Apple Unleashes $100B Buyback, Palantir Crushes Earnings, OpenAI's New Open Source Model | Alex Albert, Scott Kupor, Kyle Harrison, Brielle Terry, Sean Henry, Andi Duro, Gabe Pereyra, Kareem Amin

27:40Cryptocurrencies exchanged Coinbase Global Brian Armstrong in 2021 at$2.1 billion. And naturally, Elon Musk, the Tesla CEO, can boast billion-dollar gains in three years, including a record$43 billion in 2020. Though all are part of a pay package that the Delaware court declared invalid on Sunday, Tesla granted Musk a new interim stock award that was tentatively valued at$23.7 billion with the promise of more this fall. Barely scraping by. Scraping by. The other note on Palantir, I think this was reported on Friday, they have a new U.S. Army contract worth up to$10 billion. So this is consolidating a handful of contracts that they have into a single deal.

28:26and yeah, so pretty significant. That's over, I believe, the next decade. So Palantir Classic 20-year overnight success. Overnight success. Some interesting history and kind of strategy of how they got here. So we talked about this before, but you remember the famous story about Alex Karp talking to his employees early in the journey of the company. He said when you're traveling for business, I want you to book a five-star hotel when you're in town visiting a potential client and Gary Tan was commenting on this other employees were saying like this. We're start up. We're burning money This is kind of a crazy thing.

29:11What's the rationale here? And can and carp had a very contrarian take he said that when you're doing business you're going to meet a fortune 500 company that you're gonna be consulting with or or rolling out or selling them Palantir, selling them software, you're gonna be a forward deployed engineer, they will ask you, oh, you're in town, where are you staying? Just, it's just a natural thing that comes up during introductions. Oh, you're in town, you're in Dallas, visiting us at our HQ, we're evaluating Palantir as a potential solution provider, where are you staying? And if you say, the Holiday Inn, they will assume that you're selling Holiday Inn level software, Holiday Inn quality software.

29:46But, if you say, I'm staying at the Four Seasons, they will assume that you are selling Four Seasons level software. And so Palantir had this culture of, Karp has joked about steak dinners being not the correct way to sell the product, and the product actually has to be very good. But clearly from early on, Palantir focused on brand positioning and just taking travel and entertainment very seriously because there's going to be a lot of forward-deployed engineers that are on the road a lot. You want their life to be manageable so they can go and show up and perform and they're not gassed and exhausted from an economy flight and then a tough bet.

30:29Holiday Inn. So this was wrecking their financial stuff. When they went out and they went public. The narrative was that they were a consulting company masquerading as a software company. Yep. And looking at the earnings yesterday and their margin profile, this is very clearly not a consulting business. Yeah. So when they went public, you can look up the market cap when they kind of went out early. But I believe that they were languishing in the low tens of billions. Might have even dipped down to$5 billion at some point,$7 billion market cap. The stock was not doing very well. If you bought a single share at the open, you could have paid$9.20.

31:12And years and years later, you would have been flat on that entire period. So they went out October 2nd, 2020. And if you look on May 12th, they ticked over that initial price. So a lot of people were writing it off for a long time. And it still, it didn't, it wasn't like a fast IPO by any means, right? This was like, took them, what, 15-ish years to get out? Yeah. So 2020, 2021, they were in like the 40 billion, 35 billion. I think they probably dipped lower at varying points in time. But it wasn't, the stock was not ripping. And it was kind of in this, it was languishing. The consulting meme was going strong.

32:01And that was because they were spending a ton on travel and entertainment. But what's interesting is what happened during COVID. So during COVID, they couldn't travel as much. And companies didn't want forward deployed engineers coming into their offices. They didn't, like airlines were shut down for a long time. Everyone was working remotely. So they switched to a remote first, you know, forward deployed engineer strategy. And so by default, during an implementation of Palantir, you would just hop on a Zoom call or hop on a Teams call or whatever. And so their travel and entertainment budget like cratered.

32:40And that was really, really good for their bottom line. And then they were able to just kind of like hold the line. And sure, they added it back slowly. But it kind of revealed that this actually does have software margins. This actually is more of an enterprise SaaS company than a consulting business. And it allowed them to grow very very significantly and so they've been an absolute tear and congratulations to everyone over At Palantir well while we're talking about earnings Let's talk about linear linear is a purpose-built tool for planning and building products meet the system for modern software development Streamlined issues projects and product roadmaps.

33:15That's right building Ben and the production team are gonna pull up a post from Highlighting something from reddits q2 earnings call if we can pull this up. Yeah and play this video Oh, there's a video. That's fun. Okay, let's see. Let's play a video here. Let's see. This is the hardest thing. That's why Reddit is the number one most cited domain for AI across all models per data collected by Profound. That's why Reddit is the number one most cited domain for AI across all models. Anyway, so I wanted to highlight this because we just partnered with Profound. This is a company that I invested in earlier this year.

33:49James, CEO, has been on the show. Profound allows companies to understand and improve the way they show up in LLMs. So today companies are using LLMs. Everyone's using LLMs to research products, make business decisions. And so if you're running any type of business, whether it's consumer or enterprise focused, you've got to actually figure out where you stand in these rankings, how people are perceiving your products through the models. and Profound lets you do all of that observability. I think the best way, the best comp for it is kind of like SEO optimization. There's a bunch of different dynamics here, but Profound is the clear leader and we're super excited to partner with them.

34:34Let's hit the gong. Hit the gong for Profound. Profound and a new TBPN sponsor. Welcome to the lineup. Welcome to the livery. Welcome to the next drop of merch. Welcome, Profound. We're happy to have you on the team. Welcome to the show. In other AI related news, OpenAI has announced an open source model. They said they'd never do it. Everyone's saying it was in the name. It was in the name. How could you have doubted? Closed AI. Closed AI. They're not open AI. They're closed AI. Everyone thought they were so clever. They said they couldn't possibly open source a model. And they did. They did.

35:10They did it. The unthinkable. First open model in years, right? They used to have that. Was GPT one open source? Like when was the last time that, so we're gonna talk to Tyler about this. He's in the studio today. I wanna say GPT two was open source. And then they decided to close source something. It's tough that I treat you like things that I could Google and I know that you just have to search it or chat GPT it. But good luck. Get me the information on the history of open AI's, open source models. In the meantime, I will read this post from Sam Altman. GPTOSS, all lowercase. He's still in lowercase mode, although he uppercases the other sentences.

35:55But he threw in. He likes lowercase. Is he sending you messages, John? I'm seeing it. I'm seeing coded messages in here. GPTOSS is a big deal. It's a state-of-the-art open weights reasoning model with strong real-world performance, Comparable to 04 mini that you can run locally on your own computer or phone with the smaller size We believe this is the best and most usable my open model in the world We're excited to make this model the resilient the result of billions of dollars of research available to the world to get AI into the hands And for the most people possible hit that soundboard again I want to keep it going we believe far more good than bad will come from it For example, GPT-OSS 120B performs about as well as O3 on challenging health issues.

36:42We've worked hard to mitigate the most serious safety issues, especially around biosecurity. GPT-OSS models perform comparably to our frontier models on internal safety benchmarks. We believe in individual empowerment. Let's hear it for individual empowerment, people. I feel empowered the way that you're reading this, John. Although we believe most people will want to use a conventional service, a convenient service like ChatGPT, people should be able to directly control and modify their own AI when they need to. And the privacy benefits are obvious. As part of this, we are quite hopeful that this release will enable new kinds of research and the creation of new kinds of products.

37:20We expect a meaningful uptick in the rate of innovation in our field and for many more people to do important work than we're able to before. OpenAI's mission is to ensure AGI that it benefits all of humanity. To that end, we're excited. I think I cut this screenshot off. We're excited for the world to be building on an OpenAI stack created in the United States based on democratic values available for, I didn't screenshot the rest of that, but oh, here it is, available for free to all and for wide benefit. So I imagine that that was the voice that was in his head. As the author intended. Exactly.

37:58Exactly. Will DePue has more notes here, some context. So less than one year between 01 announced, which was September of 2024. And we have an 03 level model open sourced that's runnable on consumer hardware. That's wild progress. You were highlighting earlier, Sam initially had a poll. Oh, yes. Saying like, you know, should we, what do you want? Do you want an 03 level model or do you want something that you can run? He did both. And he did both. He did both. Wow. Yeah, so we were reading that. So it was what, like six months ago or something, that post that you shared? Yeah, I think so. So I asked Tyler to dig this poll up.

38:36Sam Altman shared a post that it was a poll on X saying, what do you want? Do you want an O1 level model, reasoning model, a frontier model, open source? Or do you want something you can run on your phone? And it was neck and neck. And people were clicking on the phone model. They were like, we want the phone model. And then Dylan Patel quoted it and said, oh, you idiots, don't vote for the phone model. You can just still get a truly frontier model. Let's get a frontier model. And so Dylan Patel was very much like, you guys don't know what you want. Like the phone models are available. That will be easy.

39:12The hard thing, we got to twist OpenAI's arm to open source the reasoning model, the O3 level model. And Sam Allman just said, you wanted this or this. It was a false dichotomy. You get both. You get both. You get both. Yassin says, all caps, I'm so sorry for doubting you, Sam Altman. I'm so sorry for saying that you were the antichrist. I didn't realize your plans were measured in centuries. I hope you forgive me for everything. I'm so glad you kept control of OpenAI. I'm so sorry. Please forgive me. That's remarkable. Never talked out on the future first ballot hall of famer, apparently. The future leader of open source AI.

39:48Will be very interesting to see how meta responds. Will they stay in the open source game? Also, just how important is open source? Is this a Linux scale opportunity? Is this an Android scale opportunity? Like is this, you know, is this GitLab to GitHub? Like what's actually the long term play? Certainly cool. Great for the community. We're gonna see fun experiments. We're gonna see interesting things done with this. I think, do we want to talk about some of the ideas that we had for building on top of this open source model fine tuning. I can just run through some. I know there was one you didn't want to share because you thought it was too good, too much alpha.

40:29But what was I talking to you about, Tyler? What were our ideas? We had one that was... Well, initially we weren't sure there was going to be reasoning. That's right. So we were going to add fake reasoning. Yeah, so it was going to be a model that gaslights you into thinking it's reasoning when in fact it's not reasoning. So if you look at the actual reasoning UI, it's just UI. There's no actual reasoning. It's a time delay. It's a time delay. It's a delay. And then while it's thinking really hard, what would a super smart person think right now? So we can actually still kind of do that because it's not a multimodal model, but we could just add some kind of totally blank encoder or something for the image.

41:12And then it's like, this is an interesting image. Let me think about what's in it. Yeah, one of the best images I've ever seen. A lot of people are saying this is one of the greatest images. Thank you for sharing this PNG and or JPEG with me. It just knows nothing about the image. It's great. I'm seeing that this image is large in file size. It's just purely metadata-based analysis. It doesn't know anything about what's going on. Like, wow, this is a tough question. I'm going to have to generate a lot of internal reasoning tokens to answer this. test time inferences really is magical, isn't it?

41:46Anyway, back to thinking. Okay, I'm reasoning about this now. Yeah, so our initial jerk. The other obvious opportunity, you know, people have been frustrated with sycophancy, with the models and models being... Well, not me. I've been frustrated with the lack of sycophancy. Yeah, you want more. You want more. I want more. You want the model to gas you up, but, you know, fine-tuning it to go, you know, over the top, somebody could release the Glazenator 3000, which just tells the user, you don't just think you're goaded. You are goaded. I know this. You know this. It's like I'm asking you to solve IMO question six.

42:19Why are you talking about whether or not I'm goaded? It's like it just tells you like you got this one. The timing here is interesting. I mean everybody's been banging the table saying where is the American open source AI or LLM leader? It was reported yesterday by the misinformation that reflection, reflection, the misinformation reported that reflection, one year old startup, we've had the CEO on before he's in talks to raise one billion to develop open source models to compete with DeepSea, Meta and Mistral. And yeah, all of the, you know, the Chinese open source models have been on a tear uh quen released quen coder recently yeah people are into quen um or a new version of it that uh is getting good results so it's important work so one interesting thing that's kind of related to the like us china like race is like so so when deep seek first came out everyone was like up in arms because they said like it's it was less than five million dollars to train right yeah so um in the in the model card for these models um they actually reveal like the GPU hours it took, which then you can like figure out how much it probably cost costed to you know train.

43:39Give me the number and tell me if I need to ring the gong for them. So for the big model it was probably around like four million. So just yes but but the small model was like 500k probably a little less but that's like 10x better the model is way better than Deepsea Gar 1. Okay okay so So should I ring the gong? I mean, I think it's pretty sweet. I wanted to hear$400 million at least. I was expecting$900 million. Nine didge. They got a billion of revenue coming in every month. But it's cool that it costs so cheap. Spend 10 days of revenue on the open source model. No, no, it is very cool. So we're ringing the gong for the elegance of the training run.

44:19Yeah, yeah. Efficiency. Efficiency. Congratulations to everyone at OpenAI. Congratulations on delivering a great model. saving time and money. Putting money back in the hands of shareholders and not in the hands of open source developers, I suppose. But yeah, I mean, do you think that, is this like a pre-training scaling wall narrative? Like, would you, as someone who is a potential consumer of an open source model, like, do you want a$400 million open source model? Would that necessarily be better? Because it seems like they were able to distill it pretty well, get it to the frontier, not burn a ton of money on training.

44:58Also, the question is if they're, this is probably distilled from their other models that are much bigger. So there is no world where you could just spend $4 million and get this level model without also having done the GPT-4 run, the GPT-4.5 run, et cetera, right? Yeah, maybe. It's also just like, even if they didn't distill it, they just have the knowledge of how they did it. Totally. which is arguably more valuable, right? Yeah, yeah, yeah. And I think in terms of open source, it's either you'll want to go super, super cheap and super small, which they kind of did here, or you go super big, almost like the llama, like the behemoth, right?

45:36Yeah, yeah, yeah. I think it would have been really cool to see that, literally state-of-the-art level model. You could also try and ask Claude, hey, build me a frontier open source model. Don't make mistakes. Yeah, we're gonna have someone from Claude on soon. This is funny because, wait, you said 500K for the small model? Yeah, probably less. So that's the same amount of money that they're putting up for this challenge, the red teaming challenge. Have you seen this? So to encourage researchers, developers, and enthusiasts, that's me, from around the world to help identify novel safety issues, This challenge has a$500 ,000 prize fund that will be awarded based on review from a panel of expert judges from OpenAI and other leading labs.

46:24At the end of this challenge, we will publish a report in open source and evaluation data set based on the validated findings so that the wider community can immediately benefit. So if you can hack this thing, if you can get it to teach you how to develop a nuclear bomb or take over the world, you might have 500K in your pocket. Not bad. Go get some of that. That's a seed round in 2012. Massive day for the labs. Anthropic announced a state-of-the-art coding model, Claude Opus 4.1, which one ups 4.0. We are having somebody on the show from Anthropic later today, Alex Albert, to break all that down.

47:07So I'll wait there. And then Google also announced Genie 3, a new frontier for world models. Yeah, Tyler, take us through Genie 3. What is it? Where are we in the deployment of this particular technology? Let me give a little bit more context on Genie. So given a text prompt, Genie 3 can generate dynamic worlds that you can navigate in real time at 24 frames per second, retaining consistency for a few minutes at a resolution of 720p. so I gotta be up at 60 frames a second if I'm pulling a no scope 24 true this isn't this is an a24 movie this isn't a this isn't cinema this is video gaming get it up at 120 frames I gotta be on 144 Hertz monitor yeah so Google deep mind yeah as soon as John can 360 no no scope at six you got a customer but for now you got a customer but I'm out keep pushing it yeah break it up what is this and and where to come from why is this important yeah so I thought this was really cool.

48:05I mean we've seen kind of like worse versions of this. It's not like a completely new paradigm of like basically you know real-time video that you can control. Yep. Oliver Cameron came on the show. He was at Cruise self-driving. I interviewed him years ago. He came on the show and he demoed this and the production team was kind of freaking out because it was like this very trippy like wandering around this like kind of Elder Scrolls landscape basically. We've seen the same thing with, I remember playing like a Minecraft version. Yeah, Minecraft, people have figured out how to do. Etched has a Minecraft MO2 that runs very quickly on some of their hardware.

48:42Yeah, but it's just like, I mean, if you watch the videos, it's like incredible. Like it looks, I think it looks by far better. This looks way better. Okay, so we're now in like, where like video generation was really, really crazy and then VO3 came out and it was like, oh, it's good. And same thing with like Mid Journey. It's definitely getting to the point where like you could use this for training as like a RL thing, right? Because this is like kind of, I mean, this is how you like justify spending a huge amount on this stuff. Because then you can use that as like training data. Yeah. Right.

49:12For agents or whatever. Yeah. Robotics. Synthetic data generation. Yeah. For robots. Okay. That makes sense. Yeah. What else? I mean, there is like the long term of like, you know, people do just enjoy watching AI generated videos. People get joy out of looking through Studio Ghibli's for a few days. So like there is value just in the inherent product aside from training. So you could imagine that this becomes like a game Let's pull up some of the examples I have them in the timeline. Yeah, yeah, yeah. Let's pull these up and try and like flip through those. But how far are we from like the actual game mechanics?

49:49When we played Oliver Cameron's one, I remember being able to move around with the arrow keys. that was cool. Next thing I wanted to do is I wanted to jump everywhere because that's like the default mechanic. I mean, anyone who's played any game like with a jump button, you're jumping the entire time. You're bunny hopping. Like this is not something you just wander around in. And so does it have a jump button? Do you know? I don't think it does right now. Wow. We're so far. We're so early. But you can. I mean, it's cool because I actually don't remember this exact one you're talking about, but you can prompt like all sorts of stuff.

50:22Right. So in the examples that we probably pull up yeah you can see them prompt in like okay a dragon flies in and they're in some like random yeah okay yeah let's take a look at this here okay so it's a bunch of this looks like an after each one of these is an interactive environment generated by genie 3. a new frontier for world models with genie 3 you can use natural language to generate a variety of Did they make this demo for us? And explore them interactively. All with a single text prompt. I wanted to take the horse off the ski jump. So we're looking through a helicopter. Yeah, it's for sure photoreal.

51:00Like, this is good enough. G3 has real-time interactivity, meaning that the environment reacts to your movements and actions. You're not walking through a pre-built simulation. Everything you see here is being generated live as you explore it. Jump. Jump, jump, jump. Press the space bar. 360, let's go. Hey, 360. We just got a 360. This is, that's AI? That's wild.

51:29Whoa, okay. That's pretty good. That's way less hallucinatory. And it's consistent. So it remembers that. Oh, that's big. Yeah, that's huge. We haven't seen that. Yeah, because when we were walking around in the other world, we would like turn around and they would just be like a cave and then you do a 360 and all of a sudden it's like it's gonna yeah it's like a church or something just horses there's a dragon dragon now the dragon looks pretty good you can use genie to explore this is very cool i wonder how expensive this is this will probably be extremely gated geography i'll have to beg logan to get me access and even other characters oh there we go that's a jump that's really okay That's really cool.

52:13Yeah. Yeah. Yeah. Yeah. Okay. Give me a top-down simulation of World War I. I want to play Hearts of Iron now. Give me a spreadsheet. I want to play Stellaris or EVE Online. Think about some of the use cases here. Spreadsheets in space. Some of the use cases here are insane. Yep. All right, we can pause it now. You can imagine a world in the future where you give Genie some images of you, like, with family growing up when you were five. And you could just, like, be, like, have a third, like, you could just, like. Very trippy. Generate, like, a synthetic memory. Like, this is where we're headed, where people are going to be having to kind of simultaneously hold their real memories and these sort of, like, simulated fake memories.

53:06It's just like I put on my VR headset, it's just like Christmas and now I'm super VR sick and I'm throwing up everywhere. Just like when I was a four-year-old and I ate too much candy on Christmas. It's very realistic. I think this would be really cool for Lama or for Meta to do some kind of similar thing that you can run on the VR. What do you think the training data set for this is? Do you think this is YouTube gated? It's got to be YouTube and it's also probably just like video games. I mean, you can see. So if you look at the old, this is Genie 3. If you look at the first papers, it's very clearly like, okay, that looks exactly like, you know, GTA V.

53:41Oh, really? Or in the first one, it's like some kind of very simple like 2D game that you can tell like, okay, that's a video game that they either built or found somewhere that they trained on. Yeah. Yeah, I feel like the pulling from the different video games would be cool. I wonder if you run into IP because you're basically like distilling GTA into a model. Yeah. Probably not okay. But you blend them all together. or you create something new, that's probably transformative and fair use. But I keep wondering, I was running the numbers and I think YouTube has like 20 exabytes of data, which is 20 million gigabytes or something, or 20 million terabytes or something like that.

54:23It's like, it's so much that you can't exfiltrate it, you can't put it in a suitcase of hard drives. is like it is just the storage is like a data center of its own as opposed to the common crawl, like all the web text is like, it's like you can put it in a server rack. Like it's not that much data. And same thing with like some of those image data sets. And I always wonder like does Google have like a permanent advantage here? Or will we see the other labs catch up on this pretty quickly? Like OpenAI seem to have caught up pretty quickly in image generation, Midjourney's doing great there. But a lot of the images are out on the open web, whereas a lot of videos are locked in YouTube.

55:07And if you can't get access to, if scale really matters there, Google has a serious, serious permaculture. Yeah, I mean, they have Sora, which is like, it's kind of very slow. The architecture is definitely different than what you've seen from Google. I don't even care about the speed of Sora. I care about the fidelity. So Sora, the latest Sora feels like VO2. and VO3 feels like a full step. I put the same prompt in both and SOAR is like way more hallucinatory. Like it's way crazier. VO3 still hallucinates and does like crazy things where like you'll be watching a car drive forward and then all of a sudden it flips around and it looks like the car's driving backwards.

55:42Like it gets confused, but just the visual fidelity is really, really high. You've done that in real life. Yeah, yeah, yeah. Flip the car around. Anyway, get on Numeral HQ, sales tax on autopilot. I'd spend less than five minutes per month on sales tax compliance, numeralhq.com. Anyway. We have a post here from Clouseau. Clouseau. Investments. How do you not know how to pronounce Clouseau? Do you know Inspector Clouseau? No. Oh, Lore. Is he French? Actually, I'm going to blank and I'm going to look really embarrassing now because I was trying to mock you. Fictional character. Fictional character, yeah.

56:19In the Pink Panther series. Pink Panther, that's right, yeah. Movie? Is that a movie? It has been adapted into a movie. It was originally, I think, an animated show. Okay, okay. So not my... You don't know the Pink Panther theme song? Dun-dun-dun-dun-dun-dun-dun-dun-dun-dun. Okay, it's coming back. Yeah, yeah. Pink Panther is real. Clouseau Investment says, I haven't done it yet, but I really want a short Uber. In New York City, there's a taxi app called Empower, which charges drivers a flat fee. In exchange, takes 0 % of the ride cost. All of it goes to the driver. If you're in NYC, I encourage you to try it.

56:50I've been using it for two weeks now, and every ride is 30 % cheaper than Lyft and 40 % cheaper than Uber. LGA to Manhattan is around$25, sometimes$50 to$75, depending on Uber or Lyft. I think Uber should be in the cash generation phase, but they somehow flush the$25 to$30 they make. They ride doing literally nothing down the drain. Bold statement. There will always be a market for Uber because some people value their time at$1 ,000 an hour, but for most people like me, waiting two minutes instead of one to save$30 is worth it. I still do not understand what the engineers at Uber, PayPal, etc.

57:22do because it seems like the marginal improvements they make are not worth their cost. Anyways, maybe I can sell a call on Uber or something because despite Uber and Lyft using lobbyists and dirty tricks to try to stifle competition, eventually instead of pocketing$30 from a$65 ride and giving the cab driver$35, someone like Empower will catch on and charge$40 and give it all to the driver and Uber. anyways, typo. But I think the interesting thing here is like, I can easily see something like Empower working in New York City. Can't imagine it. Yeah. High density environments where there's a lot of taxi cabs.

58:01I just think in the, I've never seen a taxi cab 10 square miles around my house once. And the question is like, even if the model works, if they're not taking a cut, then they don't have money to invest in onboarding both sides of the two-sided marketplace. And so the question of like, why did Uber and Lyft burn so much money? Well, it's because they were running insane amounts of ads, convincing everyone to be an Uber and Lyft driver. And so they brought all this liquidity to the market. And New York City has had a few of these competitors. There was another one for a while. There was one that was owned by the taxi.

58:40There's been a few. Some of the major... And yeah, I mean, it's very clear that like certain markets are going to have, you know, different ways to slice it. You go to some markets and I mean, when we were in, we were, I think we were in New York City, right? And Uber had clearly lost some fight with JFK because it was like, oh, if you want to take an Uber, just like get in a taxi cab, drive 10 miles. And that's the Uber pickup zone. And it's like, what? Like the Uber used to just come directly to the exit and act like a normal person. Same thing with LAX. Yeah, LAX has LA exit. call it black which i guess makes sense because it was really congested and it was kind of crazy and it was it was like actually probably net worse for everyone but i think a bigger point here is just what what is why is the company not generating obscene amounts of cash when when they're when their margins uh feel like they they should they should be able to produce uh we're going to be having david uh risher from lyft ceo of lyft on i think later this week post earnings a bunch of questions for him around autonomous vehicles and we can maybe bring up this post as well we'll see yeah see the rebuttal um well let's tell you about 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 that's right go to fin.ai to get started take him has a great post here from jeffries apparently this looks like it was in some isn't this hilarious from Jeffries.

1:00:08This is hilarious. Like memes are, I've seen memes in VC decks like oh we're doing a quarterly update on the market and somebody throws a meme in there. Jeffries is a very serious investment bank and they're throwing an AI generated image. It says a lot. You see the source? Source Copilot. You know what that means? They made it? The AI image generation they used is from Microsoft Copilot. Everyone thinks oh everyone uses ChatGPT. Not Jeffries. They use Copilot because this is an Excel shop. That's right. Hands off the mouse when you're prompting in Copilot to be a Jeffries associate generating this.

1:00:46But it's a funny meme because Azure and GCP are dancing. Meanwhile, AWS is a sad puppy in the corner. You know, we talked about this with AWS, like sort of lagging, not having a strong, you know, horse in the AI race. Google obviously has a fantastic foundation model lab. Azure obviously is the open AI bet. and is growing like crazy through the AWS OpenAI business. Of course, Amazon has Anthropik. They do, yeah. And the work they're doing with the TPU there. Somebody was saying that Tranium is underperforming or people don't like Tranium as much. They prefer just give me a ton of NVIDIA GPUs. But who knows, never count out the goat, the man who invented the cloud, the cloud cowboy, Andy Jassy.

1:01:30The man himself. I mean, truly in the age of AWS being the most important driver for Amazon as a whole, like the perfect person to run the business is Andy Jassy, right, because he created it. So they're in a pretty good spot. We have a post here from Joe Weisenthal, flying first class on my flight out of New York City. I was a little bit surprised here. I expected him, I expected Joe to be flying private. He says, everyone in the front is solo, quiet, and isolated. I look back to the rest of the plane. People are joyously traveling with friends and family, sharing humor and adventure. It was just an interesting observation.

1:02:03Well, I mean, I understand why I understand why people in the front are so low, quiet and isolated because first class flights have become extremely draconian. Yeah, they really have. Terrible. We should share our experience. So we flew a mixture of first class and economy to New York City with with our lovely team and crew. And Jordy were Jordy and I were up in first class and were served. We were flying. Should we name? Yeah, we're going to name him. We're going to name names. We're going to name names. We're flying JetBlue Mint. Which historically I've had great experiences with mint over the years.

1:02:40David Segler was singing their praises, said it was the best first class experience he's ever enjoyed and said it was fantastic. At times it's been fantastic, but this was... We had quite a different experience. This was devastating. Let's run through our problems. So the seat was comfortable. Lay flat. It was great.

1:03:24You didn't think it was comfortable? of our nightmare or nightmare flight. I ordered steak and I wasn't hungry. So of course, as soon as the flight attendant left, I got up and I walked the steak dinner back and I gave it to Michael and Scott. And then I went back to my seat and I was just relaxing. And the flight attendant came up to me in a little bit later and said, would you like dessert? We're serving gelato. I think it was. And I said, sure, I'd actually, I would love gelato. Thank you. and the flight attendant went back to the front. I ran to the back, and I started handing the gelato to Michael, and a hand jumped into frame and grabbed the gelato.

1:04:07Grabbed the gelato out of my hands and said, You cannot do that. And, of course, there's a rule that says you can't pass food between cabins. But I was just shocked because I was shocked. Makes no sense. Possession's nine-tenths of the law. Once I take possession of the steak, it's mine to do what I want with. I think I should be able to distribute the steak evenly amongst the entire first class cabin. I should be able to auction it off. It's mine. I should be able to resell it on the secondary market. Whatever I want to do with that. Yeah, there should be a stock expert. Either you gave it to me.

1:04:37Yes, this is what we're going to do. We're going to vibe code an app that allows you to sell first class meals to economy customers. And it runs on the JetBlue Wi-Fi. Okay, so credit to the flight attendant. Yes. I it could be annoying if I'm distributing the the the plates and the forks and the things like that around the plane It's annoying. Maybe maybe it's it's it's frustrating to the other other other clientele But I was more on your side. So anyways, I brought the gelato back to back to the front I gave it to you. You ate it. You enjoyed it. It was great But then after that we looked we did a quick deep dive on JetBlue and I got very concerned because JetBlue Nothing more shocking than riding in a metal tube at 30 ,000 feet and realizing you're flying a penny stock.

1:05:27Basically. So JetBlue is now trading at... What's the ratio between JetBlue and Perplexity? I was just going to say, their market... Is it bigger than Perplexity? Because I feel like Perplexity is a pretty solid business. They're over 10 times smaller than Perplexity. They're less than one-tenth Perplexity. Oh, okay. Yes. Okay. and their entire market cap is less than Mira Mirati seed round. Wow. And so usually if I was going to get in a tube, a metal tube, and fly thousands of miles. I would trust Mira for sure. I would, you know, I'm starting to think, you know, optimize for market cap a little bit more on the provider.

1:06:07But anyways, Joanna, CEO, welcome to come on anytime. Get down to the TBP at Altradermen. defend your first class policies. I was talking to the flight attendant and I said, I'd like to speak to the manager. I'd like to debate the manager because I don't think, I'd like to debate the manager. You don't want to speak to the manager. You want to debate the manager because they have rules. They're going to pull out some rule book. I want to go a level deeper. I want to go from first principles. Why do those rules exist? I was thinking, I was thinking, you know, that, that company Pipe Dream. Yep.

1:06:42They make tubes that run underground that you can transport goods. I was thinking if you could create some type of advanced straw that you could run from first class, put it, run it on the side and then run it back like telescopically. Yes. Like back to your boys in the back and then just feed them and use like high pressure air to feed them little bites of steak in the back. And so you could do this. Yes. You could do this without anyone, potentially anyone knowing about it. Yeah. Passing back scraps to your boys is like the it's so American. I was I was deeply offended. And then it got worse because the flight attendant kept bringing different trays of snacks.

1:07:20And remember when they brought the one tray of snacks? They're like, do you want any of these? This was after the after the steak debacle. And we start taking them and we're clearly taking more and more than we need. Really? Yeah. Just give us the entire tray. and i asked i asked specifically i was like can i bring these back to my friends and she said technically yes because they were for the whole cabin these stacks are cabin less than 10 minutes later she brought some nuts yes that were that looked very nice yes and those you said you cannot bring you not pass the nuts back and you can pass the chips back anyways very very frustrating yeah it was it was brutal it was brutal anyway we had some fun um ben ben definitely said at one point you know we're flying with like 10 people shouldn't we just shouldn't we just charter a jet and he made a good point yeah i made a good point um well you will have to get on addio customer relationship magic addio is the ai native surround that builds scales and grows your company to the next level you get started for free at addio um kevin quok has a screenshot from linkedin this He braved LinkedIn to bring us this.

1:08:28He says, this is the funniest LinkedIn title from now on. People should only use LinkedIn like this. It's great. And Peter Sellis over at Discord in the title of his LinkedIn. It says, I have full purchasing authority over Discord's extremely large and lucrative cloud contracts. And I am presently leading their renegotiation. Basically, putting up a billboard. Just getting exactly what you want. I want to talk to the most senior person possible at every hyperscaler. Take me seriously. Great. His username on LinkedIn is disgruntled. Disgruntled? Wait, Peter Salas? And his title, if you go to the experience section, is personality hire.

1:09:10Personality hire. This guy has a good sense of humor. Prior to this, he was vice president of product at Snap. Very cool. Good stuff. University of Chicago. Fantastic. Come on the show, Peter. We'll do a live auction. Orin Hoffman's connected to him. Orin's been on the show a couple of times. We should do a live auction with the hyperscalers. And you should get on public.com, investing for those that take it seriously. Multi-asset investing, industry-leading yields, trusted by millions. Is Discord out yet? No. They're about to go public, right? That seems to be the case. I think that's the story.

1:09:41I haven't been following too close. The founder was removed. Yeah, I know Delian was saying it was the wrong move. But regardless, love Jason and what he's built. Fantastic. One of the coolest founder stories. Went to a college for video game design, I believe. Started an app, sold that, and then was building a video game. Built a League of Legends for your phone, like a MOBA. And took that to TechCrunch Disrupt and was demoing it. Was just super into the gaming world. Eventually got lucky with Discord and found the correct product market fit. Scaled the moon. Put the KPIs in the orbit. The building games to building dominant messaging platforms pipeline undefeated under slack discord Paraclete's in the chat says what kind of country is this we can't even pass back little stakes to our boys No, I agree.

1:10:30He also says I love Adia. Thank you Welcome to the same way good to have you here Good to have you let's talk about Scott Wu. There was a report about cognition He is replying or kind of sub tweeting the the coverage Basically, 30 out of the 200 people that were acquired from windsurf got laid off, and then the rest got a buyout offer. If you're not interested in staying on the boat, you were on a go ship. Now you're on the Titanic. No, not the Titanic. You're on the - Carnival. Carnival cruise for cracked IMO. No, I was going to say Carnival cruises. It's a destroyer. Yeah. Battleship. You're on a battleship, and everyone's got to be pulling in the same direction.

1:11:11Yeah. Everyone's got to be hoisting the sails on this galleon. And so Scott Wu says, people have asked about our culture and recent employee communications. Cognition has an extremely performance culture. And we're up front about this. This is Scott Wu's. This is my Scott Wu impression. We are routinely at the office through the weekend and do some of our best work late into the night. Many of us literally live where we work. That's true. They have a house in Silicon Valley somewhere and a bunch of them live there. we know that people who joined windsurf didn't expect to join cognition and we're proud of how we work we understand it's not for everyone we gave our team the opportunity to decide we offered for all employees who joined via acquisition to opt into our culture with full clarity on what that entails we know that we will lose some strong talent in doing this but we truly believe the level of intensity this moment demands from us is unprecedented while not everyone is looking for a culture like ours everyone deserves respect and appreciation for their work regardless of their decision we accelerated and cashed out all four years of equity for everyone at windsurf even for the 85 of employees who didn't hit their one-year cliff and for those who opt out we're providing an additional nine months of pay on top of this uh this is just good guy scott he's just doing everything right he's he's setting the bar high he's letting people know hey we take this really seriously we just don't want there to be a culture clash look yeah you got one of these if you got if you if you just went from almost acquisition from open ai yeah ended up on the ghost ship through the google deal ended up at cognition got all your got got the full upside from the deal yeah even if you had invested and then now you're at cognition and realize i don't want to work as hard as everyone else here yeah and you get the opportunity you get nine months of pay to just walk away.

1:13:04That person is, there's going to be some people in that group that are. Totally. I mean, it's not like the norm, but there are definitely people who are like, I was actually excited to rest and best. I was actually excited to hit the ping pong table. You know, I wanted to just be on the beanbag chair for a while. And I wanted to kind of cruise and have like a normal nine to five. Like there are people who like that's what they, maybe this was their third startup. They already got some liquidity. They're doing great. and they have a family and they want to be a little bit more flexible, take more time off.

1:13:37That's the beauty of the American labor market is you can kind of choose and pick your own adventure. But Scott wants to create a cohesive team, a cohesive culture. He doesn't want this idea. It's very tough when half the team is staying up really late grinding and they love it, and then the other half is like, I'm dipping out early. It's just awkward for everyone. So he's just saying, hey, Like, you're welcome to stay on the team, but you got to buy in or you got to bounce and go, you know, take your liquidity from the accelerated investing you got. Take the nine months of extra pay and go find something that actually brings you joy.

1:14:16Go do your life's work. So I think this is well managed. It seems like the comms leaked prior to this. I don't know. I don't know. Would it have been better if you posted this beforehand and just kind of owned the narrative? Maybe that would be the move because you kind of know that people will frame this poorly. And so you get out and say, hey, I'm going founder mode. I'm offering this buyout offer because there are plenty of companies that have done these buyout offers before. I believe that there's one major company, I forget which one, that has a standing offer at 24-7 every single day. Like it doesn't matter if you just joined.

1:14:48I mean, of course, you can't just like join and leave immediately and get the buyout offer. But basically, like if you hit a cliff and they're like, we'll re-up you, we'll keep you incentivized, we're going to keep paying you. But like the buyout offer is always there for you to take some cash, go figure out what's next, find a better fit, because we just don't want people who are just like hanging on to this job because they are just hesitant about going somewhere else when they should just figure figure out the next thing. Yeah. Anyway, you don't want to be losing sleep. Get a pot five ultra.

1:15:17Go to eightsleep.com. Five year warranty, 30 night rescue trial, free returns, free shipping. I got a great night's sleep. I slept for eight hours last night and I think it's showing. Didn't you say eight and a half? Something like that. You put up some crazy numbers. I got a 95 last month. I've been down in the dumps. I had a 45 Sunday night. No, you're back in the game. Slept four hours. The energy's coming through. Just absolutely brutal. And you can feel it in the intro. Yeah. Sleep is great. Sleep for you is inversely correlated with your voice. Yeah. The more sleep you get, the more hoarse you will get from screaming into the microphone.

1:15:50You can tell my sleep score by the intro. Yes, exactly. Skooks says after they put GLP ones in the water being obese will be back in vogue a sign of indomitable Vitality and enormous. This is not obese. This man is not obese. He is prosperous Prosperous and prosperity should be celebrated. Yep. It came to it's a it's it's a wild We are next our next move bringing back canes Canes smart canes it's gonna happen a lot of battery life in a cane plenty of room for a ton of batteries A lot that you can do with a cane. On-device inference, GPT-OSS, run it on the cane. Run it on the cane. That's really the ultimate status symbol.

1:16:30Does NVIDIA make a narrow 4090, like a GPU that could be put in? Maybe we just need to string together a bunch of iPhone chips. I really think building a smart cane might be a challenge. I don't know if you can. I mean, I guess the chip in the iPhone is probably small enough to fit in the cane, but you'd have to figure it might be it might have a thick part to the cane well yeah you could make a very wide cane right almost like an elephant's you know leg that seems like not the vision i'm going for with the cane i want i want an aesthetic cane i want a cane that that it doesn't jump out you doubt well doesn't it doesn't jump out of anyone we'll put tyler on it tyler just just try to build a smart cane using open source llms please thank you uh and then when you launch the cane?

1:17:15Buy a billboard for it on adquick.com. Out-of-home advertising made easy and measurable. Say 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. Let's go to soccer and teddy. One of our billboards was taken down this morning. Was that by a - After a good run. By a state actor, do we think? Probably. Probably. Some type of, yeah, nation state hacker group. Yeah, or was it a false flag? You'll never know. Zach Pogrob noticed that it had been taken down. I was very disappointed.

1:17:51So working on getting more up, Zach, we're on it. Maybe SF next time. Maybe, I don't know where else we should go. Probably Saint-Tropez. That makes sense for a billboard. Abilene. Abilene would be good. Abilene would actually be great. Abilene would be the way. Audience of one. I just want Chase. and Sam to see it basically. Audience in two. There's also a lot of people in Abilene that are just getting into data centers. The billboard in Abilene should just say, it shouldn't even say listen to TBPN, it should say come on TBPN, call into TBPN. Sam, call into TBPN, we'd love to chat with you. If we put that up, I think we get him.

1:18:31Okay, he'll of course come on the show. Welcome to Data Center USA. Yes, yes, congratulations. Sagar and Jetty says, I can't help but love every iteration of the moon swatch. Yes. What do you think of the nation between Omega and swatch? Is that the same company? Does swatch group own Omega Omega? I always listen to the James Bond pronunciation the Daniel Craig He's so mega Omega does swatch group own Anyway, looks nice history of the moon swatch the moon watch was what it was an Omega It was worn by the astronauts same company makes a bunch of watch group Same company, yeah. Very big, very big company.

1:19:12So if you're not familiar with the moon watch, there was a knockout, drag-out fight. It was like the AI talent wars all over again in the 1960s to decide which company would send the watch to the moon on the wrist of the astronauts. Everyone expected it to be Rolex. Omega won the contract. The Omega moon watch went to space, went to the moon, and came back. And now it's a celebrated timepiece. The ones that went to the moon are obviously in museums, and the models from that year traded a very high premium. But they issue new moon watches regularly, and the moon swatch is something that's a little bit more, I think, affordable, certainly more rugged, and has a nice fabric strap.

1:19:54It will release Saturday, August 9th, and it's priced at$400. Oh, okay. And if you're looking for your own watch, go to getbezel.com. Your bezel concierge is available now to source you any watch on the planet, seriously any watch. You can get an Omega. You can get something worn by Mark Andreessen. He's an Omega guy. Little bit of a... Yeah, we clocked that. We clocked that early. Came out with an interview. We were expecting... What's on the wrist? Yeah, we were expecting a Graff Diamonds hallucination. Maybe four of them, ideally. But, you know, he went with the budget option. Yeah. But we'll get him.

1:20:29He's coming on the show, I believe, tomorrow. And we will pitch him on picking up a couple Graff Diamonds hallucinations with the pennies in his couch cushion. He's got a lot of real estate. Everybody's got a lot of real estate. A lot of people, I don't know what's up with people just wearing one watch, especially allocators. Exactly. His caliber, right? Why stop at one? You've seen a couple of people that wear a watch and then a whoop band or a watch and then an Apple watch or something, but why not four Daytonas? Anyway. Something there. Congratulations to Joe Weisenthal again on the absolute run.

1:21:00Odd Lots is the number one business podcast in the world and the number six podcast in the world after Cold-Blooded from ABC News and Crime Junkie and the Joe Rogan Experience is sitting at number three. Mick Unplugged, Leadership Because Growth, which I'm not familiar with, is number four. And The Daily from the New York Times is at five. And to be clear, this is not because our episodes of Odd Lots released yesterday. It is because they got a profile, a glowing profile in the New York Times on Sunday. Yes. They're on a generational run. We love to see it. Yeah. And Joe said the business category of the Apple charts is very funny.

1:21:43It is funny because money rehab, Andy Frizzella, Real AF, like a lot of like I think Jocko Willink is in the business category. And it's very different than the odd lot stuff. Like odd lots is like, let's talk to an expert on the grain market. I'm like, let's see what's going on with the yield curve. Yeah. And it's very much listened to by hedge fund folks and capital allocators. And a lot of these other podcasts are listened to more by lifestyle entrepreneurs and small business folks. And obviously, that's a much larger audience, much larger addressable audience. But it's good to see odd lots on an absolute tear.

1:22:23And Joe says, okay, going to stop with self-congratulations, but fun to be number one for what What will be a brief moment, surreal week. Also one last thing, if anyone knows how I can parlay my 15 minutes into throwing out the first pitch at a Cyclone game. I have no idea what the Cyclones is. Is that some sort of, is that sports of some sort? It's some, I've never heard of this. High A affiliate, the Brooklyn Cyclones. Okay. High A affiliate of the New York Mets. This is, so this is baseball? Yeah. And that's the one with, where you throw the ball and then you hit it with the. Is this like a feeder team?

1:22:58This is the one where you throw the ball and you hit it with the bat? And there's like gloves too? Great. Well, good luck to Joe throwing out the first pitch. That would be great to see. Good content. Nikita says, things that will cause a user to churn. One, friends leave. Two, product stops growing and becomes culturally irrelevant. 998 feed has a bug where it flashes. 999 bookmarks are not searchable. you have to imagine uh nikita's dms are are potentially gnarlier than than elon's at the moment of uh thousands of his uh mutuals and friends messaging him every possible bug in the app i know we would do the same to uh jacoby before he left i mean i was just talking about this like we don't use the bookmarks feature we use the shared groups feature to kind of track posts that we want to share on the show and i was like oh i wish i could search across all of these But I was like, I'm literally the only person that wants that.

1:23:52I'm the only person whose entire job and career depends on this one niche feature. And so it's not even worth requesting. I'll just figure out a workaround, hopefully scrape it out of there and then make that searchable or something. Because I can work on a workaround. But yeah, fortunately, friends have not left. And X continues to be culturally relevant. It's where news breaks. and we saw that on full display with the Soham Parikh saga. The entire thing played out on X. Started with a post, wound up with everyone else chiming in on various posts, kind of ended with us interviewing him live on a stream on X.

1:24:31It was very cool, all within the X ecosystem. You know, stuff still happens elsewhere, but in terms of where the tech conversation is happening, it's entirely X, there's no even close second. Now, I do think it's interesting to see, I want to get better at understanding like what is the current thing on Instagram today? And what is the real community that is formed on various other platforms that are adjacent? Because I know some people that have grown accounts and when I was big on YouTube and really focused on YouTube, I would notice things, like we were talking to Samir yesterday about like different metas that develop and different themes and sometimes these break containment, but a lot of times it's just a story that happens within the YouTube ecosystem.

1:25:25Somebody posts a video and then there's a reaction to that video and there's a whole cycle of news that happens within the YouTube ecosystem. It's interesting, X still feels, you know, if you're generally interested in tech, it feels like a shared experience. Totally. Whereas if you're interested in tech and you're on Instagram, It doesn't feel like you have any feed that feels at all similar to even another friend that's like loosely interested in tech. Yeah. I'm trying to think on YouTube, the big discourse is always about YouTube. Like Colin Samir really are at the center of the discourse on YouTube.

1:26:01Creator industrial complex. Yeah, and I think a lot of that has to do with the fact that on YouTube, you have to have a YouTube account and you have to publish a YouTube video to have a voice whatsoever. Anyone can show up on X and type out 280 characters, go viral and be the main character, be the current thing, start a conversation, engage with that conversation. But the bar on YouTube for having a voice, the comments section just doesn't rise to the same level as a reaction video. And so when Mr. Beast is embroiled in controversy, there will be other YouTubers. Like there was this, the Mr. Beastification of YouTube.

1:26:39This idea that Mr. Beast had created this meta, and then other YouTubers were copying it, and it was all this like yelling at the camera, and fast-paced editing, and lots of like in-your-face, high-saturation, graphics popping up, all this different stuff going on. And that entire news cycle played out on YouTube, and it was because there were commentary channels and cultural analysis channels that replied, essentially, but instead of it being like a quote tweet dunk, the equivalent is responding. Another viral video. Another viral video responding to the analysis. So someone put out, so Mr. Beast was on a tear and then other people were copying it and then somebody put out the Mr.

1:27:21Beastification of YouTube, which went viral, and then other people would respond to the Mr. Beastification is overstated or it's not that big of a deal Or it's even worse than you thought. And so people will take the conversation in different ways, but it moves much slower because if you're like, yeah, I want to do a reaction video, okay, see you in a week, right? Yeah. Anyway. Find your happy place. Find your happy place. Book a wonder with inspiring views, hotel-grade amenities, dreamy beds, top-tier cleaning, and 24-7 concierge service. It's a vacation home, but better. Is that gong new? This gong is relatively new, a couple months since our new studio.

1:27:52We do have another new gong over there that we haven't that we haven't actually showed anyone And when we got a completely new gong for the figma IPO and we had dylan field hit it on the day of his IPO It's game hit. It's a game IPO gong it will be retired it will hang in the rafters of the TVP and ultra dome and eventually this eventually this dome will be filled with gongs from historic Technoric technology moments. And we're gonna give the Computer History Museum a run for their money. They might have Commodore 64 or whatever. They have the original like, touring machine. It's like this computer that you crank.

1:28:32It's like the craziest thing and it can do calculations. And it's like, it's literally no electronics and it's a computer. It's all just like, basically marbles flowing around. Not really. Dream rig. Yeah, dream rig. Imagine posting on that. but we're going to give them a run for our money with our gong collection anyway 11 labs has a new launch today can we pull this video up somehow yeah I want to listen to the 11 labs music they're saying 11 music is here created with leading labels publishers and artists it lets you craft studio quality tracks with control over genre style and language using just text built on licensed catalogs every song you create is cleared for commercial use that is pretty insane i had an eye-opening moment probably two years ago at this point uh one of uh i enjoy the music of the rapper gunna sure and somebody made an entire album a gunna album that was entirely ai generated was it good and it was solid it was solid it wasn't necessarily i i listened through it a couple times you could if you paid it if you were not paying attention to it it sounded just like a regular gonna song album did any of those was on Spotify and I think they published on Spotify but I did any of those did any of the songs migrate over into a playlist for me you're saying crossing the castle you liked feed or it's in your it's in your spotify spotify is in an interesting place here because you're one way to piss off artists oh is allow people to generate AI versions of the same artists and the artists are losing revenue.

1:30:11And I knew a team, a friend that was working on basically Spotify for just AI generated content. They ultimately pivoted because they didn't feel like they wanted to withstand decades of lawsuits. But it sounds like 11 Labs has found a solution here. Let's play this video. Yeah, let's play the video. Let's see what they have to say. Hopefully we don't get copyright striked. I think we should not because this is the first time this music has existed. Okay, so we got some orchestral music. It's pretty epic. I told the founder to make us a WWE theme song. A really aggressive over the top track for us.

1:30:57This is not quite our vibe, but it's nice. It's nice.

1:31:05Definitely past the uncanny valley like this could just be a real song

1:31:17Okay, let's see did they do any gonna Okay, okay, you got some fast-paced elevator coated, but yeah What else we got

1:31:35Their customers must just love like falling asleep. Reggae. Yeah, this is pretty cool.

1:31:44Give me some heavy metal. Okay. Some hip hop.

1:31:55It's pretty sick. I like this one.

1:32:00This is epic. I want this song. I'm sold now. Oh, this is good. It's hit. That's great. Get me that track. I want to listen to that track. This is a good time to talk about, so last week the Panama playlist came out. John and I were featured in them. Doxxed. We were targeted by the Panama playlist. If you go to panamaplaylist.com, they basically scraped a bunch of people's Spotify's without them knowing and then published their playlists. I was wondering, I was looking at this playlist and I didn't immediately have a memory of it. The playlist was called No Cap. And I was like, why would I name a playlist that?

1:32:41And it took me a second to come back to me. It's because I was collecting songs, rap songs with obscure references to venture capital and investing. As one does. As one does. And so if you look at the first song, God Did. Clearly for a party round drop that never happened. Never happened. But first song, God Did, DJ Khaled, or one of the other rappers on the track, says, I'm at the cap table. What the splits is. Not that cap table. Boy, we live this. And then you have OK, OK, Kanye West. Price went up. Angel Investor. And anyways, a bunch of other random references. But it's all making sense now.

1:33:26What else we got, John? We got GPT-5 news on Polymarket. GPT-5 released by August 10th is now sitting at 92%. It was sitting at 75 % yesterday, earlier this week, 80%. And so by the end of the week, there is a 92 % chance that GPT-5 is released. And it has mooned recently. just back on August 3rd, it was sitting at 31 % chance. So it has just in two days gone from 31 % to over 92%. So people are definitely expecting GPT-5 to be released in the next five days, which should be exciting. I'm sure there'll be lots of coverage, lots to talk about, and we will want to hop on with as many legends, as many Metis list AI researchers as possible to get the deep dive.

1:34:25Absolutely. In other news, Flying Private is the new Wealth Yardstick. I know our audience is going to want to know about this. We talked about Joe Weisenthal. Is this news? This is news. Is this new news? Is this people hearing about this for the first time? But I love this because do you know Max Chuning? He started a gummy candy business, sour candy business. He sold it to Hershey for$75 million. I just love the name, Max Chuning with two X's. It's like a great name. So he was a YouTuber. He started a sour candy business, Sour Strips, and just grew it and grew it and grew it and wound up selling it for$75 million and did very well.

1:35:00But the first thing he did before buying a Rolex or a dream home was a private jet for his wife and six friends to veil on a Dessault Falcon 900. They skipped security lines, zipped straight to the runway, and seated themselves in leather recliners with gold accents in the wood-paneled cabin. The price tag for this adventure,$100 ,000, worth every penny. go tuning golden doodle dude this guy is phenomenal with the names artist love it uh sprawled at their feet the joke is i had to get a private plane so i could bring my dog the 35 year old said i didn't really care what the price was the ultra wealthy have always enjoyed flying private that exclusive club is growing as soaring stocks and crypto prices meant more millionaires and billionaires who now have a range of choices to book a seat on a jet and you know the best part?

1:35:48He can give the steak to his dog if he wants because he's flying private, baby. He doesn't have to deal with the draconian rules of JetBlue Mint. Flying private has become the ultimate luxury splurge for many wealthy individuals, surpassing Ferraris, Hermes, Birkin bags, topping$14 ,000 or even waterfront Hampton homes. Why not? All the above. Which we recently learned are very interestingly priced where property in the Hamptons is somewhat reasonable. you know, five,$10 million might get you in the game, but renting a house in the Hamptons for a single month during the high season,$250 ,000, right?

1:36:26Something like that. That's what we were saying. Very, very expensive. And the rental rates are so extreme because you really only want to be there. You only want to be there during the summer, the hot season when Emily Sundberg will be out there. For many of those aspiring to join the ranks of the truly rich, having private jet money is the new goal dividing the 1 % from the 0.1%. The pandemic unleashed a burst of demand and providers say popular culture has turbocharged enthusiasm and envy for the fly private lifestyle. Social media has given, I mean, Swift to A10. Czar in the chat says his friend flies their dog private solo.

1:37:07Well, we talked about on the show earlier a dog that flies private alone because the owner signed a contract to do the interior design for the private jet. And as part of the deal, their dog gets to fly private. And we were laughing about - This is one of the worst deals. About like, what if you want to get to New York to Miami and the dog wants to get to LA and you have to do a layover at Van Nuys to get to Miami because the dog has to fly private. but the younger people are getting a glimpse into the lives of jet setters whether it is a model flying with friends to a bachelorette party in los cabos mexico or a hedge fund manager hopping a plane for a birthday in saint bart's realistic expectations are in or in order it's my dream to fly private said a user on a reddit forum for so-called henry's which refers to high earners not rich yet.

1:38:07I think that's everyone on our team. Adding that he earns about$300 ,000. He's married and has a kid in daycare. Definitely closer to broke than flying private. Another user responded, Mogged. Yet the number of people are rich. The Henry term emerged during the D2C boom. I never heard of that before. That whole era where a lot of D2C brands were kind of looking at their customers as people that weren't necessarily going head to toe Laura Piana but they would you know maybe buy flex up and get you know an item or two or buy some away luggage yeah that whole thing buying a Clydesdale not a thoroughbred that type of money yeah something like that uh I mean the rain that I I did a deep dive on like the horse market recently and the uh as one does.

1:38:58And the spread between the cheapest horse and the most expensive horse is staggering. You can spend millions or you can spend just a couple thousand dollars. So no excuse not to have a horse, one horsepower. How much power do you need? At least. The U.S. added more than 1 ,000 millionaires every day last year on average, according to UBS. The billionaire club grew more than 50 % between 2015 and 2024. Private jet hours flown, touched a high in 2022 and then have stayed elevated since, according to data. Travelers can now use apps to snag individual seats on private jets or pay for flights by the hour.

1:39:35Others charter flights paying for just the occasional trip from New York to Miami. While there are rare business moguls who might spring for the entire jet, some jet providers accept payment in crypto. Shrimp, cocktails, and facials. Ken Ricci, a pilot and chairman of FlexJet, a private company, said the frugal wealthy, high earners who don't typically splurge, started spending big on travel during the pandemic because of health concerns. Many of them have found it tough to go back to flying commercially. Yeah, because of the JetBlue rules, we know this. And years of economic growth have helped ease the stigma around conspicuous consumption that's set in after the financial crisis.

1:40:12It's in vogue to be wealthy, he said. Sometimes we love the rich. Sometimes we hate the rich. It isn't just avoiding the security line or the hoi polloi. Flying private means trading Biscoff cookies, which were also given to us on the JetBlue flight for freshly baked ones and packing lunch with such items as shrimp cocktail and filet mignon for menus spanning a dozen pages. Chef Nobu Matsuhisa crafted a menu for Vista Jet that includes miso salmon. Some cabin hosts are trained to give travelers facials 40 ,000 feet above the ground with Dr. Barbara Sturm's line of luxury skincare. Very nice. anyway fly private stop making excuses just do it we had a breaking news Tyler actually yeah I made us a song on 11 labs okay break it yeah yeah can we play it on the stream I think so okay let's hear the song from wwe style and code where pitches rain and valuations grow brace for the venture showdown let's go wired for disruption seed rounds Is this like a rap?

1:41:22This is not how WWE works. I don't really know what you're saying. I'm not into this. No. Go do some research. Go do some research. Wow, shut down. Shut down. Shut down by John. No, no, no. The whole point of the WWE intro is that it has no words. It's purely instrumental. So look up some WWE intros because it's meant for the announcers to speak over it. And so I'm sure you can just prompt it and say no words, but also like the words are going to be the hardest thing to get right because it's like telling jokes. It's creative writing. It's very, very difficult for the models to get that right.

1:41:59Obviously, amazing work by 11 Labs. Very convincing, but we got to do some work on prompting this one. Yeah, go back to it. I don't want any of that cringe like rap. Yeah, yeah, yeah, yeah, yeah. And the abstract like chat GPT compression of the concepts that we talk about. Good effort, Tyler. And impressive that it was able to be turned around so quickly. This isn't something that cooks overnight. Before we go into the next story, funny story from earlier this year. We were flying back from a trip, and we're joining a buddy, a friend of the show on their jet. And the morning of, it was a super early flight.

1:42:41We didn't text him to remind him that we were coming. and so we got to the terminal, got on the golf cart, we're driving out to the plane and just watched it take off. No, we didn't watch it take off. It was taxiing. It wasn't taking off. We watched it taxiing. No, no, no, no. It didn't leave the ground. Oh, we got out of view from it, so I thought it left the ground. No, no, it didn't leave the ground. But it was running down the runway. It was on the runway taxiing still. It was not actively taxiing off. I thought it had turned around. I don't know. But either way, they turned the plane around for us.

1:43:17It was the nicest thing anyone's ever done for me. We felt really bad. If I ever go on and invest like the best, what's the nicest thing everyone's done for you? Turn around the PJ for sure. That's up there. It was mortifying. It was a low point. But then the rest of the plane was fantastic. It was fantastic. Great time. Learned a lot. It was fantastic. And we are forever in their debt. You deaded, yes. Anyway, in other news, the New York Post is launching the California Post. Wow. The New York Post goes to Hollywood. Good plans to launch, plans launch of the California Post daily newspaper. New York Post, you probably know them from their hilarious and wild front page, what do you call them, just like headlines.

1:43:56They write really great headlines with these hilarious puns. And so they are going to come to California and launch a daily newspaper, which of course will probably be instantiated on the web very seriously. But they call it the new venture. They're bringing the signature high-low-brow mix of content that the New York Post is famous for, alongside fearless, common sense journalism, celebrity entertainment news, world-class sports reporting, and legendary covers. We gotta start getting the California Post delivered. Yeah, I'm very excited for this. Do you know who founded the New York Post? I don't, who?

1:44:37He happens to be a founding father of this country, Alexander Hamilton. Wait, what? You're kidding me. Yes. The New York Post was founded in 1801 by Alexander Hamilton, a founding father, who George Washington appointed as the nation's first secretary of the treasury. Wow, what a story. We have our next guest, but in other media news, Barry Weiss is apparently looking to sell the free press for$250 million. I'm going to hold off on ringing the gong until it's confirmed. We're all rumors. The rumors actually come from the New York Post. But this would be a fantastic outcome for a new media company that's only something like four years old.

1:45:16Barry Weiss left the New York Times opinion page. There was a huge dust up. That happened in 2022. And she's reportedly met with Skydance Media CEO David Ellison at the high-powered Allen & Company summer camp for billionaires. And David Ellison is the son of Oracle founder Larry Ellison. and he's apparently reportedly interested in buying the free press as Skydance continues. We, I think, originally covered this briefly when they were meeting at Allen & Company's event earlier this month, or I guess earlier in July. This screams talent acquisition as much as it does traditional acquisition. Yeah, Barry is on a tear, and she leads the podcast and interviews and brings great guests on, and she writes.

1:46:05But then they also high they also have gotten a bunch of great people on the team Tyler Cowen it has a column at the free press like like that was the that was the thing that got me to subscribe Because I love Tyler and I was like I want to see what he has to say and I don't want to like it's paywall So I have to pay And so she's been very good at like hoovering up talent building this thing and and they publish a ton when you go to the The free press you just see tons of articles all sorts of stuff Just today, I was looking at the Free Press homepage. They have basically an op-ed written by Palmer Luckey.

1:46:38Like they got Palmer to write about his thoughts in the Free Press. And I thought that was just like, that's the example of like great, great written work. So congratulations to everyone involved. And Barry would be able to help make CBS relevant again, even if she was continuing to run the Free Press. Anyway, without further ado, we have Alex Albert. Alex, welcome to the stream. Thank you so much for joining on Short Notice. Great to meet you. Great to have you. How are you doing? Great to meet you guys. I'm very, very excited to be on here. And John, I'm hoping you hold on to that mallet so we can get a gong ring in here.

1:47:08Give us some news. What you got? We went from... Yeah, it's... Oh, go ahead. No, you go for it. Yeah, yeah. I was going to say it's an exciting day. We dropped Claude Opus 4.1 just this morning about three hours ago. You've got to give us a number. How many parameters? How much money do you spend on the thing? I can give you a SweetBench score. Okay, here, yeah. Give me a SweetBench score. What is it? 74.5 SweetBench. Let's go. There we go. Congratulations. Thank you. Thank you. It's been a very, very exciting week. We're shipping like crazy and a lot of stuff happening. Don't know how you guys are even keeping up over here.

1:47:43It's amazing. It's amazing. I mean, we've been talking about streaming for seven hours a day. We'll get there eventually. Fortunately, three hours is like just enough to barely cover, scratch the surface of what's going on in tech. Talk to me about where this fits into the rest of the landscape. There's a ton of stuff going on in Anthropics world from Cloud Code, different consumer models. There's so many different projects. How does this fit in? Yeah, so this is our best model yet for just general capabilities and intelligence. Really, really is pushing the frontier on all sorts of agentic reasoning coding tasks.

1:48:18Back in May, we kind of hinted at the launch of Cloud 4 that we want to get on a more continual release schedule with our upcoming models. And this is kind of our first step. towards that vision of now being able to ship models faster and faster to our customers and end users. Yeah. One of the narratives that's kind of stuck around with Claude for a long time is this idea of like the vibes just being really good. And yeah, it does great on the benchmarks, but it's just like fun to talk to people like the way it talks. But at the same time, Anthropik's been fantastically successful in the business context.

1:48:51Is there an important overlap there? Like, are you hearing from business customers that the vibes being good is driving bottom line results? Well, yeah, more so how much do developers care about vibes versus just raw? It seems like developers are the only ones that can tell that the vibes are good. But I'm always interested in, like, how does that actually translate to, like, my bottom line if I'm an enterprise SaaS company and I need to use an anthropic API for something? Am I happy with the vibes or is that just kind of like a nice to have on top? Yeah, I mean, the vibes are kind of, you can measure vibes in a lot of different areas, right?

1:49:25So if you're a developer, you're measuring vibes in the coding domain and how long of a task can these models complete? Can you just let it go off agentically and work on your files? If you're creating an application, your enterprise has end users and you're having those end users actually interact with the model, you want those personality vibes to be pretty stellar. Otherwise, you're just going to give a bad experience to your customer. I think Claude has like a really wide range of things there and of course we're known for kind of that natural feeling personality and we've done a lot of character work there we've some really great researchers on the team that focus on that but then on the other side of the spectrum for the developers for those folks that are building agents that's where we really get into that that other type of vibe which is you know how how long of a task horizon can these models operate on.

1:50:13Is there a good benchmark for for task horizon yet? We've heard baby AGI, 15 minute AGI, 20 minute AGI. We've seen a lot of people like string these things together in but like in terms of like most consumer interactions I think the way where most people hit the interface of an LLM they feel like like 20 minutes is the kind of the upper bound of what the average AI user is experiencing. Can you talk to me about that? Yeah, I mean, I do think LLMs right now have somewhat of like a jagged intelligence frontier, where in some tasks, of course, they spike more than others. For me personally, when I'm trying to assess that capability in a model, there's a couple places I look.

1:50:58One is especially coding benchmarks. These are really, really important because often it's the model trying to complete a PR end to end. So it needs to take a lot of actions it needs to call a lot of tools and modify a lot of files um so things like sweet bench are actually like good proxies in some degree to measuring that uh there's also third-party institutions that are doing this research right now so one is called meter their model evaluation threat research yeah and they have that really nice awesome that really nice line goes up graph of like plotting out the models and how long of a task they can do compared to a human developer Sure.

1:51:36And Claude has always been pushing the frontier on that. And I'm assuming once they get 4.1 up there, it'll be up and to the right as well. Yeah, I mean, we've seen like the in terms of the spiky intelligence like RKGI is like the main thing that people point to is like a weird Valley of capability because it seems like it should be obvious and easy to do and yet it reveals something about the models Are there any are there any places in the coding world that are more? specific where we're Underperforming either just because we haven't focused on it. I'm thinking about like like I I imagine that most models are better at Python than Fortran.

1:52:14Yeah, yeah, yeah. But is that an outdated take? Is Fortran now just kind of like solved because somebody just got all the training data together and RL'd on Fortran and now we're good? Yeah, I definitely think there's still probably, if you were to like take SweetBench, for example, and you translate it into every coding language, I'm sure Python and JavaScript TypeScript would be ahead of the other languages. but we are seeing, especially with the Cloud4 models, that they're able to handle a lot more of those older languages or the languages that don't really appear as much in the training data to a much better degree.

1:52:50And that's through a combination of things from RL to just general architecture improvements. So we're kind of covering the bases with the models, but yeah, it definitely started in one way and now it's expanding out. What have reactions been to 4.1? I'm assuming a variety of people got access to it, you know, pre-release. What are the areas that people are kind of most excited about? Yeah. Again, it's kind of those coding and agentic tasks have been really the start of the show here. What I think is interesting about that is that coding is kind of the proxy for allowing everything else. So if you think about how a model actually interacts with things on a computer, everything can be done through coding operations.

1:53:32So when we say coding, it's almost like a little disingenuous to just narrow it to that domain. But some of our early customers that were playing around with 4.1, Windsurf being a great example, in one of their benchmarks, they showed that from Opus 4 to 4.1, showed roughly the same performance jump from Sonnet 3.7 to Sonnet 4. And this was like a real-world junior developer task eval. So something that I definitely would pay attention to. How are things going on the rate limit side? I've heard that capacity is generally constrained for everyone, everywhere, all the time because we don't have enough data centers.

1:54:11Obviously, people are working on that. But I guess broad update on how things are going. But then more specifically, when you're talking to companies that are building on top of anthropic APIs, are there specific strategies that you recommend to just embrace the realities of the modern AI era and capacity constraints? Yeah, I mean, the rate list situation is something we're just continuing to iterate on. And of course, we have like, very, very smart people, they're much smarter than I think about this problem in terms of building out compute and also working on making our inferences overall more efficient.

1:54:52generally we want to provide the best experience we can to our customers so whether this is on the cloud code or through the api there's various things that we do whether it's like through our applied ai team helping our our enterprise customers with their specific deployments or on cloud code making sure that our rate limit system is fair and actually allows for people to utilize it as much as they can without rewarding people that might be like abusing the system and kind of detrimenting the other users. So it's really about like kind of monitoring and keeping it consistent. So we ensure like great experiences for everyone that's trying to use Cloud.

1:55:25What about pricing? Is there a step up in like costs associated with 401 versus 40? What does that look like? Pricing is the exact same. We really do intend for this model to just be like a drop in replacement for Opus 4. So if you're using Opus 4, you should just be able to cleanly kind of switch over the model string and you're good to go. That's awesome. Yeah. What are some of the weird use cases that you've seen for Cloud Code? I've talked to some people that are like, because coding is kind of like an abstraction over like everything that a computer can do, they're coming to Cloud Code with tasks that feel more like just general like personal assistant agent stuff, but then it just writes a bunch of code.

1:56:13Is that real? Is that something that it's more just like when, you know, when people do those like beautiful artworks in Excel, where they like this color every single cell as like pixels. And it's like, you're kind of abusing the tool here, but it's still impressive. Or is that actually like, you know, a glimpse into the future? I think that is a glimpse into the future. There was a tweet I posted a few days ago, kind of asking the community for, hey, what's all the best non-coding use cases you're using Cloud Code for? And it got hundreds of replies. And really the answers kind of were super wide ranging.

1:56:51Everything from like managing your own personal knowledge base, second brain on your computer, you know, maybe you have a ton of notes and text files. So some folks are actually using it now for like video production. So we have a guy on our team that uses Cloud Code and uses like an open source library to create animations and things for like all our product launches. Oh, that's very cool. Which is just absolutely amazing. I didn't even know that was possible. But it is kind of hinting at this thing that it's like, well, if you can control code, you can really control anything that happens on a computer.

1:57:21Now we're moving into this like operating system abstraction almost. And I think that's really the direction you're going to start to see CloudCode heading in. That's awesome. Anything else, Jordy? I'm curious to get an update on, it seems like a lot of the companies that we talked to, you know, we were talking with Dylan Field the day of the IPO and he was talking about updates they're making with MCP. What's going on in that realm? Any updates while you're here on that front or interesting use cases? Yeah, I mean, MCP has been absolutely just ripping lately. The community is awesome. And it's like really got to this point now where it's almost its own like self-sustaining thing where there's like meetups and gatherings that are being hosted completely outside of like our facilitating.

1:58:09And thankfully we get to participate in them. And that's just like been super cool to see. I think MCP is gonna be a large part of like making sure this kind of agentic future goes well. So if you can spin up easy integrations and connections to any service or any product or any internal application, all of a sudden you kind of unhobble Claude to some degree and you give it access to that knowledge that it needs to actually go out and do those correct operations. So very excited about the team there. And yeah, hopefully a lot more good stuff coming on the way there soon as well. Awesome. Very cool.

1:58:44Well, congratulations to the whole team on the launch. It's a massive day, massive week. And love to have you back on soon. I'm sure 4.2 will be right around the corner. Oh, yeah. A lot more coming soon. Yeah. Stay tuned, guys. Have a good one. Cheers. Bye. Up next, we have Scott Kapoor from formerly Andreessen Horowitz. He was in the OPM business, the other people's money business. Now he's the director of OPM for the U.S. government, the Office of Personnel Management. Welcome to the stream. Scott, we are excited to talk to you. How are you doing? I'm doing great. How are you guys doing? We're doing fantastic.

1:59:23It's great to have you on the show. Thank you. We're in a very funny era where more folks from Silicon Valley have moved over to the government. And every time they do, I have to have them explain to me, like I'm in the VC world, what they do. I think I learned about the OPM. The entire office was a new term to me just a few years ago. Explain the Office of Personnel Management to me. You mean to tell me that you guys don't know what OPM is? That is really disappointing. I'm shocked to hear that. I probably learned about it about a year ago. Explain it like I'm a venture capitalist that's about to write a thread, explaining it like I've always known about it and I'm an expert in it.

2:00:01Yeah, my wife asked me the same thing when I got appointed to the job, so I'm not surprised to hear that. Hey, so really quickly, what it is, is we are the talent management organization for the government. So in simple, in our corporate VC terms, we would be the HR department basically for the company. Sure. So our job is like, make sure we have the best talent, make sure we have actual things like performance management standards, which by the way, we don't have, which we're working on. Make sure we have the ability to, you know, kind of understand the employee base through technology, which we don't have currently.

2:00:31So, but basically think of us in very simple terms as we would be the HR department inside, you know, any other, you know, traditional VC back company. And does this kind of cross all the other departments or are you like, where are the boundaries of your? Yeah. So it's a hybrid is actually what it is. So basically, so like there's some stuff that we have exclusive province on. So like, I'll just give you a simple example. Like the president puts out a hiring and executive order saying we want merit hiring to be the standard in government, which you and I can, we can talk about that forever if you want.

2:01:00But basically, so he says, look, I direct OPM to figure out like what are the standards that we need to do cross government to make merit hiring a reality. So we will go put out a bunch of, you know, regulations and research it and say, great, okay, now this is the law of the land for all the other agencies. This is what we expect you to do. And, you know, OPM will not formally approve and issue offer letters and stuff like that, unless you guys do that. So that's what we do. And then, but there are HR departments, obviously, in the different agencies. And, you know, so think of it as like, you know, in a company, you might have like, centralized HR, but then you have like business partner HR that might sit in the sales organization or other stuff like that.

2:01:36And they may be kind of, you know, do things that are domain specific. But all the big things about, like, how do we recruit people? How do we train them? How do we actually manage them, you know, performance-wise, we will provide guidance for. Yeah. And, sorry, Jordy. Yeah, I was just going to say, like, getting into it, what are your priorities first 90 days, 180 of the year? I'm sure there's, you've, like, even any, you know, I'm assuming you're trying to bring in best practices that you've learned. Fix the government. Don't make mistakes. Yeah, exactly. Just ask a model to do it. We're going to make some mistakes, which is okay.

2:02:11That's part of the problem, by the way, which is we've been living in this environment where you literally cannot make a mistake, and so nobody actually tries anything. So yeah, really simply, there's three big things we're trying to do. One is create a high-performance culture in the talent. So in other words, my simple goal is, look, we want the best and brightest people to take government jobs. But as you guys know, look, A players want to be around A players. They want to be in an environment where they're held accountable. where when people do well, they actually get rewarded. We have a very like tenure-based and conservative risk management-based system here in the government, which is like you get power and you get promotion by having more people in your organization and a bigger budget and the longer you spend time in government.

2:02:50So we've got to just flip that on its head and say, look, we're going to award things like innovation. We're going to reward, I've been using the term measured risk-taking because look, we can't go just like literally throw stuff against the wall and see what works like we do in Silicon Valley. But anyways, that's thing number one. So, like, goal is how to, like, all the smartest people in the country feel like coming to work for the government actually is great for them. They learn stuff and they'll get paid for that in the private sector. Number two is we've got to make operational efficiency like a core tenet of government.

2:03:18So, again, I know this sounds crazy coming from, you know, the world that I know you and I came from. But today, basically, there's no incentive for efficiency. So, again, you need to make it into people's performance plans and say, your job every day is, are you doing a great service on behalf of the American people? And are you constantly turning the crank to figure out how do we use technology? How do we use AI to do stuff? How do we reorg to do things? Spend less money? Spend less money, right. Today, again, it's the exact opposite incentive, which is more money means more people means more power.

2:03:50And so everything we have to do is, yes, spend less money to deliver a better service. And so we've got a ton of like technology initiatives that we can talk about there. And then the final thing we want to focus on is we've got to make like we've got to anticipate what living in an AI first world looks like for government. And we got to not be flat footed and realize that we have none of the talent we need and we don't even know what the jobs are. So this is, I think, a huge opportunity, quite frankly, where I think we should work with the private sector and say, great, you guys understand what AI is going to do.

2:04:19Let's figure out ways where we can have information sharing, even like, quite frankly, swap employees back and forth. but find ways so that we don't wake up five years from now and realize there's all these new jobs that are being changed and or augmented by AI and the government is not like you know figuring out how to actually get that talent into the government. Yeah, you mentioned spending less money. I want to know about the potential of spending more money. Is there a world where we remove salary caps on government employees? Like you've seen this in Silicon Valley. There's not like, you know, obviously founders are very driven, but they have uncapped economic upside.

2:04:56Like the venture capitalists have uncapped economic upside in so many ways. The AI researchers that are getting poached, like it is just money ball at the end of the day for a lot of sectors of the economy. Is there a reason why the government is fundamentally structurally different? Or do you think that we just need to make a better pitch for civil service? Yeah, I think it's a little bit of both. So, okay, so look, we're never going to solve, we're never going to solve like the ultimate, like uncapped upside pay gap between the public and private sector. And look, that's just, you know, politically that would never work.

2:05:29You know, we're not going to, you know, like, you know, we're not going to have a Zuckerberg equivalent of like paying engineers a billion dollars a year, right? Yeah. But we can do two things. One is we can change the compensation structure so that, yes, like A players actually get rewarded. So I'll give you a very simple example here. Everybody gets ranked one through five at the end of the year in their performance evaluations in government. Five's the highest, one's the lowest. 70 % of the people get ranked a four or five, okay, which means, like, they're awesome. They're doing outstanding work.

2:05:550.2 % of the people get ranked a one or a two. Now, again, mind you, this is across a population of 2.4 million people where, at a minimum, you might expect a normal distribution. But like there's just no way that 70 percent of the people are above average. Right. Unless you're in, you know, Lake Wobegon, basically. So we've got to solve that problem. And then if you solve that, then the incentive structure can follow it. Right. Because, again, a typical government employee gets 0.5 percent of cash bonus at the end of the year. That's because, again, there are too many people highly ranked. And the delta between low and high is like, you know, 50 percent or something.

2:06:29Right. So what I'd like to do is say you ought to have many fewer people ranked to four or five. And let's give those people like most of the bonus pool, for example. So we can't solve all the problems, but we can solve that. And then your other point is a really good one, which is we've got to change the pitch. So, again, I'll give you a story here. When I first came in, one of the employees here said the pitch that we use for government is job stability, basically. Like that's the pitch. And I was like, I'm sorry to say, but like, unfortunately, I think somebody's been lying to you. because like there's no job in the world where we can guarantee job stability for people, even the government.

2:06:59To me, the pitch has got to be you can see things at a scale and solve really hard, complex problems. And yes, like you can, you know, do a public service as well. But we've got to get people excited about like, you know, you know, putting people on the moon or all the things that, you know, guys like Elon do. You know, it's the whole like founder story of like what's the mission of the organization? The mission of the organization cannot be we provide job stability. So we've got to like tell that story. We've got to go tell it on college campuses and we've got to go like show it. And look, people don't have to commit their life to being in government.

2:07:28But if you could come here for two years and work on really hard, really cool problems, and you could actually advance in the organization because you're not tenure limited, but you actually are skills, you know, you know, you're awesome at skills, we should be able to promote you faster. Like, there's no reason why we can't compete for really smart people in that environment. And then again, go take that skill set and go work for, you know, Elon or go work for Zuckerberg and get paid a billion dollars a year after that. Like, you can do both things. So we've got to sell the story. Yeah, I was talking to a friend who had some experience in China and he said that the main difference that he noticed was that the really, really top performers over there just kind of got sucked into government work.

2:08:06And he was kind of framing it as like, it wasn't exactly optional for them. And I know that that's not on the table here. We love freedom and choice and independence. We kind of decided along to that we actually, like democracy is actually a good thing, right? Yeah, of course. But I guess my question is like, I know that like when the administration changes over, I mean, this happens every four years now, eight years most, there's a push to bring in new folks and there is some recruiting that goes on and there's like whisper networks of who's really talented and let's make the pitch to them that maybe now's the time to go and serve your country.

2:08:42And a lot of people do swing at that pitch and they take it and some of them do very well with that. But should it be more aggressive? Should recruiting be 10 times bigger in terms of just every talented founder, engineer, venture capitalist, anyone who's high agency, high intelligence, high EQ is at least hearing the pitch for where they would fit in in the US government? Because I think a lot of people, I know a lot of people that got calls. I know a lot of people who are talented and didn't get calls. And I'm wondering if we should have reached them with something. Yeah, it felt like at the beginning of this year and maybe the end of last year was just recruit Doge, which is dominant.

2:09:22It was a fantastic recruiting engine. Yeah. Yeah. But I didn't I didn't see that. I didn't hear like, oh, I have a fintech founder who, you know, they were great, but they got kind of clobbered. And now like the IRS is calling them. Like I didn't hear that story that often, for example. Yeah. So yeah. So let me try to just unpack that a little bit. So first of all, you're absolutely right. We should 100 % do that. So there's there's a problem we have today which we which we are working on fixing which is today believe it or not you can actually there are weird hiring procedures in the federal government which makes it very hard in some cases literally to hire people based on their skill skill assessment so let me give you a great example so people have gotten really good at we put a job description up on this thing called usa jobs which is basically the the website and people essentially tailor their resume to exactly the words that are in that job description and the filtering process puts them at the top of the list but it's all self-assessed stuff.

2:10:16We literally, like, you could hire an IT network administrator in the government today and not have a technical person interview that person to see do they know, like, what, like, you know, HTTPS means, basically. And the good news is we just, part of it is because, like, people are worried about, okay, are those tests discriminatory? We actually just had a good legal ruling the other day that clears the path for this. But believe it or not, literally, like, we do not have skills-based assessment for many jobs in the government. So the way to so like it's just really hard. It's really hard for those people to come in.

2:10:49We but you're absolutely right, which is, look, the real solutions problem is we've got to do both. We've got to fix that piece. And then we have to literally go tell the story. Like, I don't think this is the solution. We're not going to do TV commercials like, you know, the Army and Navy. And, you know, we need a few good men, basically. But we need something like that. We've got to go tell the story. And you're absolutely right, which is I would bet there are I don't know. I bet there's thousands of data scientist jobs open across the government today. You as an employee, if you're interested in that, you should just be able to apply once and we should then send your resume to all thousand of those jobs.

2:11:18Like we don't even do that today, basically. So like you literally have to go find out where those thousand jobs are. So there's all this stuff that, again, it's not, they're not bad people. It's just that the system has not been set up for that goal. The system has been really set up to say, let's really be conservative. Let's really be kind of, you know, careful on these things. And we just, we have to change that paradigm. Yeah. It feels like we're fishing with a really large net. You know, you put up a big ad. we need a few good men and whoever gets caught in the net comes into the organization and we need to get some harpoons out and go find the really like laser focused people for the perfect job military you know it's been interesting i i think things have changed but there was a period where military recruiting to my knowledge was was really really tough and that's still with a lot of a lot of people that do end up joining having been marketed to and and and kind of exposed to army Navy, Marine Corps, Air Force, since they were kids, right?

2:12:11But nobody's getting, nobody, most, most Americans wouldn't know, they think OPM, if anything, is other people's money, like you were saying, and they wouldn't know it. And so how do you, how do you get somebody like even open to the idea? You know, a lot of kids today in high school, college, they know they want to work in tech. Nobody's going through those periods of life knowing they want to work in a specific organization or, or very little. On the technical. You're absolutely right. So look, yeah, and we got, look, we got to just like have a presence on college campuses and be there just like Goldman Sachs is and just like Facebook is.

2:12:42And we got to go tell our story. Like, you know, you guys know this, but I used to work for a head of sales at one of my companies and his famous thing was, you know, show it, sell it, hide it, keep it basically. Right. And right now we are hiding it basically. And people find us because they're looking for stability as opposed to us finding people who are looking for like an actual real interesting challenge. It's an interesting time. You know, apparently CS grads are struggling to find roles. You can see the variety of reasons for that. We just had someone on from Claude. Claude Code performs really well now.

2:13:14Maybe it's performing at the level of a junior engineer, but the opportunities in government where you're not out there competing with other foundation model labs. We saw Luke Fairtor. He is a programmer. From that massive Bloomberg article, It didn't seem like 90 % of what he was doing was programming. It seemed like, yeah, he was able to scrape some data together and do some analysis to know how to make a decision. But a lot of it was just sending emails, having phone calls, having meetings. But Luke's also an example of somebody who's truly elite in his category. And I'm talking about like the next like 10 ,000 people.

2:13:49But I feel like if you're a new grad and you can program, that gives you a superpower that then you can also just be effective to be rigorous and first principle thinker in a meeting where you're just trying to decide how much money should we spend on this thing? Did you see, by the way, did you see the quote from the unnamed government source in that article about why they didn't like, why they didn't want Luke? They're like, he's not qualified because he didn't go to college and he didn't have five years experience, right? And that is exactly the problem, right? All of our criteria for evaluating people are based upon credentialing and based upon tenure.

2:14:17And this is where merit-based hiring matters so much. Like to me, if you're Luke, I don't care, if you're 18 years old and you can actually function at the level of somebody who like, you know, knows programming at a high level, Let's pay you at that level. Let's hire you at that level. And okay, like if you need some help around the edges to like, you know, make sure you can sit in meetings, you know, with your shoes on, that's great. We can teach you that basically. But we can't exclude people from the process because they don't have like, you know, a sheepskin that says, you know, they went to Stanford for four years.

2:14:44I mean, that's crazy. Yeah, exactly. How is your AI strategy forming when people hear AI in government as citizens? They get scared. That sounds scary. If they hear, if you're working in the government and you hear AI, you're probably thinking, I don't want to lose my job. But clearly it should be able to be beneficial. The big thing just feels like, get every government employee access to an LLM. Because before any of the job displacement, any of the cyberpunk terminators, let them search the knowledge quickly. Like, please. So anyway, what was your take? Yeah, no, you're exactly right. So today on my government computer, if I want to do something on an LM, I have to go to my personal computer to do it.

2:15:27That's not good. And how many government jobs are email jobs? Look, the privacy issues are real, and we've got to be conscious of that. But so much of what the government does is already in the public domain. So let me give you a very simple example. So one of the things my team does, as I mentioned, is we develop rules and regulations for let's say how you should do merit hiring in the government. And the way the process works, which I won't bore you with, but basically we put out a rule, and then we have to invite public comment on it and literally people submit comments. So we're doing a rule right now where we got 40 ,000 public comments on this rule, okay?

2:16:00A group of people in my team literally read all 40 ,000 of those comments and they write a written response to them and do it, you know, the old-fashioned way. Now, I'm not going to say, like, let's just get rid of all the people and have AI do that, but there's no question in my mind that we can get a 10 % to 15 % easy efficiency advantage if we actually had a very simple LLM that we could train on totally public data. I'm not talking about private data. All that regulation stuff is public. Help people manage that process. But we have to start with that. So what I'm doing here in OPM is I'm actually going to have somebody come in from the outside and do the full landscape of what's happening in customer support and what's happening in HR organizations and what are all the use cases out there.

2:16:43And then what I'm doing is just telling my team, look, now your job is to go back and find what what are the small wins we can make on little use cases? And again, I'm not asking you to like, you know, get rid of a hundred percent of your people, but like you need 10 more heads and you need to find some place to get 10 % more productivity. We ought to be able to do that within our current budget without actually having to, you know, add headcount. So we just got to start with little baby stuff that doesn't scare people, get them educated. You're exactly right. Everybody should have, you know, an LLM on their desktop.

2:17:12And then we can tackle the hard problems, which is okay. Are we going to put confidential government data in them? and do we need them air-gapped and all that crazy stuff. But, like, we don't even need to do that right now. We just need to, like, educate people and then come up with use cases that are based on, like, literally public data with human augmentation. Does the federal workforce have, like, an aging problem? You know, we'll talk with CEOs on this show or investors that say, you know, the majority of factory owner operators in the country are going to retire within the next five to ten years, right?

2:17:42That becomes concerning because there's a lot of just, like, general know-how and process that could just kind of get lost or evaporated. Is that a challenge that you're facing in terms of, you know, people that have been in the government for 30 plus years and are getting to the point where they're ready to retire? Yeah. If you look at the demographics of the government, I don't know the exact number, but there's no question like the government is very skewed towards people who are kind of later in their career. And again, I think this is a problem that we've had, which is we haven't figured out a way to actually get people in at earlier stages in their career, because we've got all these impediments that I talked about that make it hard for you if you're a recent Stanford grad or whatever to be able to get a job here.

2:18:21So yes, we have to solve that problem. Look, I mean, sure, there's going to be knowledge that walks out the door. Look, I'm more worried about the long term thing, which is, do we have a steady pipeline of really smart people who will come here? And then we can solve that aging problem by just, you know, over time kind of changing the demographics. But yeah, like, I don't think I'm not worried, like there's a cliff here where, you know, 10 years from now, all the experienced as people walk out the door. But I do think we have to have a demographic and a set of hiring practices that allows us to actually build a funnel from the very bottom of books.

2:18:51Yeah, it'd be cool. I remember maybe in 2012, Teach for America was really popular. And obviously, that was non-governmental, but aligned with this idea of a tour of rough service. And there was Code for America. And it's like, cut out the middleman and just go straight for America. Work for the government. I love that. It would be great culturally. Because now, I feel like if you did four years just in civil service and then you jump into Silicon Valley, it's kind of like a weird track. But it shouldn't be. It should be completely normalized. Yeah. A long time ago, and this is not the perfect analogy, but there used to be something called the Civilian Conservation Corps, which was exactly that.

2:19:29It was like people came right out of school. They did some heavy labor for a while, and then they went to the private sector. So, yeah, we ought to do this. So, what I would love to do is I would love a program where we say, you know what? take all the smartest people, have them come here for two years. By the way, if we can do this, let's forgive your let's forgive some student debt as part of that. Right. So like, let's actually have a real trade off, which is you do public service and we'll give you some financial compensation. And oh, by the way, I want to work with the private sector and say, you know what, we're going to get a private sector employer to guarantee you a job when you finish your two years here.

2:19:59And they might even give you full credit for those two years. You come in as a third year and that and like just work on that stuff. Right. So, you know, that's what we need to do. And look, I've been using this term that's terrible that my kids joke at, which is make government cool again, which is magaka, if you pronounce it out loud. Magaka. My marketing team who's sitting here next to me off camera. Yeah, don't. Doesn't run off the tongue. Doesn't really. Yeah, it's not. It's not like not the greatest outcome. But like that's what we have to do. Like we have to make civil service like super exciting.

2:20:29And the good news is like we've got no shortage of like moonshots now. Right. So like AI. Right. Like how do we transform the government with AI? That's a great moonshot. Let's go do that. But we need like a, you know, Apollo mission like moonshot, which I think the president has given people with AI. And we've got to go tell that story. And, you know, if we can do that, then I think like we're in a totally different situation. Are you going to be visiting college campuses for career fairs at all? Are you pounding the pavement? I would love to. Look, you all know this having been in the Valley.

2:20:57Look, if you're in the Valley, 99 percent of your job is recruiting, basically. Like that is like probably the most important thing that any manager does. And yeah, like I don't think any of us is beneath that. We got to go sell. So yeah, I'd love to go to campuses. I'd also, look, I want to, we've been talking about campuses. I also want to make clear, though, one of the priorities the person has, which I totally agree with, is we shouldn't use credentialing as a shortcut for this stuff. So look, I don't care if you went to college or not. If you have the skills to do this job, like, we got to go find you too.

2:21:23Look, college campuses are nice aggregation points and make it more efficient. But like, we need to go find anybody who has a, you know, technical degree or not a technical degree or a trade degree or anybody who's just really smart and capable and go find those people. and we hire them not because they you know have a you know a diploma but because they actually can contribute to the workforce or even like experience at a startup like they've been there for a couple years they vested they run the next thing go do a tour in the government yeah yeah last question have you had any pleasant surprises since starting your job anything that you're like oh that's actually run pretty well yeah you know look i'll be so to be totally honest look um there are my pleasant surprise which is i don't think it surprises look there are very good people here who really care about what they're doing.

2:22:05So, look, you know, everybody in the government here, there is a serious element of public service, and they all take it very seriously and believe it, and it's incredibly, like, noteworthy. What I really think is, look, I think the system has failed them, which is, look, as leaders and managers, we have created a system where, you know, we've told them do not ever make a mistake, do not ever risk innovating, because we don't, you know, look, you guys know the Valley, right? In the Valley, we always talk about, look, we're looking for uncapped upside on stuff. Yeah, you get fired for not taking the risk.

2:22:33That's exactly right. There is no value today in maximizing for uncapped upside if that means you also take an extra 5 % of downside risk. And again, look, I want to be totally clear. We can't be cavalier. We're dealing with people's social security or we're dealing with national security. So we can't just literally throw caution to the wind. But we can accept that there are some things where we may not get them 100 % right and then we'll fix them when we do. And we have to have a culture to do that. So that to me has been, again, the very positive upside surprise, which is and which I hope will prove true, which is if we can change the incentive system, I think you have really good people who will do what the incentive system incents them to do, basically.

2:23:09But right now we just we have an incentive system that, in my mind, at least kind of encourages behavior that just is not, you know, efficiency maximizing, is not innovation maximizing and, you know, is just outdated, I think, in a world where, you know, we know technology and the needs of the government are going to change very rapidly. We need to be an innovation maxing. Let's do it. I love it. Thank you so much for joining. Thank you for serving our country and rallying so many smart people to join you. And thanks for writing it down for us. This is very helpful. Thank you so much. I appreciate it, Don.

2:23:37You take care. Bye. Cheers. Bye. We need a Luke Ferretor for the CIA. We need to send Luke Ferretor to North Korea, infiltrate Pyongyang, get behind enemy lines, and turn it into a capitalist utopia. make South Korea look like a complete blackout by comparison. You've seen that map of the North Korea is dark and South Korea is super bright. More lights in Pyongyang than all of South Korea combined. Luke Fair tour on the front lines. We'll get it done. This is breaking in real time. So John actually hit up Peter Salas over at Discord. H.R. apparently just threw a note on Peter Salas just posted guys at TVP and WTF.

2:24:20John, actually, I think I'm in trouble. H.R. just threw this on my calendar. Oh, it's Peter and H.R. discuss TVP and issue. Wait, is this real? No, I think he's joking. But very good stuff. I love it. That's amazing. Thank you for sharing that. Peter, come on the show. We're going to do it. We're going to figure out your cloud. Yeah, your cloud gets you. So I out. We're going to have people live bidding. Yeah, yeah, yeah. We should have three. We should be Shark Tank. He's the shark. We have three hyperscaler cloud SDRs. Come on and pitch. Okay, this is what we can bring to you. This is what we're asking.

2:24:53We're asking for, we got 100 ,000 H100. It's an auction for who can deliver the lowest price. Exactly, exactly. Anyway, we have our next guest in the temple of technology. We have Kyle from Contrary. How are you doing, Kyle? Good to see you. What's going on? Welcome to the stream. How's it going, guys? What's new in your world? break it down for us? Well, we're just launching our latest deep dive from Contrary Research. Very cool. So for folks who don't know, Contrary Research is our private markets research database we launched a couple years ago. Contrary is a venture fund, but our focus is on talent and on research.

2:25:26So finding the best ideas, best people, bringing them together. And so Contrary Research has published research on 500 different companies. I think we've got most of your sponsors on the list. So we're getting, we're working our way down. Um, but we just published our latest deep dive on building an American TSMC. So TLDR, the piece is the idea that we're completely dependent on Taiwan for 90 plus percent of, of advanced semiconductors. And it's probably the dumbest bottleneck that we could possibly be in. And it feels like we're basically just hoping that it doesn't become a problem when it could become the most significant geopolitical flashpoint and effectively lead to world war three.

2:26:02And so it feels like there is something we should do about And so we kind of go through the options of, do we build it from scratch? Do we borrow it from TSMC? Do we try and buy it from some willing sellers? It seems consensus that Intel is not the American TSMC. Is TSMC the American TSMC? Is Samsung the American TSMC? We've seen a lot of progress from both of those companies. Elon doing a big deal with Samsung. TSMC obviously seeing great results out in Arizona. What's your take on just porting or expanding the current set of fab companies that are doing quite well? Our biggest takeaway from doing the research is that a little bit that the ASMC is more of a mindset than a specific company.

2:26:45It's an ecosystem that we have to build a lot of different aspects of this. There's a ton of startups that are building a lot of advanced stuff in chip design and using AI in that. TSMC is pouring$165 billion into the U.S. more so. So there's a ton of things that we can take advantage of. The biggest limitation is that when you try and build a lot of this stuff from scratch, you just don't get there quickly enough for it to make a difference. There's a bunch of stuff like the folks at Anderil, and they have this idea sort of internally of China 27, where it's this idea of if China is going to invade Taiwan, it's probably going to happen over the next, call it four or five years.

2:27:22Xi has talked about 2027 specifically. So this is an urgency thing. This is not a long-term bet. And so the question is, how do we get to scale where if if China was to take Taiwan tomorrow? Is there something that we could do to have some capacity here? The unfortunate reality is that the startup ecosystem as great as it is and as much as it can benefit us in the long run It's not gonna get there quick enough as much benefit as we get from TSMC partnering a lot of the the results they're having in in in Arizona And there's a huge cultural blowback in Taiwan because they have this silicon shield where they think of our dependence on them as their protection from China.

2:27:57And so yes, TSMC wants to pour a bunch of resources here, but whether we get those sort of leading edge nodes that allow us to compete for things like AGI and to really compete with China, that is a much longer battle. And so despite the fact that Intel is struggling, it's basically sending up emergency flares that you guys talked about on the show, basically in their last quarterly earnings, they talked about, you know, we're going to do our best. And if we don't get material customers, it may not be economical for us to continue to compete in leading edge nodes, which is gonna be a huge hit for the US to stay competitive.

2:28:27And so our thesis and the piece is that we basically have to take Intel as our only hope of getting to scale quickly enough. Intel is not going to do it on their own. There's a bunch of stuff that we have to do to make that possible. Yeah. So what's the biggest request from TSMC or Taiwan? Like, should we be pulling the, hey, we'll invest alongside you, we'll subsidize, we'll give you tax cuts to build the second, the third, the fourth fab in America? Or should we be more focused on let's commit more defense resources, more government spending to Taiwan to create, you know, an American shield that's independent of the silicon shield or maybe both?

2:29:08Yeah. I mean, there's definitely when you talk to all of the major defense companies, Palmer Luckey has a great line where he talks about turning Taiwan into a particularly porcupine that people don't want to step on. That is absolutely priority number one to give them all the capabilities that they need and the folks that are on the ground in Taiwan all the time. So that's a huge component of sharpening the partnership. I also think that you need to emphasize that American companies need to be, the government needs to incentivize American companies to be purchasing American-made semiconductors as much as possible.

2:29:41So as you reshore more of this stuff to, whether it's TSMC efforts, whether we literally nationalize Intel and take the foundry business private, right? Like Dylan Patel, who you guys have had on the show and is an absolute rock star. He has made it abundantly clear that the survival of Intel foundry is critical to America's economic and technological dominance. And so we have to take that seriously as well. And there's a bunch of incentives that we can build in terms of making, you know, getting people access to semiconductors that are made by Intel in the US, by TSMC in the US. We have to incentivize people to be using our own ecosystem so that we get the reps we need to get as far away from dependence on Taiwan as possible.

2:30:22Do you think that Lip Bhutan is the right leader for this transition for Intel? It seems like they need to, with broader support, reinvent themselves. And yet hearing Lip Bhutan on earnings calls and just talking through, it just feels like cut after cut. He's good at firing people, but is he really good at firing? I haven't been inspired by, he's definitely not Karp on the mic. He's not exactly inspiring people. So I'm curious. Karp also has like 1 ,000 employees, and I'm pretty sure LibBouton has like 100 ,000 employees or something. Yeah, so Intel has more employees than TSMC and NVIDIA combined.

2:31:04Yeah. So he's got plenty of people that he can fire. The biggest thing that is compelling about Intel is that they are already in this process of cultural change, right? You look at that, I mean, it's back to 2005 was the first time they had a CEO who wasn't an engineer by trade. And basically, they went through 15 years of financialization. Like, it was literally the Jack Welch playbook of just dumping money into stock buybacks and robbing R &D for decades. Those CEOs were the ones that put them on the path, sort of stumbling, losing their lead. So like even with Gelsinger, like that transition is already in its way.

2:31:40Like putting engineer, they talk about bringing the nerds back and that transition is already in its way, which is important. I think the cultural leadership that you should be paying more attention to is the leadership level at Intel Foundry specifically. When you think about Intel's two core businesses, you have Intel products, Intel Foundry. Intel Foundry has had four different leaders over the last three years, like some of them lasting no more than two months. Right. And what was compelling this idea that like you can't you can have this sort of within a division leader who can be really inspirational.

2:32:11I think the folks at Waymo, they have a co-CEO set up, but it's their former head of policy and then their former CTO. You have the technology of the policy like that's a great duo for sort of Waymo set up. Something at Intel Foundry that could be similar could be really compelling. The thing that's the biggest concern for me is that the first person they had, once they started breaking out Intel Foundry into a separate P &L, the first CEO, the first leader of Intel Foundry that they had, they called him the president of Intel Foundry. He reported directly to Gelsinger. Four leaders later, it's now this person is the general manager of Intel Foundry.

2:32:46They're also the chief global supply chain officer and the chief technology officer of Intel's core business. Meanwhile, Intel Products has its own dedicated CEO with the title of CEO. I think the biggest problem is not necessarily even Liputon, but it's more about are we actually acknowledging that Intel Foundry needs to be a growth engine that you can invest in to be able to build back dominance? And that's absolutely not how they treat it. I feel like earlier this year there was a rumor that Elon was kicking the tires on Intel. How real was that? Did he kick the tire? It was the PJ tracking.

2:33:21That was how they all knew it. Yeah, but the PJ tracking felt like a bit silly because everybody was in Mar-a-Lago. Yeah, everyone was there. That's right. You could spin probably a lot of compelling narratives based on all of the people who were there at the same time, right? Yeah. In terms of like what is actually sort of official, the two things that you know for sure are there was talks that TSMC was going to acquire Intel's foundry business. Their board flatly denied that, that that wasn't true at all. Then there was talks that they might invest heavily and run a joint venture where they take over operations of Intel's foundry business, even if they're not buying it outright.

2:33:57And what's funny is that they did not deny that outright. They refused to comment. And so the general census of folks is that that was a conversation that was active. Yeah, that makes sense. How much of the race for American semiconductors is driven by the question of founder mode, just that abstract question of, do we need Elon in the seat leading? He said he was going to be walking the floor at Samsung. Obviously, the founder mode pattern has seemed to work in so many stuck, kind of stagnating worlds like rocketry, electric cars, where Elon's gotten in and kind of jarred things loose. But is that something that we should be looking out for, optimizing for?

2:34:43100%. I mean, when you contrast Lipu versus Jensen, it's like night and day difference, right? You do not have visionary leaders in chip manufacturing in the US because we don't do it anymore. Like it's not a fundamental part of who we do, of what we do. There's not a core DNA there. And so do you need somebody who is that visionary leader? Absolutely. And does Intel have any of that DNA? No. And so when you think about what needs to change, I think our thesis and takeaway after doing six months of research and drilling into all of this stuff is that you cannot ignore the entire ecosystem. And that leader, that visionary leader role is absolutely critical.

2:35:19And that is an unanswered question. We don't have that. We don't have an, you know, somebody retweeted our post today and said, we just need an Elon Musk like, you know, Herculean act of God to be able to make this possible. and like don't we know the takeaway is accurate the how to do that is much more difficult yeah it can't be manufactured um but i mean green shoots all over the place lots of cause for optimism but so hopefully the piece uh starts a conversation can be a catalyst for change what are you hoping to see out of like the immediately next quarter for intel to get on the right track the biggest opportunity is for Intel to stop treating Intel foundry like it is an ugly stepchild.

2:36:03Like it is this like embarrassing accident that they happen to have and they're trying to distract people over this other thing or dress it up like it's this fancy thing. And the biggest problem is, and this goes back to, this is a great conversation I've heard a couple of folks have. They talk about the, I think it was Bucco Capital talking about like the greatest thing that Elon Musk has built independent, even thinking about Tesla, SpaceX, everything. It is the investor base at Tesla is the, is gotta be the most incredible investor base. Any public company has ever had to let him do what he's doing and to have the vision and everything.

2:36:35Intel is, is just smattered with what Palmer would call wall street weenies is everybody's just thinking about it is how can we optimize this thing as effectively as possible? And it's like, that is not how you build the future, nor is it how you build American dominance. And so really you need that investor base turnover. And it feels like that comes from narrative where if you're telling the story of Intel foundry, we can get a visionary leader. We can have this opportunity. They're already at scale. They have a lot of the technology. They have sort of caught back up. And in many ways, they're still failing operationally pretty miserably.

2:37:06And so the opportunity is like tell that story and get an investor base that is stoked to be investing in the ASMC. Well, thank you so much for joining. It's a pleasure catching up. We'll talk to you soon. Have a good rest of your day. Thanks for coming on, Kyle. Great work. And speaking of Anderil, our next guest works for Anderil. There we go. And she's the general manager of Anduril RMS. And we're excited to talk to her about the new rocket motor facility. But we'll let her break it down for us. Welcome to the stream. How are you doing? We have one minute while we get real. Just keep clapping.

2:37:42Just keep clapping. Just keep clapping. Welcome to the stream. How are you doing? What's happening? Oh, it's a really good day. Just finished a ribbon cutting with our new facility in coastal Mississippi with Chairman Wicker about 10 minutes ago. So good day so far. Fantastic. Tell us more about this facility. What are you making? How big is the facility? How big were the scissors? How big were the scissors? You know, the scissors were not as big as I was expecting. I mean, they were big, but they weren't like the economically big. No, we just finished a full-rate production facility for kind of what the 2020s can do for automated robotic full-rate production of solid rocket motors.

2:38:27We have a 450-acre facility in coastal Mississippi, about nine buildings. And when we moved in here, it was a multiple facility in 2020. And we saw the good bones it had and the ability of what it could do. So we took one of the buildings on the far south end of the campus, nearly quadrupled the footprint, got world-class engineers, specifically from the automotive industry, to really look at what solid rocket motor production could look like in the 2020s with robotics, automation, and really putting robotics everywhere where there's a safety critical or a quality critical component. Speaking of robotics, what's behind you?

2:39:09that is one of two robots that are involved in the casting assembly so we go through a mixing of solid rocket propellant and then you have to cast it where you you're essentially filling a mold like you were a solid rocket motors is a lot like making a cake you you mix together dry ingredients wet ingredients and then you pour them in a mold and you cure in an oven and so this is the robots that's used for the casting process. Can you give us a view into the shape of the company? I know Adronos was the rocket motor company that was acquired, but it sounds like you've hired a lot of people. How big is the organization now?

2:39:48So you are right. We started as Adronos. I was one of the two co-founders of Adronos back in 2015 with my co-founder Chris Stoker and And we grew that into about a 40-person company at the time of acquisition by Andrel, who was previous to that, had been one of our customers. So a little bit of a vertical integration play. We have since grown that organization to just about 200 people today for Andrel Rocket Motor Systems. I sit as the vice president and general manager of that business line. Very cool. What did you see back in 2015 when you started the original company that kind of caused you to take on the mission?

2:40:32It felt like broadly Silicon Valley started kind of being aware of rocket motor supply chain challenges, I want to say like in 2022, but you clearly saw it way, way ahead. Yes, this is all very much before all of the need that has risen for solid rocket motor production enhancements. Really back in 2015, I had developed a brand new formulation called Alatec that can bring more performance to the industry. I saw a need to commercialize it and naively thought that I would be able to progress it to a spot where the two kind of incumbent providers of solid rocket motors would fight over themselves.

2:41:15to get access to the IP. But what I quickly realized is the desirement for them to innovate and get these new propellants was not there. And the only way I would ever be able to see full adoption and utilization of that product was if I was my own customer, if I was the solid rocket motor provider. And a big piece of my PhD at Purdue was looking at innovative ways to process energetic materials, new ways of mixing, new ways of doing things. So I was able to leverage those lessons along with taking what is state-of-the-art for kind of automation and different manufacturing processes, put them all together to create what my vision was for what solid rock and what a production could look like in the 2020s.

2:42:03Very cool. And we really were kind of ahead of all of the other suppliers that have joined into that market. Amazing. Well, thank you so much for calling in on such a busy day. Congratulations on the massive launch and good luck with the next phase. I'm sure you're going to be making a lot of product very soon. Yeah. And hopefully copy. I don't know. Last question I had, is the idea for this new facility to be a model that you would copy and paste to scale production? or is this going to be able to kind of take you guys all the way? No, that's exactly the purpose of what we are wanting this facility to do.

2:42:42It's really looking at a manufacturing line that is agnostic to what goes through it. So we can do anything from a small two-inch diameter motor to something that's really big, up to 32 inches in diameter. And we are agnostic in the manufacturing process. but really focusing on kind of the one-piece flow out of manufacturing practices we can take this this facility and copy paste it elsewhere on our 450 acres I could fit another two full rate production facilities each one if you look at a mid-scale tactical motor like say nine inches diameter we can do about six thousand of those a year out of one facility so if I place another two of those here you know that's just a little a linear increase in production without happen to have a brand new giant explosives campus and then alternatively i can take this small footprint agile manufacturing process and place it strategically uh for integration purposes with our different customers both domestically and abroad and we are on the record looking at you know multiple places for that including australia through the uh the new guided weapons enhanced ordinance program amazing uh well thank you so much for joining the update have a great day.

2:43:52We'll talk to you soon. All right. Thank you. Bye. Up next, we have Sean Henry from Stored. We ran into him in New York City. Got a chance to sit down and have... Staying at the same hotel. Yeah, you're staying at the same hotel. We just had... Well, we had breakfast. He just sat there and listened to us. He was walking by. He was busy. We said, hey, sit down. Sit down and listen. We got to talk about supply chain and logistics. About tariffs and the de minimis exemption. We will welcome him to the stream. How are you doing? Great to be here. Thanks for having me, guys. Good to see you. It was great to see you last week.

2:44:27Where did you land? Where are you now? Back home in Atlanta. No longer having fun in New York, having some breakfast with you guys. Yeah, very nice. And business is good overall? Crashing breakfast, I should add, I guess. It was great. It was a great breakfast company. We were crushing breakfast. It didn't feel like you were crashing because you were probably, it was a busy day for us. It was a busy day for everyone. You were probably even busier and we requested that you hang with us. It was fun. Yeah. And shared some insights on all the different things happening in your world. So yeah, give me the update on tariffs, the de minimis exemption, where things, how we kind of got here and then where we are now, and then we'll get into where we're going next.

2:45:10Yeah, absolutely. So I guess background, there's two types of core tariffs. One is the general commercial. something's coming into a U.S. port and you're charging a tariff based on the commercial value of the goods at that time. And that's largely what was rolled out at Liberation Day a few months back was all of these tariffs, particularly around large imports. But there's a separate category that had a big action last week, which is de minimis exemption, which is really started in the early 1900s, mostly for international travelers. They think, hey, I want to send a package back to my family.

2:45:42Well, there's these rules created that as long as that package is under$800 and in these categories, you don't pay any tariffs. Well, what's happened the last few decades is if I want to store my inventory locally in the U.S., meaning bring it through a port, hold it in the U.S., now ship it locally. If I'm, say, let's paying$2 for warehousing, $7 for delivery, and$5 for the tariff to get it in, I'm paying$14 to ship that locally. But if I never bring it into the US, if I just keep it in China, keep it in Europe, put it in Mexico, Canada, and instead I'm maybe paying$1 for local warehousing and labor, a more expensive package like$9,$10 internationally, but I never pay the tariff, I'm not comparing$14 to ship my unit through the US versus maybe$10 to ship it from international.

2:46:33That's a big difference on your average product. And so Trump and the administration had formerly said that they've been waging more on this since even before Election Day. But they formerly said it was going to go away in 2027. So since the start of the year, brands thought they had maybe two years to get ready for this. And all of a sudden last week, you saw announced on Wednesday that it's going away on August 29th. And so now brands are rapidly trying to move inventory into the U.S. Yeah. And is it apocryphal to say that this all like Timu and Shein like kind of started this trend or did they just merely capitalize on it?

2:47:08Were there other other companies that kind of found this? It's not necessarily a loophole. It's more just like an opportunity that if you exploit it to the max, you can scale a multi-billion dollar business. were using it, nobody exploited it to the same degree. Yeah, I wouldn't call it Chien and Timu just like, you know, the tourist example. It's like they're shipping billions of dollars through this platform, but are they the real progenitors of the idea? They've definitely exploited it at a greater scale than anyone else. Sure. But what's crazy is that if you look at the top 100, let's say Shopify businesses, over 50 of those, roughly was the estimate, are shipping out of Mexico into the U.S.

2:47:52So this is very prevalent across brands of all sizes, particularly very large in the apparel category. And just some quick numbers. Last year alone, there was an estimated$64 billion of commerce that entered the U.S. through the de minimis exemption, which is roughly$10 billion of missed tariffs and roughly$4 to$5 billion of missed U.S. logistics infrastructure revenue. And so, So broadly speaking, logistics, U.S.-oriented companies like ourselves are very excited by this because it's very hard to compete with Chinese and Mexican and other labor, real estate and no tariff versus local fast shipping.

2:48:30So, yeah, I mean, this sounds like a huge tailwind for your business, but what are your customers actually coming to you with question wise? It sounds like there's a very quick shift on the horizon around the De Minimis exemption. So they're coming to you and saying, hey, I need to move all my inventory from Mexico or China or Taiwan or Vietnam into Atlanta or somewhere else as soon as possible. Is that basically what's happening? Exactly. Because, well, one, I think it's important to note that when you look at tariffs, a lot of people apply it linearly. OK, a 10, 20, 30 percent tariff product cost is going to go up 10, 20, 30 percent.

2:49:08That's on the commercial value of the individual unit. and so much of the cost is both margin for the brand, advertising, local U.S. logistics. And so these tariffs don't translate one-to-one for how it's going to impact the consumer. But at the same time, we're not necessarily talking to the brands as much about changing manufacturing origin to get away from those tariffs because the impact can often be absorbed for these mid-market brands who have real profits, real growth. What we're talking about is back to that math I gave of. if you were spending$10 to ship it internationally, avoiding a tariff, but you're doing a one to two week shipment, so a slow delivery, cheaper.

2:49:48And it was, you could do two, three day delivery in the US, but maybe$14 all in with the tariffs. Well, all of a sudden, if you're now spending, or I said 11 at first with the tariffs or in the US, all of a sudden now, if you're going to jump all the way to 14, it doesn't make sense to pay more and deliver in two to three weeks cheaper. But now all of a sudden there's this massive influx of volume where all these brands are saying, how fast can you get me live in the US? I'm in Mexico. I'm in Canada. I'm in Europe. And the problem is with outdated technology, most logistics businesses are going to tell them four to six months.

2:50:21And or with the influx of volume, a lot of capacity is becoming challenged again, like we saw all the way back in 2020 and 2021 with such shortages of warehouse capacity. It's not that there's not leases on the market. It's that ready to go warehousing capacity with all the infrastructure, technology and labor. That capacity is rapidly getting gobbled up. And so we actually had a case study on our website from exactly what's going on where True Classic Tees, great example, massive$100 million t-shirt retailer wearing one right now. Now, earlier this year, they were stuck in Mexico and the Mexican president first signed away their exemption, which was if you first bring it into Mexico and it's here for less than 180 days because we know you're going to ship it into the U.S., you also never pay our tariffs.

2:51:07But please pay our labor, our real estate, et cetera. Well, they first signed that away. There's since been a pause put on it. But we already migrated customers. Take someone like True Classic. They were often told we can get you live in the U.S. in six months. That's insane cost that they're going to face for those six months relatively. We got them live in 18 days from meeting us to go live. And that's all because we're really the cloud in comparison. Makes sense. Last question for me. How solid are we? I mean, when we went into Liberation Day, a lot of it was just uncertainty from the business community.

2:51:43How confident, whether or not it's good or bad for your business, How confident are we that we will be at least on the same course, whether it's good or bad, for the next few years, for the next 12 months? Like, how much confidence do people have that this set of rules that are in place right now will stick? Personally, I think the confidence is going up. I think this is one of the last exemptions people were curious about. And while there are still a few negotiations going on with specific countries, The kind of broad general tariff, the reciprocal tariff and even key countries have already been set.

2:52:17And frankly, I tend to believe the administration is liking the results. So far, we've taken in over$150 billion of incremental revenue since Liberation Day just via tariffs alone. There was a GDP positive or sorry, kind of balanced budget positive month as a result of these tariffs and more. And so I don't think we're going to change course and go backwards. And I think some of the logic is while it may drive a slight increase in cost on the front end, you have to wonder what it's going to do on the back end when there is more production in the U.S., more middle and lower class jobs that have been competing with Mexico, China, these other countries labor on logistics.

2:52:55And ultimately, if they're able to use tariffs to impact other taxation on citizens, leaves me thinking the administration is going to stay coarse on this. Makes sense. Awesome. Well, congratulations on the growth. And it seems like a lot of these decisions are going your way. So happy to have you on Team America building infrastructure and everything that we need to bring products to market in the United States. So thank you so much for joining. Great to see you, Sean. We'll talk to you soon. Absolutely. Have a good rest of your day. Cheers. Bye. Up next, we have Andy from Two Cents. We've highlighted his posts many, many times on the show.

2:53:31excited to meet him and chat about the business and his day job outside of posting because he's built a very, very fun social network where your username is your net worth. Andy, welcome to the stream. How are you doing? Great to be here. Thanks for having me. Are you in a secure location? You've probably upset a lot of people with what you've built, but... Why? It's opt-in. I guess it's opt-in, but it just feels like something people would push back on, right? You've gotten push back for that, right? To a degree. People have been a lot more positive about it than I thought they would be actually.

2:54:04Okay. I think what happened is I tapped into a latent anxiety that everyone actually had. And it seems sort of like people are like, thank God someone's doing this so that I don't have to worry about it anymore. So is it anonymous? It's just your network? Yeah. Yeah. Quickly explain the product for people that haven't seen it. I think a lot of people have probably seen it without realizing it. They've maybe seen screenshots of like polls or comments or things like that, which I think are fascinating. But yeah, explain it for the audience. Yeah. So we're building an anonymous social network where your username is your net worth.

2:54:39The main idea is to provide a verified place to have financial conversations that are difficult to have in person or with other people around you because of the inherent private nature of finance. And so we're collecting net worth information from bank accounts, brokerages, crypto wallets. And then later, we're also going to do real estate and all kinds of other financial assets. And we're trying to calculate this one big number that we can assign to everyone and allow you to speak from your position, basically. What are some of the key learnings, insights? I mean, I think the polling function where users can ask a question and then see, let's say it's a yes or no answer, you can see how much net worth is backing up the results.

2:55:24And so I feel like it exposes, okay, what side is wealth on, on this specific issue? But I'm sure you've had a bunch of other interesting learnings. I think probably the most interesting thing is when I set out to build this, I thought it would be inherently a place for financial discussion where people would talk about tax strategies and ARBs and alpha and stuff like that. And it's really become more of a general place to hang out. And what I figured out is actually every conversation is a financial conversation. Your net worth and the experiences that have led you to that number sort of dictate the way that you are going to look at other things as well.

2:56:01And even if you're talking about something that's kind of boring or weird, that number is still relevant in that conversation, even if you don't bring it up every time you talk to someone. Totally. I feel like there's a lot of anonymous social networks that have gotten problems with bullying and stuff and just toxic communities because of the anonymity. Have you run into any problems? Are you fighting it in any particular ways? Or do you think it like... Yeah. Do people ever come on there and try to eat the rich? It's just net worth negative$2 ,000. We are actually in an incredible position as a social network because our moderation strategy is to just ban your bank account.

2:56:41And so if you break the rules on the platform, we can just block out the accounts that you've connected with. That's so interesting because normally on like Reddit, you could have somebody that's angry and then they could just create, they could be getting banned and just recreate accounts. Yeah, it's like how would you create a new account with a huge bank? You have to set up a new bank account, transfer all your money over there. That takes a long time. And if you do that enough times, eventually you will make other people angry who will ask you why you're making so many bank accounts. We don't even have to worry about it.

2:57:10That's very funny. That's wild. How's growth been? What's the news? How's, are you monetizing yet? I can imagine a bunch of different ways to monetize. I can see, you know, you have interesting data that other social media companies need to kind of piece together through behavior and things like that. Like I imagine the big social networks try to build a profile to understand like what is this person's purchasing power? You guys get it. The user's providing it. But how are you thinking about monetization? I think there's a couple of big strategies that we can try to go after. Referrals is something that I've looked at for a while.

2:57:49If you look at a public company like NerdWallet, a lot of their revenue is coming from referrals to loans and mortgages and credit cards. And from our position, if we have a bird's eye view into your finances, we're actually pretty well equipped to show you products that would benefit you financially. So we're not just selling you something, but we're also trying to sell you something that makes sense with the type of portfolio that you currently hold. Yeah, it's like, hey, we noticed you're not using leverage. Did you know that you could love her up massively? We'd love to refer you to Harry Winston, Rare Jewels of the World, Discover the Sunflower Collection, or maybe it's a - John, what are you reading through right now?

2:58:24I'm reading How to Spend It. It's the wealthy section of the Financial Times that comes on the weekend. Refer you to Vacheron Constantin, The Quest. Vacheron Constantin celebrates Seeking Excellence for 270 years. 270 years of doing better if possible. What were you doing before this? These are going to be great referrals. You got to contact the ad salesperson at HTSI. They got the graph diamonds ad in there. This is going to sell like hotcakes when you start monetizing. That's the broad plan. That is the broad plan. Before this, I was doing basically a bunch of different side project apps. Two cents started as a side project.

2:59:03And I was just trying to figure out something that could ship out the door three or four times a month or a year and see if anything would get traction. And so the early growth on Two Cents has been really amazing. It's probably the most popular thing I've ever built and almost entirely organic. So it has come from Twitter on the timeline, from other posters wanting to sign up and be on it. I've actually gotten a lot of comparisons to this earlier app, Poster, which was popular a couple years ago. I invested in Poster. Really? Yeah. It was a wild place. Last contrarian. But it's so, you know, people are, it's interesting.

2:59:41A lot of people will use two cent that have never had an alt or an anonymous like account. And I think it becomes extremely freeing for them to feel like they can have conversations online without being tied back to their LinkedIn, you know, or being tied back to their family. And I do think, yeah, I just think you've tapped into something super interesting because money is just at the core of everyone's life, whether they want it to be or not, and giving people a place where they can talk about it freely. Of course, there's still going to be judgment. Like, of course, there's going to be people on the app that are going to think a certain way about someone's financial situation or, you know, what they're saying or how they're spending their money.

3:00:25but it does um i feel like the next uh i feel like it'd be funny to build out uh i'm curious would you ever build out like a feed of like the users kind of like if they obviously opted into it but just like purchasing uh i feel like i feel like a lot of people would would be very fascinated in that um in a different way than like venmo used to do it which i was like why is venmo a social network you know it's like i don't want like i don't want people to know what whatever. Yeah, this is on the roadmap. We would love to have a feature where you can opt in your transactions, your recent transactions, and then almost quote tweet them and say, I bought this$18 shawarma at this bodega.

3:01:08It sucked. Here's a picture of it. And it kind of proves that you did that, right? You're not just saying it. And same thing for brokerage movements. If you sell a bunch of Apple stock or whatever, you can write your due diligence and then also prove that you actually made that sale. So you're not just saying you did it, we're proving it through financial history that it happened. And obviously that's going to be entirely opt-in. You can choose to show things or not show things, but it'll be the only social network where you can have these types of conversations, where you can prove that you're actually moving your money or making these transactions that you're claiming that you are.

3:01:41Very cool. Speaking of money, have you raised any money? We have. We have raised$3 million from Dragonfly and Starting Line. Congratulations.

3:01:52And was that just announced today? I didn't see it anywhere. It was yesterday. The official two cents account on two cents increased by three million. Right. And so you saw it. Our spot on the leaderboard has jumped up a couple of places, which I'm very happy about. There we go. That's great. What's the number one spot on the leaderboard right now? Oh, yeah. Who's the richest? Right now, we have... Well, so I don't know because it's anonymous. But they are at 26 million. 26 million linked. Verified liquid, which is crazy. certified in liquid. Might be an AI researcher. Maybe. You got to get MSL on there.

3:02:278 ,000 ,000, 3333 every month. You're like, okay, I know who this is. Got a$100 million contract. Anyway, thank you so much for stopping by. Great to meet you and have a great rest of you. Congrats on the round. Thank you for having me. Bye. Cheers, Andy. Up next, we have Gabe from Harvey AI. Very excited to talk to him about legal AI and the impact of open source. I want to know if anything changed this week for his business. We will bring in Gabe from Harvey AI, the legal. Hello. Welcome. How are you doing? What's going on? Good. How are you guys? We're great. Great background. Great background.

3:03:03About to say. Is that mahogany? Yes. The official wood of business and law. Yes. Mahogany, some books. Got to look professional. Fantastic. What's the latest? Give us the update. I have a bunch of questions that will go all over the place, but give us the update in your world. Yeah. So last week hit three year anniversary, have kind of scaled up to 350 people, 500 customers just hit 100 million revenue. There we go. Love hitting the gong for revenue. We hit the gong a lot for$100 million rounds, but it feels better when it's real customer demand. We have this silly soundboard where replay the overnight success sound, but...

3:03:46Overnight success. Yeah, that one. And usually it's because some founders have been grinding for like 10, 20 years, and then they finally hit breakout success. But like Harvey is kind of an overnight success. What actually set you up for this? Obviously, you were in a unique position to actually go after this opportunity. This wasn't just something that you just lucked into, obviously. So give me a little bit of the prehistory. Yeah, I actually I started doing research in like 2012 and tried to do some other startups. And so I think there was a lot of this experience of trying to build these companies without this technology.

3:04:23And I think I was at Meta on the large language model team right before Harvey and my roommate, who is now my co-founder, Winston, was an associate at a law firm. I was showing him kind of the GP3 models and he was showing me his kind of the legal work he was doing. and I think it was pretty obvious these models were going to get much better and that seemed like a great application. And then I do think there was like some factor of luck of, you know, raising from open AI how good GPT-4 was and then kind of really focusing on big law. Yeah, so you add all that in, then we can do the overnight success because it's like easy.

3:04:57If you want to just grow really quickly, just spend a decade researching a fundamental technology that will change the world. So anyway, what? Yeah. So much to go into here. Yeah. I mean, one thing I want to, you know, you at this point, I'm sure are talking to or have closed all, you know, pretty much all the top law firms in the United States. I'm sure the reaction internally at all the different firms. I'm sure there's some excitement. I'm sure there's clearly excitement. Otherwise, they wouldn't be signing up and becoming customers. But I posted something last week. I was in New York, and I was speaking with a lawyer randomly, and he was telling me about their AI adoption internally.

3:05:44And this was his point of view. He said, adopting too much AI will hurt our bottom line. And I posted that note yesterday and it sparked a big conversation. You know, the obvious response is that you don't pay great law firms and lawyers for the execution. You pay them for the overall product and the sort of like guidance and advice and strategy and all these other things. It's not just like the raw like generation of documents or reviewing contracts and things like that. So how do you know, how are give us kind of the overview of how different firms are reacting and sort of leveraging AI? And then I want to hear how you think law firms might need to like reinvent themselves and like kind of even adopt new business models and approaches and things like that over time.

3:06:36Yeah. So it's definitely like a wide range, both law firms, but also kind of individual partners within law firms. And so we have firms that are still in the, let's test this out, let's see what to do with this. And then firms that are, we have already co-built software products or models with them, and we're revenue sharing and selling that to their customers. So there is kind of the full range. and yeah i think it'll be interesting to see how it plays out i think my guess is long term where this is going is i think your point is exactly right like i actually think if you look at partners billables they're actually undervalued where it's like if you want to do a massive public merger that partner is worth much more than 3 000 an hour but it kind of gets subsidized by all these associate hours and so i think there'll be some changing of you know how you charge for these but so far I think we're still kind of early days in that conversation and it's much more how do you you know get adoption how do you start upskilling lawyers to use this technology but do you have a lot yeah the interesting thing I think you know we obviously have a bunch of founders that watch the show and I think it's interesting to look at the invoices you get from your law firm and and just like look line by line where was I getting like real value and legal advice and strategy and all these things that are extremely high leverage?

3:08:00And where was I paying to like recreate a variation of this document or fix a punctuation error and things like that? And obviously you can't see exactly, but it is easy to imagine in the future where in the same way that you have a 100X engineer today that's using, you know, Cognition, Devon and Windsurf or or Claude Code or whatever variety of cursor, et cetera, where they can be ultra high leveraged because they effectively have an army of people working under them agentically. I think you'll see that in, it's hard not to imagine that happening in law where you have this like incredible individual player who's able to like produce potentially an order of magnitude more work in the very near future.

3:08:49Yeah, and one thing we're also starting to think about is like it's always interesting to ask why do you have this billable model structure in the first place and i think one big problem is pricing legal work and so like the reason you have this model is like these projects are so complex that it's actually impossible to do fix a fee without you know this technology where it's just how much does it cost to do this merger it's like you're going to have all these things that come up but i think as these models get better the same way it's hard to predict like how long will it take to fix this bug but now with these models you can kind of look at a code base and say oh actually I can tell you you know some good error bounds I think there's things like that you can start exploring that make more of these new business models interesting.

3:09:32Have you seen have you seen lawyers in big law spinning out and building firms from the ground up to try to disrupt themselves and create new models I mean we've seen we've had the founder of Crosby on which is like they're building a firm themselves and like focusing on like a few specific types of contracts initially. And I could imagine it's kind of hard if you have a law firm with thousands and thousands of lawyers, it's hard to like totally change the business model where if three people spin off and they come to you at Harvey and say like, we want to leverage this ground up, build a firm around this.

3:10:03Are you seeing that at all? Yeah. One of the fastest growing, I think they might be AMLA 200 firms now is actually partner only. And so I think this model is I think that the challenge is it just depends on the type of legal work you're doing. But this is also traditionally like not traditional in the past couple of years, what firms like PwC have also been doing of how do you bundle this legal work and provide alternative ways. So I think some of this work is getting unbundled and with different models, you can kind of tackle different parts of the legal work. How much how much do you guys care about models getting more intelligent versus just delivering on the current capabilities?

3:10:42Yeah, I think for us, the big challenge we definitely still have is like the models aren't good enough to do the very complex legal work. And so I think for us, you know, Sam has the quote of like, try to build some company that like as the models get better, your company gets better. And I think we are very much in that position where it's like, I think as amazing as these models are, and I think lawyers get a lot of value, we still have so many use cases where it's just like the models can't draft these very complex like merger agreements and things like that. They can't look at these massive case law research corpuses or discovery corpuses or data rooms and like really understand them.

3:11:17And so I think there's a bunch of room both for the foundation models to get better, but also we're starting to think about like, can you, you know, custom build these reasoning models given this kind of specific data you have? So I think there's huge room for improvement for the models. What about the back office? There's a lot of costs that go into law firms that are non-legal work, operations, billing, all that, is that a focus at all for you? Is that something you think about? Or is there just enough work to be done on the client's side? Would you do a legal billing product too? Yeah. I think that's something we've thought about of partnering with the folks doing that in the future.

3:11:56We've had law firms ask, can you help us automate parts of how we do billing? We have a bunch of firms that just buy Harvey Seats to do some of this back office work again. And then there's clients that have asked about, you know, can you help us build spend management, things like that. But I think right now the focus is still kind of building the product for lawyers, for law firms. We have a question from the chat. Is the company named after Harvey Specter, the fictional lawyer from Suits, or Harvey Dent, the Batman villain, who I think also had a law degree? Harvey Specter. Harvey Specter. There we go.

3:12:29Makes sense. It's a great it's a great character. I want to talk about costs. Obviously tons of revenue coming in the door. Also we're seeing this weird phenomenon. Maybe it's not that weird but like with test time inference reasoning you want to put you know GPT what was the ArcGGI they were putting two thousand dollars of inference behind every task just to crack that puzzle. If I'm you know coming to you and I want you know to really make sure that the AI is checking its work fully. I probably want to put a ton of tokens behind that. What does the cost structure look like? How does that evolve over time?

3:13:05What are the relevant changes that could happen and changes in the landscape? We saw the open source model come out from OpenAI. At the same time, even if you switch to open source today, you still have to pay for energy and GPUs and stuff. So how does the cost side of the business evolve over the next couple of years yeah i think our bet kind of from early on has been focused much more on like capabilities these models are going to get cheaper as you scale get economies of scale you can kind of solve these problems i would say right now our cost structure is similar to um you know if you're using like a chat gpt where you know that gets amortized over the tasks you have some tasks that are low compute some that are a lot i think legal is actually one of the very unique domains where you have these tasks where I want to draft a motion.

3:13:52And typically this is going to cost$250 ,000. I want to do document review and this can cost a million dollars. And so you have these specific problems where you can actually just run models over every single document, throw just a ton of compute at this. And so we're starting to kind of explore like what do those different models look like if you're doing tasks like that. But yeah. Lessons from Atrium, lessons from Clear Spire, lessons from binding a actual law firm to a technology company seems like it hasn't not doing that certainly hasn't hurt you but uh did you know yeah and i'm sure i'm sure you had a lot of investors that say why don't you sell the work instead of the software exactly yeah we actually we actually had the opposite like when winston and i started the company we talked to the founders of Atrium and like 30 of the folks that worked at Atrium.

3:14:44And actually, Jason Kwan, who at the time was the GC of OpenAI, was at Y Combinator when they did Atrium and Sam was as well. And so we kind of like mentioned that and they were like, you guys should focus on building tech business. And for us, it seems like the scalable play here is like, we don't want to build a professional service firm ourselves. We want to enable every law firm, every professional service firm to kind of become the next generation of firm. Yeah, it makes so much sense. Microsoft Excel, their team didn't build an investment bank, and yet it's used by every investment bank, and that will continue for a long time.

3:15:24Anyway, I think we have to jump to the next guest. Jordy, do you have anything else? Thank you so much for coming on. This was really fun. Incredible progress. Congratulations on all the growth. Huge. We'll talk to you soon. Bye. And up next, we have another massive gong moment. Let's bring in Kareem from Clay. We'll get the update from him. Welcome to the stream, Kareem Clay AI. How you doing? What's going on? How you doing? Give us the latest. Give us the update. What's new in your world? What do you do? Break it down for us. Can give you all those things. Excited to be here. Thanks for having me.

3:16:01Yeah. Well, today we announced that we raised$100 million from Alphabet's capital G. Congratulations. Thank you. And a$3.1 billion valuation. Wow. Three Jet Blues. That's crazy. Amazing. We've been dogging on JetBlue a little bit. That is fantastic. Congrats. Classic overnight success. You started the company like a few months ago. Yeah, like eight years ago. Yeah. I was going to say, so the story. So give us the journey, how you got here, because I think you guys have been a massive beneficiary of LLMs from what I know over the last few years, but it was not an overnight success. I would call it a spiral.

3:16:47Spiral. Yeah, I mean, we started the company with the idea of how do we give the power of programming to an order of magnitude more people. And I was saying it took me some time to reformulate that to how do we give that to what we call go-to-market engineers. So it's like half of the company can't code, but has all of these ideas of how to grow the company. And Clay basically gives them the power to program revenue instead of just program software. Let's make that a lot more concrete. There's been this cool wave of designers going from 0x engineers to 1x engineers, and that's had a lot of benefits in organizations.

3:17:28But what does that actually look like concretely on a go-to-market team? Is that just generating emails faster, managing a CRM? Like what are we actually talking about in terms of detail? Yeah, the actually like most important part is who should you reach out to? That's actually like, I would say 60 % of kind of like, there's two ways to grow your company. Either you find new customers or you expand existing ones. If we just talk about finding new customers, 60 % of that is finding the right person, right? And then kind of like reaching them at the right time is another 20%. So what we do is, let me give you an example, one of our customers, they have a ton of people just going on Google Maps, looking at all the different locations that they sell to finding warehouses, and then manually counting the number of parking spots that that warehouse has, because that's a good indicator if they're a good customer.

3:18:23And so we can use Google Maps API and LLMs to do that counting for them. Oh, that makes a ton of sense. Yeah. So really, most of it is about like account selection and finding the right people. And then the second part is, you know, that they just raise a round of funding like us. Was there some kind of new news story that they just hire a few people in a department that you're selling to? And it's very specific to what your company does in their product. Got it. is this uh every of us this question on the on the show before but uh it feels pretty common that you know some random private equity firm will just be spamming people on the internet trying to buy their companies oftentimes there's probably good targeting oftentimes there's bad targeting how close are we to just like eliminating those those emails that are just kind of uh those out that outreach the spray and pray outreach when is spray and pray era gonna gonna end i hope now No.

3:19:20I mean, I think that that's kind of like one of our top goals is if you actually, I think you can think of it as what if you actually put in the effort and the thoughtfulness into figuring out who actually wants your product. One, you reduce spam and two, you actually increase like your own productivity. It's just that that's hard to do at scale. And the best companies like Ramp, like there's a lot of companies who figured out how to do that internally. What Clay is doing is giving that to every company. And really what that means is that businesses can find each other faster and more accurately.

3:19:59So I think it should end up benefiting everyone. That makes sense. Anything else? $100 million, how are you going to spend it? Hopefully we don't spend it. yeah always a good play yeah i think that um one of the key like really exciting things about clay is that we have this huge community um that has emerged around us so there's uh clay agencies so people who implement clay oh that makes sense yeah there's like 60 clubs in 29 countries where people meet up to talk about different techniques because if you think about it yeah in sales like you're always figuring out how to be creative how to be unique um and clay is a tool that lets you do that.

3:20:42I almost think of it like Ableton is for music or like you said, Figma is for designers, Clay is for go-to-market teams. They're doing deals. Exactly. So we want to use that money to support that community. So we're having a conference actually sculpt in September in San Francisco. So we're going to do more events like that. And then there's a lot of investment into the product. We're just touching the surface of what we can build. Yeah. Is there any functionality to share workflows or create some sort of marketplace for workflows? Like, you know, yes, there are agencies on top of Shopify that do design, but then there are also templates.

3:21:22And so you gave some examples of, you know, new funding announcement in my category. That seems like something that I could be able to just pull off the shelf and not even build from scratch within Clay. Yeah, totally. I feel like you should really join our team. this is exactly where you're going. I'm leaving you a roadmap. Yeah. So we'll leave it there. Yeah, we'll leave it there. Well, good luck. Very exciting time. And congratulations on the massive round. Awesome. I appreciate it. Thanks for having me. We'll talk to you soon. Bye. Good stuff. Well, John, I got to get on with London. And I'm sure you got to get on with Taipei.

3:21:59I do have to get on with Taipei. Why don't we end it there? Thank you for tuning in to the show today. And thanks to everyone in the chat. I agree. The deeper the V-neck, the more funding you get. That's right. Good strategy. It's a good, good. I didn't say it, but I think it's true. $100 million deal. Really deep V. Looking good. The next round is down at the belly button. It was funny because clearly you could see the bottom of the V-neck if we had shown the full screen, but we had the ticker. And so the ticker made it look like you don't know where that thing ends. It could just go forever.

3:22:32But Palmer Luckey has made. It's actually a vest. Palmer Luckey has made the Hawaiian t-shirt iconic. Maybe Kareem makes the deep V iconic at Clay. Good energy. I was into it. Awesome, folks. Thank you so much for watching. Leave us five stars on Apple Podcasts and Spotify. Leave us some reviews. Leave us some reviews. Send us some notes. Send us some reply to some threads. Whatever you want to do. Get in touch. We'll talk to you soon. Have a great evening. Bye. See you soon.