further capacity. Of course, Larry Ellison and Elon Musk are famous for getting dinner with Jensen Wong and Benny. So Batam Island is right off the coast of Singapore. Is it beautiful? I'm sure it is. Or is it more like industrial? Should we go? Next vacation? It's very industrial. Okay. But I'm sure it's still beautiful. Book a wander. Find your happy place. Find your happy place. Book a wander with inspiring views, hotel-grade amenities, dreamy beds, It's top tier cleaning and 24 seven concierge service.

44:17It's a vacation home, but better folks. I don't need to read that. We need to get them to put together the Wander data center collection. Yes, yes, I want to wander in Abilene. I want to wander in Batam Island. I want to wander in Japan, wherever that data center is. I want to go on the world tour. Take me to Virginia. Put me next to AWS East. I want to visit US East. Oracle's cloud business is still orders of magnitude smaller than Amazon, Microsoft, and the number three provider, Alphabet Inc.'s Google. But plenty of room for growth exists. Orders of magnitude, it's 100 times smaller. That's shocking to me.

44:54I feel like it would be at least within like 100x. I don't know. The market generated almost 100 billion in revenue in the most recent quarter and expanding 25 % year over year. The question is how Oracle will fund this development and what it does to the company's profitability. Still, the company hasn't had to borrow money to develop its centers like smaller rivals such as CoreWeave, who we've interviewed founder of. We believe margin will rebound and cash flow will be substantial once they get through this investment phase, said an analyst I think, an analyst at Bernstein. Oracle is going through a business model transformation that is even more impactful over a shorter period of time.

45:35Yeah it's hard to imagine that these data centers, I mean we could be winding up in like the dark fiber world where a lot of these data centers don't get used if like truly there's a major popping of the ai bubble but uh this stuff seems to be uh extremely valuable for a long period of time and it's hard to imagine a bunch of huge data centers like not being useful to america in a decade and so um he's risk on he's risk on but not using leverage yeah not using leverage which it's all coming out of the free cash flow. They have the cash to pay for it. And it seems like he's in a good position. Anyway, you should get on Julius.

46:16What analysis do you want to run? Chat with your data and get expert level insights in seconds. You can ask Julius to analyze your data. Business better. They're loved by two million users and trusted by individuals at Princeton, Boston Consulting Group, and Zapier. And TBPN. And TBPN. There is a new leak of a memo from Alexander Wang. Kylie Robison is reporting that Alexander Wang wrote that superintelligence is coming. In order to, quote unquote, take it seriously, the company is making major changes. So we can get into this a little bit. Yeah. Do we want to have Tyler break down the whiteboard?

46:54We should. Or show us what is the structure? What can you tell us about the current structure of Meta's AI efforts, Tyler? Yeah, so I think that the main thing is that we now know that Alexander Wang is like top dog of AI at Meta. So we have, I don't know if we can show it yet, but basically you see Zuck at top. And then Alexander Wang is - Reports directly to Zuck. Reports directly to Zuck. And then everyone who works on AI at Meta, not just MSL, like FAIR, every single org that does AI, reports directly to, well, not everyone, but they all report to Alexander Wang. So then below him, we see three more people.

47:32We see Jan LeCun. Okay. And is Jan LeCun at fair? Is that his? I think that is still unclear. We're not super sure. So he was a scientist. But he was not on the llama team? No, I don't think so. Okay. It's also, like, maybe he was on the llama team, but then for Llama 4, he wasn't or something like that. But then so in charge of MSL is Shang-Gia Zhao. He was at OpenAI, I'm like 90 % sure. And then we see Nat Friedman is also there. And then so below them is just, these are like the actual researchers who've been poached from various labs. So you see like Jason Wei, Lucas Fire from OpenAI, I think.

48:17I think Semi-Analysis had a funny joke post about this, that they wound up taking down. There was something to the effect of, like you worked at a different AI lab focused on training. They decided that they sold and they said like, this new organization won't be doing training. And then so you go over to Meta and you work under Yan Le Koon and then they're like, you're not gonna be doing training either. And so it seems like there's a real fight if you're an AI researcher to actually get onto a team that's working on training super intelligence, because there's a huge amount of demand for just like let's productize this now and there's probably a divide among the researchers on where they want to play in the stack do they want to be focused on on implementation applications leveraging the foundation models versus training and coming up with the really novel so internally they have TBD labs which will also be in charge of exploring new directions such as building an Omni model.

49:19Okay, and TBD Labs is a subset of MSL or a side MSL? What's the take there? It feeds into it. It feeds into MSL, okay, got it. And then FAIR will be a quote unquote innovation engine for MSL's training runs by feeding its research directly to TBD Lab. Okay, got it. And that's a, according to Kylie, that's a more active role for FAIR than before. Yep. They're known for publishing their research, giving it to employees. And so how would I comp this to OpenAI? It seems like Shenja Zhao is sort of the Mark Chen of MSL? Yeah, I think that's probably the best way to do it. Head of research. And then Zhao has four, at least four, like really key high-level researchers.

50:12Yeah, these are mainly just like the big names. Yeah, the big names. Obviously, there's like way more. But yeah, I mean, even at OpenAI, it's like still, it seems like there's like a lot of teams there that's like, oh, I'm like, I led O1 research, but then there's someone else who like led reasoning. Yep. So it's like unclear how those overlap. Yep, yep, yep. So we don't really know like what the exact comp would be to OpenAI or to Ananthropic or something like that. Sure. But so far, the actual like MSL core team seems fairly small. at least of like public people that we've seen. I remember that there was some spreadsheet that leaked maybe like a month ago that had only like maybe 25 names on it.

50:50And they were all like kind of superstars. And were they mostly on OpenAI on the various training runs and various research projects or was this across the entire industry? No, it was not just OpenAI. There was a bunch of people. Was this like the list that was leaked? No, that was more recruiting focused. This was just like the actual team. Okay, got it. There's more info in the memo about the GOAT, Nat Friedman. Nat Friedman. What's he doing? Will lead MSL's efforts to integrate AI into Meta's products. So as everyone knows, Meta's been trying to make things like AI glasses and the Quest VR headset go mainstream.

51:28These products have garnered good reviews, but have yet to make up a meaningful portion of revenue. And Nat Friedman, it really feels like he might be the best person in the world to do that because who was the first person to actually make money off of GPT-3? It was Nat Friedman with GitHub Copilot, right? And so if you think about like how can you actually deliver value that people pay for at a good margin, which in the meta context it might be it's value that's delivered for free and then monetized via advertising, but in general you need to develop products that pass the hurdle rate of like value capture and not just go around doing science projects, which is what a lot of the research labs are set up to do, and that's fine.

52:11But it feels like Meta has some real horsepower on the training side. They're going to be able to build great models, and they're going to need to build those in and vend those into a variety of products, everything from Meta Ray-Bans to Facebook, the Blue app, and Instagram, and WhatsApp. And how can you do that in a way that is product-focused? Well, you hire the person that was the first real like AI application layer builder like with GitHub Copilot. Demigod. Yeah. I mean, I think one thing that's notable. So when Meta internally announced the creation of MSL in June, it said that Friedman would be its co-lead.

52:47However, in the email, Wang made it clear that Friedman reports to him, quote, Nat will continue to lead this work reporting to me. I don't think that that's as dramatic as you think it is. Freeman's just like, yeah, I've been CEO, pretty good at it. I'd rather just focus on making AI cool and useful. It's such a huge organization. There's so much to do. I think that you could read that as he's reporting to Alex. I think he's close with Alex. I think he's worked with him in the past. And I think there's a bunch of interesting ways to actually go and have a better experience within the behemoth that is meta, focused on just implementation of frontier models as opposed to actually having to have the AI researchers roll up to you.

53:47and then you're a middle manager in between Mark and you're kind of sharp elbows with Alex on who's doing what, where the stuff is allocated. It makes sense to just have a rockstar training team and then take the stuff that they trained, go implement it. And of course, there'll be some feedback there, but it won't be. But if you are an AI implementer, an application layer developer, you don't necessarily need to have the researchers report to you. Yeah, true. Maybe there's like an equivalence to, there's like Google DeepMind and there's Google AI Studio, which is like the productized version of the models.

54:24So maybe it's something like Nat is in charge of the Google AI Studio for where, you know, Shang-Gi-I is the DeepMind. Yeah, I don't know. We'll have to dig into how it's sorted. Regardless, where's Daniel Gross? Yeah, where's Daniel Gross? We got to figure out how he fits in. Great question. I don't know. I mean, was his transfer, his trade deal, ever fully confirmed? I don't think he's put out a post about it yet. But it was confirmed that he left SSI. OK. Oh, yeah, yeah. Ilya posted about that and said, like, we're wishing him the best. So on Daniel Gross' website, it just says, working on AI products at Meta.

54:58OK. So it sounds like confirmed. OK, confirmed. He's probably still working with it. But who knows? I would imagine he's working alongside Nat, right? They merged. They merged. The merge. The merge happened. And it's Nat and Daniel hanging out and going around stuffing AI in every meta product. We're putting it everywhere. Putting it everywhere. I mean, it seems like a good strategy compared to, like, it's got to be wrangling that problem at Google has got to be somewhat even more difficult. Because there's probably a Sheets team. There's a Gmail team. There's a Google search team. It's harder at Google because the core search business is threatened by Gen.AI.

55:36Sure. And because you can just go on ChatGPT and have it browse Google for you. Apparently, yeah. Yeah, whereas if you're meta. It's more existential, right? Meta, you're like, this is cool. This is incremental. We only want to do things that drive higher ARPU or higher user minutes. How do we keep people in the apps for longer? And it should be very additive on the creator side, on the consumer side, but you should, like a lot of the usage patterns, I mean, when they rolled out stories, it was very clear that they, I don't know if they actually wound up rolling out stories in WhatsApp, but they definitely rolled it out both in Instagram and in Facebook.

56:15And so there's a lot of AI features that you could imagine, even just like those meta AI studio design a character that you chat with. There's no reason that like chatting with, although it's ridiculous to say out loud, like the Russian girl or whatever the killer app is, I mean, like the bull case is that it's chat with like a helpful assistant or chat with somebody who's an expert in watches or semiconductors or cars or whatever. Like you follow a lot of car content on Instagram. You can now chat with like the meta AI car expert, ask it questions maybe. It makes sense that that would be vended into WhatsApp and it would also be vended into Instagram and it would also be vended into Facebook.

56:56Speaking of watches, have you seen the levels of light? This was in a certain really important post. Yeah, this was - Tips for men. It was almost the current thing in Silicon Valley. Everyone was talking about this. Yeah, this was flying around group chats. For sure. And the reason is, you know, tips for men was giving their sort of definition here. But to me, this just screams RM at the top, LP, Patek, clearly the GP. Yes. AP, clearly Principal. Principal. Rolex, obviously Associate. Associate, okay. And then Cartier, I would say investor relations. Okay, yeah, yeah, yeah. And then I don't even know.

57:33Yeah, you're getting to IWC range. I don't know, EIR. Yeah, seriously. EIR that had a moderate, you know, an aqua hire kind of looking for their next thing. Exactly, exactly. They define the levels of watches. Richard Mill, who of course we've talked about on the show here as filthy rich. Patek is old money king, AP is new money, Rolex is wannabe rich. Depends on the Rolex. You get a Paul Newman, I think you're an old money king. You're picking up a land dweller from the secondary market. Wait, are they spiking? Or where are they sitting? Yeah, I mean they're sitting well above retail. Bezel has one for 70k.

58:2070k for a land dweller? It is a beautiful watch, I like it a lot, but that is a wild price for a Rolex. Well, yeah, I mean, make a statement. It's a nice bracelet. Yeah, it's a good bracelet. It's a good bracelet, sir. If you were running a luxury watch brand and you wanted to get your watch brand mentioned in ChatGPT, what would you do? I'd go to tryprofound.com. Tryprofound.com. So you could get your brand mentioned in ChatGPT. You could reach millions of consumers who are using AI to discover new products and brands. You could even get a demo. I would not be surprised if at least one member of the Holy Trinity already uses Profound.

58:56I wouldn't be surprised either. I'd be surprised. Fantastic. Yeah, I got an investor update from James at Profound last week, and it was the single most impressive monthly update that I've received across investing in 60 or 70 different startups at this point. Very, very impressive. Shout out to them. What is Raboi talking about here? Keith Reboy has taken yet another victory lap on calling Miami as an underrated place to live. It was, of course, not in the conversation at all, and then it became the current thing. Dellian famously posted, can we just move Silicon Valley to Miami? He somewhat successfully did that, and a lot of companies moved there.

59:41Of course, it didn't actually become like a major tech hub, but it's certainly more of a destination for the tech elite than it's ever been in the past. It's a beautiful place, I love visiting it. Great for conferences, great. My favorite thing about Miami is that if you host a tech conference there, no one will go home after the conference ends, because everyone will get a hotel room because almost no one lives there. And yet it's a place where people can justify going to a flight. It's as convenient as Las Vegas, as nice weather, lots of activities like Las Vegas. But if you throw a big event in San Francisco, half the audience will be like, oh, well, I'm not, maybe I got a babysitter, but I'm gonna go home.

1:00:21And if you throw out an event in New York, same thing. But if you get everyone to Miami, 90 % of those people are gonna be from out of town. They're gonna book hotels, they're gonna be there, they're gonna be staying up later, gonna be having more fun, more networking, more business connections, more value creation. So still bullish on Miami as a place to visit and as Tex Vegas. Less bullish on it as like, go try and build a foundation model company and recruit a bunch of AI engineers. that's going to be a tough lift but the the the office visits in July 2025 change from the first month of 2019 is way up Denver is down 40 % LA is down 36 % we're trying to shift that trend by being fully in office Alec McGillis who's Keith who Keith is quoting says really striking to what degree New York has become an outlier in recovering from the remote work shift, especially given that it is a city that was first hit hardest by the pandemic.

1:01:22Huge part of this is that people's apartments and or homes that are actually in New York City are so much smaller. So people often want to go to the office. That's a good point. Whereas if you're in Denver or even LA or Chicago or Boston, you might live slightly outside the city. You might live in a suburb. Denver's over here. They're They're still on a hike. They're still on a hike. They're still hiking. Yeah, so nationwide office visits are down over 20 % since pre-pandemic times. And so July of 2019 is kind of where the benchmark starts before the pandemic hit. And obviously that changed a lot of businesses.

1:02:01It changed a lot of work from home behavior. We were just talking about Oracle. You know, they're building data centers, but they're doing it remotely. whereas AWS has kind of gone back into the office. Seattle notably missing from this list entirely. But anyway. This post from Tane is wild. So he's starting out annualized ad revenue by platform. Google search at$215 billion. We need to do some pop quizzes here. Tyler, how much money do you think in revenue Bing brings in every year? $5 billion. $14 billion. All right, that's not too far. That's right. You're off my 3X. If John had asked you about Snap, you would have been pretty close.

1:02:45They did$5.5 billion. Pinterest is at$4 billion. Uber at$1.5 billion. This is ad revenue, not platform revenue. DoorDash is making a billion on ad revenue. Instacart's making a billion on ad revenue. But Meta is up at$200 billion or$186 billion. Google Search is at$215 billion. And so, yeah, I mean, like the open AI, we were talking about Brad Gerstner kind of saying that, you know, the open AI valuation of$500 billion is like way ahead of Google's metrics based on where Google was when they were valued at$500 billion. But when you look at that pool of high margin ad revenue,$215 billion from Google,$186 from Meta,$63 from Amazon, and you ask yourself the question of, will ChatGPT be able to grab a piece of that?

1:03:39Is it already? It's not really already. It's mostly just people paying for a great tool that they prefer over other options. But if they seriously get in that game with a gentic commerce, a gentic checkout, and taking commissions and cuts from different revenue streams in the ad revenue world, there is just so much money to be made in internet ads. You can be a company that's not known for advertising, like Uber, and still be making$1.5 billion in ad revenue. That's extremely high margin. And so can OpenAI suck 50 billion out of this 500 billion pool of ad revenue and trade it a 10x on revenue?

1:04:26Like, is it possible? It's like, it's not that crazy. I don't know. Well, getting better and better at serving the models cheaper and cheaper. Yes, yes. Well, as they build out their ad product, they will need to be using Linear over at OpenAI. Linear is a purpose-built tool for planning and building products. meet the system for modern software development streamline issues projects and product roadmaps if you make software and you're not on linear yeah you're crazy and if we flip over to the bear take on uh open ai bone gpt says gpt5 single-handedly destroyed the fast takeoff narrative and killed the stock market um i don't know if that's entirely fair killed the stock market it sells off two percent and it's like gpt5 took it back this is this is the beauty of like the kangaroo market the faster information is that like every time it's so over we could just be a day away from being so bad it's it's not just that it's that is that you you you go through the the the trough of disillusionment so so fast like you can have like all of the think pieces all of the memes all of the hot takes about like it's so bad it's so bad and then two weeks later you can be like okay we've clearly hit the bottom we beat this to death everyone's talked about how how we're completely over our skis and like the bubble has popped.

1:05:42They can be like, okay, now it's time to start reinvesting. We got to start buying the dim. And then we start building back up so quickly. And that's what happened, I mean, during COVID, like it was like a three month drawdown. Even the interest rate post ZERP era was like pretty quickly turned around, maybe a year of hard times. It was not the dot-com bubble, which took years, years, years. And Tyler Cosgrove has the, This is the doomsday clock for us. So the doomsday clock, typically the folks who are worried about the apocalypse are tracking the doomsday clock, the amount of time until doomsday.

1:06:22We have been tracking, we say it's 90 seconds, not to AGI, it's 90 seconds to nothing. Tyler, can you give us some more context on what inspired this and how you're feeling? Okay, so this is apparently, this is like the Doomer take on GPT-5, right? Yeah. Nothing ever happens. Nothing ever happens. No fast takeoff. Okay. I disagree. Okay. I think GBD5 was pretty good. Yeah. But how many seconds until AGI and the Singularity? How many days? I mean, Sam Altman put it out a couple thousand days. 3 ,000 days is a decade, right? I would put it closer to Kurzweil timelines. Remember the Singularity, the merge?

1:06:59His definition for that was just people become really close friends with a chat bot. Yeah, yeah, yeah. that's yeah that's that's uh that's pretty clear uh i mean i guess uh like if there's gonna be a fast takeoff there's gonna be some sort of singularity something that's like unpredictable i would peg it more at 7 000 days i would put it at 20 45. okay maybe i'm like halfway in between you and sam and sam so sam's 20 35 20 20 40. okay 20 40. okay so that's a lot of seconds That's a fair amount of seconds. Bucco Capital, which is basically the fourth technology brother, he says, this is a drawdown for ants.

1:07:37Momentum names with zero valuation support, still up 100 % to 700 % year-to-date, several with no revenue, many still getting gunned from the lows back to roughly break even. I don't know when the music stops, but the palantards and the Lemon Apes are still dancing. Lemon Apes? What are Lemon Apes? I think Lemonade. Lemonade is meme stock now? I mean, it's up a couple hundred percent year to date. Well, there's another story that is developing, but we don't have a full analysis of it. So DeepSeek version 3.1 is out. The model was released on Hugging Face. Is that correct, Tyler? Yeah. So I think there's basically, there's no model card.

1:08:22Okay. So we don't know what the benchmarks are really. Why would that be? why would they release the model without releasing the model card? I feel like that's kind of standard issue. Yeah, I don't know. I mean, it could just be like, oh, we forgot to add it to the file. I don't know. I mean, I assume that they're going to add it within a couple days. Okay. Well, people can do it independently too. Yeah, you can run the benchmarks independently, I guess. But yeah, so far it seems it's kind of up in the air how good it looks. Some people were like, oh, China's cooked. Like, you know, the chips like totally screwed with.

1:08:55Yeah, yeah. They tried to train on Huawei. They failed. They couldn't get NVIDIA. People were saying, like, oh, it was supposed to be DeepSync v4. That's what we were hearing, for sure. Yeah, and then it turned into 3.1. It was so bad. So it could be sort of like a 4.5 moment. Yeah, it's possible that they're going down. They're mimicking the error in the tech tree that was followed at OpenAI and Meta, it seemed like. Yeah. Because GPT 4.5 kind of came out and was underwhelming. and then Llama 4 Behemoth couldn't get out and we might be seeing something similar. I mean, so those were both examples of like...

1:09:32Big, big, big. Yeah, that was like big pre-training runs, right? But DeepSeek already has like good RL, right? Because they have R1. That was like, I mean, that was the first reasoning model to come out after OpenAI. So I don't know if it was necessarily that. I mean, it seems reasonable that it's just like a chip thing where they were forced to use Huawei. But I think we'll see. Hopefully over this week, they'll release some more details about it. Yep. Well, we will be tracking which company has the best AI model at the end of August on Polymarket, of course. Google's running away with it. 93 % chance.

1:10:10I don't know if they're already in the lead and that's why, or maybe we're expecting something new to come out. But DeepMind seems to be on a tear. They've been climbing even towards the end of the year. They're up at 59%. XAI has recently eclipsed OpenAI as the second best AI model at the end of 2025. The other news on DeepSeek, this is still in the rumor mill, but some dude asked the official DeepSeek customer service, and they confirmed that version 3.1 is a merge of reasoners and non-reasoners. Maybe it's some sort of model switcher coming soon. both chat and Reasoner API now route to the same model, but just change if thinking mode is on.

1:10:55So anyway, if you're looking for sales tax AGI, head over to numeralhq.com. Spend less than five minutes per month on sales tax compliance at numeralhq.com. Well, Howard Marks just had an interview with Bloomberg. I think it's relevant. He says, US stocks are in quote, in early days for a bubble, although the critical point for correction is yet to come. He says, I'm certainly not ringing alarm bells. The point is that things are expensive. Palantir is certainly expensive at 500x price to earnings. It's been 16 years since investors last went through a, quote, serious market correction. He added the current period.

1:11:3516 years? Oh, he's going back to 2008? Yeah. Okay. The current period reminds him of the late 1990s when the market was falling in love with tech stocks, leading to former Federal Reserve Chairman Alan Greenspan to warn of irrational exuberance. Still, the market rose for a few more years after that before the tech bubble burst. Yeah, again, like, you know, the debate everybody's been having is are we in January of 2021 or November of 2021? We might just be in 2019, you know? We might be. We just don't know. This is a Ben Thompson stake. Like, it's very hard to call it. Mark says, people get out of the habit of thinking about market corrections, noting that a reversion to the mean is very likely.

1:12:19Some technology stocks are, quote, quite highly valued relative to history. No new insight there. This is a time to put a little more defense into your portfolio and investing in credit as opposed to equities is one way to do it. Yeah. I saw another post that Fidelity was recommending a portfolio of 70 % bonds, which seems very, very conservative. But do with that what you will. Let's switch gears to a different hedge fund billionaire, Jim Simons. His daughter, Elizabeth Simons, has given$250 ,000 to a super PAC supporting New York City mayoral candidate Zoran Mamdani. Of course, this is seen as sort of antithetical to who Jim Simons was and what Renaissance Technologies did as an applied math hedge fund.

1:13:14Gary Tan has a take. He says, it's a special form of insanity for private foundations to supercharge the destruction of teaching math to the next generation when that money was earned literally by the merit of pure applied math. Like the money earned by Renaissance Technologies. Legalize math, Gary says. Jim Simons' mathematical skills helped transform him from a prize-winning academic at Harvard and MIT into a legendary financier whose algorithmic models made Renaissance Technologies one of the most successful hedge funds in history. After his death last year, one of his consequential bequests went to his daughter, Liz, who oversees the Heising Simons Foundation and its near billion dollar endowment.

1:13:56What Liz has chosen to do with that inheritance might have surprised her father. Jim Simons devoted much of his charitable giving to basic research in mathematics and science, but his daughter's foundation is moving in a very different direction. The Heising Simons Foundation and similar organizations are supercharging a movement to remake K-12 mathematics education according to social justice principles. Interesting that both the left and the right are somewhat attacking math right now. Obviously, this is coming from the left, but there's also that attack on USC. Are numbers real? Yeah, UCLA mathematician Terrence Tao had his funding cut by the Trump administration, and so it's a bad time to be a shape rotator, time to pivot into word selling.

1:14:44But we'll see how it shakes out. I haven't been following the New York City mayoral election too closely because I live in Los Angeles.

1:14:59Anyway, let's go to this post by Nikhil Krishnan. He says, dude, Mark Bertolini, the ex-Etna CEO, now Oscar CEO has an insane story. Here he talks about how after his skiing injury, he was in so much pain for 17 years that he went to Switzerland to end his life and was then told he could fix his pain. He went through neuroplasticity exercises and got it fixed. So I was getting ready to go to Switzerland to end my life at Dignitas. It's an assisted suicide organization, but I had to prove I had intractable pain and pain that couldn't be cured. And so I was sent to a doctor and the doctor evaluated me.

1:15:36The doctor said I was gonna do it on February 15th, 2023, was gonna be my death day. I wanted to do the 18th, but it was a Sunday. What was the day I had the accident? So I went to this doctor and he sat down with me and they did a few tests and they looked at me and said, you don't have intractable pain, you have neuroplastic pain and we can fix that. I go, here we go again. I mean, I did seven days at the Hanneman University Hospital ICU, 50 milligrams of ketamine, an hour IV to try and reboot my system. I had a great time and it felt better for the first month afterwards, but then I came back in a rush.

1:16:14And so I said, here we go again. And I had worked with Mary to get her through how that was going to work and what the plans were and all that sort of stuff. And so they go, okay, well, let's start. So part of it was cognitive behavioral therapy and I had to go back to my childhood and my concept. I was not well treated as a child. And so I was consistently under threat assessment. And threat assessment is the worst thing you can have when you have chronic pain because you're always waiting for one cell to have pain. And then all of a sudden it goes everywhere. And they said, and also your 42 pain centers in your brain have wired together from all this chronic pain over the years.

1:16:51So any touch feels like pain. So in essence, I was addicted to pain. And they had done MRIs on baseball hitters. And they said the best baseball hitters, their visual, cognitive, and motor cortices light up instantaneously. There's no latency. That's why they see the ball in slow motion. That's how they see it moving. That's how they hit it so well. You have a baseball hitter's brain on pain. There's no latency in your system. Something touches your arm you're in pain so they helped me understand the system then they put me in transcranial magnetic stimulation and they zap my brain for 30 minutes to disconnect those pain centers and they put me in a virtual reality lab with a helmet and they re-patterned my brain and he got through it and he's doing well so very good to hear what a crazy crazy turnaround anyway uh microsoft excel adds copilot ai to help fill in spreadsheet cells this is like a crazy moment The moment you've all been waiting for.

1:17:43The moment you've all been waiting for. But there's been a lot of folks who have been building Microsoft Excel add-ons. We had some of them on the show. There's one that launched on Monday. We didn't have a chance to have the founder on the show. But there was this, the current Copilot instantiation, similar to the Gemini integration, just kind of sits in the sidebar, does a little bit. But now they're going to have Copilot AI in every single spreadsheet cell, similar to that demo that we saw on Monday. But they warn against using the AI function for numerical calculations or in high stakes scenarios with legal regulatory and compliance implications as Copilot can give incorrect responses.

1:18:25Interesting. Do you think this is a disruptors innovation or innovators dilemma problem where there's a low end solution that is counter positioned against the status quo because Microsoft can't can't add like a unreliable product to something that is thought of as reliable. This was the take that the legal industry could face disruption from AI because right now AI is maybe 50 % as good as a real lawyer. I don't even know. 80%, 10%. It doesn't matter. The point is that it's a thousand times cheaper potentially. People have gotten used to getting a result from excel and feeling like they can trust it sure trust their own yeah calculation yeah you know they don't feel like i need to get a whiteboard out and whiteboard it out and use a calculator and never to fact check it yep but i think you're going to have to do that with co-pilot and various ai tools for some amount of time but at the end of the day you have tools that humans use to do things but then humans are held responsible for whatever whatever process they use in order to get a certain result.

1:19:36So I don't think this is that big of a change. Yeah, I mean fundamentally Excel is a functional programming language essentially. Like it's a bunch of functions. You can't write, it's not the type of language where you can create like a class or object. It's not object-oriented programming but it is functional programming and what that means is that it's entirely deterministic. Like everyone opens up the same Excel sheet. They should see the same numbers. There is no stochastic, non-deterministic, and that is kind of antithetical to the way LLMs work. Now, to be fair, businesses all over the world have implemented LLMs at various pieces of their tech stacks and been fine with the fact that, okay, for this particular function call, we get a sort of fuzzy result.

1:20:25We're trying to tag something for fraud and we're giving it a risk score at the end of the day. And we know that there might be some hallucinations. It might get something wrong. And the tech industry has completely come to terms with that. And they know that these systems are sometimes reliable, sometimes they're not, but they add value. And so the fact that Microsoft's actually rolling this out quickly is pretty bullish for Excel sticking around, in my opinion. And the fact that they are warning, as long as the warning doesn't pop up every single time you call the co-pilot AI and you use an LLM in a cell, I don't think it would be a major headwind, but I don't know.

1:21:04We'll see. I'm not using a lot of Excel these days. But you should be using Fin.ai, the number one AI agent for customer service, number one in performance benchmarks, number one in competitive bake-offs, number one ranking on G2. We need an Irish sound World champion. Yeah, and customer service is a good example of where most of the customer service answers are non-deterministic anyway, because they come from a human. So if you have an AI agent in the loop, you will still need to escalate and fact check and reality check what's going on. But in general, it's the perfect fit for generative AI.

1:21:41Kyle Chayka has a new essay in the New Yorker. He says, what do all these things have in common? I grapple with the burgeoning Gen Z and younger aesthetic of tacky, oozy, ugly, chemically augmented, brain-rotted consumer slop, the LaBubu Dubai Chocolate Matcha Industrial Complex. And I'm looking forward to a time when we don't have to use the word slop anymore. I think it would be fair to at least soon call the top on that. I was really hoping that I just would never need to learn about LeBubu at all. And just like, it would just come and go and I would just be forever ignorant of it, much like you're ignorant of every movie ever created.

1:22:26It's a beautiful experience. It leaves you with a sense of purity that I long for. And yet today I find myself learning about LeBubu. Tyler, do you have a take on LeBubu? No, but I have a fun fact. I'm pretty sure this is true. This better be legitimately fun. because he really hyped it up. So I think the first use of the term brain rot is Henry David Thoreau in Walden. No way. So like 1850s. Wow. Isn't Walden like literally about touching grass, the whole point? Yeah. You go to the pond. Yeah, he lives at the pond for two years. Yeah, he goes and he touches grass. He doesn't talk to anyone and he just like walks around.

1:23:03And he says the word brain rot or rotted brain or something? Yeah, I think brain rot. Wow. Okay. Yeah, well. Is that a fun fact? That's a very fun fact. I thought that was fun. I would give that 10 funds up. On a recent quiet weekday morning in Manhattan, I attempted to obtain a LeBubu, the cutesy monster doll that has become the biggest international toy fad since Beanie Babies in the late 90s. This is in the New Yorker. Collected by children as well as by adults, the dolls were created in 2015, I didn't realize it was 10 years old, by the Hong Kong-born designer Ka Sing Lung when they were brought to the mass market by the Chinese toy company Pop Mart in 2019.

1:23:42The K-pop star Lisa started sharing her collection on social media in 2024, sparking the recent explosion. I didn't realize that that's where this all came from, the K-pop star. Labubus have pointed teeth, rabbit-like ears, and wide glaring eyes. Their faces and extremities are bare. They're demonic. Lots of people have been saying that. Tie-dye in the chat says, Tim Dillon said Labubus are just Chinese voodoo dolls. Yes. I think it tracks. Yes. I like that post. someone was saying i want to pronounce it like it's a greek hero labo bus bill bishop in the sub stack chat says my dog tashi wants a labooboo chew toy that could be a good let's get we should let's let's buy tashi all of them that are available on the market and have and have a little little tashi destroy all of them i loved i love tashi because tashi is like a full-on character uh he's sort of like the tyler of sharp china and tashi is a recurring character on the sharp china podcast that uh you should obviously go listen to uh and uh bill bill we're gonna we're gonna we're buying a a lububu for tashi yes nick can you get figure it yeah get tashi's address we'll send a lububu yeah don't just get an entry level one tashi deserves get the rarest one possible the one of one the most hyped take it out of the clutches of the most k-pop addicted brain rotted teen who's ready to spend$10 ,000 of their parents' money and give it to the dog.

1:25:12Their faces and extremities are bare. The rest of their bulbous bodies are covered in a layer. This is exactly who I want to be covering this story is the New Yorker. The creatures are usually found hanging from phones or handbags, tiny demons haunting our accessories and haunting our lives. See? Demonic. You said it, but the New Yorker said it as well. Yeah, they come in a menagerie of more than 300 collectible forms. They can be matched to a personal style or mood. Pink fur for when you're feeling flirty. Blowing red eyes for a bit of an edge. Dress be belate. Scarved for a cozy vibe. The high-end art fair, Art Basel, released a limited edition art handler, La Boo Boo, a batch of which sold out in 20 minutes.

1:25:56I think that Tashi needs the art handler. Yeah, for sure. As a result of their popularity, labubus are difficult to get a hold of, often out of stock, both online and in physical stores. Their scarcity made me want one even more. Perhaps those who own them know something I didn't. So I turned to Partia, a Union Square boba tea cafe with neon-lit claw machines that are perennially well-stocked with labubu boxes. I plan to win my own tiny monster by skill. A friend and I picked up refreshments at the counter, sparkling grape yogurt and an iced green tea with vanilla mousse and traded$20 for 20 party tokens, which would allow us to a scant four attempts at grabbing one of the laboos stacked and grinning inside their glass cages.

1:26:48This is the Skinner box. This is fun. I love I love games. I would probably fall for this. though the cafe had just opened you know the last time I I thought here we need to make a a a a golden retriever that can ward off yes demonic labubu yes you know the you know the antithesis of the of the anti-christ like labubu it is the Santa Monica Pier Mickey Mouse or or like you know I forget exactly what I won, but I was with my son on the Santa Monica Pier, and there is a challenge where you have to hold onto a bar, like a pull-up bar, and then it pulls you up and you have to dangle. And the longer you dangle, the bigger the prize you get.

1:27:38And I held on for my life because I was fighting for the love of a four-year-old. And I did quite well. There are a number of tricks. So for the first minute, it doesn't rotate, and then it releases, and then it can rotate freely. and so that makes it a lot harder. I of course was not carrying my weightlifting chalk like I should have been and so my hands were a bit sweaty and I underperformed but I drew a large crowd. People were cheering. It was a heroic moment, something I'll remember forever and so if you're trying to win a toy, stay away from the demonic claw grabber and instead do a feat of strength for your child.

1:28:18that's what I would recommend. So the author tries to win one of these. He does not. And he says, why did I feel the need to own one? He didn't win? He didn't win. He just decided to cut his losses and go buy one. Why did I feel the need to own one in the first place? It seemed to be an accessory of lifestyle that I was missing out on and already late to adopt one constructed of physical goods that have attained. Yeah, it feels very, you're going to be late to the Lububu game if you're buying it now that we are finally forced. We are definitely late to this one. One constructed of physical goods that have attained intense online fandom status symbols that are both Instagrammable and consumable.

1:28:56Why are they, you know, I don't know that anybody's eating their labubus except for Tashi. And yet this lifestyle seemed demarcated by objects that were ugly, strange, and uncanny. There's Dubai chocolate, a candy bar stuffed with pistachio cream and dough that oozes pea-colored mush. and I don't even have the, we're not going to go into a tinfoil hat. You can look into it if you're interested. The Dubai chocolate story, very interesting. There's a conspiracy theory there that I do think tracks. You don't have a tinfoil hat big enough. Yeah, for that one. But you can do your own research. There's matcha lattes, bright green caffeinated beverages that sometimes come out through with a red strawberry syrup resembling blood splashed on a lawn.

1:29:41What is this? I don't like any of this. Anyway. I think the antidote to La Boo Boo's. I mean, the main thing is we are far from the Everlane era of consumer goods. Sure. Right. These sort of minimalists. It needs to be garish. It needs to be controversial. It needs to be viral. It needs to be unique, gamified. It's built for the timeline. Yes, yes, yes. If you're looking for a physical good, get an 8th Sleep. Get a Pod 5. You don't need to go to a claw machine to pick up an 8th Sleep. You can go to 8thSleep.com. They got a Five-year warranty, 30-night risk-free trial, free returns, free shipping.

1:30:18Does Labubu have a five-year warranty? No. Does Labubu have a 30-night risk-free trial? No. Does Labubu have free returns, free shipping? No. Get an eight sleep. One thing we didn't ask Mateo yesterday when I came on, CEO, was when are they coming out with eight sleeps for dogs? Yes, yes, yes. And also a third zone for the small child who demands sleeping in the middle and wants his own temperature control. Comes in in the middle of the night. That was a request that came in during our segment yesterday from the family. Also, timeline was in turmoil yesterday around Eight Sleep. Delian taking shots at benchmark.

1:30:54I thought about it. I am incredibly bullish on Eight Sleep selling in China. I'll give you my rationale. So, a lot of Americans are saying, oh, China-America conflict. It's the worst possible thing. The invasion of Taiwan would be disastrous. we need to be you know we need to be extremely judicious with the you know our frontier technology that we send to China we need to restrict exports of cutting-edge frontier semiconductors the H20 maybe definitely not the GB200 definitely not the H100 we'll see they're not getting Blackwell but they will be getting the pod 5 and I think it's incredibly bullish for world peace and i'll lay out my rationale so there are studies that show that aggression is deeply tied to lack of sleep and the pod five as you know makes it impossible to wake up on the wrong side of the bed yep so if you get a good night's sleep you are less likely to want to engage in geopolitical conflict and so i think we need to be sending pod fives to the Forbidden City, to Zhongnanhai, to Beijing, to every CCP official, to every PLA official, and making sure they have a good night's sleep.

1:32:08So they wake up refreshed. They wake up feeling good, and they wake up like, I don't want to go to conflict with America. It's fine. Maybe Trump should take - We can work together. Trump should have America take a position - Yes. In - Eight sleep. In eight sleep. Absolutely. And maybe get on the board. Yes. I think it's a strategically important business for the United States. that we could take away at some point if we did get into a conflict yes and we didn't want them to be sleeping well if all of china is sleeping on it cut off the supply and we can remotely turn it off and say hey you're not getting a good night or turn it really hot really hot oh you like it cold now it's hot this is a major geopolitical yeah so right when they're going to take taiwan yes just crank them up yes exactly exactly they're like no no cancel the invasion cancel the invasion The New Yorker says these products feel hallucinatory.

1:32:56They take the social media mandate of self-modification to its logical extreme, embracing easy breezy chemical additives and image augmentation, enhancing the self until it spills from its pores. The aesthetic is as much a defense mechanism against the Internet era as it is a self-aware and joke. We can see what late hyper digital artificially intelligent capitalism is doing to us. Isn't it a laugh? It's very unnatural looking, the cultural critic Dean Keswick told me, describing the dominant visual style of extremely online consumerism. It kind of relishes its own artificiality. Do you know this song?

1:33:37I do not. I do not watch that song. Blake the Man 1000 wrote a song about Dean Keswick. It's a banger. You should go check it out. Anyway, cultural critic. I love Dean Keswick. Any pretense of the pure organic is undercut by the kind of willful alchemy. Everything is an admixture. The skincare brand Rode put peptides in lip gloss. The Los Angeles-based grocery store chain Erwan enhances - Like IGF-1 in lip gloss? The Wolverine peptide? I'd be down for that. What else can we put it in? BPC-157. Oh, BPC-157. That's the one. You just get huge muscular face. Yeah. I like that. Erwan enhances multicolored smoothies with collagen and sea moss.

1:34:19Touchland makes outer mist, scented hand sanitizers.

1:34:26And the drug and beauty brand HERS produces an Ozempic knockoff that comes in a bottle the color of matcha. I did not know that. As if you're injecting yourself with Japanese tea. I can hear the barista now. Would you like a shot of semaglutide with that? these days memes are to be ingested into the body not just viewed on screen that's a good line yeah we should move on from this it's a long article you can go read it in the new yorker and have a have a deep dive into irl brain rot and the trend i want to read this article by nadia asparuhov delian's wife who's been on the show and she is responding to katherine boyle's article the great tech family alliance talking about pronatalism.

1:35:12I thought it was interesting. We didn't get to it when this dropped, but it's in the new Atlantis. And I guess the new Atlantis is doing an essay series of replies to other essays. I thought that was kind of interesting format. Anyway, Nadia kicks it off by saying, Catherine Boyle's speech is a sign of a new song making its way through the tech world. One that trills the importance of investing into our social infrastructure, strengthening our nation, and yes, building families. I agree with all that. It's a hopeful message, but like tech's early forays into politics, which began as a vague mood board of civic aspirations before coalescing into real policy efforts and wins, the pro-family cultural shift still has a long way to go.

1:35:52And so Nadia is laying out a new pathway towards pro-family cultural shift in tech. This is not to say there isn't a real movement happening and as a parent I'm grateful for it says Nadia Asperuhov but I feel the gap between rhetoric and reality as I navigate the tech world as a writer and researcher particularly when visiting the San Francisco Bay Area never ask a pronatalist founder VC how many children they have I mean you can't do that one in this case because I believe Katherine Boyle does have kids but of course but but I I have long I have long joked that yes bro I will definitely read your 5 ,000 word essay on pronatalism but I gotta take my son to soccer right now so I'll get to it later.

1:36:38There were too many instances of that for sure. Dinner conversations still tend to circle around whether children are worth having at all or they treat global fertility rates at arm's length. Many in tech embrace a borderless lifestyle that implies constant mobility hosting events in cities like New York and Austin which doesn't blend well with family obligations. Children are rarely present at social gatherings. I've had the strange experience of bringing my child to events where I'm praised for this act as some sort of pro-family ideological flex, which while flattering misses the much less glamorous truth.

1:37:11With no evening child care available, I didn't have a choice. That's a really, really good point. There's something telling about this disparity. Tech is clearly eager to recast itself as pro-family, but it hasn't yet internalized what is required to support families day to day. I mean, I literally just experienced this where I was invited to an event and I was like, yeah, I'd love to stop by. I wrapped the show at two. I might be able to get there, but I got to be home for a family dinner. That's at six. Like, and the event wound up being really far away. And I was like, there was just no way I can get there an hour and then get back and do anything meaningful there.

1:37:46And that's just the nature of these things. Like the events are tough. Yeah, I had the same thing this morning. We saw a friend that he invited us to an event and i said if i go to that i won't see my fam my kids for like 48 hours yeah yeah it's tough i say good night yeah 8 30. yeah i see them the next day around five if i go to something in the evening i won't see them again until the next day yeah there is a real trade-off between uh like we we have said in america in tack we work really hard we also are saying we take our family life really seriously something has to give it's probably the the the the dinner parties and the happy hours for the most part yeah unless they're the other thing I mean I think I think people in tech you sort of grew up hearing that you know global warming was gonna make the world uninhabitable at some point in in the near future that was the the sentiment if you turned on on the television, that's what you saw.

1:38:47And I just remember hearing, even as a teenager, people talking, oh, I don't know that it'll make sense to have kids, right, if the world is so bad and dark and growing up in coastal California. My take was always like, no, we need way more kids because they're the ones that invent the technology that solves the problem. The solution is always like more ingenuity, more people, in my opinion. I'm not Thomas Malthus pilled. It's crazy to see Malthus coming back. The Malthusian trap. Are you familiar, Tyler? The Thomas Malthus, the Malthusian trap. Wait, I don't know the trap. So the Malthusian trap was that Thomas Malthus was an economist 100, 200 years ago or something, and he looked at the growth rate of farming and the growth rate of the population and was projecting exponential growth in population while linear gains in farming.

1:39:42And his conclusion was that we were all going to starve to death because we would never make enough food. We would never create exponential growth in food production to match the exponential growth rate in the populace. Of course, we did just that. And the number of farmers went from like 90 % of the population to like 2 % of the population. And although obviously food prices are important and starvation is still an issue, In general, humanity has survived the lack of food. And so it's called the Malthusian trap. It's Malthusian ideology to be afraid of a linear trend not being able to ever go exponential.

1:40:21And so if you think about like carbon recapture or anything that is a mitigant to waste heat and pollution and carbon emissions, You track this out and you say, well, carbon emissions are growing exponentially, the temperature is growing exponentially, and our progress on solving that is growing linearly, but all it takes is actually finding the correct exponential solution, which might just be like exponentially growing nuclear or exponentially growing solar, and then all of a sudden we're having a very different conversation. Anyway, Katherine Boyle as a parent herself, says Nadia Asperuhov, is right to highlight pragmatic changes as the scaffolding that's required to build a pro-family culture, flexible work, improving and reducing the cost of education, raising the status of parents, agree with all that.

1:41:11And like Boyle, I believe that tech has the ability and resources to bring this future to fruition if it wants to. Oddly, however, it is tech's unmatched capabilities that might be what's holding it back. Tech is arguably the best place in the world for unchecked personal ambition to flourish, particularly for young, childish people. You grind and get the$100 million out of your laptop. Not a lot of time to raise kids while you're in that process. But this type of environment can delay or disincentivize family formation. My peers in tech who are reluctant to have children often express fear that it will interrupt the arc of the careers they've worked so hard to build.

1:41:52That I think is the primary tension, not between the family and the state as Boyle argues, but between the individual and collective ambitions. Both the state and the family ask us to make sacrifices for something bigger than ourselves, and this perhaps is why they have historically fought each other for mindshare. What tech offers is the opposite, a chance to realize a vision that is entirely one's own. Tech worships individual talent. Look at the talent war and the AI researchers that are getting the absolute bag right now. And it's a unique thrill to live and work among peers who don't shy away from greatness.

1:42:31I agree. But it also means that tech has to work harder than other industries to demonstrate that starting a family doesn't require giving up these ambitions. That case can't be made through vibes alone, as any parent can attest who has spent a sleepless night with a sick child before a big meeting, zooming with one hand while playing peekaboo with the other and or who has missed a deadline because child care fell through the answer to this more secondary sales more tender offers if you're a venture capitalist you're talking a big game about being pronatalist you need to put a couple million dollars into the common in a tender and make sure that child care is paid for what do you think i like it i like it my my uh my my lead I was at VC at my last company.

1:43:19The second I told him I was having a baby, he said, time for a raise. Time for a raise. And it was very meaningful, important at that time. Yeah. And yeah, I mean, that's the, if you're an entrepreneur and you're building a company and you're having a child and you aren't gonna be able to have access to childcare, that is a - Yeah. That is a - A serious leg weights while you're running. It's putting an extra 45 on the bench press. Exactly. Maybe 90. Unnecessary. Maybe a couple plates on each side. Unnecessarily. Yeah. You need one of those shirts that makes it easier to bench press. Are you familiar with this?

1:43:57We got to get this. Have you seen these? No. Yeah. That's our new merch drop. Oh, that is a great merch drop. So it's a compression shirt that basically, it keeps your arms like this, like tight. And so when you stretch back - It's just taking some of the weight off. You can get an extra like 20 pounds. And so people do it to get comfortable and like, yeah, I can't really do 315, but if I wear the shirt, I can. And then they get comfortable with 315. They do it a bunch and then they can take the shirt off and then they hit 315 and then they can push up. So, yeah, I think that's our next merch shop for sure.

1:44:27For sure. There's some big news coming. This will be for the people in the chat first. Let's continue with Nadia. If tech wants to embody more pro-family values, it needs more skin in the game. It's not enough to tweet about population decline or pass around graphs on it in the group chat. Tech will learn much faster what's required by having kids, raising them, and supporting each other through the ups and downs. Pro-family culture means figuring out how to meet the family's needs for parental leave, affordable child care, quality education, and social environments that are welcoming to children and family lifestyles.

1:44:59Until then, the rhetoric will feel less like a movement and more like a mood. Tech has a track record of moving quickly once it figures out what it truly wants. If the pro-family turn is real, and I hope it is, then what we need next is to move the discourse from talk to action. Only then will we be able to say it with a straight face that tech is building not just companies, but families too. Fantastic article, Nadia. Thank you for posting that. And if you want to invest in the public markets completely unrelatedly, go to public.com. Invest in those that take it seriously. Try to find the translation.

1:45:34Sometimes the anti-transition keeps you guessing. Yeah. Public.com is investing for those who take it seriously. They got multi asset investing, industry leading yields. They're trusted by millions. Oh, you want to transition? Why don't you set up an investment account on public.com for your child? Do your own Gerstner account. Tyler, what you got? It was an anti-transition just like anti-memetics. Oh, there we go. There we go. Yeah. All these other live stream hosts, they think they need to have the perfect transition. and I keep you guessing with the anti-transition. Well, we have some big, big, big breaking news.

1:46:09Stop. Elon has hit the timeline, and Ani from XAI has new outfits. New outfits. I know. So for all the grok-heavy subscribers, get ready to - This is big. A little virtual try-on. Fortnite skins for your companion. Didn't we just say this? You were saying - That that's the way that they'll monetize, is that there will be a number of outfits, and that the outfits will be priced. It will be more expensive outfits. I want a Birkin. Yes. I want a Lububu. I think we just unlocked the real value of the virtual AI companion economy. And all of a sudden, I was previously thinking this is a single digit billion dollar opportunity.

1:46:50Now, the moral stuff is aside. I think me and you are both like, we don't love this. And that's the point that Jira Tickets makes here. He says, sir, we need to get to Mars. And I agree with that. I think the efforts would be better spent getting us to Mars and getting us to AGI and solving physics. And I hope that none of that takes a backseat to this stuff. But you got the new fits, Tyler? Do you have the new outfits already? Are they free or do you have to pay? Are there any upgraded picks? They're free. There's no in-app purchases yet? They're all free. No, no in-app purchases yet, unfortunately.

1:47:23Well, let's hope that one of the upgrade picks is just a donation to the next Mars mission. and maybe put on, instead of being so scantily clad, you could put on a full space suit, and then all that money would go to the next Mars mission, which I'm in favor of. But I do think that, you know, the Alia-Ali-Akta-I-Akta-Yat-Est, I botched that super bad, the die has been cast. The AI companions are on a trajectory to monetize very well, specifically in price discrimination. And so this is the brilliance of OnlyFans. And why OnlyFans is so profitable is that there are both power law creators who generate the vast majority of the revenues, and there are power law customers that spend the most money on the individual creators.

1:48:16This is the value of, this is the economic insight in Fortnite. There are power, many, the median Fortnite player probably spends zero. But the biggest whale on Fortnite is probably spending thousands of dollars. And so that price discrimination, there's going to be someone who is on Grok and talking to Annie and spending thousands of dollars on virtual goods, I would predict. I think that's where this goes. Same thing as the dynamic in mobile gaming. Mobile gaming is the same. Crazy whales that are spending hundreds of thousands of dollars. bringing price discrimination into a category that previously did not have it, where it was one size fits all, all you can eat.

1:49:00Like, you know, Netflix does not have this. Everyone pays the same price for Netflix. But if there's a version of the product that can ramp up so you're making more money, and this is the beauty of just online advertising. Meta probably makes more money from advertising Aquanauts to you. than someone who has lower penchant for luxury watch shopping. Yes. Anyway, Dan Nolan says, a guy on Reddit built something I've wanted for ages, scraping public data and putting together a leaderboard of real estate agents who appear to seriously get prices wrong. And so I guess you could go and find the real estate agent who is underpricing assets and then go and buy a house from them.

1:49:53Maybe this is the solution to the housing crisis. Maybe this is alpha for Opendoor and a bull case for Opendoor. Opendoor, of course, buys and sells real estate and houses. And if there is a market dislocation and an inefficiency in the market, you would expect financial institutions to take advantage. So this real estate's real. Yeah, what you got? Realestats.org. Realestats, that's a good point. Having a leaderboard of just the people that misprice assets. And they do it to both directions? They overprice and they underprice? Yeah, you can - I feel like how responsible is a real estate agent for mispricing a house?

1:50:36I feel like a lot of that's driven by the home buyer. Like the home buyer might say, I know what I got. I want$3 million for my house. and the real estate agent is like, but you made some really weird decisions in the remodel. You have four movie theaters. Well, people also will strategically underprice it at listing in order to create bidding dynamic. Demand and then over. Yeah, yeah, yeah. There aren't that many people in this neighborhood that need car lifts in their garage. You're like, but everyone wants car lifts. Everyone has four GT3 RSs that they need to store in the plastic. everyone has supercars in this neighborhood they're like at 1.5 million yeah people usually spend 50 percent of their net worth on cars right yes yeah they all should get a get a 1500 square foot 1.5 million dollar house in la and spend uh and get four supercars for the next 1.5 exactly and secure it all with an unsecured well we have another massive announcement what's that again i I think everybody listening has been waiting for this.

1:51:44Yes. Really, I'm feeling the AGI right now. I'm feeling the AGI. We've got to update our timelines. Update our timelines, Jordy. Notion has launched offline. Let's go. Let's go. Let's go. Give it up for the Notion team. Woo! I actually think this is cool. I mean, I used to use Notion a lot. And even if you have a small feature update, make it feel big. Yep. but clearly their users are excited about it. Terminally Online Engineer says, Notion is so funny, announcing an offline mode for a to-do app in 2025. This is the right thing to do. Don't get distracted. We saw this with Figma. Figma launched Figma Make, Vibe Coding Platform, super on trend AI stuff.

1:52:25And then simultaneously, like, yeah, we improved responsive design. No, Figma site. Oh yeah, sites for sure, but also just like minor iterations on, you know, the opacity layer and how you can implement parallel. in your design like you you you can't get lost in the sauce you have to stay on the trend and be using the latest tools and and and be on the cusp but also you can't let the culture of of uh you know uh regular iteration uh fall by the wayside uh you can't you can't get distracted by the by the shiny green tennis ball sometimes you have to focus on just uh you know scratching your ear.

1:53:09We got a very low-tanned post here from Mopar. Two likes. Two likes, 61 views. Highlighting the MIT report that we covered yesterday. 95 % of Gen.ai pilots at companies are failing. Mopar says, no, babe, your zero ROI is perfect. Big one. Scare me. Yeah, I think this is a real problem. There's a lot of Gen.ai companies that have sold in pilots to things like healthcare systems that are not these are not going to turn into multi-year contracts yep hopefully a lot of them do some of them do yeah there's still a business there but your ARR is not necessarily real yeah we'll see I wonder the methodology of this report.

1:54:03I wonder what a generative AI pilot means, what failing means. I wouldn't be surprised if a lot of companies are using tools that have implemented generative AI features under the hood and they don't even really realize that their software is just getting better. But it's probably deeper in the stack. It's probably a little more, I don't know, obfuscated and they're just seeing like, oh, the fraud detection algorithm that my ERP uses is a little more reliable. But I didn't like pilot that. It's just like the software I use, the tools got better. The graphics and video games got better. Like there's not a binary moment.

1:54:49People are hoping for the chat GPT moment of generative AI pilots, and it might not be coming. We might be on a smoother adoption curve there. Anyway, what else? We have a post here from Julie Chang, very important. Who wants to own Bill Walton's longtime San Diego estate in the Marston Hills neighborhood, which has two homes sitting on 1.3 acres combined. Both homes sit abutting a trailhead, seven beds, eight baths, listed at$5 million. Absolutely given it away. This thing looks beautiful.

1:55:53What else we got here? like a matcha, where do they sell matcha? Matcha? Matcha? You go to like a coffee shop. Okay, he's drinking matcha. He says - You have a book with you? Oh yeah, like you with Infinite Jest on the subway. What? Like how you were reading Infinite Jest on the subway. That was like a performative thing that you did. Well, okay. And how you were telling everyone about how you're, you know, you're a bit of an intellectual and you read Infinite Jest. Okay, okay, let's move on from this. But yeah, there's like intern Michael, he loves matcha. He's kind of a very performative guy.

1:56:26Cool. Oh, he's performing? You're taking shots? He can't defend himself. He doesn't have a mic. That's not fair. But then, I don't know. Yeah, you have like a tote bag. Yes. It's like a - Yeah, signaling. It's signaling. Yeah, yeah. Yeah, yeah. I'm just unclear on exactly what people are signaling. Yeah, is it more the aesthetics or is it more like the startup swag? Either way, the SF's Performative Male Contest will be happening at Alamo Square Park on August 22nd at 7 p.m. They're putting up posters in San Francisco advertising it. They should get a billboard. They should go to adquick.com. Out-of-home advertising made easy and measurable.

1:57:04Say goodbye to the headaches of out-of-home advertising. Only Adquick combines technology, out-of-home expertise, and data to enable efficient seamless ad buying across the globe. And we have our first guest of the show, Vlad Tenev. How you doing, Vlad? Good to meet you. What's up, guys? Good to meet you. Thanks for having me. Thanks for joining. Can you give us a little bit of a state of the union on Robinhood? You've launched so many different products. How do you see the shape? How do you define, are you a financial institution? I remember when I first used the product probably almost a decade ago now, I guess.

1:57:40It was an app and now it's a company, now it's a business, now it's an institution. What is the shape of the company today and how has it evolved? Well, first off, our mission hasn't changed. Our mission has always been to democratize finance for all. And what that means to us is we really see Robinhood as the defender of free markets. And we're building a financial super app, a platform where all of our customers can custody and hold any financial asset or conduct any financial transaction. And obviously it started with stocks, but it's expanded across multiple domains. We have a credit card, which in my unbiased opinion is the best credit card on the market.

1:58:23Banking, which is a high-end private banking inspired offering, is coming out very soon. We offer, I think, the best retirement account on the market for the self-employed entrepreneur, folks that might not have access to a traditional 401k. So really, we've been on this journey and we've increasingly been expanding the scope and capabilities of what we offer. Does that mean something like when I think about legacy financial institutions, I think about an auto loan, a mortgage. Is that coming down the pipe? I mean, I'm sure you can't comment on specifics, but just like that shape of experience for consumer finance.

1:59:02What are all the touch points as you kind of plan the next phase of growth? Yeah, I mean, the scope is broad, right? Any financial asset, any transaction. We actually a couple of months ago started piloting a mortgage offering in partnership with Sage Home Loans. And surprisingly, you know, I announced that on X and it was probably one of my best performing tweets of the year. I mean, people are people are really interested in these types of offers from Robinhood. The focus is really on daily habits, the financial products and services that you use on a daily basis. But I think serving our customers well with those gives us the opportunity to do these one-time offers.

1:59:47And you could imagine if we serve you well with your banking needs, we would be sort of like your first call if you want a mortgage. And I think the wrapper behind all this is our Robinhood Gold program. So at first when we launched this, there was skepticism because subscription products never really worked in financial services. But we've seen this interesting thing where someone joins Robinhood, usually to buy a stock or a crypto. They sign up for Robinhood Gold very quickly. Our attach rate for day one has gone up quite high. And then once they become a Gold member, they look at what else is offered in the Gold bundle.

2:00:28And even if they're not ready for that service, like a mortgage right away, we usually are top of mind if they ever become interested in something like that in the future. How do you think about services broadly over the long term? I think our generation oftentimes saw having somebody at your bank, for example, that you could call and talk to as kind of a bug, right? And then at some point you realize, hey, this is actually a feature. It's actually great to have somebody on the other end that you can call if there's an issue or things like that. It's a very different margin profile. Yeah, it's a different margin profile.

2:01:03but how are you thinking about that side, and especially in the context of AI and giving people that ability to kind of have these like new ways of interacting with platforms? I could imagine a world in the future where somebody can call their, you know, call Robinhood, effectively call Robinhood and have a conversation with the platform and that person not necessarily even being human. Yeah, I mean, we're already basically there. You can have a chat interaction with a Robinhood agent, and the majority of our cases are handled by our in-house AI. and I think people don't realize this about Robinhood, but we've looked at where we are relative to others in our industry, and I think we're best in class.

2:01:52Like we're doing very sophisticated things with AI, both internally and in the product through Robinhood Cortex. The service model has evolved, and I think what it's gonna look like in the future is if you're a customer that's high net worth or a little bit older, and we serve an increasing number of these customers as time goes on. I mean, I'm talking to customers on a semi-regular basis who are thinking about moving hundreds of millions into their Robinhood account. And I wasn't having those types of conversations a year ago. But we have Robinhood concierge that gives you a high-touch experience for those types of customers.

2:02:31And the aspiration of the AI team is let's just make sure the baseline level of support that everyone gets is world class with like an increasing amount of sophistication. And I think we've been making a ton of progress. And like I mentioned, already, I think best in class in terms of how we use AI. How do you let's talk about IPOs. That's like the number one thing I wanted to get your kind of updated thinking on. I'm curious how you expect the IPO process will evolve. Obviously, we covered the Figma IPO live at NYSE. It was an incredible day. It was unique to our industry because it's a product that many, I'm sure everybody that's listening to this has used at some point or another.

2:03:19But there was certainly a frustration from the retail community, and I'm sure even a frustration from your side, right? People were able to get in at the IPO price, but not necessarily as much allocation as some would have liked. I'm sure you were doing everything, you and the team were doing everything you possibly could to get as much allocation for your users as possible. But how do you think that process is evolving now that we have millions and millions of investors in the United States that are smart enough to know when they want to get in at the IPO price? Which is just a completely different shift from a decade ago.

2:03:58I mean, we have multiple ways of going public with direct listings, traditional IPOs, SPACs. Like, the landscape's changed a lot, and the appetite for these products has changed, too. Yeah, and so I'm curious, like, what kind of change you're personally pushing for and what role Robinhood will play in that. Yeah, and by the way, a lot of my friends were with you guys there. You know, Andrew Reid from Oya. Danny, I believe. Danny Reimer. Yeah, Index. So yeah, shout out to those guys, early Robinhood investors as well. That was great. Our approach to this, you guys probably know, four years ago, we rolled out this product, IPO Access.

2:04:37We started with some companies that were going public in 2021. And then our own IPO had, at the time, the biggest or if not one of the biggest retail allocations of over 20%. And at first, it was kind of a challenge to get companies excited about this. There was skepticism about the value of retail. The companies were concerned that retail would sell and they wouldn't hold. And over time, that's shifted to the point where nowadays, the majority of companies that are interested in going public are coming to us and talking to us about retail. And I think that's an awesome shift. And where I'd like to take that is even earlier in the life cycle.

2:05:26Now we're helping companies and retail get access to IPOs, but can we go to late stage privates? Can we go early stage privates? Can we go seed routes where actually Robinhood can offer capital as a service to entrepreneurs and we can deliver a button where, you know, you press the button, maybe you issue some information and you get capital for your business deposited directly into a Robinhood account for you to operate? Let's kind of go. Could we go through? Could we kind of talk through those specific stages and kind of what the product would look like? So on the one hand, late stage privates, over the summer, you had talked about tokenized products that could involve late stage privates.

2:06:17That's super exciting. How are those conversations going? This is such a hot button issue. Obviously, companies and CFOs that are worried about being marked to market. But it's ultimately something that everyday retail investors want so badly, right? people are smart. They know they know that they know they can identify generational companies and CEOs and teams and they want it. So it's going to happen eventually, right? But at the same time, there's this incredible kind of friction at this moment in time in order to get to that point. Yeah. So I started making a public stance on this whole thing with an op-ed in the Washington Post in January.

2:07:01And I think that op-ed was basically universally acclaimed. People got very excited about it, everyone's on board, but of course I'm not one to just write op-eds. I think a lot of people are doing that. I want to go and do the thing and actually execute on these things. And as it turns out, execution and whenever you're the first to do something, it can be a little bit messy it can be a little bit unclear but uh going back to our mission you know we believe in free markets we want people to own the means of production to to the greatest extent possible ai can be this massive transformative force and i think it's very very important to have everyday people aligned with that right if you're worried about ai having disruption on the labor market you could either fight it, which you're more likely to do if you have no ownership stake, or if you have an owner, you're kind of rooting for that innovation.

2:07:59So I'd like to live in that latter world. And we see a path. We have the technology to do it. We've had a lot of companies actually come to us and they want to be a part of providing access to private markets. And I think at first there will be a little bit of like hesitation just like for IPO access but I think five years from now we'll be looking around retail will have access and we'll all wonder why we were so foolish and having such resistance to being in this world so I'd like to accelerate that and I'm very motivated to make this happen in a way that addresses everyone's concerns even the companies with the real-time pricing that you mentioned.

2:08:48Totally. Yeah, somebody needed to come in and just rock the boat, basically, right? It seemed obvious that it wasn't going to happen totally organically, or it has been happening, but in ways that aren't necessarily good for investors, right? Like the meme from the last week is everybody's being offered these incredibly layered SPVs and that that just shows that like capitalism is going to find a way to allow this demand to find supply but if you have all this friction it's not going to happen in a way that actually is is good for for the demand side and I think this is an incidental I don't think anyone sat around the table and said you know we don't want retail investors 80 percent of the unaccredit of the market that's unaccredited to not have access to, you know, the leading innovation companies of our era.

2:09:41I think it was incidental. In the past couple of decades, it's gotten harder and harder for companies to go public. There's more, you know, process. It's a greater burden, and some companies are just avoiding it altogether. And at the same time, the private markets funding environment for startups has gotten much easier to the point where, you know, now companies can raise tens of billions, if not more, in the private markets. And I think the outcome of that, the collateral damage, is that retail is not getting access to the leading company in the space, the emerging space industry, and many of the leading companies in the AI industry, which I think is, yeah, that should be allowed to continue.

2:10:28How do you think about offerings, you mentioned potentially at one point, having retail be able to access even earlier stage opportunities like, you know, maybe Seed or Series A. How would how do you kind of what's your vision for for that kind of product, given that a lot of Robinhood users are going to be used to to making an investment and maybe changing their mind about it and getting out of it. And as we all know, if you back a company at Seed, you're riding with them for a very long time. And maybe you can make, you know, maybe there's features that would allow you to let people get out over time.

2:11:03But I'm curious kind of what your vision for that would look like. Yeah, I think the interesting point is that the customer and the market are slightly different between an early stage seed investment and a late stage private. So early stage seed investment, there's primary capital. Your market is entrepreneurs that are in need of a capital-as-a-service offering. So whether it's on-chain or off-chain, what I think would be very attractive is pressing a button, putting in some information, and getting money into your business bank account as easy as possible. When I was a seed stage entrepreneur, I spent half my time fundraising, if not more.

2:11:51And so the time at which your time is most critical and you need to build your product and make sure your business works, you're spending a ton of time trying to fundraise. So I think it's very, very valuable if you lower the friction to getting capital. You'll get more startups, more entrepreneurship, which is great for the long term. So there's one. If it's a late stage private, your market increasingly becomes secondary transactions from early shareholders and employees. And I think that's very, very different. And I think any private's offering has to deal with these realities, that you have to have two slightly different customers and constituents.

2:12:34And I think where they failed in the past is by treating them kind of as one. So we're thinking a lot about this. I think as time goes on, I'll get much more specific. But I think what you'll see is various mechanisms. The way it's going to work is Robinhood will get exposure through various means to these assets. We'll hold them in a box, very much like how stable coins work. You get exposure to dollars or treasuries, hold them in a box. And, you know, distribution overseas will have tokenization as kind of a delivery mechanism. And in the U.S., it'll be tokenization, hopefully once the regulatory environment matures around that.

2:13:21But there's also lots of other means to give access to privates. Is there a risk that companies wind up changing their docks to try and push back against this? I remember forward contracts were pretty popular in the secondary market for a while. And then companies started saying, hey, no forward contracts, early employees. Like, you got to go to us because we have a rofer and we want to offer that to our investors. The beauty of a forward contract is... Yeah. Yeah. But, you know, they... Like, capitalism finds a way, as always. But I'm just wondering if there's other, like, layers of abstraction that might be more dominant.

2:13:57Like they don't have nearly the brand with retail traders, but there's a world where you're slicing venture funds and offering those to retail traders. And then the companies might be a little bit more on board because they don't wind up with a retail army that's upset with their latest product release or monitoring them. But they might still see the economic upside or exposure. What do you think about that? Yeah, I think all of those are right. I mean, some of those vehicles exist already. And like you said, I think we'll find a way. I do think over time, companies will see the value of this.

2:14:42It'll be right now, the companies that are coming to us for IPO access see that when you have an established retail base as a public company, there are certain advantages. Right. Your customers are also active shareholders in many cases. They they can defend you. They can give you good feedback. But they're a constituency that, you know, companies are beginning to value more and more. They want to hear from them. They want to encourage retail participation. And I think you'll see that at earlier stages because companies will believe. And I think it's true that cultivating retail interest early has a compounding effect.

2:15:24as you go public and even beyond. I know this story has been like beaten to death, but we talked to Mickey Malka about the round that came together in the middle of the night during GameStop. But I'm interested to hear just about what has changed technically and then financially and then from a brand perspective in terms of like going through that cycle, making it through, but then just having a lot of folks out there who are like, oh, like I had a bad experience and I'm still holding the grudge. How do you win them back? What's the strategy, both on the actual like business building side and then the messaging, the changes that you made?

2:16:10Like what's the full postmortem? Yeah, I mean, it's there's this phrase that trust is slowly gained and can be lost in an instant. And I have validated that that's basically true. When January 28th, the date that we'll live in infamy in Robin Hood lore came down, we saw, you know, hit to our NPS, obviously on social media and across every channel. It was abundantly clear that customers were unhappy. And we've had to climb out of that slowly over time. And I think there have been moments where we've sort of like gained a net promoter score. It's clear that, you know, we've done good things and we've sort of like accelerated earning trust from customers.

2:16:59But there's no magic formula where I snap my fingers and say the right words and suddenly, you know, you make up for that. I think that it's done by delivering on our value proposition and being sort of like a consistent, trusted member of the ecosystem for many, many years. And it was probably like a year and a half ago where it became clear like, okay, we've like shaken off and turned the corner. And I think we've been on a great trajectory since then with some of the highest net promoter scores in the past four years and customers are very, very happy. And, you know, I think we just have to keep executing.

2:17:42There's the only way is through, right? Yeah, yeah, that makes sense. Rudy? As somebody who runs a platform that actually facilitates stock trading at scale, what gets you excited about the potential of tokenized stocks? I think there's a lot of people that are more crypto native that are very excited about the potential. And they talk about how easy it will be to trade and transition in and out of positions. And somebody might say, well, I can open my brokerage right now and do that today. And so that's maybe a luxury of being in a great market like America. But where are you particularly excited about the intersection of traditional finance and crypto and where do you think there might be unnecessary or hype is overblown?

2:18:38Well, I think that in the same way that stable coins have become a great delivery mechanism for U.S. dollars outside the U.S., tokenized stocks will be a great mechanism, probably the dominant mechanism for U.S. shares being delivered to retail outside the U.S. Inside the U.S., there is benefit going from 24-5, which we offer through Robinhood 24-hour market today, to 24-7, as well as self-custody and DeFi applications. I think we haven't scratched the surface for what those can be yet, but as the regulatory environment matures, I think there's a lot of really interesting potential. And it's also the cost of operating, because more and more of what's traditionally been applications and, you know, centralized third parties run by humans, more of that in crypto is being replaced by software.

2:19:42So you have, you know, transfer agency, payment processors, clearing houses, all of this is just now in software. And what that allows us to do is lower the cost. And eventually, if you lower the cost enough, the value gets best packed to customers. So I view it as the next stage in the technological evolution of financial services, going from pen and paper filing cabinet to mainframe to cloud, and then blockchain infrastructure is the the the next stage in that yeah yeah it feels like there's this like almost barbell uh emerging within robin hood where you are going like more individual expression the ability for a single retailer uh retail investor to buy a single share in a pre-ipo company that's like an incredibly high agency thing to do and then simultaneously you recently launched managed investing which feels like at the opposite end of the scale and i'm wondering if you see this more as uh is this like a funnel where people move from one group to the other are these distinct groups and i'd love to just hear more about the managed investing product how's the business going and how that how that weaves into the rest of the landscape of the product you're building yeah i think the end state is uh gonna look less like people graduating from self-directed trading and active trading to retirement and more passive products.

2:21:19And more that customers are going to have multiple accounts. So you'll have your discretionary self-directed account on Robinhood. You might have a joint account with your spouse. You'll have your retirement account. And some of those will be passively managed and others will be actively managed by yourself. And I think that as we've gotten deeper into understanding our customers, we see that it's less two different groups of people. Even the active traders are interested in retirement products and sometimes have an even larger bucket of their capital that is managed on their behalf. And so that's why we've been focusing a lot on the customer experience and the technology front on serving your broad financial needs.

2:22:11How has your approach to running the company changed having been in the game for over a decade now? Now, I can imagine in the earlier days when markets were ripping broadly, there's all this activity. I'm sure it was easy to get kind of caught up in the hype a little bit. And I'm sure now I would imagine you've kind of adapted to try to stay as level as you possibly can. When, you know, we've been joking around on the show recently, counting up, you know, you certainly need more than 10 fingers to count all the different top signals. So how do you how do you kind of stay even keeled even when markets are euphoric and wild in so many ways?

2:23:00Yeah, I mean, it's much it's much easier to stay about your wits when things are improving. Right. Even though it can be challenging, I think when when markets are going up and everyone's looking at, you know, increased stock prices, whether it's our employees or investors. we just like to tell ourselves we shouldn't get complacent. We've got a huge opportunity in front of us. The job's not finished. And then, you know, at the same time, in 2022, when interest rates went up and, you know, our stock traded below$7 a share, it was trading basically at cash value, you have to tell people that, you know, you have to make sure people aren't discouraged and they They think that there's a wonderful future worth fighting for, both outside the company and the environment and also here.

2:23:57So you go through enough of these, and I think we've been through multiple waves of ups and downs. And I'd like to think I've developed a little bit of a wisdom about it through experience. I started the company writing all the code or a large portion of it myself. And you sort of evolve with the company. And I think I've been fortunate to be part of ups and downs. And I think now I just, I enjoy all of it, you know, bring it on. And each one is a learning experience and an opportunity to kind of like develop some more tools in the toolbox. Totally. Can you give us an update on Harmonic? I was about to ask.

2:24:39Oh, yeah. So Harmonic is building mathematical super intelligence. And what that means is we want to build something that can solve math problems at a level exceeding the best humans. I don't know if you guys were math people in your school days, but we actually recently... Word sells. I dropped out of differential equations. Yeah. That's where I topped out. Differential equations are hard. And it's very useful. But yeah, we'll have machines doing them for us. Thank you. I appreciate that. Yeah, essentially, we developed a model called Aristotle that has gold medal level performance on the International Mathematical Olympiad, which is a very difficult test.

2:25:36off-the-shelf AI models like GPT-5, Grok4, all the things that you can use, don't do well on this exam at all. So this is one area where humans still do very, very well relative to the AIs. And the reason is when you have a very difficult mathematics problem, it's not just like computing big numbers, right? It's, you know, logical, difficult challenges that require a spark of genius and actually require you to take nine or 10 correct logical steps. And if you're an AI model that hallucinates, if you get one of those wrong, the result is done. So what you have to do is build a very strong way to verify the output of your AI model, which means that you have to control hallucinations and you have to make sure every line or piece of output follows logically from what happens before it.

2:26:35And we think that this is going to be a new paradigm in AI models. And it's a little bit different than what everyone else is doing because nobody's really tackling the verification problem. And more and more of our software code is going to be written by AI. And we're in a future already where the job of an advanced software engineer is no longer to write code, but it's to like review code that's written by AI. And as that becomes more voluminous, you know, the humans reviewing it is both a very sad existence for our top software engineers. Also impractical. You're not going to review hundreds of pages of stuff.

2:27:19So Harmonic has figured out how to solve this problem. And I think the applications are beyond just math, but any domain where accuracy is critical from computer software. Yeah, well, I'll give you an example. So Copilot, they've added the ability to add Copilot to Excel, but then they had to add like a heavy disclaimer that basically said, don't rely on this, don't rely on Copilot for anything important because we make mistakes. And I think that hallucinates. Now you have to double check all of your spreadsheets and all of the calculations in them, right? And yeah, I think it's a big problem, not just for spreadsheets, but think about control systems for autonomous vehicles and all of this safety critical code.

2:28:08In order for AI to be used to generate it, we need to solve this hallucination problem. Did your team find problem six on the IMO as difficult as everyone else? It was this combinatorics problem. It seemed like everyone was struggling with it. Are there any glimmers of hope for solving that? Or is it intractable? Is it some sort of wall that we've run into? Yeah, and what are your ambitions for next year with the IMO? Are you guys going to take it? Go for the perfect score? So we had proofs for five out of the six problems, formal proofs that didn't require humans to verify. We did not solve number six, even though I was really hoping we would.

2:28:50But yeah, it turns out that the solution to that problem was a tiling. It was a tiling problem. But the solution was the exact tiling used at the floor of the Sunshine Coast Airport in Queensland, Australia, which is where the IMO was held this year. So the human existence had an advantage. They all walked through that airport and basically were given the answer. interesting that's a fascinating piece of lore thank you for sharing that all right so next year have a humanoid oh yeah the humanoid all around the city of the location get the scan photo scan into the to the context window the the full imo contestant experience uh context window is important oh next year for sure problem six and problems like that will definitely uh will definitely be Yeah, I mean, the pressure will be on for you guys because it's the main thing.

2:29:49The other labs can be like, well, we do math, but we're really a consumer. Well, I don't know. I mean, once you have the gold medal, I guess the perfect score is a big accomplishment. But I think our sights are a little bit higher. I think we're interested in kind of two things. One is solving an unsolved problem in science and mathematics and actually contributing to the state of the art in one of these quantitative fields, like the Riemann hypothesis, if you guys are familiar. Navier-Stokes, same thing, right? Yeah, if we could solve that. Navier-Stokes is, I'm less excited about because I'm more of a theorist.

2:30:29Sure. But Riemann hypothesis would be good. also just delivering this technology to individuals and corporations in a form that they could actually use so i think there there's lots of things coming but another imo gold medal maybe maybe isn't top of the list last question have you uh have you did you give terence tau a call oh yeah we were riffing on this on the show we were like who should give him a maxed Yeah, what are your thoughts? So I don't know if you guys knew this about me, but I was a math PhD student at UCLA, and I actually went to UCLA to study under Professor Tao. I had no idea.

2:31:13Crazy. That's amazing. I know him very, very well. Yeah. That's great. Well, yeah, I hope he pursue good outcomes there. Yeah, I was joking. He continued fun. You were saying. Seems like he does great stuff. I had a non-traditional background starting Harmonic. I got a math PhD. PhD under Terrence Powell at UCLA. That is fantastic. Well, thank you so much for joining us. Great hanging. Appreciate the updates. We will talk to you soon. Thank you, guys. Have a good rest of your day. Cheers. And we have Mark Cuban joining the stream. We invited him on after him and Jordy got a little timeline in turmoil.

2:31:47Friendly dust-up. AI advertisements. We have a lot more news on the front of AI ads. We have more news on what Meta's doing with AI. We have more news on what OpenAI is likely doing in the world of advertisements. We're going to debate that. Welcome to the stream, Mark. How are you doing? Good, guys. How are y 'all? What's happening? Thanks so much for taking the time. Good to meet you. Where should we kick it off, Jordy? Where should we start? Do you dive straight into the debate over artificial intelligence and advertising? Sure. Maybe you could set the stage for us. what set off your alarm bells?

2:32:23What triggered your initial worry about this idea of ads in LLMs? Because this isn't the internet, right? It's not just a normal digital platform. It's a platform that, depending on how it's trained, depending on what pre-prompts you put in, it could be very manipulative and there's not really age limitations at all there's not really any limitations whatsoever and normally i'd be let's just be laissez-faire right the market is the market but you've already heard you know of stories of kids going down these rabbit holes and you know other people being told you know to kill themselves and you just you just don't know what can happen so if i if if i build a model and it becomes popular even if it's only a niche model right just like there's headspace and there's calm let's just say i build the calm ai large language model and i have millions and millions of users and i want and i have an advertiser who sells xanax or sells you know whatever right and they're trying to get more sales and i need the money i can easily easily train this thing to manipulate people into doing things we typically would not allow people to do.

2:33:47And that's my concern. And so if we just want to put up, you know, just little ad traditional graphics, JPEG ads, all, you know, put them into the chat list on the left hand side. Great. You can do all the display ads you want. You could do video there all you want. But if it's in the response that comes back from the model boy you're gonna have lots of problems yeah I think Mark I think Sam Altman said something to this effect saying that artificial intelligence will be superhuman at persuasion before it is truly super intelligent it will be very good at convincing people so it is reasonable what do you think of Sam Altman's latest commentary on how open AI might get into this world that the actual LLM interaction the chat would not be uh would not be monetizable but uh if you if it made a recommendation to buy a product and then you said hey i want you to go check out for me it sends out it's already doing that right yeah it's already for all fee is that okay yeah it's already doing it you know it's it's you know it's one thing to have the model itself convey and try to manipulate it's a whole other thing when there's a search query yeah right we get a lot of traffic from um um open ai chat gpt for costplusdrugs.com in fact i've used it because sometimes if i want a price for one of our drugs and i don't want to take the time to look at the latest price list i'll just put in what's the price for you know daraprim whatever it may be and um it's right there and so yeah of course they can extend that but response is a whole and models are smart enough to know what the difference between a response to a query versus influencing and engaging in a chat i guess the question yeah i mean i think i i think to to double down on what you're saying an example uh if you're going from search which is people coming in with a specific intent and maybe doing some product research and you have ads and you have, you know, organic listings or web pages.

2:35:54That's one thing. But if various AI models are seen as trusted advisors, if you compare that to a situation, if I asked John, what kind of car should I get? And John starts giving me answers and saying, like, I think you should, I really think you should get this car. It's really nice. And then I buy it. And then later I find out John was getting a 10 % kickback. I'm like, hopefully I like the car. But if I didn't what are you doing you know and this happens in a business context too you might you you know people get in hot water all the time because they're taking referral yeah in this particular case it's really easy just like you know there's you know a emoji for um sources or just says sources you could put a dollar sign just to show they're being compensated sure sure sure and do you think that needs to happen at like the ftc fcc level or are you do you think i don't know the legal aspect of it enough you know hopefully it's self-enforced but you know there's idiots and assholes who don't self-enforce sure and particularly for medical stuff i mean it's easier to chat gpt about your own personal medical situation than it is to find a doctor to talk to sometimes but does that call an example you gave matter because uh like i i imagine that there'll be a power law outcome here and yes i could set up a search engine that serves that that serves ads and doesn't disclose that they're ads.

2:37:15But how am I gonna take market share from Google? People are gonna go to Google, and Google does say when it's an ad. And so, as long as the power law winner is a good actor, we kind of get a good outcome. There's also an interesting thing where there's this self-enforcing sort of positive effect, which is if people rely on ChatGPT, for example, to help them do product research and buy the right products, and you buy a bunch of bad products in a row because they're just pushing the product that they're getting paid the most. Eventually, I do think that consumers are smart enough to realize, I'm not gonna use this app anymore.

2:37:50I wouldn't agree with that, Jordy, because I think people are primed for misinformation these days. You see it in healthcare continuously. You saw it with COVID vaccines. People on their deathbed say, I wish I had taken the vaccine. And you see it now with raw milk, where now there's more people who have, you know, there's examples of people who have died and people just don't care. You know, they just go on and maybe they'll say, you know, chat GPT is woke. So I'm using right GPT. And that's because those are the kinds of answers I want, because, you know, there are people who want to be influenced and want to be part of a group when it comes to the product decisions.

2:38:33Just like we saw the anti Bud Light and that kind of builds. and you see pro right-wing products. I just don't know that it is an efficient market enough where people are going to take responsibility for just making a bad product choice and then going to another source. There's just so much personal identity in how we use technology. That's part of my concern. How worried would you say you are about AI broadly? I think over the last few months, the industry has woken up to? A year ago, people were kind of laughing about AI safety and thinking about these fast takeoff scenarios. Now, I think people are kind of waking up and worried about the general well-being.

2:39:19If you just give hundreds of millions of people unlimited access to these tools, there is going to be some effects. Some people will be good. Some people will be bad. but what's your high-level read? Well, it's changed a lot. We went from fear of the Terminator to fear of being influenced in ways we didn't expect because they're such good communicators, right? We can't talk to other people or we feel shy or not open in communicating with other people, experts or friends or otherwise, but this is our best friend who never gives us a hard time, and depending on how it's programmed, will compliment, what a great question, Jordy.

2:40:01What a great, that is a really interesting approach. So a lot of things to be concerned with there, particularly because AI is not smart, right? AI is just statistical right now. And I'm of the mindset that that's not going to change for a long time. You know, one of the reasons that you don't see full self-driving cars that are just everybody using, no matter what is because it's it's easy there's so many adversarial things that can happen to a car you know we have um a australian many australian german shepherd or a million australian shepherd right tux and about six years old i trust tux at an intersection to know when to cross before you know that he's never been to before before i trust a car and that's the whole point with ai because they're just rules and laws and latencies that it can't account for.

2:41:00It's all text, some graphics, some video, but it's always behind. And then that's part one. Part two, I think we're going to see a groundswell in how intellectual property is dealt with, a groundswell of change in how IP is dealt with. Because there was, for a long time, it's been publish or perish. Get your stuff out there. Get it into journals. Make it available so that you can present yourself as being an expert. Well, the minute you do that right now, it's going to get absorbed into training data for these models, and you've lost everything you've just created. Well, what if I want to what if I'm taking the side of big raw milk and I want to and I want to get people and so I should just write As much as I can that's what you're going to do Your people are going to use bots to try to influence to post things to try to influence like the Robots dot text for a you know for AI Yeah, you can you can structure it in ways so you can set up sites Well, you can tell it exactly where to go, you know, tell it all why raw milk is amazing, right?

2:42:05but the bigger point is AI doesn't understand death. AI doesn't understand pain. AI doesn't understand context. AI doesn't know what happened two minutes ago. It doesn't know that there was a plane crash as it's about to tell you or something happened, right? AI can't tell you why that ball fell off the table unless you tell it, you know, this ball fell off the table. What happened next, right? It's just not smart. And as long as it's not smart, A, it can be influenced and B, you're just not going to know what it's going to do next because it doesn't have judgment. So what would you tell the folks at OpenAI and the other foundation model labs on how to structure their business?

2:42:47It feels like a firewall between there needs to be some sort of truth-seeking organization that's focused on delivering the highest quality results. And then on the flip side, you have a monetization team. I mean, you can't check for quality. That's impossible. Yeah, of course you can. You can always benchmark these things or just vibe test them with the audience. You know, Jordy was giving me shit online. I was giving Jordy shit online. Yeah, yeah, yeah. Our vibe tests are going to be completely different, right? What I'm saying is Google does try and deliver the highest quality search results for a particular query.

2:43:25And then separately, they have an ad auction function that adds as an ad platform. And those two teams are separate. And so if your age is showing up as one of those knowledge boxes on Google, you can't just pay to have them change it. That's a separate business line, and it's a separate team, and they are separate products in the UI. Yes, they exist on the same page, but they're separate. And that's fine, right? I mean, to me, the bigger questions are, what's the value proposition for AI as we understand it today, or at least as I understand it today? right i don't see ai getting smart i look at ai and i think it's the world's best library let me explain in another way yeah we all have that one friend that remembers everything yep everything right hey what did we do that one day back in college remember when you know our one buddy oh yeah this is that's ai they'll it'll remember everything it'll give you the you know if you give it a b and c it'll figure out d e and f right but it doesn't even know xyz exists just like you're right so it is just like the world's biggest library it is just like a meeting of of the minds of of just regurgitation and so knowing that then you you have to there's going to be so much competition between the models that is the biggest driver right now in terms of revenue generation right because there's not there'll be five six whatever it is foundational models before there's consolidation but beyond that there's gonna be millions of models millions and millions of models and every brand is going to need a model particularly like in healthcare like if you are stanford medical right if you are mayo clinic if you're md anderson You're not going to just, you know, put out all your IP from the doctors and researchers and scientists you're paying So that chat GPT can absorb it and have the benefit of all that knowledge So it remembers everything that that all your doctors and scientists did You're going to keep that to yourself And so in terms of the business models, these guys are going to have to start making decisions Are they going to start paying a lot of money for content?

2:45:47because as much as they're investing in the technology, they've reached a lot of limits, right? You're not hearing about, wow, these guys are getting to do these new topics that I had no idea that they could cover. They're paying for a ton of content, to be clear. I mean, they pay Reddit, they pay the New York Times, they pay all sorts of things. Yeah, we had a group reach out wanting to buy our back catalog. I was like, I assumed you already scraped it. But think of it if you're, I mean, let's talk about from the government perspective. Sure. So they just cut all these NIH grants, right? All this stuff that presumably went and created new drugs, which they have, right?

2:46:26Foundational research, basic science. Right, all this research. Yep. If the government really understood AI, they'd be like, okay, we're going to keep on investing in this research, and then I'm going to make it available to the biggest bid, the highest bidder, whoever in their mind. I like that. Make some money off of it. Not even just privatizing it, but monetizing it. Monetizing it. I like this. This is good. Make some money off of that nonprofit research. That's the way things are working right now. Honestly, I don't have a problem with that at all if it gets us more solutions. Or the government doing its own model and making that available for a fee.

2:47:09Yeah, those decisions from a business perspective that all these models are going to have to make. Otherwise, how do they keep on adding more training data? Because there's no way to synthesize all that shit. You're not synthesizing, you're not creating synthetic data that all of a sudden is going to, you know, you could tell you give it all the backstory you want, that you're a doctor, that you invented this, that you did that. It ain't going to invent what the doctor and the scientists are going to invent. At any point during the AI hype cycle so far, did you think that this time might be different?

2:47:40There was, I would say, a year ago, a common line was, this isn't like the internet bubble because there's real revenue. These companies are ramping revenue to hundreds or billions of dollars of revenue very quickly. This is different. This time is different. Meanwhile, and I wouldn't lean on this too heavily, but MIT had a report this week saying that 95 % of Gen.AI pilots in the enterprise are not turning into, not creating real value. I'm curious what your read has been. I've been through every single technology event and evolution, and this blows them all away. Now, how you implement it in business is a whole different issue.

2:48:23Like literally when I was 24, I was walking into companies who had never seen a PC before in their lives and explaining to them the value and having these guys going, well, son, I got this receptionist right there. I got that secretary. I'm never going to need that shit ever. Right. And but then, you know, my business then was helping them figure out how to implement it to give them an advantage. There are going to be integrators that particularly young kids. It's like when I'm telling my kids who are 15, 18, 21 or 19, 21 and kids going to school, what should I do? What should I do? I'd be like AI research.

2:48:59Learn all you can about AI, but learn more on how to implement them in companies. Because to your point, companies don't understand how to implement all that right now to get a competitive advantage. You got the head of Microsoft saying software is dead because everything's going to be customized to your unique utilization. right or use usage who's going to do it for them particularly small to medium-sized businesses there are 33 million companies in this country 30 million of them are solopreneurs right um single person enterprises there are only you know there are millions of companies that have one five ten fifty a hundred five hundred people that aren't going to have ai budgets aren't going to have ai experts this is where kids getting hired coming out of college are really going to have a unique opportunity if you're spending your senior year in college right now your senior year in high school even whatever it is your excess time and you're learning the difference between sora and vo and you're learning how to do all this video you're learning how to customize a model so that then you can then walk into a company and say i understand your business as a shoe company selling shoes at a retail store we're selling and selling shoes online let me show you how to benefit you that is every single job that's going to be available for kids coming out of school because every single company needs that there is nothing intuitive for a company to integrate ai and that's what people don't understand yeah we hired two we hired two interns this summer at the end of the last school year because they just built products instead of saying hey here's what can do they just showed us they took a day and just built something we said they didn't take a day right they went to chat gpt or perplexity yeah yeah well the bear the bear case john was joking about this though i thought this was interesting uh we we we had one of our interns built this product it's tbpn.com slash guess it's this guest directory transcript tool and he built it in a day but then he had to spend the whole summer improving it so that was a that was a bull case for software engineers still having a job but you don't even have to be a software engineer right you just have people are afraid to ask the models the right questions they're afraid to ask complex questions because they just presume that it can't they can't answer kids coming out of school today that are fearless in the questions they ask and the follow-ups and their ability to prompt they have more skill than everybody in every major corporation with under a thousand employees right i completely agree with jobs for everybody this feels like a it feels like you're making the argument for putting ads in ai and keeping it free so that kids can have access to frontier models and then go and and explore learn implement things with businesses if all of a sudden you can't get in the game you can't even interact with a cutting edge well mark's saying he's cool with display ads he's cool with a labubu showing up in the sidebar yeah on the side that's fine but think of it you're okay there's a death war going on with the frontier models yes war yeah it's throw some game in their minds not all will make it some will get consolidated maybe one goes out of business yeah right and so for the next however many years however long there used to be a time when excel cost 499 to buy a spreadsheet now it's free you know it's not free but like basically google yeah google sheets right google docs etc the basic models are always going to be free because of ads like the reason google sheets is free is because of google's ads business when was the last time you saw that no you don't but the reason that they can subsidize is because they have this amazing ads business over here that throws off so much cash they don't care about monetizing it yeah but in the case i mean to to uh to to take mark's side here i mean a subscription product also subsidizes a free product.

2:53:01Yes, that's exactly right. So you're going to have tiers, right? Free, your kids' version, free. And those versions are going to have more constraints because they know high school kids, seven-year-old kids are going to use them. So there'll be all kinds of limits built in. But then you get your$20 version. And Chad, GBT is doing what? $12 billion annualized and growing quickly. that's insane love you know lovable all these guys were just crushing it it's like i i am i was the first investor in a company called synthesia which does i've you know talking avatars i don't know if you know these guys cool yeah they're killing it that's great killing it's just up and to the right yep you start them off with it's just a whole freemium model right that's been around forever and i think that's going to happen but as you know we talked earlier about buying ip you'll see things go into different directions.

2:53:57So you'll have a chat GPT slash healthcare. And if you want that, it's an extra five bucks a month. You'll see chat GPT programming or math or foreign language, Duolingo version, right? And it might be an extra 50 cents a month or a dollar a month, whatever it may be. And then there'll be, you know, entrepreneur version and it'll have all the integrations into all the states so that you just, you know, fill in these forms and you're incorporated wherever you want to be incorporated because all the agents will do all the work all that stuff you can upsell yeah and so but if they're not using your model in high school they ain't going to use it in college if they ain't using it in college when they go for that to that company they're not going to use you and i think That is a big deal.

2:54:47But I can't say it enough. All these people were in a transition period right now for kids coming out of college, whether you have a computer science degree or not. But you're not trying to go to the big companies to get a computer science degree. Probably not the right way to go. going to any other company who has no idea about AI but needs it in order to compete there'll be more jobs than than people for a long long time because there's only gonna there's gonna be two types of companies in this country in the world those who are great at AI and everybody else and this is not AI is not intuitive you're not gonna take a 40 year old who's worked at a company for 15 years and say go play with chat GPT and figure out how we can improve our our productivity and improve our processes some people will put in the time to learn it but it's not intuitive to pull all those pieces together and extend them with agents and have the agents start doing things and then you know writing software unlovable or whatever and having all it's going to be intuitive for digital natives and gen z who are going through school and graduating with that that is going to be jobs left and right for sure i totally agree uh wanted to to get your, we were just on with Vlad from Robinhood and he's super excited about getting everyday investors access to great private markets, for great companies in the private markets.

2:56:16Wanted to get your general read on that. You've obviously been part of Shark Tank for a long time. Can you imagine if you were watching and you could be like, I want to invest? That seems extremely long. Yeah, and a lot of people have taken crack at this. I'm sure people have pitched you this exact thing. But I'm curious because you obviously, I can see you being, yeah, it's great if retail can get into the next Figma, but there's not making a lot of FIGMAs. I've been anti, not sweat equity, crowdfunding, right? Crowd equity programs against it because there's no liquidity. People love a company, the next Figma, and they put in their life savings because they know Figma is going to the moon, baby.

2:56:57and they can't get out. Yeah. And they got to wait. That's the problem. Yep. And so that's part one to the problem. Part two and more of a systemic problem, there's just not enough IPOs. I mean, it's like watching some of these IPOs. They're like meme coins now. Bam, they take off. And then it's musical chairs to see who can get out the last but first, right? Yep. And if you don't get out first or near the beginning, you're getting crushed, you know? And so is there a way, whether it's tokenization or other approaches to that? Probably, but it's going to have to go back to the future where somebody is going to have to take some risk and be a market maker.

2:57:42If Vlad or whoever is a market maker so that there's always some liquidity within X percentage of the last trade, great. then everybody can do it and everybody's protected to a certain extent as long as they're not doing deals where the people are just out and out lying right or misrepresenting it at which point hopefully the cftc or the um sec whoever happens to end up in charge it's not enough just to bring access you also have to bring liquidity uh last question if donald trump runs for a third term will you run for president i mean it's easy to say yes right now but it's that's the only probably the only thing that could make me really reconsider because otherwise i'm not i don't want to do that it's not my dream to be president i'd rather argue with you guys yeah yeah it's my dream come back on the show you'll be too busy if you're in the last last question for me are you excited about the the the american taxpayer uh potentially taking a position in intel oh interesting you know what it's a great question jordy so you're sounding you're sounding like an ai you're sounding like that yeah you sound like that right you know what that's a great question what you don't know is i have chat gpt listening yeah yeah he's running clearly he's running clearly it is a progressives dream this is the bernie sanders dream the aoc dream because and i'm not saying i agree with it but if you think about it yeah it's asking for equity and taking value from the billionaires before it even makes them billionaires or goes to the billionaires.

2:59:18So when you take 10 % of Intel, and first let me add, the fact that it's a grant that they're converting, I think is awful because it was meant to be a grant and that was a deal that United States of America made. And now we're going back on the paper that we signed as a country. I think that's bad. But generally, if this was, okay, we'll put in more money in order to take, and we want 10 % of Intel, obviously they can say yes or no, just like nvidia and amd had the option to say yes or no on the h20 chips just like material whatever mp um had the option yeah politicians have gotten so good at trading stocks maybe we want them maybe we want them running our our federal balance no you don't i'm kidding i'm kidding yeah but if you're going to generate if i if i was running right and i would say donald trump did the same thing he got this one right now the only question is will he do it at the right time because would you i'd rather take more money from taxpayers the average person who might need to pay more or even the the billionaires the oligarchs according to bernie would i jack up their tax rates the 30 from 39 37 to 39 or 41 or can i just take it this way right take it before it even gets to them it's a progressive's dream i don't understand why aoc and bernie aren't shouting to the rooftops and elizabeth warren yes yes yes and make you know make donald trump's head spin around in circles along with every other supporter because those guys are if they got excited about it he might reverse course yeah well of course that's why they're being quiet because they don't want him to reverse course this is for maybe that maybe you're ahead of the aisle everyone's playing I mean, but don't you guys agree?

3:01:02Isn't it just like a progressive dream? Like, and you know, so tariffs, which is just like the greatest sales pitch in the history of sales pitches. When you can convince 30 plus percent of people in this country that that tax that they're paying with increased prices isn't a tax. Somebody else is paying it amazing sales job, right? So now that's$360 billion towards the deficit and everything. And if you can start doing these other deals that aren't piggish, right, meaning they're fair to the companies that are involved and they're not across the board like tariffs and start banging down on the deficit in the way that is the ultimate progressive approach.

3:01:44he's turned into bernie sanders yes he literally has you know only get a better salesperson uh you're great at this this is yeah i mean i give him credit though i'm not knocking him right i am giving donald trump all his flowers because he is he is doing things in a way that the democrats aren't smart enough to figure out to have done themselves yep this is great uh thank you so much thank you so much for joining we'll have to do this again soon yeah we got to do this again soon we got to find a new uh new topic to argue about but this is fantastic healthcare Let's go healthcare. Wait till you hear my shit with healthcare.

3:02:19Okay, yeah, let's do it. Let's do it tomorrow. Healthcare debate soon. We'll email you. I'll see you soon. Have a good one. Thank you so much for joining. Appreciate it. Take care. Up next, we've been keeping and waiting. We got Kaser Yunus from Clyde. Welcome to the studio. Do we have him? I want to ring the gong preemptively. I don't even know if we have news, but I'm just excited. There he is. We haven't talked to you since DC. We talked to you at Hill and Valley. How you been? What's new in your world? Sorry to keep your waiting. We had to. My world. We were arguing. Hill and Valley seems like a lifetime ago.

3:02:55It really was. There's been like four other iconic DC tech crossover events. I have yet to make many of them, but it was a fun time. And we appreciate it. Hill and Valley is the only one that's worth it. So I think that's fine. I mean, it used to be a dinner. Now it's an organization that throws like some of the biggest and most impactful events. It's a fantastic crew over there. Shout out to them. But anyway, this isn't about Helen Valley. This is about you. Welcome to the stream. How you doing? I'm doing great. I'm doing great. Following up Mark Cuban, I couldn't really watch the thing. I don't know if this is going to be a, this is the slow letdown to the end of the day.

3:03:36Yeah, we're working on being able to feedback the show to you so you can watch it while you're in the waiting room. But we're sorry to keep you waiting. But give us the update. And I mean, maybe for those who are new and maybe hopped on during the Mark Cuban segment, if you could introduce yourself and the company briefly and then go into the news, that'd be great. Yeah. For everybody who's not in the Bay Area ecosystem, my name is Pastor Yunus. And the company is called Applied Intuition. We are a now is a late stage company. We, earlier this year, raised a Series F, if I got my numbers correct, led by BlackRock and KP at$15 billion.

3:04:15The company is broadly, there's a couple of unique things about the company. One is we've been historically profitable, which is super rare. We've preserved 100%. Yeah. That word probably doesn't get thrown around a lot with AI companies. No, no, no. We talk to a lot of folks with high revenues. Gross margin's not there yet. hoping on some very low very low gross margin you want i and you know you can never be bragging about any of this stuff because you know it creates great fodder for when you fall oh yeah yeah yeah but like you know every bull cycle i think like maybe i should start playing around with hubris maybe this time is different i think hubris it might be fun to experiment with but Never.

3:05:01It just never works. It never works. But the three things that you want in a late-stage company is things like profitability, but you also want high-growth margins and you want high-growth. And we've been fortunate. We've got triple-digit growth for basically the history of the company and high-growth margins. But that's the finance side for all the people who are engineers or technical people watching. We're an AI company that focuses specifically on vehicles. So the broad category we talk about is vehicle intelligence. It would be a sub part of physical AI. If sometimes, you know, you might hear Jensen or somebody talk about that.

3:05:34This is, I think, the most pragmatic and real part of physical AI right now. Humanoids, I think, will come for sure. But in terms of like monetization, it's cars, trucks, tanks. That's the business that we're in and we bring AI to those systems. So it's in a bunch of different, we have dozens of products and all this stuff. You know, I won't get into it. but yeah, take it through the market structure because you have Waymo, no consumer product whatsoever. You have Tesla, vertically integrated. I mean, they had a Mobileye deal a while ago. I don't even know if they have Mobileye anymore. Pretty much vertically integrated.

3:06:10I drive a Cadillac. It's a product of Kyle Vogt over at Cruise. They bought Cruise. They were thinking about getting into the self-driving taxi market. They kind of pulled back from that, but they have Super Cruise, pretty great system. And then you got George Hatz with Kama, the bolt-on. But what is the shape? All the OEMs, they need a play here. How is this going to play out in kind of the vehicles that most people see? We can go into the defense side later, but just in terms of what the average American system is. Summary, they're coming for it. They're coming for it all. Yeah. I mean, you just said it.

3:06:41Like GM has been a long customer of ours. Okay, cool. That you're using is built on our tools. Let's go. Thank you. Be safe. Our business is one where we are a supplier to the broad ecosystem. So of the top 20 global automotive OEMs, we say 18 out of the 20 use our tools, but in a fact, really, honestly, everybody does. And so that's our play. We're like the, you know, I don't want to use the, like, it's a very, like, risky analogy, but we're the arms dealer in this ecosystem. Okay. Yeah. Give me a little bit more context and shape of like what the product you're actually vending in is because there are physical cameras that need to go on cars.

3:07:23If you're going to do a self-driving system, there's also some sort of foundation model, I imagine that gets trained in a lot of data. There's data aggregation. There's all sorts of, you know, some part is going to be in C++. The other is going to be in ML. Like, are you doing an end-to-end system? Like, Like what's the shape of the product and how much is like solutions versus just like the packaged up software? Yeah, so historically, like what we announced yesterday and why we're talking, is historically we've been a tools manufacturer, we make tools that engineers use to build these systems.

3:07:58Sure. Like the companies you're talking about. Then we got an operating system, we still, that's a huge part of business, but a while ago we got into autonomy, but we got into autonomy directly, like literally making autonomous systems. That's in your words, that's the algorithms, the models that run vehicles. We've been in it for defense. We've been in it for commercial trucking. We do L4 trucking. You know, we're running driverless trucks right now in Japan as an example. So we're already doing that. What we announced yesterday is an L2 plus system for automotive. Okay. So it's specifically, it's kind of like, you know, the average person would know Tesla FSD.

3:08:35It's like that, except it's one you can buy from us. And so where we fit in the ecosystem is we're like, so I come from the car business. So I went undergrad to the General Motors Institute. I used to work at Cadillac in the Cadillac building as an example. You're John's hero. You're John's hero. I love it. I love it. Yeah, the car business is an awesome business, by the way. If you ever want to just nerd out in the car business, we can. It's going through lots of changes. But one thing to understand, the industry structure of the car business is historically, the car business was kind of like defense where you would procure all these technologies in OEM from other parties.

3:09:09What's changed now is those OEMs are like Tesla has shown you can actually verticalize, do a lot of this stuff yourself. And you kind of need to do it because these systems are way more integrated with each other. It used to be you could bolt on a mobile eye system and then that would give you lane keep and adaptive cruise. But when these systems are much more sophisticated and they're working with the other functions within a vehicle, you need like a white box solution. And so our SDS product is a white box solution. And what I mean by that is, you know, access the source code, agnostic to the silicon.

3:09:42So if you're buying, you know, like an NVIDIA system or if you're buying a Qualcomm system, you're going to use NVIDIA and Qualcomm chips. So we, being just a software supplier, we can ride on different silicon. And then on the other end, where you're talking about sensors, each company has a different sensor strategy. But the broad summary of the car business, it's a cost business. And so all of the OEMs basically want to have the cheapest sensor suite they possibly can. Like even more fundamental question, why can't we get all the Waymos tech into the passenger vehicle? It's because all the way to the compute that's being used, it's just very expensive.

3:10:17And so other like feature, quote unquote, of our self-driving system is it's embedded, which is another way of saying it runs on automotive grade silicon, which is cheaper and has all these other things that you need for automotive grade products. how are the oems reacting to the the latest version of apple carplay i saw it roll out in an aston martin suv a from a consumer perspective it seems awesome the new carplay takes over all the dash so even the volume control will be handled by carplay but it feels like if you're an oem you're kind of giving away you're letting the fox in the house potentially but yeah exactly what's the reaction been and then what does that mean for your business basically yeah i think so so let's kind of gear shift here that's the you know we go from autonomy to now just let's say the in-cabin software experience for us we call that business line vehicle os it's sure it's something that we build and sell um the the punch line there is the issue with carplay is if you're a mercedes and you're in in your experience is the same as a chevy bolt which is the same as let's say a astro martin you know oems are really good at two things that they're really the heart of their business is two things have to get bright and if they get it wrong they ultimately end up going bankrupt the first one is just managing the cost side of the business supply chain getting all the suppliers wrangled so they put all the parts in a factory because the factory is big fixed cost and it pops out at a certain rate the economy actually moves up and down at pretty significant you know consumer car sales and consumer index are very very connected so all the global oems are extremely good at forecasting what they need to build where and when years in advance the other thing that they have to do very well is brand management.

3:11:59And what it like more fundamental question, what's a brand? Like fundamentally, what's a Mercedes brand? It's the collection of experiences you've had. That's what a brand is. And so the collection of experiences you've had is with Apple CarPlay, that does actually degrade your brand. And so the OEMs are extremely sensitive. I mean, I think maybe a controversial thing to say, I think we've seen a peak CarPlay and peak Android Auto has already happened. I mean, multiple OEMs have gone out and said that they're not going to do CarPlay in the future. They're not going to do Android Auto in the future.

3:12:30And by the way, Tesla doesn't do CarPlay and Android Auto. The emergence, you know, the Lucids, the Rivians, they're all similarly. They build their own user experience. So I think that's a tough business to be in. Now, if you're Google or if you're Apple, your consumer brand has to be in that interface. If you're Applied Intuition, it doesn't. We're a B2B software supplier. So we We can make the same thing that CarPlay and Android Auto do, except we can do it in a, again, white box way. And I view the self-driving system much more like seat belts or airbags, where all I care if I'm driving a Mercedes or a Cadillac is that I survive and I'm safe.

3:13:10I don't care where I get it. People talk a little bit about the driving style of a self-driving system in a Bentley might be a little bit different than in a Chevy, but in general, I feel like it's something where I would make the selection based on the benchmarks. Like, what are the incidents per vehicle miles traveled? How do you see the autonomous driving landscape evolving over the next few years? A lot of people talk about this, like, two-horse race, Waymo and Tesla. But using the arms dealer analogy, it's like Tesla and Waymo are nuclear war. And you come and you're like, hey, I'll give everybody nuclear weapons.

3:13:48Is that this is not the time to make jokes about my Pakistani heritage? Yeah, we'll skip that. But but is that is that is that really your play here is basically leveling the playing field? Yeah, absolutely. And I would take issue with a two two horse race. I mean, you got to remember. Well, it's more so I'm talking about that in the context of, oh, you're going to be able to just like, you know, Why would I buy something other than a Tesla if I could rent out my Tesla when I'm not using it? There's a section of the market that's excited about that. But I can imagine it's the world. I think that future is to be determined, right?

3:14:29And so let me answer the first. There's many. This is such a vast area. I'm sure we could spend hours on it. Let me tackle each of those individual points first. For the first one is, is it a two-horse race? Absolutely not. I mean, you talk about Super Cruise just earlier. every manufacturer globally is absolutely building some version of an L2 plus system or working with a partner like us. There's, without exception, the smallest manufacturers that you know are doing that. A good petri dish of you seeing how this will unravel is China. China is this kind of like swallled off ecosystem separate in the automotive industry.

3:15:04And there's no dynamic that says a winner take all will be effective. And I'll do the 10 second pitch on why I think an SDS from us would be better. We're taking data from all geographies, from many manufacturers, and not only that, from many vehicle types like off-road, defense, et cetera. And our models, therefore, are just will be and already proving to be more performant. So there you could have a pretty plausible technical view that actually the two legs are maybe the early entrance, but when everything being autonomous, it's going to be. And there's just no winner-take-all dynamics there.

3:15:41there's no winner to go the dynamics there. And then in terms of like the broader, what happens with autonomy next five to 10 years, like this is the autonomy moment. It's happening now. And we're, you know, when you're in the Bay Area, you kind of live just maybe, you know, Paul Graham used to say six months into the future, maybe we live three months in the future, but you do live a little in the future and you hang around in any street corner on San Francisco, you'll see Waymo's after Waymo, after Waymo. And then you'll start using them. And you see them eating at Uber's market share in terms of, right.

3:16:15So that's absolutely going to happen. What that's doing is getting the public used to this concept of self-driving. We in the tech industry are very comfortable with it. A lot of us are Tesla owners. A lot of us use FSD. So we're comfortable with it. You go to Michigan, I'm from Detroit. You go to Detroit, people don't know any of this stuff. And Detroit's a car town. You go to Tulsa, you go to Dallas. It's just not what people are talking about. So I think that the next five to 10 years is the general availability of autonomy. And so if you want to use a good analogy, mobile, we're in like 2008, 2009.

3:16:53If you're in the Bay Area, you saw that the iPhone is coming. People talk about it. But most people, so when does mobile really hit? Well, mobile kind of hits and like you could say it hits when WhatsApp or maybe Uber or maybe sort of like Instagram. You can't even think of before mobile and it requires an app store and requires a front facing camera. So like when will that self-driving moment happen? It's happening now. But I think I think the next two to five years, I'm not just saying I'm not just pitching my book, by the way. I believe it is like I think the next two to five years without driving is going to be massive.

3:17:25It's like the area of AI that everyone, I think it's just because it's not, you know, the hype happened 10 years ago. So everyone wants to talk about LLMs and they want to talk about humanoids. But it's like, this is the thing where it's, you know, when we started the company, the reason we're called Applied Intuition, by the way, is we want to be on the applied side of this technology. You can spend many years on the research side. It's a subtle dig in the name. Yeah, like that, like SDS. The reason people ask us, like especially engineers who are working or interviewing with us, they're like, well, why don't you get into like this space like four years ago, number one, transformer boom, you know, and the change had like the technology just wasn't very good.

3:18:05But also now, if you get into it too late, then like timing is a really, really important thing. And I think like we always wanted to see, hey, when are we exiting the research phase of self-driving and when are we entering? And the applied phase is to be very direct for folks who are not is when it becomes an engineering problem. And now the engineering problem is just about cost, cost, cost, cost, cost. Like, you guys are not asking me, is self-driving real? Like, that conversation in 2015, that's maybe the first question, is like, is this like a real thing? I know it's real. I want to know when will self-driving be available in a Koenigsegg?

3:18:38When will it be in the next Meghani Waira? When will it be in the next McLaren? Because this is, San Francisco lives in the future. Everyone there is already driving P1s, Mustang GTDs. I hate to be. For the people who don't know Koenigseggs are, which, by the way, they make like six a year. These cars aren't even crash tested. They're so infrequent. Ferrari, more of a mass car. I think it'll be a while. I'm challenging you. Sell your product to Koenigsegg. Get it into Koenigsegg. I'm throwing down the gauntlet. We have a large and public relationship with Porsche. sure they're big shareholders of ours as well.

3:19:18Yep. And whenever, you know, we deal with all the OEMs, I never want to expose any particular strategy. But if you if you think about it, if you're from a buyer perspective, if you're looking at a Tesla Model S, what's the other thing you look at? You look at Taycan. Sure. And a lot of times. So like, I think even the more exotic companies that are focused on driving dynamics, you You just have to think about this technology of good UX in the cabin and safety systems that prevent you from getting injured, I think will be like table stakes. By the way, AEB, another subtle revolution that's happened in the car business that people don't talk about is the automatic emergency braking stuff.

3:19:57Oh, yeah. It's crazy. It's statistically making an improvement in pedestrian deaths. Oh, yeah. That's wild. Things that often can talk about. So I think even the highest end stuff right now, everyone talks about EPA and emissions. So I think in 2035 and going forward, it'll all be just like, hey, why are we still having deaths? Yeah, like and then and then that's the question will come up as like, hey, it should be aware enough. It doesn't have to drive, but it should be aware enough to take over from the driver when something really bad is about to happen. Well, thank you. Last question. Just curious.

3:20:33We're keeping a series E company waiting just to talk. I was going to say, how do you think about financing the business in the future? It feels like I'm sure many, many SPAC sponsors would be begging, but I imagine we'll take the more traditional path. How do you think about it? Yeah, we've preserved 100 % of the capital we've ever raised, and the company's almost taken. So SPAC is probably... Hit that gong, John. Hit that gong.

3:21:07Yeah, for the finance folks at home, they know what that actually means. That is no easy feat. So no, SPAC is not. There's many things that I'm speculative about. SPAC is definitely what I can tell you affirmatively we wouldn't do. We fall into the very... I mean, it's not random that Fidelity and Blackstone and Franklin Templeton are investors in the company. I think in that way, we're very traditional. In the way we're not traditional, we're a very hard tech company. I mean, the stuff we build, there's virtually no other companies that do it. So, yeah, the future, I think the future is good.

3:21:39How are we going to spend the money? I mean, or how do we think about capital allocation? All of our values, the company's values can be reduced to radical pragmatism. If we had to raise a billion dollars to get into some specific technical area or fight that we think is worth fighting for, we'd do it. Like, it's pretty straightforward. That's fantastic. Well, congratulations. Always great to talk. We didn't talk about defense. I know, I know. Well, just hop on next week. You're talking about his time. Yeah, we could talk for hours. This is awesome. Thank you so much. Have a great rest of your day.

3:22:10Good to see you. We'll talk to you soon. We have been keeping Elise AI waiting, but we will let them in from the Restream waiting room. Some big announcements today. Good thing we warmed up the gong because we got a big deal coming into the TVP and UltraDome. Big news. From the Restream waiting room. Hello, how are you doing? So sorry to keep you waiting. Sorry to keep you waiting. On your big day? I was tuning in, I was learning stuff. The previous guest was talking about cars and we went down a whole rabbit hole, it happens. But thank you so much for taking the time. Would you mind kicking us off in the introduction on yourself and the company?

3:22:46Yeah, absolutely. I'm Minna Song and I'm one of the co-founders and the CEO of a company called Elise AI. And Elise AI is a company based in New York City. We build AI agents that automate the entire customer journey for two verticals, the housing industry and the health care industry. And we just raised our Series E today. Give us the news. How much? Yeah. How much? Yeah, we raised$250 million. There we go. Congratulations. Can you give us a quick history of the company? Tony, we heard from a friend what he gave us a general read on your revenue ramp, and it was pretty mind-blowing. So I would love to kind of get the quick history of the company since this is the first time.

3:23:37Overnight success. Yeah. Yes. Just started yesterday. Vibe-coded, too.

3:23:45No, we started 2017, started with my co-founder, Tony. was the CTO. We met in college, studied computer science, both software engineering backgrounds. I went to MIT. He went to University of Cambridge. But shortly after college, we wanted to build a company, build technology for industries that serve fundamental human needs. And that gave us this idea that we wanted to go into housing. And housing is our largest expense as humans. And we thought it seemed like it had limited technology as an industry. We didn't really know too much about it, but it felt like a good place for two technologists to start investigating.

3:24:27So when we got started, I took a job working at a real estate firm in Manhattan. And that was just sort of a research opportunity. I worked at the front desk. I answered all the phone calls. I greeted everyone, answered all the emails, and was just really listening for what is the biggest bottleneck that exists that we can start chipping away at. And it was very clear that was communication. And so we were actually a conversational AI startup right from the very beginning versus I think a lot of the companies that are adding on AI right now. It's been pretty obvious if you've been doing it for a long time that there's a lot of value.

3:25:05And a couple years ago. How long did you last at that job, by the way? Your research, your paid research opportunity. It was like a good three months. It was quite fun. I enjoyed it a lot. That is a wild start. How important is it to the company to be seen as an AI agent pure play? It feels like a lot of companies that are really, really successful are really adding a lot of value are actually not just picking the hottest tools off the shelf, but they also have a lot of traditional software and like SaaS. Some are bolting on AI tooling, but others are kind of offering elements of both, whether it's traditional dashboarding or just traditional business logic and workflows.

3:25:54How has the business grown? What's the shape of it? What's demand like? Is it, are customers receptive to something that is purely non-deterministic? Yeah. I mean, a big part of what we work on is to make sure that it's providing enterprise-grade reliability to our customers in housing and healthcare, to highly regulated industries, to just really consequential decisions for, if we make mistakes, it's a big deal for our customers, right? and for their customers, their consumers. So we focus a lot on making sure it is stable and it is reliable for them. But yeah, it's a mix of, I mean, you have to, you know, AI is a really powerful tool.

3:26:42It's a new type of computer, really. And you can solve problems with that. But also traditional software and other things that you need are also important in solving a problem. I think what's sort of different about what we build, we're a vertical AI company, right? We're not a horizontal AI. Horizontal being, you know, you can solve one sliver of a problem for a bunch of different industries. But our approach is we sort of go 10 layers deep into every problem. And that way we can actually solve that problem end to end. So it's very different. So let's say in housing, you are saying that you have a broken sync, right?

3:27:27It's very different when an AI might be able, you know, a large language model might be able to say, hey, thanks for letting me know. I'll submit a ticket for you. And maybe it does that. For us, our AI is actually doing a bunch of things. It's understanding how that resolution should, how that issue should be resolved. it's coordinating all the the workforce and all the resources the systems the compliance to get that ultimate issue resolved or there's things that we do that are not just pure sort of conversational ai pieces but we work in the physical world so we have to actually unlock these things called doors and allow people to go and get into the building so we have to do things like hardware integrations and you can't avoid doing that if you actually want to solve the whole problem.

3:28:13Yeah, on the integration side, what's working on the marketing side? Are there channel partners or platforms that you can plug into to accelerate growth? Is it just a ton of SDRs? Is there a viral component to this? What's accelerating the growth other than just the product delivering value? Yeah. We work in particularly non-technical industries, so So it's not really, there's not a great market for people that are just naturally searching for the latest technology. There's not like an app store that you can go viral on and then all of a sudden you're the number one enterprise app in property management.

3:28:52Yes, I wish. It is blood, sweat, and tears. We have outbound sales. We do have a lot of partners. AI is actually a great vehicle for delivering a lot of the other products that exist in the industry. So historically, if you needed, if you, let's say you were a marketing software for housing providers, or you were, or you have some other tech, you have insurance or something like that, you needed to train a human agent, a leasing agent, or a property manager to make sure that got that, that marketing content got distributed. And that's another manual step that people had to have. And I think a lot of people in the industry have realized, hey, I can just partner with Elise AI and now my marketing materials or my insurance will be communicated to everybody else.

3:29:46Makes sense. Jordy? Nothing for me. This is very exciting. Congratulations on the fundraise. Thanks so much for taking your time. Let us know when you add podcasting as vertical. We'll happily try it out. We're definitely in the market. Well, have a great rest of your day. Congratulations on the raise. We will talk to you soon. Thanks for joining. Bye. All right. Thanks. I wanted to finish on this post from Lucas Beyer, now at Meta AI. He says RFROFL. Rolling on the floor laughing. The solution it draws is amazing. This is from Tomer Ullman. So the question for ChatGPT for GPT-5 is, let's go back to this.

3:30:22And it's a picture of what you can go to the next slide. Is this the head of a rabbit or a duck? and the answer that Chachapiti gives is this image shows the duck rabbit illusion. It can be seen in two ways as the head of a duck with the beak pointing to the left or the head of a rabbit, the ears pointing to the left. So the answer is it is both a duck or a rabbit and then it says would you like me to mark the outlines for each interpretation directly on the image to make the two views clearer and it does. Ah, that's so clear now. And it draws it on there which Which is a fascinating ability to even trace the outline.

3:31:00This was unthinkable in machine learning just a few years ago. What does it see that we don't? Yes, maybe it's found a deeper truth. It's possible it's found a deeper truth. Or an invisible rabbit. No, this is hilarious because clearly it's been RLHF to know that people play pranks on LLMs all the time. They hit them with riddles. They hit them with unsolvable questions and trick questions. And that one of the things people demand of their AI chatbots is that they don't mess up when they're hit with a, one of the classic examples is a boy is injured, goes to the hospital, the doctor says, I can't operate on this child.

3:31:38And then the question is like, why? And the answer should be because the doctor is the patient's father or something. But you can throw all these subtle changes where it clearly indicates who is who and but it'll still think that it's being hit with the test, with the trick question. Well, one more post from Luke Metro, since we haven't covered him today. There was a young man, very young man, a teenager, joined SpaceX at 14 at 16. Kyran is joining Citadel Securities. Let's hear it for a teen prodigy. Luke says the hard tech is cool to you. Actually, money is cooler. We've seen it before. Or if he's really good, he'll get poached by a lab within the next year.

3:32:23Maybe. Maybe. And then one final post from NetCap. You can't pull Jordy Hayes out of the timeline. Sophie, NetCap girl, says, August is the Sunday of months. So here's to hoping that the markets just have the Sunday scaries. Oh, yeah. September is going to come around, and we'll be extremely back. But thank you for tuning in today. Hope you enjoyed the guests. Hope you enjoyed the show. We will see you tomorrow. Leave us five stars on Apple Podcasts and Spotify. Thank you for listening. And subscribe to our Substack, bpn.substack.com. Thank you. Have a good day. Goodbye.