Tokenmaxxing, SF Street Name Auction, Corporate Retreat Gone Wrong | Riley Walz, Aditya Bandi, Zach Shore, Hongwei Liu, Zak Kukoff, Thomas Laffont

You're watching TVPN.

Today is Tuesday, April 7th, 2026. We are live from the TVPN LTO, the temple of technology.

The fortress of finance. The capital of capital.

Let's pull up the lineup because we have a great show for you today, folks. We have Riley Walls, the jester of Silicon Valley. You had a different name for him when he came on. There was something else.

Internet rascal.

Internet rascal. That was it.

He is a rascal.

He's coming on to talk about the auction to name a street in San Francisco. We have a whole bunch of funding news. Aditya Bandy from Noon, Zach Shore from Hermias is coming on. Wang Wei Lu from Mapadin is coming on. And then Zach Kukoff and Thomas LaFonte are joining us live in person in the TVP and Ultradome. Well, there is a whole bunch of news to run through. The first story is that Meta employees are apparently token maxing and competing on an internal leaderboard called Clawdonomics for status as a token legend. This is from The Information. Over a recent 30-day period, total usage on the dashboard topped 60 trillion tokens. And this sparked a huge debate over how much is Meta actually spending with Anthropic. Of course, the other big news is that Anthropic just passed $30 billion in run rate revenue with one of the— probably the steepest revenue growth chart in human history. Absolutely legendary.

Yeah, this chasing status as a token legend reminds me of —kind of, maybe it was a year ago at this point, you were saying, like, will tokens ever become like eyeballs the way eyeballs were during the dot-com era? Just optimized for eyeballs. Obviously, not every eyeball visit to a website is created equally. But people were optimizing for eyeballs. And now, you know, the reaction to this, I think, has been generally at least online, like been, I guess, reassuring. A lot of people are saying, Gary Basin says, "Ew, why?" Marty says, "Goodhart's Law, when a measure becomes a target, it ceases to be a good measure." And so who knows what's actually going on internally, but we do know Zuck is pushing the entire company to be as AI-native as possible. And this guy loves spending money.

Too, right? I have a crazy bull case here that I will run through. Let's get through some of the story. First, we gotta pull up this comic from xkcd in the comments here. When a metric becomes a target, it ceases to be a good metric. It's right under the leading post. There we go. And says, and the other counterparty says, sounds bad. Let's offer a bonus to anyone who identifies a metric that has become a target. It is good. I don't think that's what's going on here.

Max Leiter was texting a friend at Meta. And sent the post we just discussed on token maxing and said, true. And the person said, yes, it's pretty sad. But I mean, imagine, so Meta has been, there's been rumors of Meta layoffs for a while now. Unclear how many, if any, if any have happened. But if you're sitting there, the company, Zuck is saying like, we need to get AI native. Boz is saying we need to get AI native. And then suddenly there's a token leaderboard. You do not wanna be at the bottom of the list. I will say that, right? You don't wanna be the guy who's having to explain like, no, well, I'm actually getting the most out of each incremental token. And the other guy's just like set up an agent that just counts 1 to 2 million over and over and over.

Just checks out every single line of code that you've written or something. Yeah, I mean, you have to measure the actual output, the impact on the business. I mean, fortunately Meta has been a huge beneficiary and a huge winner of AI. The ads are getting better targeting. They're seeing, they're delivering more ads and the quarterly earnings have been strong. The headline number here that sort of took everyone by surprise is that Meta's staff used, this is from the information, Meta's, the story claims that Meta staff used 60.2 trillion tokens over 30 days, which would pencil out to about a third of Anthropic's ARR was the number that was thrown out. But both of these claims are pretty questionable. And so Tyler did some back of the envelope math to show, that the 1/3 revenue estimate is way, way too high. And I don't know, do you want to take us through some of the reasoning there? And then we can talk about the knock-on effects of all this. Yeah.

Okay. So 60.2 trillion tokens is the number. Like, we just assume that's true. So basically, I'm going to assume that all the employees are basically just using Opus 4.6. Yeah. So then there's basically 3 numbers you need to look for in like the API cost. So there's like input. Yep. There's cached input and then there's output. Sure. So for Opus 4.6, it's $5 per million tokens on input. It's $0.50 per million tokens on input cached. And then it's $25 on output.

Yeah. So if you multiply that 60.2 trillion tokens at the highest possible rate, $25 per million tokens, then you do get like a billion dollars in a month, which is crazy. That's not what's happening.

That's a crazy number. But you have to think about it like, If you're using Claude Code or any of these coding agents, the vast, vast majority of the tokens used is input. Imagine you're working on some coding file, right? There's 1,000 lines of code in the file. Maybe the model's only changing 10 at most, right? That's a very small percentage. The output tokens are going to be a very small percentage of the total tokens going in, right? So OpenRouter publishes like a lot of this data. So you can kind of use those ratios to figure out what is actually like, what are the actual numbers of the input versus cache versus output? Yep.

So just to get sort of like market standard averages, like baseline benchmarks. Now Meta could be using these tools differently, but if we are to assume that the shape of their agent decoding efforts are similar to the average, this is what the numbers would look like.

So maybe there is some bad incentive where people are just saying to the model, count up to a billion and then do it again. So then it's totally skewed. But if they're doing it relatively normally— so on OpenRider, it's about 98.9% of all tokens are input. And that's including cached ones, right?

Because you're stuffing the context window with all your code base or a huge amount of context. It's going and finding some stuff.

And that's not changing every time, so you can cache it. So then that's like 1.1% is output. Yep. So basically, if you basically get all the numbers, that means like the kind of mean token, the mean million tokens is going to be $2 and like around $0.26. So that'll get you to something like $136 million a month for the 60 trillion tokens, right? So that's like way less than the 900. Yep. So that would be $1.6 billion a year, like run rate. Still huge. That's a lot. But that is still in the max.

Assuming they're in the top.

Yeah, that's assuming that OpenRouter, the kind of breakdown of how they're using the tokens is the same as OpenRouter, which I think it's not. But if we assume that, that's like $4,500 per engineer if there are, I think, 30,000 engineers at Meta every month. $4,500 on tokens. $4,500.

That's actually in line with what I've heard a lot of other people spending in terms of their token budgets. Yeah, that's not like absurd. That's not absurd.

You're trying to incentivize people to use that. Yeah, yeah, no, not at all. But so you can actually see the breakdown on OpenRadar of how people are using tokens. So 17, the biggest plurality is OpenClaw, which is 17.6%. Yeah. And then Claude Code is 16.8%. Sure. So I think if you think about Claude Code, you would imagine that like in Claude Code, there's the kind of percentage of cash tokens is gonna be higher than in OpenRadar. Claw. Yeah. So I think Meta's usage is actually going to be more heavily based on the, the cached tokens. Sure. So if you do it just based off like Claude code usage, you'd actually see a higher percentage of, of the input tokens be of the total tokens. So it's only like 0.8% is the output. Yeah. So then if you get all those numbers through again, it's only like 55 million a month, which would be $669 million a year. Yeah. And each engineer would be like $1,800. Yeah. Yeah, that's actually pretty low, which is like, I think, very reasonable.

John Chu over at Coastless says, plenty of my Meta friends told me folks have been building bots that just run in a loop burning tokens as fast as they can due to this policy. It is an app. It's an absolutely stupid policy and is similar to how Meta uses lines of code to measure engineering output. Managers are supposed to use it as a proxy and dig in to understand work complexity, but plenty of managers are lazy and just don't. That was in response to Christina over at Linear saying, "Ranking engineers by token spend is like me ranking my marketing team by who spent the most money. We may not have hit our KPIs, but Joe spent $200,000 on a branded blimp that only flies over his own house, so he's getting promoted to VP." I'm pro-branded blimps, though.

I like that idea. So my take on this was that, yeah, it sort of ties to what Jensen Huang was talking about at GTC. He was saying that an engineer that's making $500,000 might soon command something on the order of $250,000 a year in token budget. Andrej Karpathy had a similar line. He said, "It's all about tokens." He said on a podcast last month, "What is your token throughput and what token throughput do you command?" And so Meta actually has two different harnesses internally. They have a version of OpenClaw called MyClaw, and then they also, of course, acquired Mana harness, but it appears that they're running Claude, maybe Opus under the hood to actually generate the tokens that come through those harnesses. The interesting thing is that at $250K AI budget per engineer, you're at like $20,000 a month. And so based on Tyler's math, this feels like, okay, there's gonna be another maybe 4x to get to Jensen's prediction. The baseline that was going viral around, oh, Meta spent to like maybe $1 billion last month with Anthropic, that would work out to like $83,000 a month, which is absolutely insane. And I don't think anyone's really thinking that that's what's going on here. But this felt, a lot of people were saying like, this is, you know, there's a lot of negative things.

The other thing is like ARR is the annualized run rate version, right? Which does come down. We don't know how they calculate it, right? Sure, sure, sure. I'm assuming they're not choosing like a Saturday. Yeah. Like multiplying that, right? Sure, sure, sure.

But the bigger question is like a lot of people, you know, John Chu there is saying that, you know, oh, this is like a bad metric and maybe it is, but I think it makes clearer the strategy with Meta Superintelligence Lab because if you're looking at, you know, it's clear that they're spending hundreds of millions of dollars on this just for internal code gen tooling, like running their business, they are going to, spend an inordinate amount of money on frontier inference. And so training a model there, they will be able to amortize the training cost of the next model that they build, not just over can they get a product out that goes viral and becomes its own standalone chat app that people pay for, or maybe it's ad-supported. Like just on the internal usage, they could be running a you know, multibillion-dollar token bill that they would have to pay another lab. And so if they develop that internally, it's pure vertical integration. And then you also have everything that's happening on the actual ad targeting and content delivery side. And when you add up all of those, all of a sudden the big question has been like, does Meta— is Meta going to be able to launch an entirely new AI product like Vibes or something like that? And This is a data point that to me says they don't need to, because just from a pure vertical integration story, the investment in MSL can pencil out. What, are you laughing?

I just want you to get to your schizo theory. What's the schizo theory? That this whole token maxing thing is like a barrage while they distill the model. Oh, oh yeah, yeah.

I mean, there is a world where if you're running, If you're generating trillions and trillions of tokens of a frontier model.

They're like, wow, Meta is really burning through a lot of tokens.

And you haven't generated everything.

It's like, oh, we're just token maxing.

Yeah, I mean, there's another story about distilling we'll get to later in the show, but there is a question about if I write an essay and then I have a model rewrite it, those tokens, they are from that model provider. I buy them, they become mine. Can I train on them? That's probably out of terms of service. So you would think no, but you sort of wind up in this Ship of Theseus world where if Meta pays Anthropic $100 million or $1 billion to go rewrite every line of code, every email, every Slack chat, every internal message, like basically map the entire organization, rebuild it, they wind up with an incredible training corpus that they can use for their next model. But I would imagine that they can't. And I imagine that the enterprise contracts go both ways. They can't, you know, the lab can't train on the corporate information. That's standard in all of the enterprise contracts. And I would imagine that the opposite is true as well. Although it is this fuzzy Ship of Theseus world where, If you're using coding agents to upgrade your infrastructure and then you wanna run and train some model on your infrastructure, do you have to pull out the tokens that were revised by the AI lab that you don't have the right to train on? It's all very interesting. Apparently startups that have gone out of business are able to sell their corporate histories for something like $1 million to data brokerage firms and AI labs now. Have you heard about this?

Yeah, I heard about it. Skeptical? I'm skeptical. I mean, certainly there's a market for it, but—

Basically all the code that a company built over a few years, maybe they raised—

Code, but also usage within different enterprises.

Yes, all sorts of different stuff. Basically, like an RL environment or something that can help whatever business process they were doing. So if you're— and you could imagine a data broker buying the dataset from a startup that has their go-to-market motion that can be RL'd on, and then also their code base that can be folded into training on how to write better code and all of their marketing messages. And basically everything that they did on internal, external comms, everything is tracked. Maybe they've been using Granola or something. They have like very detailed notes of everything about how they built the business, even though it wasn't successful. There's gonna be a lot of lessons there that can be folded into the next trend of the next thing.

Anyway, in other news, yes, Intel is joining TerraFab. Yes, let's read. Intel is proud to join the TerraFab project with SpaceX, XAI, and Tesla to help refactor silicon fab technology. Intel says our ability to design, fabricate, and package ultra-high-performance chips at scale will help accelerate TerraFab's aim to produce 1 terawatt a year of compute to power future advances in AI. And Robotics and throwing up a post of hanging with Mr. Musk himself. Intel up on the day, up almost 3%, unsurprisingly, and just continues to be on a tear, up almost 15% in the past month and 167% over the last year.

Let's go through The Wall Street Journal's coverage of this. Elon Musk is partnering with Intel on his ambitious TerraFab project, which aims to build specifically designed chips for SpaceX and XAI, as well as for Tesla. In an announcement Tuesday, Intel said it would work with the companies to design, fabricate, and package ultra-high performance computing chips at scale. The company shared a photo of Chief Executive Lip-Bu Tan shaking hands with Musk, CEO of SpaceX and Tesla. The partnership is a win for Tesla, which has struggled in recent years— Intel, which has struggled in recent years, leading the company to cut production capacity when demand was surging for data center chips and when competitors like Nvidia and AMD have thrived. That was always just such a tough pill to swallow when you would talk to the ASIC companies like Cerebras and you would say, hey, like, you're doing something new. You're not, you're not doing Nvidia chips, is there any way you could get off of TSMC? And they're like, no, like we still need to be in Taiwan. Obviously there's a huge geopolitical component here. We can get into all that. But last year, the Trump administration reached a deal to acquire an equity stake in Intel for around $9 billion to help secure the American chipmaker's business. The US government held 8.4% of Intel shares outstanding in March as of March 20th, according to securities filings. The figure doesn't include warrants that could increase the government's equity stake in Intel. So as you mentioned, Intel shares gained 3% in Tuesday trading. Terafab represents a step change in how silicon logic, memory, and packaging will get built in the future, Liputan said on X. Tesla and SpaceX confirmed the partnership in posts on X. In March, Musk unveiled the plans for a single facility in Austin, Texas, to make chips to be used by SpaceX and XAI, which merged in February as well. As by the publicly traded Tesla. He pitched the project as an opportunity to quickly experiment on chip design by designing and manufacturing the chips in one facility. The fab will make chips for use in Tesla's robotaxis, which they're already fabbing, I believe, at Samsung, although they do have NVIDIA Dojo chips, I think, that are TSMC. So they've been using both of those fabs, but Optimus will also need chips, and they are planning to use Intel for that as well. So these are two areas of priority for the electric vehicle maker as it shifts its focus to artificial intelligence-enabled products. It will also make chips optimized for use in space, where SpaceX is planning to deploy huge numbers of satellites capable of handling computing tasks. You know, yeah.

So who else do you think they need to get involved here? Because just the two of these, the two of these got, you know, Intel, Yeah. And Tesla coming together. It's good to have more involvement. But still, I think the entire project—

No, we've seen a few of those like AI leader gatherings in DC where you see Tim Cook and Sundar and Sam Altman and Dario and all the leaders are together. And I was always hoping that at one of those dinners they would say, okay, Everyone's going to try and say the biggest number, but this time it's going to be how much you're committing to Intel and how much you'll, how much you'll buy from them if they come online with a competitive product. Because the demand side has always been a big problem for Intel, that they have the capability, they have the plans to build the 2-nanometer, 3-nanometer plant, like a frontier plant, leading-edge fab. But every other company has been so tied to TSMC But we— but I think everyone now acknowledges that TSMC is not— is not investing super heavily in CapEx. They're not going, you know, they're not scaling up as much as— as the industry would like them to. And so lots of folks have— have sort of signaled towards a chip bottleneck coming in the next few years. And Intel has the opportunity to communicate that this seems like the first step in that— in that chain. So companies including Tesla often design their own semiconductors but need a supplier to actually make the so-called— make them in a so-called chip fab. Musk's companies have sourced chips from a wide range of suppliers, including Nvidia, Samsung, Taiwan Semiconductor. Oh, I got it. Musk said that TerraFab is needed because his company's demand for chips is slated to far outstrip the supply it gets from partners. I was listening to Chuck Robbins from Cisco talk about data centers in space and the heating issue came up and he was like, "Yeah, I don't really have like a solid answer for that yet." But I do think that if you are bullish on data centers in space, you have to start with the fact that Starlink works in space currently because it is doing compute.

You couldn't possibly put, let's be honest, John, we couldn't possibly put a computer up there.

Yeah, like there are computers with like, they can't inference frontier models, they can't, you know, it's not gigawatts in space yet, but there are, I believe across the entire Starlink cluster, megawatts of compute in space with solar panels and they do heat up because you are running a chip that routes packets across the internet from one, satellite to the next to get you your internet via Starlink. And so it's not that it's a solved problem, it's that we are actually, we are on a path to, you know, deploy some level of compute in space, Tyler.

Yeah, I mean, we've seen like Philip Johnson, like there are chips in space right now, like there are GPUs, I think, aren't there? He said there were like 5 or 6 H100s, right? Yeah, yeah. So like they do work. It's like the, I think most people's problem with, space data centers is that it's like economically it doesn't make any sense.

Well, so yes, that is the correct angle, but a lot of people are getting— Not that it's like hungrily— No, no, there is a whole conversation about like, it is impossible and you need to like move past that into the economic equation, which then gets you into timelines and actually thinking about what needs to happen to dissipate that heat. But clearly, yes, you can. I mean, you can put humans in space. On the ISS and cool that. They're like, we have created ways to move heat around in space for decades. It's obviously a new challenge. But I think starting with the baseline of like, there is compute happening in space right now. We're going to try and, I mean, Elon wants to like 1,000x it, 100,000x it, 1 millionx it. I don't even know what the scale is, but orders of magnitude. And so there's new engineering challenges, but at the very least, it's worth acknowledging that there is computation happening in space at scale with the Starlink cluster.

Anyway, speaking of space, looks like Elon is going to use SPCX as a ticker for the SpaceX IPO, which he had to acquire from Matt Tuttle, hence the ETF's ticker change shown below. Eric from Bloomberg says, we predicted this could happen in a December note. Nice catch by Will, who famously gave the Meta ticker to Zuck. I did not know that Will Hershie had the Meta ticker previously. So, we know somebody that squats on— Who had the Meta ticker? A guy named Will Hershie. Oh, interesting. He has a company called Roundhill. But we know somebody who's—

I thought it was Matt Ball.

We had somebody here come in outside of show hours and say that they were squatting on a bunch of tickers and the idea— It seems so. I think what might be the reality is that you actually, that it needs to be further along than just reserved. I don't know that having it, you can go, if you're a startup today, you can go reserve your ticker today. But I'm not sure that that actually gives you enough leverage to when Elon comes knocking ready for an IPO, you actually, have priority over?

I think you have to actually be doing something with it. So, if I have it correct, I'm looking it up to make sure that I have the facts straight, but Matthew Ball started an ETF based on the metaverse, which was— he wrote a whole book about the metaverse and had a series of blog posts outlining the broad trend and the various companies that would benefit from that. And so, ticker squatting. The team does not like— oh, okay, Roundhill squats on lots of tickers. That's interesting. I don't know. Apparently Meta Materials was in that ticker for a little bit. There's been a few others. But I think— Roundhill creates ETFs. ETFs. Well, I think if you have an active ETF, that's a lot different than just having a reservation. Although, I don't know, maybe you do get paid off.

So this fund, it was only like $10 million. So it's very small. So you actually could like— yeah, yeah. I mean, $10 million is still like a lot of money, but not, you know, it's still a fair amount.

But you could potentially launch an ETF with $10 million. Yeah.

Like, like if you have a really good ticker. Yeah. You know, you could— you can definitely make an argument that like, yeah, it's worth over $10 million and you can somehow marshal enough capital to, you know, do something with it.

There was a— there was a YouTube influencer who launched an ETF. And it sounded so good on paper, because it was like, wow, like his audience came together and put like $10 million into this ETF, but it was like generating something like 3 basis points of fees or something. So regardless of how it was like, regardless of what the fund did, like the actual flow would be like less than what he was earning in YouTube ad revenue. Like a few thousand dollars a month would be like the net gain from the ETF effort. And I think it was something that got wound down pretty quickly. Anyway, I had a broader take on Elon and Intel. So back in 2022, I was trying to imagine this divergent path, the road not traveled of Elon buying Intel because Intel—

Yeah, we talked about this before that there was a moment where people were tracking like, Intel's corporate jet, Elon's. And GlobalFoundries too, but yeah. Yeah, and it seemed, but I think it ultimately ended up just being election related.

Yeah, yeah, I think so. Yeah, they happened to be at Mar-a-Lago maybe at the same time. But it was such a cool idea because Elon had, you know, I mean, he'd run PayPal and stuff. He had some experience in like software, but he hadn't run a social network. So there was a lot of questions about like, what would happen to Twitter? Would he be able, would he need to change the business model? He wound up doing subscriptions. It wound up being a good exit because he rolled it into XAI, which rolled into SpaceX. So everyone did fine. But there was a lot of like, okay, $44 billion for Twitter, which after the ZIRP era ended, all of those companies traded down 60% or something. So there was a lot of chatter around, is, is this a good deal? Is this a good use of his time? He's known for running incredibly intense engineering operations. Making self-driving electric cars, rockets that land themselves, like semiconductor fab feels more like that than social network feels like, like rocket factory, right? It is a factory at the end of the day. And so I was trying to work backwards from like, could that actually have happened? I don't know. The numbers I had was, were, so at the end of 2022, Intel was a $110 billion market cap company. Like, that's a huge company still, even though it had been beaten up in the public markets. But Elon had just put together $44 billion to buy Twitter, and the CHIPS Act had put in 2022 had just put together $280 billion, including $53 billion specifically for U.S. semiconductor manufacturing.

Also, just think about where Intel would trade if Elon had been able to own it or roll it in.

I mean, it would be Just putting together the $44 billion was a Herculean effort. And there was a ton of debt. And there were all these different parties involved.

Yeah, it's just a much, much different business.

$110— and it's not like you can just go buy the company at its lowest possible value. You have to make an offer that all the shareholders will approve of and potentially be higher than what the market cap is. Because if you're an Intel shareholder and you're looking at a $110 billion market cap, you would imagine that it could go up and it did. It's now over what, $200 billion Intel market cap? What are they at right now? $260. So like if you were a shareholder in 2022 and you were at $110, you probably are happy that you didn't get bought out at that time. But it was possible. It wasn't that much of a stretch to imagine something happening there 4 years ago. We didn't wind up on that path. But, you know, my hope is that this is the first step on the long road to generate enough demand for domestic chip manufacturing to really move the needle, because I'm rooting for Intel and I'm rooting for American semiconductor supply chain. Tyler went over to the TSMC plant in Arizona. It seems like it's going well. Yeah, it's huge. It's huge. They didn't let me in. They didn't let you in. But we need more very clearly. And Intel is working on that, fortunately.

All right. We got to talk about a corporate retreat that went badly wrong. Okay. Technology company Plex took its 120 employees to Honduras for a week-long bonding experience. It was a disaster from the moment they arrived. Senior executives at the tech company Plex were eager to treat their 120 fully remote staffers to a week-long corporate getaway in a tropical paradise. Pop quiz, Tyler. Do you know what Plex is? I don't know about Plex.

No. Have we seen Plex before?

I don't know either. So we all failed, but now it's your job to figure it out. I will continue. The plan for the Honduras trip was simple. Company meetings and team building. Is this a streaming company? By powdery soft beaches during the day and island fun at night at a cost of roughly half a million to the company. They'd build the trip around a Survivor theme with teams and challenges, but it'd be fun, not too physically grueling. The CEO of Plex, a free streaming platform, would play a role similar to that of Survivor host, Jeff. Perhaps the executive should have taken it as a sign that just as the first bus of staffers pulled up to the resort, the chief executive was already in his hotel bathroom experiencing the initial waves of a violent stomach infection. What followed was a comedy of errors, including military drills that outpaced anything this group of office workers had in mind, a rogue porcupine, stranded airplanes, and one syringe to the butt of an employee. Corporate retreats are generally assumed to be torture, or at least a semi-stressful chore, what with their forced fun activities and hybrid work-play environments that leave workers confused about boundaries. Is that like the industry standard? That seems wild.

I don't know. I don't think I've ever been on a corporate retreat. I've been on some like Founders Fund events, but those aren't really retreats. Those are more just like conferences. But I don't know, corporate retreat seems, I don't know, unexplored territory for me.

It's no wonder the new season of Jury Duty, a comedy series that tricks an unsuspecting non-actor into believing his off-the-wall fictional circumstances are actually happening, is set at a corporate offsite. But in real life, PlexCon 2017 beats anything on TV. Here's the story of an all-staff company getaway told by 6 people who were there, a trip where most everything that could go wrong did go wrong, nearly a decade later, they're still working together and still talking about it. It's crazy that they— It had an effect.

It was bonding experience. Yeah.

Well, yeah, it's crazy that this is now coming out. So Shawn Hoff, 42, founder of Moniker Partners, an independent corporate retreat agency that planned the trip. About 3 weeks before we arrived in Honduras, we got an email from the hotel's general manager that said, I will be departing. I wish you the best with your retreat. I knew something was off. 3 days later, another email. The head chef was no longer gonna be at the hotel. Scott, 52, Chief Product Officer and Plex co-founder. We get there. We've gotta take the bus from the airport, dirt roads. You start getting closer and there are guard towers around the property, people with machine guns and stuff. A lot of people were like, where are we going? Keith, CEO of Plex, 54. We usually go a day early and we set up. If there's any little thing, we have to get it right just so the employees have the best experience possible. Keith woke up the day that people were coming in Sunday morning and he is sick as a dog. Everyone there is fried. Basically, people are telling me, don't eat the vegetables, don't eat the vegetables. That's like the same thing. No, no, no, 'cause they clean it, they wash it in water. It's usually not filtered water, right? 'Cause it would just be kind of crazy to. Yeah, yeah, here it is. So, I've got to have a salad, just one salad. So, I got E. coli, which may be the worst thing you could get possibly ever. Just as people were arriving on the buses, I was like, I had lost 8 or 10 pounds. They had a doctor come to me, which apparently is pretty standard. They nailed an IV bag to the bedpost. Just people are arriving for a party that night. The next day is Survivor-themed kickoff. There's not one person on the planet more excited about Survivor than Keith and his wife. They have watched every single episode. My wife and I met Jeff, the host of Survivor. What, what I wanted is when everybody shows up, I do a Jeff, welcome to the island, here's the theme for the week. But Scott got to to do it. The opening survivor thing was a contest where people on their different teams open up a platter. You have to eat what's on the platter. Sean, Sean, who's the Plex head of business development. Are you gonna call him? Yeah, somebody is cold texting me, pitching me their startup, and they've called me a bunch of times today. Wait, is it actually them or is it their AI? I wish I could pick up.

It's just like a little bit too much to— Yeah, it's mostly chance it picks up.

But yeah, cold texting somebody, like getting their number, I don't think that's the new meta. No. It's bold.

We heard from an executive in tech that they are getting dozens of emails every single day trying to recruit them. And every email comes from a new Gmail account that's unregistered, brand new, but it's all like, you know, LLM written, very different, like doesn't really do all the research but has a few keywords in there. And it's clear that someone is building sort of like a next-gen recruiting agency that's basically just a lot of spam. Feels like the end result will be like a return to relationship building and not like broad top of funnel.

I should read the cold text from this morning. And I have nothing against cold email and just being bold, but I did read this out loud to you, John, so I'll read it to everyone. So I got a text from an unknown number today at 7:00 AM. All right, Jordy, good news or bad news first? This is blank, and I'll leave the name out. And then I just get a PDF of a deck, and then a text. All right, Jordy, the bad news is this was an unplanned introduction, and on the surface, probably lukewarm outreach. The good news is that there's zero doubt you're now in touch with the founder with the most grit of anyone you've interacted with the past 12 months, and likely anyone you'll interact with over the next 12 months. 50,000 seed round passes over the past 10 months, here to make 50,001. So you should be coming in being like, I've been passed on 50,000 times. I'm hoping this is the one that gets through. That seems like a rough estimate, though. Months of feedback and iterations have made us better, so you're seeing more quality presentation than rejection 10,000. Looking forward to your message.

The chat wants the builder to pitch. They want you to hear this out.

Everyone's in favor of this. Wait, wait, pitch who?

The chat wants you to get on the phone with them. Do it live. I mean, they wanted to do live. I don't know if you should do live, but you should take the call and get to the bottom of it.

I will take the call.

I will take the call, but— Let's go back to the corporate retreat.

Okay, so they hire a former Navy SEAL. To basically haze the team on the beach. And you can pull up a picture, an image here.

The quote is, this is not a super fit group in general. One of our biggest mistakes was hiring a former Navy SEAL to pump the team up. As I'm in my room dying, I could hear them out there doing all the drills and yelling. And so I'm in here thinking, this is terrible, but it sounds terrible out there too. Uh, we're doing army crawling on the beach. It was 100 degrees. I bailed out partway We went through, I went into the ocean just to cool off. I went in probably on all fours because I was tired. It's not a fit group, not a super fit group in general, the ex-Navy SEALs. Like, we can tone it down, no problem. We get up there and it's hot and humid and people are passing out. I don't think he'd ever seen quite such an unfit group. We ended on, I guess, what's probably a golf course. On command, everyone had to hit the grass. Everyone's silent. We're pretending we're Navy SEALs, but I happened to land in the wrong spot. I'm just like, oh God, what is happening? I was sitting on a fire ant hill. I was wearing shorts. I jumped and had hives and bumps from the bites. This is ridiculous. Someone saw an alligator on the golf course. Sounds like a ridiculous—

There was a porcupine that fell through one of the ceilings.

This is like a fire festival for corporate retreats.

A fire festival of corporate retreats.

It's fun that this porcupine is horrifying. Wow, look at this guy. Rick Phillips discovered this in his room shower. The hotel got, uh, pretty much just got the porcupine and left. I guess for me it was a good thing because being a non-talkative software engineer, I got some notoriety. It's a beautiful resort. There are sand fleas they had to fumigate every day. What a weird quote. Uh, we did a nice dinner down by the beach and everyone got bit by the sand fleas that weren't supposed to be there.

We all got matching tank tops. Real life real-life episode of The Office. Anyway, you can imagine Michael Scott.

There is some breaking news. Anthropic is set to preview powerful Mythos model to ward off cyber AI threats. The AI company is partnering with Amazon, Microsoft, and others to offer the new model to find and patch software bugs. This is from The Wall Street Journal. Anthropic is taking steps to arm some of the world's biggest technology companies companies with tools to find and patch bugs in their hardware and software. The company is making a preview of its new AI model called Mythos available to about 50 companies and organizations that maintain critical infrastructure, including Amazon, Microsoft, Apple, Alphabet-owned Google, and the Linux Foundation. Cybersecurity researchers and software makers worry that artificial intelligence is becoming so good at exploiting vulnerabilities that it could cause widespread online disruption. Security experts have predicted that AI models will discover an avalanche of software bugs, and the effort is set to help companies stay one step ahead of cybercriminals and other threats. This feels like a very good rollout strategy generally, both because we've seen a huge amount of cyberattacks and hacks and accidental releases. Like, even if it's not, you know, there's been— we had a member of the security team from CrowdStrike on the show last week, Talking about the rise in cyber attacks broadly. And so getting the most frontier models in the hands of big companies early, great from that perspective. And then also just great as a product demo, which will get the entire organization excited about deploying the technology broadly. So very good as like a, as a B2B go-to-market motion. This makes a ton of sense. When measuring the dollar cost to find a bug, Mythos claims to be about 10 times as efficient as previous AI models. Details of Mythos' capabilities were previously reported by Fortune. While Anthropic has no immediate plans to release Mythos, other models will likely match its bug-finding capabilities within the next few years. Graham said, "We basically need to start right now preparing for a world where there is zero lag between discovery and exploitation. So very carefully rolling these out. The Cloud Opus 4.6 found more high-severity bugs in Firefox in 2 weeks than the rest of the world typically reports in 2 months. And so good, good news there for cybersecurity.

There is some— some news yesterday in the New York Times. Shots fired at Indianapolis Council councilman's home after a vote that was backing a data center. No one was injured, but Councilman Ron Gibson called it deeply unsettling. This broke on X yesterday, and Alex writes in The New York Times, bullets hit the home of an Indianapolis city councilman early Monday morning, leaving shattered glass and holes through the front door, and a handwritten note reading "No data centers" was left under the doormat. The councilman, Ron Gibson, was among the city's leaders who voted 6-2 last week to to approve a rezoning measure that would allow Metroblocks, a Los Angeles company, to build a data center on the northeast side of Indianapolis. Local residents had protested the proposed data center for months, citing concerns about environmental impacts and changes to a historic neighborhood. Dozens of people filled the city council chambers, city county council chambers last week before the vote, holding signs and speaking in opposition to the data center, the station reported. Gunfire at his home crosses a line, Mr. Gibson wrote in an emailed statement. I understand that public service can bring strong opinions and disagreement, but violence is never the answer, especially when it puts families at risk. So very, very dark situation. Mr. Gibson wrote that he and his 8-year-old son were awakened by the gunshots between 12:45 and 12:50 a.m. Monday, and he rushed to reassure his son that he was safe. He said 13 rounds were fired at his home. With bullets striking just steps from the dining room table where his son had played with Legos the day before. So incredibly, incredibly dark. Rune says, just so everyone is clear, this is evil. You were justified in thinking it's morally bad. Tons of apologetics happening for bad people. If you think behavior like this is just desserts for the tech industry due to some hobby horse you have, you've gone insane. And Noah Smith says, we may still be underrating how big of a political issue AI is going to be. And there's a video here from CBS we can put the sound on.

Oh, yeah. Just days after voting in favor of building a new data center in Indianapolis, local council member Ron Gibson says he woke up to the sound of gunfire overnight.

Gibson said 13 rounds were fired at his home while he and his 8-year-old son were asleep. Some of those bullets landing just steps from their dining room table where his son was playing with Legos the day before. When he stepped outside, he says he found this handwritten note reading "No data centers" under the doormat.

There are real benefits tied to this development.

Construction is expected to support roughly 300 jobs over a 3-year period While some councillors argued the data centre will bring revenue and jobs, there was pushback from residents over environmental and quality of life concerns. The vote to move forward passed 6-2.

Indianapolis police are calling this an isolated and targeted incident, and police have yet to identify a suspect.

And one group that protested that data centre last week released a statement today saying in part, violence Violence has no place in our community or in our advocacy. The FBI now assisting in the investigation. Tony Chanel, thank you. Yeah, Noah Hershkeld in the chat says people have gone insane. Uh, it is, it is extremely disheartening to see. Um, I mean, I, I would hope that the, uh, the case to be made for data centers is much more complex than 300 jobs over 3 years. Like the ratepayer protection pledge, the environmental concerns, all of that needs to be addressed and communicated, not because of this, this is horrible, but just in general from the public that has pushback and reticence about the data center development.

Yeah, local tax revenues.

Yeah, and it should be a boon to every community, and every community should feel like feel like the net benefit is truly positive. And I feel like people, many, many community members do not feel like it's in their favor. And so there's a lot to do on that front. In some other more positive news, OpenAI, Anthropic, Google are uniting to combat model copying in China. This is a bigger discussion around AI safety. We've talked about this. Would you look at that? Some faith in humanity. Who knew?

Who knew they could get along? Yes.

I mean, I'm sure people in the chat have seen the New Yorker article where there's just tons and tons of quotes from various AI leaders all upset with Sam Altman. And the inter-AI drama has been, bubbling up since the dawn of OpenAI. Like OpenAI was started as a reaction to Google and then Anthropic leaves and teams up with Google and then Elon doesn't like Anthropic and then Ilya Sutskever and Mira leave, but they don't join Anthropic. And so there's been so many personalities and so many disputes. I feel like the, the takeaway is that this is all extremely high stakes. There's a technological transition happening, a huge amount of money on the table, a huge amount of influence on the table. And so everyone is sort of clamoring for their share and it's creating a lot of friction. But my overall takeaway from the New Yorker article was a lot of that had been already reported out. A lot of that was, we sort of knew, I knew that there were rivalries and a lot of hurt feelings between different members of the AI community, and nothing was particularly shocking to me. But if you have more comments, please leave them in the chat. Let's go through this. Pull it up. What's actually going on with this model copying in China question. So rivals. OpenAI, Anthropic, PBC, and Alphabet Inc's Google have begun working together to try and clamp down on Chinese competitors, extracting results from cutting-edge US artificial intelligence models to gain an edge in the global AI race. The firms are sharing information through the Frontier Model Forum, an industry nonprofit that the three tech companies founded with Microsoft in 2023 to detect so-called adversarial distillation attempts that violate their terms of service, according to people familiar with the matter. The rare collaboration underscores the severity of a concern raised by US AI companies that some users, especially in China, are creating imitation versions of their products that could undercut them on price and siphon away customers while posing a national security risk. And so I was trying to square this question of distillation and model commoditization with the news that Anthropic has reached $30 billion in run rate and has an agreement with Google and Broadcom for multiple gigawatts of TPU capacity. Like, clearly there is insatiable demand for frontier tokens, frontier models. They're incredibly expensive to train. We saw in the Wall Street Journal that, that these expected training costs from— yeah, it was training and inference, but it was hundreds of billions of dollars. And so The hope is that you're able to amortize that over at least a couple of years, a long time ideally. The shelf life of a model after you train it is pretty limited if you're being commoditized and copied. If you're being distilled, it's even faster. At the same time, just staying on the frontier clearly leads to an incredible ramp in revenue. Is commoditization a real problem? It feels like it's almost just more of a problem from an AI safety perspective because you can't have the geopolitical conversation like what Bernie Sanders is proposing around different labs working together, potentially pausing or slowing down or just even adding more constraints and reviews before models get released. It's harder to do that if you have a different country that's racing ahead and moving much faster and trying to close that gap. Now, if the model is delayed in America and their whole strategy is to distill the model, well then they're still 3 months behind even if we take 3 months off in America, so maybe that's not an issue. But it was an interesting development. So, I guess models get bigger and more powerful. Hopefully, it becomes easier to track distillation efforts, I would imagine. So what is the canonical example of the distillation question? Was that originally Deepseek, Tyler?

I mean, I don't know if— it's very hard to prove, right? Because it's like, if you just look at the weights, there's no way to really tell. It's just like you just ask the model, who are you? And then it says, I'm Claude. Yeah.

It's like, okay, yeah, that's the smoking gun. But that doesn't always happen because you could remove all references to Claude. Yeah.

Like people were saying when, you know, Deepfake first like came on, everyone's like, oh wow, this is so good. Yeah. People at OpenAI were saying like, okay, they've distilled on us. Yeah. But like, was there any like proof? Not really. Maybe just like vibes. If you talk to the models a lot, you see that they kind of respond similarly compared to other models.

And I mean, the smart strategy if you are a distiller lab would be to generate a bunch of tokens from Google Anthropic and OpenAI, mix them together so that maybe every once in a while it says it's Claude, but only one-third of the time because you have a whole bunch of other tokens from other models and it all blends together. So that actually underscores the need for the Frontier Model Forum and the three companies working together.

I do wonder how you like actually combat this, because it seems very hard unless you basically just say the only people that can hit the API are like trusted enterprises, and you have to sign these big contracts before, which I think is— there's a good case to be made for that being like the way forward.

I mean, it doesn't seem that crazy to do like varying levels of KYC for varying levels of token spend or enterprise contracts. So if you are— if all of a sudden it's like You're doing meta-level inference. Well, let's make sure that it's meta that's actually the one on the other side of that contract and that they're not vending the Frontier tokens into another, like reselling them basically in some way. So you need the chain of know your customer. But in theory, you should be able to have a $200 pro plan that does not ever deliver the level of tokens to fully distill the model. And then as someone ramps up on the API and they're spending $5,000 a month, okay, you do a little bit more of a check. Then they're spending $100,000, then they're spending $10 million. And once they get up, like every lab must know, okay, to distill this, you probably need to spend $100 million, $10 million, whatever the number is, set the KYC threshold there. But it's going to be cut in a third because it's going to happen across all of these labs. So whatever their threshold is for, okay, it's really time to go and do the proper KYC, you sort of need to divide that by 3 because you have to assume that it's happening across all 3. And then it's also probably happening across multiple smaller organizations that are essentially fronts and are harder to KYC. But I don't know.

Uh, let's do a lightning round of timeline posts. Okay, what do you got? Reid Wiseman, he's an astronaut. Yeah, he's on Artemis 2. Yeah, December 7th, 2016, he posted at 8:47 AM, dreamt I was in lunar orbit last night, been in that post-vivid dream that wasn't real funk all morning. And, uh, yesterday he made it real. Wow.

I saw this post and I was like, "Oh, that's just some random poster. Okay, cool. Yeah, cool story. Anyone could have that dream." I didn't realize that it was actually the astronaut who is now around the moon and making his way back to Earth. There was another cool video from the from the moon mission. Astronaut Victor Glover discussed what it means for him and the entire Artemis II crew to be observing Easter Sunday from space during their historic mission. We should pull this video up. It's on X. It's about a minute long. Let's see if we can play this.

I think these observances are important, and as we are so far from Earth and looking you know, the beauty of creation.

I think that for me, one of the really important personal perspectives that I have up here is I can really see Earth as one thing.

And, you know, when I read the Bible and I look at all of the amazing things that were done for us who were created, it's, you know, you have this amazing place, this spaceship. You guys are talking to us because we're in a spaceship really far from Earth, but you're on a spaceship called Earth that was created to give us a place to live in the universe, in the cosmos. Maybe the distance we are from you makes you think what we're doing is special, but we're the same distance from you. And I'm trying to tell you, just trust me, you are special in all of this emptiness. This is a whole bunch of nothing, this thing we call the universe.

You have this oasis, this beautiful place that we get to exist together.

I think as we go into Easter Sunday thinking about all the cultures all around the world, whether you celebrate it or not, whether you believe in God or not, this is an opportunity for us to remember where we are, who we are, and that we are the same thing and that we gotta get through this together.

I love it. Powerful stuff. Great mission. Creatine cycle. Yeah. Having a GF is insane because it's literally unlimited chat with no tokens spent. There you go.

What did Sucks say? "Is 3 weeks too young to give my baby Retrutiide? Not trying to raise a fat porky butterball." Yes, it's too young. Be careful out there. Have you seen this image of the Instagram growth guru? This goes viral constantly and normally it's like a Reels reaction video, but someone took that, this exact message and put it on X and it went mega viral with 130,000 likes. Instagram growth guru, Gurus are so funny. He can't use his laptop because he's holding a drink. He can't drink because he has a cigar in his mouth. He can't smoke his cigar because both hands are occupied. I think this is surmountable, though. I think you can actually hold the cigar with the martini or something. But it is true, like, peak performative growth guru. Very, very funny. But I'm sure it had its intended purpose. Chase passive income. Says, that's my mentor you're talking about. He makes multiple 8 figures in passive income and only charges $25,000 for a full day boot camp mastermind to teach me his strategies. Delete this.

I yield. Harry dialing into our morning meeting. Morning. Stocks are down slightly after Trump announced a whole civilization will die tonight. Really dark over on Truth Social today. And hoping— Yeah, it's on the COVID of the Journal today.

Trump stirs fear in Iran over talk of attack. President says the US military could take out the entire country in one night. President Trump said Monday that the US military could take out the entirety of Iran in one night, and Iranian officials have told mediators now trying to reach a last-ditch ceasefire deal that they fear Trump will follow through on a massive attack Tuesday night. So we are We continue to hope for a resolution to the geopolitical conflict. It's very, very frustrating. And yeah, it would be so much better to refocus on problems at home and opportunities at home.

Anyway, moving on. What is Pirate Wire saying about Citrini and Analyst Number 3? Ryan writes, our newspapers used to put reporters in active war zones. They've stopped. So new media is picking it up. Last week, research firm Citrini put a man in the Strait of Hormuz to figure out what is going on in the oil world. Citrini's, uh, quadra— quadrilingual employee, referred to only as Analyst 3, packed a bag with $15,000 in cash, some zins, and cigars before shooting over to Oman. After managing to finagle his way onto a speedboat, Analyst 3 reported from the water just 18 miles away from the Iranian coast while smoking a Cuban. In a world where the New York Times is calling NATO the North American Treaty Organization, we've got live financial reporting from Iran. Yeah, this was just absolutely an insane story. An insane story of you can just do things.

Totally. Yeah, I'm talking to Citrini about coming on the show later this week. Obviously, he's been working working incredibly, incredibly hard with his team to report and publish this and turn things into analysis. And there's a whole deep dive that you can read on Citrini and just what a fantastic piece of reporting. Truly unexpected. I didn't think anyone would do this. We were joking about it and they just went and did it.

What's going on? We stopped doing ads but NASA—

NASA's doing as well.

He's carrying a torch. Okay, let's see. They have a number. Is that a Torch Daddy or something?

Yes. Scrub Daddy. Yes. That's the viral infomercial product, right? Yes. And then a Liquid Death as well. Is there a Liquid Death? This has to be fake, right? The Taco Bell thing, this is a joke. I can't tell. Is the Nutella in here?

These are all— It seems to be real Nutella.

AI generator or something. Yeah, Grok wanted me to tell you it's real. I don't know. It does seem like people are having fun with this.

But yeah, very, very funny. All right. Andrew Huberman says 93 years and 231 days old. And Isolde Stein dead hangs for 2 minutes and 52 seconds. That is insane. Set new world record.

We tried this this morning.

Yeah, we did this this morning.

2 minutes and 52 seconds is truly an eternity. I don't know if that's in the cards for us. Maybe, maybe we can get there. We— what is CARPA at? 5 minutes something?

Yeah, CARPA is at around 5 minutes. Absolutely insane. Yeah, these, these numbers don't sound big, but you start hanging and it is absolutely brutal. So congratulations to, to Anne. Yeah, what a feat.

What a feat. Well, we have Riley Walls in the waiting room. Let's bring him into the TD Pin UltraDome. Riley, how you guys doing? Hey, guys. Are you— We are good. Are you where I think you are? Introduce yourselves.

Explain where you are. We are at the alley. We're live at the alley. Live at the alley. This is the street right here. Okay. Yeah, life is good. This is my friend Patrick. Hi, Patrick. We worked on the alley together. Yeah, this has been a really fun project. Yeah. So we, yeah, it's, it's been a blast.

Give us the full history. Yeah. Of, of the alley, how you guys came to own it, and then we'll get into this project.

Basically, we learned that this, this alley was foreclosed on. It's kind of a long story where this, this woman accidentally bought it thinking that she was buying this house right here. No, which is, no, it is this one. Oh, it is this one. Which was, which was worth like a million dollars. Okay. And she bid $25K on it thinking that she was getting the steal of a lifetime. She won.

And is this the kind of thing she didn't— she didn't want to tell anyone, like, because she was like, okay, this is like— probably thought it was too good to be true, but then like, yeah, maybe it was real. So she didn't want to tell anyone if she had talked about it.

She knocked on the doors because this is an apartment building, and she knocked on the doors and told the tenant she was not raising the rent. Um, so she was— she really thought she owned this, and then she realized she didn't. And then, um, there was a whole news story about her mistake, and then we eventually reached out to her and negotiated for a little while, and we bought it for a little more than she actually bought it for. So she got bailed out, and we were able to do a nice thing and still do this crazy project. That's amazing.

So did you find this alley from the news article? Is that how this all started?

Yeah, we had— Patrick and our friend Theo and I, the three of us, we had been talking for, like, a while before that about buying a different alley in San Francisco. Sure. But we couldn't end up buying that one. But we saw this new store, we're like, oh, this is perfect.

Yeah.

And we tried to email her to, you know, get us the offer, but Riley had to send a letter in the mail.

And she got the letter physically? Yeah.

Physical snail mail works. That's amazing.

Okay, so it transfers to you, you own it, and then what rights are you entitled to? Because clearly it's not the house next door, but you own something. Like what exactly Do you own?

We, we spent, um, a good amount of time with lawyers to figure out what actually is allowed. Yeah, there's actually a car, it's driving right at us right now.

If you need to move, you can, you can show us wherever. What neighborhood is this, by the way?

This is in the Sunset, uh, which is like 24th and, uh, yeah, 24th and Kirkham.

Okay, sorry, middle Sunset.

Sorry. Live, live on air. There's a car coming. There's a car coming down the alley.

Okay, so it's possible.

This is actually a great question because we, yeah, live demo here. So we legally can't block the alley, so we are obliged to move. That is actually in tandem with the question you guys just asked. Yeah, we can't block it. Cars have the right to drive down it because there's easements here. But we do have the right to paint anything on the surface of the street, and we also can give the alley name. Okay.

So yeah, it's, it's— is there already a sign? Does the alley have a name?

Has it historically had a name? No. So it, um, it— Google Maps calls it Dirt Alley. We think that some editor randomly added that like a couple years ago, but it's nowhere else. Okay. It says that name.

No official name.

And then, and then is there any square footage that anyone could build on? Is it, or is it just the actual road?

It's just the road. It's literally like 8 feet wide and nothing else.

So do you have to put up signs?

And the aura of, you know, if you become the owner of this, of having an alley named after you in the great city of San Francisco. Yeah.

So yeah, take us through the process to actually auction this, set up the website, draw demand. How did all this play out?

Totally. Yeah. So when we first bought it, it was actually dirt alley, right?

Like there was nothing on it.

So first we had to pave it.

And so we found some pavers, they paved it, 80 feet of it or so, built the website in a weekend, I think, with our friend Theo, who's helping us with the project.

And then we thought about how to go live.

And I don't know if you want to say more.

Yeah, I think this has been cool because the The art side of it, like people can go to paintthestreet.com. It actually just ended 5 minutes ago. But for like the, during the weekend, people could submit little like 48 by 48 pixel drawings. And we have space to paint like 1,200 of them on the street. And that's like totally free. And then to kind of COVID the entire project and be a little more like capitalistic, we auction off the naming rights. So that ends in like 55 minutes or so.

And yeah, so the full street is going to be painted with this, this full alley, like this whole mural of everyone that picked something. Yes. Yeah. Yeah. And this was free. It was basically anyone that wanted to could go in. What was this inspired by? Reddit the place, that idea?

Yeah, that's been a huge inspiration. That's been really cool to see. So yeah, we're kind of reviving that and making it IRL in SF.

Talk, talk about the expectations because when And when we first chatted about this, we were like, well, we'll be the early bidder to at least make sure you guys get your money out. And then by the time you actually sent us the auction, it was already well, well, well above. So how's the response been?

Yeah, it's been great. 'Cause we were kind of stressed 'cause we're like putting a lot of money on the line and we're like, this will either like flop and like, or it'll actually do really well. And it has done really well. Which is, which is good. Right now the highest bidder is WordPress. WordPress Way. WordPress Way, $135K, which is, which is insane.

And how is the auction going to work at the end? Is it going to be like— I forget the terminology of it, but if somebody places a new winning bid within the last, let's say, minute, does it extend the auction? Are you expecting a— like, are people circling now? I feel, I feel like I saw Josh Browder I know that Josh is probably thinking, I'm going to wait until the final minute and come in with the top offer. But what's the dynamic going to work like? So in the last 5 minutes, if somebody bids, then it gets extended by another 5 minutes until it ends.

And then once the auction closes, what's the process like to actually rename it? Have you traced through, like, what does it take to update Google Maps, Apple Google Maps, Waze, all the different mapping features, do you think that would be pretty easy, or is it like, is there like a self-serve portal for like renaming streets? How do you prove this?

There is, this is a unique street because it's privately owned, and as the owners, we kind of, the source of truth for a name like this is the sign itself, so putting up a sign, and then we can, I think Google Maps, like, you take a photo or something and you submit the edit, and it should appear soon. I actually, when I was in high school, high school, I, um, maybe I shouldn't share this on the internet, but I, I, uh, as like a senior prank, I installed a street sign like on an unnamed alley in my hometown named after my high school track coach. And, um, it got renamed on Google Maps, and it's been like 5 years this road has been on there, and it's still on there. Yeah, yeah, 5 years later. Yeah, like people think it's real. It's really funny. Yeah, the mayor has mentioned it in a meeting before. Yeah.

Did you have to do anything to sort of, I mean, these projects, they always run the risk of like vandalization basically. I'm looking at someone clearly tried to paint wrong way on the street, although they couldn't quite get all the upvotes to land the way they wanted. So it's a little bit out of order. But did you have a process for reviewing submissions or was this all community led? Like how did you think about? The risk of someone putting up something that was like offensive?

Yeah, we had, we had some attempts at this, but, but we, we had some automated stuff to remove it and then we were, you know, the entire time it's been up since Thursday, we've been manually reviewing it. You know, as a team we have a group chat and we're just, you know, reviewing it and someone's taking the turn and wake up in the morning and there's some stuff we remove.

But it's not, it's not been so bad. It's been, you know, it has been pretty friendly to us.

It's relative to how bad it could be. It gets crazy. I remember, didn't Justin Justin Bieber put out a post on like, "I'll go wherever the audience votes me for my next tour," and they tried to send him to North Korea. It's a famous example of like the internet going wild. At the same time, Boaty McBoatface, very successfully named Boat, you know, everyone enjoyed that one. Give us an update on some of the other projects. How is the Pokémon Go payphone project going these days?

Yeah, that ended a few weeks ago. Basically, I had gotten a list of all the payphones in California. There's still like a few thousand that work. Yeah. And then made like a Pokémon Go type game where you have to go to different payphones and you can claim them by calling a number. And yeah, like 300 or 400 payphones were called from, and it was really, really cool to see people go out and find them all.

And there was a leaderboard, but no specific prize? No prize. Yeah, just the memories. Who won? How many calls did they make? How much did they travel?

This girl named Maggie won. She won by 1 point. And yeah, she was driving all over the state, it seemed like. There was a lot of phones. I think I nerd sniped quite a few people, which was cool to see.

What other projects are on your plate right now? Are you still working on the J Suite, or is that project sort of done at this point? I know that there's a whole documentary you can talk about, but what else is in your world?

Yeah, Gmail has been pretty crazy. There's like 15 or so people that have helped in some way for that project. And it's kind of died down now. Maybe there'll be some resurgence in the Epstein files. Maybe we'll see. But yeah, that's been really, really crazy to see. Project that I'm probably going to drop this weekend is we scraped a bunch of data about the US, like how much money the US government spends and will hopefully make like a, like a Spotify rap style, like, oh, here's, here's how much you paid in taxes this year. Like, oh, this much, this many dollars went towards like defense or Social Security or things like that. So that'll be, that'll be cool to see.

I'm sure, I'm sure that'll make a lot of people really happy. Yeah. Yeah.

It's crazy to see, like, I'm like, oh, actually, like, I kind of, you know, seeing these actual numbers, like where things actually get spent on, like, I feel like it just makes you think a different way.

Yeah. In general, like with something like that, Spotify Wrapped works so well because people screenshot it, they share it. How do you think about the user-generated viral loop for these projects? Is that like a key piece of the idea phase? You think about, OK, how can I actually create some sort of flywheel for generating attention? Or is this just like sometimes you get lucky and people share, and it's sort of an afterthought.

I think it's fun thinking about, like, I mean, like, the Alley Project, it's just like a fun idea. Like, we were actually talking at like a party like a year ago, and I was like, oh wait, there's some alleys that get foreclosed on in SF. Like, wouldn't it be funny? Wouldn't it be baller? Yeah, I said it as a joke, but then Patrick and Theo were like, let's actually do it. And then there was a girl that was like in law school at the party. We were talking to her about like the legality of it, and then And it just, it took a while to actually make it happen, but like, it's just kind of fun for us. And then thinking about how to actually make it go viral is also like a kind of a secondary thing. But it is, the framing is so important for these sorts of things. And like, we also wanted something cool. Like, this is kind of just a cool like thing to do for SF. Like, it's cool putting SF on the map and doing things like this, especially when there's like, yeah, I think we're really lucky that these sorts of things can get funded too. I think SF is a very special place for that.

What kind of inbound pitches do you guys get at this point? Like, I'm sure people are DMing you like, hey, I found this weird kind of anomaly. I think, I think, I think you could, you know, something like an alley. Is there an inbound flywheel yet? There's all sorts of weird things.

I don't know, like someone's like, yeah, we have like a tank in SF. Like, what ideas for that? Like, there's like a lot of weird things.

I'm expecting you guys to figure out some mechanism to effectively take over a country at some point. Maybe that's the next step.

I wouldn't put it past you. Well, congratulations. Where can people go to actually bid?

paintastreet.com/auction. 46 minutes, 28 seconds left. We're going to be keeping our eye on this as it counts down. Yeah, I'm very interested to see how people react when there's only a few minutes left.

Yeah, this will be very fun. Well, thank you so much for taking the time to come give us the update and chat with us. Congratulations on a huge success. I mean, this is well above— I mean, $135,000, that seems well above what you paid, and you should be able to recoup all the costs. Do you have a plan for what to What are you guys planning to do with the money?

Yeah, not gonna use it for ourselves at all, to be clear. We're gonna, we'll, it'll take some money to paint the actual mural on the street, but there'll definitely be something left over. We'll just keep it for the next project of this sort in San Francisco, some kind of IRL project here. That's amazing.

Well, what a fun project. Thanks for coming on and sharing with us.

We'll talk to you soon. Congratulations to you guys as well.

Yeah, it's crazy that we're coworkers now. Yeah, we're colleagues. Yeah, we're gonna be up in SF next Tuesday. Hang in person.

We'll have to do a show from the alley at some point.

That'd be amazing. Yes, awesome. Great hanging, guys.

We'll talk to you soon. Congrats.

Goodbye. Cheers. Um, the, uh, have you seen this map of the surface of Venus? Did you know that Venus has land surface? I did not. Uh, the land surface of Venus has some insane RPG world potential. Someone needs to vibe code a video game where you can go run around on this or play some sort of 4X game. Yes, someone says, "Hasbro is really messing up not releasing Risk: Venus." There's another— there's another cool AI project that launched from Netflix. There's a video here showing their new project Void. The AI removes objects from videos, but it even corrects the physics after the objects or people are removed. And there's a demo in the comments. Here. So you can see that they— that the bottom is the element that's being removed. And so if you have the kettlebell is deforming the pillow, once the AI removes that, which is often a very, very time-intensive VFX task, the physics of how the scene would have played out if that character had not been in the scene are then, you know, recalculated. And so the duck— not only is the duck not being smashed, the duck just appears like normal. And this is just a project that will— if you've ever done any of this type of work, it's incredibly cumbersome. And so I think this will be adopted very, very quickly. Nishan says he had a use case 4 years ago. A big Hollywood VFX company came to came to us, the company I was working with, and asked us to remove freckles and pimples from the face of actors and actresses' 4K movie footage. At the time, we really struggled and failed to do it. If it was now, we would have easily tackled the problem. And there's so many examples of this. The physics— wait, the physics correction part, removing stuff from video isn't new, but making the background actually behave correctly is a completely different problem. And so this is a huge, huge move for the VFX industry.

Industry in Hollywood. Also, does Netflix have a history of contributing to open source?

Yeah, I think so. They've done a lot of standard setting around cinema camera gear. They were a big proponent of the Blackmagic ecosystem, I believe. Great price per value for shooting a film and delivering in 4K, which is something that they mandate across across their entire ecosystem, but can be cumbersome for creators if they have to go shoot on a very expensive cinema camera. They've done a lot of stuff there. And then of course they acquired that AI company from Ben Affleck, I believe. Wasn't that the story? Oh yeah. And I don't know, it's possible that there were some researchers from that team that bled over onto this project, although this seems like it was in flight for longer than before. I agree, Corridor Crew needs to do a video on that feature. Be very cool to watch them road test it and see, you know, where the boundaries are, because I'm sure it doesn't work in every possible scenario. The demo footage is always going to be the best, but, uh, very, very cool. Uh, anyway, without further ado, we have our next guest, Aditya Bandy, from Noon, in the waiting room. Let's bring Aditya in. How are you doing?

Hey, hey, John. Hey, Johnny. I'm doing good. How are you folks?

We're doing great. Great. Welcome. Great to have you on the show. Since it's the first time on the show. Please introduce yourself and the company. Absolutely.

Yeah. Thank you for having me. My name is Aditya Bhandi. I am the co-founder of Noon. And yeah, I'm, you know, originally, you know, from India. I moved here to the Bay Area in 2015 as part of my first startup getting acquired by Yahoo. I'm a product designer turned product manager and a second-time founder. You know, I have a co-founder. His His name is Kushagra. He's also a product designer turned Second Act founder. Yeah. What was the first company? So we were actually, you know, back in 2013, there was this problem of like rendering PDFs and Word documents inside apps. Sure. And basically I was building this company called Bookpad out of Bangalore, and we can render 13+ formats in any app. So someone like Dropbox could integrate with us and start rendering documents. So that's the company we built and Yahoo we bought it for Yahoo Mail. So we became part of the whole attachment layer for Yahoo Mail.

Yeah, so if you're in Yahoo Mail and you need to open a PDF, it just opens in the browser natively. Yes, Word document. How much of that is a technical challenge, and then how much of that is just sort of dealing with Adobe IP? Because I imagine that Adobe has products and they maintain a standard, but you can use some of it. There's pieces that you can't? Like, what were the decision-making criteria around building the product then? Yeah, it was very complex.

Like, you know, if you— basically all of these formats were not made for cloud or for browser. They were built in the 1980s. Like, if you open up any PDF or a Word document, it's insane how complex they look inside. So we had to build a rendering engine. We had to build a conversion layer that understands these documents and and tries to recreate them for the cloud. So that's basically a very technical challenge for us to do that. Yeah. And we just were very young and happy and trying to solve something very, very complex, and that just made us very happy.

Yeah. So walk me through the decision to start the next company. What was that like? What was the, okay, this is time to run it back?

Yeah, so my co-founder and I, we've been product designers all our lives. We love building products. And then we've been observing the AI space. We've been using design tools for the last 20 years of our lives, right? Like from grad school. And then for the most part of all the tools that exist in the market, they're all graphic design tools. You can design the visuals of the product, but not the product directly. So designer is limited to designing the visuals, and then they hand off the visuals to the engineer to to build it. That's how it was always done. And the moment AI coding tools came into the market, everyone started realizing, oh, I can build, I can do the functional aspect of it as well. And then they started realizing how amazing it is to do both together. But all of that was happening in multiple tools. It's all broken. And that was the idea behind this. Like, hey, why is it happening in multiple tools? Why can't it all happen in one tool where someone who wants to design and build products can just do it all in one tool. Yeah.

So I mean, when you hear about vibe coding projects, you hear about like insane lines of code, so much functionality like instantiated just at, you know, a couple sentences dropped in a prompt. What changes about the product design workflow in a world where sort of the entire underlying structure of the product can change on a dime all of a sudden?

So the tools that are very text-based, so the problem is they're very limited in how you can say something, right? Sure. Now, describe me a painting. Yeah. Let's say you start describing the painting. You can never fully describe the painting, right? Yeah. That's the problem with text-based editors or text-based design tools.

I mean, you hear a lot of people sharing that even just for like one-off Open Claw tasks, like texting an image or screenshot from their phone along with a prompt just to give a little bit more. Hey, here's a screenshot of my calendar. I'm essentially exporting a bunch of information via PNG, which is a richer format than people thought, I guess. And so how closely do you want to link the the visual design elements, the interaction elements, the logic of how things fit together because, you know, like we've already sort of transitioned out of the, you know, visual designer in a pixel-perfect, like layered Photoshop file into something completely new and dynamic. But that barrier is bleeding and/or is becoming fuzzier, basically.

100%. And then the way we look at it is if you're trying to understand what are we trying to do, first of all, what we're saying is we work on top of your product code. So essentially, we don't do any MCP. We don't need to do any middle layer like Claude or Codex. We don't need any of that. We directly sit on top of your product code. Codebase. Yeah, codebase. And then what happens is the designers or anyone who's who's working on the product can see that code visually rendered on the canvas. Sure. So now you can understand, okay, this is what my product looks like in different screens and different components. Yeah. Then from there, you have a lot of control to refine it, make it better, create new screens, create new features. Then you don't need multiple tools. Yeah. That's the second advantage. You're not switching between multiple tools. Things are not getting lost in translation. Yeah. All the details are preserved. And the last thing is— Oh, sorry.

Yeah, yeah, yeah, yeah, yeah. Last thing.

Yeah, and the last thing is like you work on both the functional and visual layer. Sure. So it's up to you. You build the visual first and add functionality. You can control both of them in a fine-tuned manner. Yeah.

So if you're thinking about just something as simple as like making a CTA bigger and you're dragging the size of a button button. That can be quicker in a visual format than in a prompt where maybe you wind up describing the number of pixels that you want it to be or say a little bit bigger, okay, a little bit smaller. It can be tricky. But what is the flow of actions that are happening behind the scenes once I actually resize a button, for example? Yeah.

So let's take the button as an example, right? So first of all, the button that you're seeing on the canvas, it's not a rectangular shape. It's a real button. Code behind, so that's why it's getting rendered on canvas for you. Then when you're actually, you know, trying to edit the button visually or maybe using some, you know, AI there or maybe directly you're doing it, the code is getting edited in real time. So essentially you can think of us as like a visual code editor. So you edit the code visually. So we're kind of building a code editor from scratch that can be edited visually.

Yeah. How are you thinking about How are you thinking about going after customers? Do you want to be in the enterprise or more self-serve? Where's the beachhead?

Yeah, and how much of this round was based on traction to date versus long-term opportunity and scale of the market?

Yeah, no, I think for us, we're a design tool. Design tools, people want to try it first before even thinking of— thinking of paying it or getting into a team. So we're always going to be a free tier, use it for free. If you like it, then pay for it. That's our model. So we're going to be bottoms up that way. But in terms of the traction, we are currently a lot of some of the best known tech companies in the Bay Area. We have pilots with them and they're going live. So a lot of that has helped us first build the product with their feedback and then do pilots with them to make sure it's working for a team for serious work.

Yeah. Talk about some of the angel investors in the round. How much did you raise?

So we raised $44 million overall. And yeah, thank you. It's insane. It's an amazing, amazing journey for us. And some of the, some of the investors that we have is like first-round capital Chemistry, F4, Scribble, Elevation. These are the VCs. But the most exciting part for me and my co-founder is that we have some of the best design and product minds in the industry. Part of the fundraise, right? So Leo, the second designer at Facebook, investor in Vanta, Perplexity, Vercel. Then we have Katie Dill, head of design at Stripe, Scott Belsky, founder of Behance. Then we have Mike Davidson, head of design at Microsoft AI, Andy Morissette, head of design at Perplexity, Ian Silver, head of design at OpenAI. These are just some of the names. We have a lot. We didn't share the full list actually for fundraising. It's just some small— Yeah.

Yeah, that makes sense. How are you tracking the way design design is changing in the era of vibecoding. It feels like more and more small businesses, more companies are building custom software. A lot of it is sort of like internal-facing, so design is more of an afterthought, but you have to imagine that if the software is valuable and provides growth— a growth vector for the business, at some point you're going to wind up up with design challenges and trade-offs. Are you thinking that this will be like a tool that's eventually dropped on top of like a Vibe-coded system, or do you think this was more of like an entry point into a company building new tools internally?

So all three types, right? Let's say the first type, an existing company, big product, let's say Spotify or— someone else like Dropbox. They have an existing product. They can use the tool to build their next feature. The white coding, if let's say you're doing a side project, you white coded something on Lovable or some other white coding tool, you can bring in your project here almost instantly and start designing and working on it better. And you have a third one, you're starting from scratch. You can of course start from scratch very easily on tool and then and then build your product. So we can support all entry points. It's a very multipurpose tool that way. How are you thinking about business model?

This feels like very logical for seat-based, and yet there's been so much pushback around the seat-based model. You're sort of starting from scratch so you can pick your monetization method. What are you thinking of for the next couple of years?

Yeah, I think right now, at least, we're working on the pricing model. We're talking to a lot of customers to understand what works for them. We don't want to break their budgets, right? Like there are certain budgets assigned already for this. So TBD, we'll soon publish our pricing on the website when we show the product as well. So very, very soon you'll see the product go on the website and the pricing going live. Very cool. Awesome.

Thank you so much for taking the time.

Congratulations on pulling the round together.

And we will talk to you soon. One thing, like I would I've got a gong.

Oh yeah, you have a gong? I have a gong for you. Hit that gong. Because we've— I couldn't hear anything. I think Zoom noise cancellation just completely killed the gong hit. We'll hit the gong again for you. We'll hit our gong. No, it's cool. Next year, we'll just have a beautiful, may I use—

Thank you.

Thank you.

It's great to have you on. I'm sure you'll be back on. We'll talk to you soon. We'll talk to you soon.

Congrats to the whole team. Have a good rest of your day. Name a street. We are counting down the minutes. How long do we have? 29 minutes. WordPress still, still at number 1 at $135,000. Our next guest is Zach Shore from Hermious. He's returning to the show with some massive news. Zach, how are you? Where are you?

What's up? Great, thanks for having me, guys. This is your second outdoor guest today.

It's fully spring. Riley was in San Francisco. I assume you're in Georgia?

No, close. I'm in Virginia. I'm at the Defense Action Forum, the JPMorgan— oh, cool— event. So good place to be when you're announcing a $350 million unicorn round. There it is. That's what I was waiting for.

There we go. Thanks, fellas. And congrats to you as well. Thank you. Thank you.

Big moves for the TVPN fellas.

Big moves. Big moves.

What's new? What's— it's been a minute since you've been on the show. Yeah. Give us all the updates.

Yeah, guys, a lot. I mean, obviously we've got the raise, but the raise is a function of the milestone. So, you know, we flew our second aircraft in 9 months, which is pretty unheard of. The first one we flew last year in like May of '25, and then we just flew our Mark 2 aircraft. That's an F-16-sized unmanned aircraft. So, you know, fighter jet speed, fighter jet size, thousands of pounds of payload, thousands of pounds of thrust. We're slated to fly it again on Friday. There you go. Exactly. That was out in White Sands. And, you know, imminently push this thing to supersonic and really just start— really start building the heavy systems that the country needs. The second aircraft, which is a Mach 2 aircraft, is in production right now in Atlanta. And then we just announced also that we're expanding our headquarters out to Los Angeles in the Gundo. And we're gonna build our Mach 3 platform out there. So yeah, we'll be neighbors.

I'll come see you guys. That's amazing.

All right, what with autonomous jets is, how is the development and just like R&D process different? I imagine, there's a bunch of advantages because you're not worried about a human life. Obviously, there's still a bunch of risk and you don't want to crash the thing that the team works so hard to build, but the iteration cycle feels insane based on how you've described it. And so I'm curious how that works.

Yeah, I mean, you kind of nailed it, right? I mean, taking a person out allows you to take a lot more technical risk, just like full stop. I can lawn dart something, intentionally, right? Just to push the envelope on a vehicle to really understand where the edge is.

Lawn dart? Is that?

Yeah, just like, right? Yeah, I guess.

If we need to, that's not our goal, but you could, right? If I really want to find the edges of performance and there's nobody on board, you have that, you can take that kind of risk and then you can iterate faster. There's also, I can take all of the systems that exist on an aircraft that are there for human survival, the oxygen, the ejection seat, all the, command and control, you know, capabilities, the human-machine interface screens, the stick, the throttle, all that. I can pull that all out and I can put in more payload, more fuel, and just continue to drive more capability for the warfighter. I mean, we just saw that incredible mission to rescue those F-15 pilot and copilot in Iran. And ideally, you know, you have a vehicle like this, you don't have to do a rescue mission. We don't even have to we put ourselves in those positions and ask American men and women to take that kind of risk. So there's, you know, operational utility to the unmanned platform, and there's a significant accelerant to development because of the risk we can take.

Yeah. Take us through the different aircraft that you've built so far, because the— and then remind us of the goal of how fast are we going. Is like miles per hour the correct benchmark for each subsequent test? I imagine that you're trying to make each one faster than the last, basically? That's correct.

I think, you know, we speak in terms of Mach, right? So Mach 1 being supersonic. Yep. But so the first aircraft was not really an aircraft. We called it Turkey. It was sort of a, actually, no, before that was Emu, excuse me. Emu. A flightless bird, right? So that had a jet engine and it had a bunch of the avionics and the sort of radio links to just show that we could build an integrated team, build some hardware, hook up the engine, and get this thing sort of taxiing down the runway. And we did that in 2014. In '25, we built Turkey, which was a flying bird, but turkeys aren't meant to fly. And so that aircraft flew at Edwards and we demonstrated we could rapidly build a jet-powered aircraft. It was a 10,000-pound airplane that had fixed landing gear. It was a GE J85 engine, which is the same engine that you see on the jet trainers that American men and women train on. And that was in May of '25. And now we're on Mark 2 and we call that aircraft Eagle. And this is where it gets kind of fun, right? This is where you start to see what I would call product utility. So this aircraft is the size of an F-16, maybe a little bit bigger, fully unmanned. It's got a 30,000-pound thrust engine. To give you a sense of comparison, like the CCA program, those engines are roughly 3,000 pounds of thrust. So you're looking at 10x more power, about 10x more payload, and just a totally different problem space that we're working in. And the first aircraft of this series, we're building 3 of these Mark 2 Eagles. The first one we flew, that's what you showed the video of, this one will go supersonic. So the premise of this aircraft is demonstrate that the vehicle design, the shape, they call it the outer mold line, can get through something called transonic. So transonic is that window right before supersonic. And there's a lot of very unique things that happen with physics, for lack of a better term, right before you go through that supersonic window in terms of shockwaves and stability for the platform. And so you really want to demonstrate that your plane can make it through that, you know, 0.99 Mach to 1.1, 1.2 Mach window. That's a huge risk window that we're going to unlock here shortly. The next vehicle, Mark 2.2, will have some additional pieces of proprietary technology on it. Our proprietary pre-cooler, which is a technology that sits in front of the engine, will be on that next aircraft. That aircraft's being manufactured right now in Atlanta, and that will allow that airplane to do Mach 2. So now we start to really get into that really, really high-speed regime. As an example, you know, the F-15 is the fastest, uh, fighter jet in the world right now. That aircraft in a dive with nothing on it will do maybe 2.5, maybe, maybe Mach 2.5. So we'll do, uh, we're gonna go for that number straight and level with this aircraft. And then the third aircraft in this series, uh, is going to be Mach 3, and that aircraft is going to be manufactured in the new El Segundo facility and we'll be flying somewhere around the middle of '27.

Talk about the actual technology. Speed is just insane.

Yeah. Speed of R&D. Yeah. It's iterative design, guys.

I mean, right? Like, continue to build hardware so that as I am building, as I am flying the current airplane, I am building my next airplane, right? And so on and so forth. I mean, this is SpaceX over and over again, right? This is SpaceX for aviation, or as we say, SpaceX sideways. So, you know, that's how— you can take this kind of this hardware risk and continue to go into your original question. Our goal is Mach 5, right? That remains our goal. But the key technical unlock in here is actually Mach 3, because in order to unlock our next propulsion approach, we need to demonstrate that you can get a turbine engine, this case the F-100-229, the F-16 engine. We need to demonstrate that I can fly that engine at Mach 3 for a period of time. And so that's what this series of aircraft are going to do. And then We've also got some pretty exciting capabilities that we're going to be able to offer to the warfighter with this platform at the same time.

How focused are you on the, on the, on the defense industry specifically? Because there must be demand from commercial. There's demand from the AI world for this technology. How are you thinking about the tradeoffs there? Yeah, great question.

So commercial is one of those things where Yes, this technology eventually will naturally lend itself to a commercial application, but we're a ways away from that. I mean, if we think about just aviation historically, you start with the defense environment. Think about how jets and planes were even adopted. I mean, the Department of Defense is going to take more risk and going to, you know, help develop inherently just by operating these systems, these technologies to lower the complexity, lower the risk on them so that they can be eventually adopted in a commercial environment. But the other problem is when you work into commercial aviation, the The flight certification requirements for a new commercial engine or a new commercial airframe are steep. Sure. For good reason. Right. We're putting people on board. So to have the economic viability to pursue that certification process requires you to have a stable business that's got robust economics. So for us, even if we do eventually want to go after commercial, you have to build a viable defense business first just to have the economic scalability, not to mention the expertise on the platforms. And so for us, you know, I certainly want to do commercial work eventually, and I think these propulsion systems will lend themselves. But we are a defense company, right? That is our bread and butter. That's where we're focused. We're not interested in some of the, you know, the energy plays on the propulsion. We are really just true north unmanned, high Mach, high altitude systems for the warfighter. And as those systems come online, we will be, you know, arguably the best aircraft manufacturer in the world, and then we'll start lending that to other aircraft problems.

Talk about the actual technology that enables you to go Mach 3, Mach 4, Mach 5, ramjet, scramjet. What is the, what's the lineage here? How much of this has been used in the past? What are you inventing from scratch? What are you pulling off the shelf and leveraging?

Great question. I mean, a lot of what we're doing has been done before. We're trying to stay out of the world of science problems, which is where you get into the scramjets, Mach 6, 7. And stay in the world of engineering problems. And science problems, things have to be invented that don't exist yet, uh, specific materials, production processes. Ramjets have been around since the '50s. NASA did a lot of work on this. You see these on missiles currently and all over the place. And so ramjets are very well understood and well tested. And so what we're doing is taking a different approach with the propulsion system and using something called a turbine-based combined cycle, or TBCC. And This is an engine type that's a mix of multiple propulsion cycle propulsion systems. NASA sort of led the way on this, and we demonstrated this propulsion cycle on the ground in about 2022. And I think DARPA is the only other group that's done this. And so us and DARPA are the only people who've demonstrated this propulsion cycle. And it's got basically 3 components to it. You got your inlet and the air comes in, and the first thing that happens is we've got a proprietary pre-cooler that cools the air down and slows, slows the flow. It then hits the turbine. So in this case, the F-100-229, the jet engine. So we're not building a new jet engine. Now, the challenge has been with ramjets, for a ramjet to light, the air has to be moving through at Mach 3. So how do you get the system? How do you get that air flowing at Mach 3? Typically, we've seen it in missiles with rockets, and you boost it, right? You use solid rocket or liquid rocket, but that has a couple problems to it. Number one, you have to really harden the system to handle all those Gs. And you're not gonna have big wings because of drag. So you're really looking at systems that are not really optimized for flight, and they're just one way. So with an aircraft, you can have a more graceful acceleration. We are able to basically tune that jet engine, that F100, to get Mach 3 airflow. It goes through the ramjet, and then I can light the ramjet. Now, once I light the ramjet, I cocoon off the turbine engine, I route the air around it, directly into the ramjet. Now I can fly Mach 3 to Mach 5. When I decelerate, I do that whole system in reverse. I open the doors, the air comes through the turbine engine, the ramjet shuts off, and the jet, the traditional jet engine takes over and takes me from, you know, Mach 2 back down to the ground. And this is why it's called the turbine-based combined cycle, because I'm combining these two propulsion genres. And it allows us to use mature technology to unlock these sort of these flight conditions. You won't get much above Mach 5. That is, hypersonic is that line at Mach 5, and we don't need to get much above Mach 5. There's two reasons. One, survivability, and the sort of analysis we've seen is it's overkill, and you're gonna find yourself in sort of physics land that's gonna cost you more and take you a lot more time to run up that curve. And also, you get in the Mach 6, 7 world, now you're in science problem world. That's scramjets, which are still kind of more new. That's CMCs for very bespoke material sciences, right? I can have a hot end, a hot vehicle that's titanium or Inconel or steel, and I can mass produce that to handle the heat and temperatures. And so what we're doing is taking a lot of old information and we're modernizing it. And for that matter, the, the SR-71 had an aircraft called the D-21. You can Google this. It was a drone that sat on top of the Blackbird between the, uh, right on the back between the engines. Engines. And once the aircraft was above Mach 3, they turned that aircraft on. It was only a ramjet, no moving parts. The air flowed through that ramjet. They lit the ramjet and the aircraft flew off and it went Mach 3+ for about 3,500 nautical miles. And so in this way, we can start accessing those conditions again, just like we did in the '50s and '60s.

Were those, was the D-21 like a single use designed to fly like reconnaissance or deliver bombs?

It was designed to fly reconnaissance.

Yep, exactly. But then it would just ultimately— crash and burn? It would— well, it was supposed to return.

They didn't end up going the full distance with it. They had a couple successful test flights. They also had some bad flights where they had some challenges. You can look up the history. But more importantly, this concept was executed and validated in the '60s, right? And so we can take modern practices and all these learnings and sort of bring, bring the past forward and say, you know, speed is in vogue again. You know, the, the introduction of ICBMs and stealth technologies really took took America's focus away from high-speed systems. And that's why nobody really worked on these things for about 60 years.

Question from the chat. There's a claim that China has a system that can go Mach 20. Do you know anything about that? Does that seem realistic or propaganda or sounds like—

It sounds like the South China Morning News, if anybody's aware of that newspaper. There's a lot of claims. So I mean, Not proving anything, but I would call, I'd call shenanigans on that.

Okay, last question. Talk about the decision to expand to El Segundo. Incredibly cool community, at the same time, 3 time zones away. It feels difficult to manage a team. Is it about talent? Is it about resources? What's the thesis?

Yeah, it's talent. It's talent, talent, and tacos. Talent and tacos, right? You've got, The core engineering talent that knows how to iterate on heavy hardware, heavy high-speed hardware is really resident there and nowhere else. And, you know, ultimately, as much as the hardware we're building is exciting and innovative, the sort of the team we are reconstructing, reconstructing is really the unlock for this business because aviation has not seen this pace of iteration or these flight conditions in a generation. And so it's not like we can go and hire somebody from insert company that's done this before. We have to effectively take the best we can find from, you know, the closest parallel and then, you know, bring them into the aircraft world and say, take what you learned from, you know, innovating and, you know, and working quickly on rockets and now apply that to aircraft. And so if you really want that talent and you want to draw that talent, the best place in the world to be is El Segundo.

Well, congratulations on the round. Congratulations on the progress and enjoy the rest of the conference. We'll talk to you soon. Very— All right, guys, thanks so much. Have a good one. Goodbye. Up next, we have Hongwei Liu from Matipin in the waiting room. Let's bring them in to the UltraDome. We're still figuring out the transitions between guests. We're working on it. We're working on it. How you doing? Welcome to the show. Thanks for having me. Thank you so much for joining. First time on the show. Please introduce yourself and the company.

Sure, John. My name is Hong Wei. I'm one of the founders here at MappedIn, and we are mapping all of the indoors one building at a time. Okay. Up from Waterloo, Canada.

How are you mapping it? What is the sensor?

So look, there's low-tech and there's high-tech. You can download the MappedIn Scan app, grab any iPhone, grab any 360 camera these days, and just walk through buildings. Yeah. Everyone thinks that that's what you have to do, You don't. The amazing thing is that 70,000 people have now mapped, you know, all sorts of stuff, schools, their own house, offices, malls, just by scanning in the piece of paper that's on every wall. I bet somewhere in that room of yours, there's an emergency escape map on the wall somewhere in the studio. Oh, okay. And you can take a photo of that and you can get it in. The problem, of course, is it's a picture. And so how do we turn that into vector data? How do we make that interoperable useful to all the other apps that you already use and take for granted. That's what we're trying to do here at MappedIn.

So what are, what are all the different use cases for indoor mapping data?

So, so the one that, you know, I've become known for because I've been at this for a while is, you know, that touchscreen in the mall that you guys probably use at Santa Monica Place and stuff that gives people directions? Yeah, that's us. No way! So we do that for, yeah, we do that for a third of the malls in the world. We do it for the Super Bowl 4 years in a row. We do it at LAX. Where you guys are, trying to make that airport experience a little bit better. We think we touch about a third of Americans a year with our products. They probably wouldn't know it's us, but we're trying. And now increasingly we're starting to be embedded in safety applications. We've mapped thousands of K-12 schools now in the United States, unfortunately, because when something bad happens, you need a map of the inside for the good guys to know how to run inside.

Yeah. When I think about those customers that you mentioned— malls, the Super Bowl, LAX— I feel like they have reasonable budgets. So is it about cost savings to use an image based on a floor plan as opposed to just walking around with a 360 camera? Because that doesn't feel that cumbersome when I think about the price of an Insta360 or a GoPro or even your phone. Like, it feels like the data collection shouldn't be that cumbersome, but what is motivating the desire to sort of go lighter on the data collection side? Yeah, fair question.

And first, I assure you they all negotiate very hard. So to any customers listening, they got a good deal. But I think the hardest part about mapping the indoors is not mapping it once, but keeping it up to date. Okay. You know, the goofy example is like Santa Claus moves every year in the mall at Christmas. Oh, yeah. So if you want to know where Santa Claus is going to be this Christmas, the only person who knows is the person on the ground planning Christmas at the mall in that back office, right? If you wait until that's already happened and then you somehow try to scan it, it's too late. Never mind that it's private property and someone would escort you outside. So a lot of managing and mapping indoors for you and I and for everyone else who needs that information is getting ahead and being able to plan ahead and managing that constant change. You know, you can scrape the outdoors with a LiDAR sensor or with a satellite, with a car. It's good for about 5 years. Even if you manage to scrape the 100 million buildings that are privately owned indoors, it's good for about 5 days. So how do we enable those folks to plan ahead?

Yeah, so I mean, the tagline is Google Maps for the indoor spaces. Are you going to partner with Google? Are you going to partner with Apple Maps? Is there an API that you can just expose these things to? Because I imagine if I'm LAX, I pay you, I'd want the data to be as available as possible to my customers and patrons.

I wish I could use that tagline. I'm glad you did. Yeah. So I don't have to. But look, I think from their perspective, and obviously I can't speak for Google and Apple, But, you know, Google wants better data so that they can serve more users. Apple wants better data so they can sell more phones. They— I think they're generally pretty happy when someone does all the hard work of mapping the indoors and makes that information accessible and standardized. So we publish a lot of data on behalf of our many clients, the LAXs of the world, the, you know, SoFi Stadiums of the world, to platforms like that. I can't speak specifically to which ones and who, because they're all pretty sensitive about like, this is my private property, you Mr. Big Tech can't have it. But to your point, consumers need this information. It's on the wall. So we're just about standardizing the pipes and enabling building owners to publish their information.

How do you think this interacts with the potentially coming robotics boom?

I've been getting a lot of calls about that. Wouldn't surprise you, of course. So I think there's enough problems to solve outdoors for robotics that I'm, you know, I'm still kind of holding my breath for when this becomes more real indoors. There's more constraints. I'll just say that if you, you know, we're only now solving for self-driving cars outside. It's been, what, 20 years since we've been talking about it. And that's actually a much simpler problem, right? Like streets, buildings, lampposts, pylons. There's like, you know, far fewer things you have to recognize outdoors to be able to move around. Indoors, there's way more. And the training data is not actually easily available because unless you're Roomba, you can't just scan people's houses. And even if you could, it's incomplete. So, you know, we're building up, we think, a pretty large training dataset. We've seen our technology now deployed for Department of Homeland Security, various fire departments and police departments throughout the United States. But going, you know, I guess I'm sometimes accused of being too Canadian in that we don't sell ahead of reality. I think we're a couple years out of being practical.

You gotta be like, we're going to map the inside of Saturn's moons soon. You gotta think like— 50 years ahead?

How about Mach 3 jets for starters? I'll do that first.

Yeah, yeah, yeah.

Maybe airplane hangars, you know? Yeah, we're headed that way. I'll say this, if we don't pull it off, I don't know who would. We have by far the most amount of training data at this point of the indoors, but we're doing it on behalf of our clients. And I really do think that the indoors belongs to the landlord, belongs to the people operating it. And so it's about who can enable them to realize that future.

Yeah, yeah, and to help their customers. Makes a ton of sense. Tell us, how much did you raise?

I love how you just got a third of the market. Like, it is such a large— so much market share. And those early pitches, yeah, it's just a fascinating market overall. But anyway, tell us about the round.

Yeah, well, my fiancée hates that I don't like going to malls anymore. You would believe it. It's really not fun. I'm going to guess.

'Cause it just feels like working.

Yeah. Oh dude, all right, we raised $24.5 million USD. Hey, there we go. Congratulations. That's a lot bigger than the sales gong we have in our office, so thank you guys.

One day. That's a huge one. Awesome. Well, thank you so much for coming on the show. Congratulations on the progress. Thanks for having me. And we'll talk to you soon. Love you guys. Have a good one. Love you. Bye. Love you. Love you. Is a great song. Uh, our next guest is live here in the TVP and Ultra Dome with us, Zach Kukoff. Welcome.

Thanks for having me. Welcome to the show.

How you doing? Thank you. Thank you. For you guys, it has been far too long since you've been on the show. Last time, uh, we were talking—

why are you in— why are you in LA? Why is a man like you in Los Angeles?

What does a man like me have to gain in San Diego, a nice tan, relaxation.

Everyone who's joined the show has been outside. I mean, a few people, but it's— spring is fully in. It's beautiful. Yeah, in full swing. Did you get snowed in in DC? Did you go crazy? Blizzard?

We had Snowcrete, so we were stuck. What's that? Snowcrete? I had never heard of this before. Washington, the only East Coast city I've ever seen, they haven't figured out how to plow snow. Okay, the The snow, here's the problem with Washington. The snow lands, everybody just sits there, and there's a wait for it to melt. So their only plan is to like help it get warm.

Work smarter, not harder. Yeah, except neither.

They just let it sit there, and it didn't melt, and it stayed really cold for a week, and it merged into like a weird hard, soft, hard, soft, the water and the snow, and somebody who's smarter than me can do the like chemical side, but basically once it melts and freezes over, 4 times in a row, it becomes this thing called snow creep and you're basically stuck and you can't leave your house. So we couldn't like drive basically for the better part of like 2 weeks. It was brutal. Yeah, it was wonderful. But how is business broadly?

Reintroduce yourself, re-describe your role, and then we can go into some of the hot topics.

Yeah. Yeah. Okay. I run a lobbying firm. Yeah, a lobbying firm. I run the tech practice of a lobbying firm. Firm's called Louis Burke. We do science, tech, education. Wow, that's so nice. Wow. That should be in every room I walk into. They should always applaud for me. We do science, tech, education, healthcare. I run the tech side of the house.

You're like, you're really—

I was watching you guys on the way over.

The most like straightforward explanation. I really like jazzed about it.

I was like really into it. Anyway, I run the tech side of the house. So I lobby for venture firms, I lobby for private equity, I lobby for high net worth individuals, mostly who come out of venture private equity, and then I lobby for a whole bunch of tech companies. And specifically on the federal side? On the federal side, that's right. So we do everything at the federal level.

So give me your sort of postmortem on big beautiful bill. That was the last thing we talked about. Yeah, wow. What were the key decision points? I know that the government was shut down for a while. There's a lot of back and forth. Like, how did this all shake out?

Okay. Basically every— this is— it didn't used to be this way. It used to be in Washington that things operated, I wouldn't say smoothly, but like go back to the '90s, West Wing optimism. Newt Gingrich, Contract with America, Bill Clinton. Like, yeah, yeah, we got a sound effect for that one. That's nice. And things pretty functioned pretty well. Now in Washington, basically every time something happens, it sets up the seed for the next issue. It's like if you only had the Treaty of Versailles over and over and over again. Okay. So big, beautiful bill comes out. What happens after big, beautiful bill? You lead up to the big conflict, which is around Homeland Security, ICE, and enforcement. And you basically have a variety of issues that come out. Most recently, you guys, I'm sure, were tracking Christine Ohm out, right, at DHS. Others may be coming out soon, too, right? You know, AG, many others in that case, too. A lot of this stems from the inability to get the entire government funded, right? Big, beautiful bill. And by the way, reconciliation, reconciliation as a whole, guaranteed funding for border enforcement for ICE, right? So when the Dems later come back and say, "Hey, I wanna be able to stop some of these things that are happening in Minnesota and things like that," the only lever they have to pull is to stop funding DHS as a whole. That's why you get these airport shutdowns. That's why you get these super long lines at TSA and so on and so forth.

[Speaker] Yeah, but from a tech perspective, how many, like, all of that, that like flows through to whatever tech leaders want to happen in Washington just is slower. But what was on the top, what's on the top of the stack in terms of to-dos in Washington for Silicon Valley broadly these days?

I mean, I will tell you it's less of a policy to-do and the biggest political issue right now is probably the data center stuff. You guys are, I'm sure, tracking, I'm sure people aren't talking about it. You saw the thing in Indiana, right? City councilman. It was horrible, terrible. Had 13 shots in his house. Okay, putting aside how horrible that was, there is like a very broad bipartisan emerging consensus that data centers rise energy costs, water, and all sort, you know, create pollution and whatever. You saw like they, you know, increase heat temperatures for, you know, parking lots. A lot of that's not true, but because the perception of it is real, you have this It's a big, huge, armed, basically, opportunity for candidates who are populist sometimes, but opportunist always, to come in and run on that issue in '28.

So the big—

the problem with tech right now in D.C. is you have a bunch of tech money coming in. You guys know Leading the Future is the OpenAI-associated super PAC. Sure. And Public Action or Public First, something along those lines, is the Anthropic-aligned super PAC. Both have been putting tons tons of money into races like New York 12, I think with Boris, right? All sorts of people who have been either pro or anti-data center and AI development as a whole. There are now hundreds of millions of dollars of tech money going to try to arm those. That's the big issue for tech. Like if you want to get anything else done, you basically are saying, how do I play now in midterms, but also looking ahead to '28, even if I'm not talking about data centers directly, to make sure my performance preferred candidates get in to actually open the doors for the things I wanna be able to do?

Yeah, and how much of the data center question is about research education on those issues? Like there was a full back and forth on the water issue. I think that one landed in a pretty good place. With energy, less so, because no one debates that these data centers use a ton of electricity. I mean, they're measured in electricity.

That's how we refer to how big they are.

They are. It's a gigawatt or megawatt or—

well, it's a proportionality problem, right? Like the data center issue is, yeah, do they use a lot of electricity? Sure. Yeah. But in comparison to things like growing almonds, right, or raising a cow, sure, it's actually not that much. And it turns out we love cheeseburgers and we find cheeseburgers like have huge value in America. And so we accept the tradeoff that in order to have cheeseburgers, we have to pay a little bit for electricity. Yeah, it turns out there's a lot of value in AI, by the way. This is not a winning political argument. I wouldn't go up and down and make the argument that, hey, yeah, everyone's going to have to get used to paying a little bit more to pay for AI. No one wants that. No. Who would want that?

And that's the backbone of the ratepayer protection pledge. That's right. That's how is that going? Because that is a pledge right now, but it feels like it could be codified into law at some point.

I don't know that I'm super bullish on it being codified, but what I'll tell you is anything that SEC can do to get in front of the issue, like you guys know the account On Twitter, more Perfect Union. Yes. Okay. Okay. I think they are the most effective messengers, by the way, in politics today. Sure. They have one narrative they push, and it's very, very well done, which is look at these great local small-town, salt-of-the-earth people fighting back against the evil tech firms who want to do things like build data centers. Right. And so the more that you see that and the more salient the issue becomes, and it's already so salient as it stands today. You guys saw League in the Future played in the sort of two Illinois races that just happened recently during the midterm cycle. They won one out of the two. If you look at what they ran on, in neither case did they run ads saying, "We love data centers, please fund data centers," right? The ads were about saving democracy and it's great to have Jesse Jackson Jr. and the other candidate and so on and so forth. Like the more the tech can come out and ahead of these things and then change the topic to diversity a variety of other issues voters care about, the better position we are in. The more that we just try to take this on, 'cause we don't need 30 more years, right? We need to have a window of time where we can come in and get data centers built and actually start to operationalize a lot of the future value of AI. If you wait too long, there's a window where our adversaries pull ahead of us and the opportunity for American AI to dominate goes away. And if we're too aggressive, right, the other inverse of this, is we risk really alienating a lot of the communities that are local. So you need to thread the needle on getting out ahead with things like the pledge, right? Making sure we can come out and show that we're listening to constituent needs while also at the same time actually delivering on some of the positive benefits that are not just more doom around AI's gonna take your job and kill your horse and burn down your house and whatever. Sure.

What has messaging been like around the, like, I think for a lot of people they're like, okay, data center in my area. What does it do for me? They're like, am I still gonna be able to use LLMs or video models? They're like, even if it doesn't go in, they're like, yes. So, okay, so what are you gonna do for me? And so I think like one of the best arguments for it to date has been like tax revenues. Is there any that can be repurposed for a bunch of other things? Seems to be pretty significant based on some of the numbers I've heard, but. Is that kind of messaging resonating, or are people still just saying, you know, flat out, not, not in my backyard?

Yeah, it is. I would say it's very nimby. I don't think the tax stuff is breaking through to people. And the reason it's not breaking through in part is because for every person like you and me and John who says, of course, these things are going to throw off cash, how could you not look at them and think you can fund an entire school district with one data center, right? You have somebody else who goes, "Well, that doesn't logically track to me," because you have 5 employees manning these things, and they're gonna be more and more automated as time goes on. And more to the point, why not put them somewhere else where I don't have to deal with it and see the negative externalities? Like, the messaging thing that tech hasn't figured out yet, and partially we do it to ourselves. Every time you have somebody come out and say, "AI is gonna take your job, it's gonna totally disrupt society, it's gonna end," you didn't have— the people who built NAFTA weren't selling NAFTA Nafta by saying, hey, Nafta's gonna ship all of these jobs overseas and be horrible. And by the way, Nafta, whatever you think about it, did set up the conditions for a lot of jobs moving overseas. Like, tech has to get out of its own way a little bit and stop saying things, even if there is some, some tail chance of it being true. Like, we've left the context bubble, right? We're not in, I lived in SF for 7 years, like, we're not in Berkeley right now where we're having a great conversation in front of the light cone and whatever, We're not having sort of in-depth intellectual discussion. We're in persuasion mode, and you can't come out and say that.

Yeah, what about just taxes broadly? Because when I hear a data center will throw off a lot of taxes, I think, well, maybe corporate taxes wherever that company is headquartered, and they'll certainly pay real estate taxes, but it's a very small real estate footprint. And so are there any local, local regions that have figured out how to rethink the tax base to actually capture some of the value and say, look, we're down, but you're gonna have to pay and here's the actual deal and let's think about it more in those terms to make sure that we're actually internalizing the value.

In some ways you're a Georgist, right? You're like, hey, I'd like to have a land value tax. Yeah, yeah, yeah. If we just capture some of the emerging value improvement we've created, which I'm very sympathetic to. I hear the argument. The big tax conversation today, it's been a little bit back and forth, and the back and forth has been between some people who are saying, hey, we need tax breaks to lure in data centers. Like, for all that we hear about local community hates data centers, whatever, there's a huge swath of, like, offline people who are saying things like, actually, I would like to have any job in my area, right? I'd love to have any good construction job, any job running these things. So that's okay. So one side of the issue is, do we have people who are saying we need tax breaks or tax incentives to lure? Right. And that's actually a wedge issue even within Republicans, too, by the way, who say, hey, some of whom say, hey, you're putting this in my backyard, you're destroying the character of my community. Very sort of the classic NIMBY arguments we've heard. And some of whom are saying things like, gosh, I would love to have more industry in my community. Okay, so that's one. The other side of it on the tax side are people who are saying, "Okay, we already have 'em, right? And so what are the other things that we can levy?" Sometimes land improvement, sometimes by the way, it's a consumption tax on energy, right? So you can actually, like, there are people who are saying, "Look, should we craft more narrowly targeted energy consumption taxes?" Commercial in their orientation, which is not a perfect, it doesn't—

And it's not something we've historically done because if I'm using energy to do my dishes and you're using energy to not watch Netflix, We don't typically put a value on that differently, but maybe extraordinary circumstances require extraordinary—

I mean, many, many have pushed for, you know, podcasts tax, energy consumption.

Podcasts are the backbone of America, okay? We cannot tax them. And so you see this back and forth in Georgia. Like I would tell you, okay, Virginia, which is I think the largest concentration of data centers in the country, in part the reason Spanberger, the Democrat, won the new, recent gubernatorial race in Virginia is because she campaigned on the sort of very kitchen table issue of your energy cost has gone up. Yep. I am going to do the things to make it go down. Yeah. One of which is slow down on data centers. Right. And in Georgia, if you look there right now, I would tell you the best chance the Dems have of flipping the Georgia gubernatorial race in the last 20-something years is running on— they, by the way, they flipped some of the Georgia public utility commissioner seats, which are like super nerdy, low salience. but they flipped them on this question of, I am paying more for my energy. I don't want that to happen. How do we prevent these data centers from going on? So it's a big problem, and the super PACs are doing a good job, but they could be doing a lot better of a job of articulating, by the way, non-AI-related messaging to support AI candidates.

What about zooming out to just energy broadly? Because I feel like, while the AI companies might be, and the tech companies might be, narrowly focused on data center construction, a lot of them do have bets in solar and nuclear. And a lot of venture capital firms have funded solar and nuclear efforts. And that feels like potentially a more pro-bipartisan issue. How do you see that breaking? How do you see the— like when we talk to nuclear founders, we get extremely excited and then they give us their timelines and it's like, it'll be online in 2030. And I'm wondering if there's anything that the government can do all of that forward? Are there any efforts or any reframing? Like if you were to put on your strategist hat on, like, how do we— I mean, how do we just get excited about clean energy? Because that feels like it's longer term, but it's potentially the release valve for everyone, makes everyone happy.

So, okay, I'll say two things. One, and I should disclose that we lobby on basically all of these issues. So everyone should discount appropriately. Yeah, I'm talking my own book. Okay. One is you can think about what are the ways— Thank you very much. Thank you very much. Thank you very much. One is we could think about what are the ways that we can co-locate, by the way, data centers with currently stranded sources of energy, right? So if you have like— there's two components to energy. There's production and there's transmission. And the government can do a lot— the federal government can do a lot on from both the research side and some of the permitting side on the construction consideration. The transmission side is a lot more nuanced and sort of a boring conversation. But suffice it to say, like, there are very limited ways in which federal government can really make a big impact on transmission. And so you might think to yourself, okay, what are the ways where we can take places we've already built energy, right? Everything from all of the above, right, from nuclear to solar to coal to natural gas, and build data centers that already exist. That's same thing, by the way, that crypto mining did, right? For many, many years. Call that door number 1. Door number 2 is I am very bullish on nuclear and things of that nature, and partially because we do lobby for it, but also like if you're thinking about what are the ways in which nuclear moves forward, what you actually want is for the federal government to think about what are ways that we can centralize authority over permitting more and more? Because you don't wanna have like the local permitting group up in arms saying, we don't wanna Three Mile Island, what you actually want is for like a much more centralized one-stop shop authority. Sure. And so in that way, you're not— you call it industrial policy, call it big government, call it whatever you like, but you're getting back to the idea that the government plays a much firmer hand in actually driving where these things go over the short and medium term. Mm-hmm.

Is the data centers in space thing bullish or bearish for building more data centers on Earth? Because it feels like a get-out-of-jail-free card a little bit. It's like, hey, Now that we can build 'em in space potentially, like let's just not build anymore here.

Well, it's my position that America owns space too. Okay. Let's just start there for a second, or at least a little bit of space. Yeah. Okay, in all seriousness, the honest answer is I find it hard to believe that the economics of building data centers in space flip before we have a government that is more flexible and more amenable to people who want to build locally. Like part of the challenge is this, you saw the, Here's a quintessential example, okay? The Germans get very excited about decommissioning nuclear energy, right? They lobby for it for years and years. They have the famous tweet you've all seen of the Green Party candidates announcing the shutdown of the last reactor. Then a crisis happens, in this case, Russia-Ukraine, and suddenly Germany is saying, gosh, I wish we had nuclear energy. Any source of energy we would take today, okay? You have the strait getting closed. You're saying to yourself, gosh, I wish we had any source —of energy.

I think oil is at $140 a barrel right now.

Yeah, I don't know the exact, but that sounds about right.

Which was where it was at when— The Russia-Ukraine invasion.

Yeah, that's right. OK, what's the common thread across all these different things? Government is reactive, and people are reactive to crisis. And so it is— you would never root for a crisis. You don't want a crisis to ever happen. Don't let it go to waste. But don't let it go to waste, to quote Rahm Emanuel, right? And so you're far more likely to say, hey, what's more likely to me on a short medium time horizon, the economics of building in space rapidly flip, possible, but I don't know how likely, or something happens on Earth, and there are millions of things that could happen, such that people now are newly incentivized to build and allow for building at home, and I hope it doesn't require a Sputnik moment, right? I hope, by the way, another Sputnik moment, 'cause maybe Deepseek was already one. I hope that's not the outcome, but if it is, the outcome will be people are much more amenable building, I think, locally here on Earth and in America.

Thank you so much for coming on and breaking it down. Yeah, we appreciate you.

I appreciate it. Congratulations, guys, again.

And we will talk to you soon. We have some breaking news I can take you through while we bring in Thomas Lafont from CO2. The alley has officially been auctioned fully. The naming rights to Riley Walls' alley have just sold to Notion for $140,000. It's now the Notion Way. It does seem like they sniped it. The Notion— I like that. What was WordPress doing? They called it the Notion Way. Yeah, WordPress was asleep at the wheel. They let it get away from them for just $5,000.

Do not pay. Just, just absolutely.

I like that someone was trying to make it Bags Street. Core Automation Inc. The Way also was there. Someone went back and tried to keep it named Dirt Alley for $111,000. Gumroad was a good one. Well, we can bring in our next guest, Thomas LaFont from Coatue. He is here live with us in the TD PN UltraDome. Thomas, great to see you. How are you doing? Thank you so much for taking the time. It has been too long. I want to actually begin Let's begin at the very beginning. Can you tell us where you grew up?

Yeah, I was born in Paris in 1976. I'll be turning 50 in— Congratulations. Actually, about a month and a half or so. Overnight success. Thank you. And really split my growing up between the US and France. My father was an executive who kind of moved around. So between Paris and New York, did a back and forth twice, was kind of used to moving around. Settled in New York in 1988. One of the first questions I get is why I don't have an accent when my brother does. And I think it's really related to our— kind of where we grew up and our difference in ages when we moved to the US. But obviously incredibly grateful to have moved to the US. What I do remember from France is the feeling of a country that kind of felt stuck in neutral. And so coming to especially a city like New York with the dynamic economy and just the feeling of life just in the buildings, construction, was incredibly motivating. I went to a really international school in New York that had a lot of different types of people. It was a French language school, so you had your kind of classic expats. You had a lot of diplomats from all over the world since French is a diplomatic language in a lot of countries. So I was really grateful to that exposure. And then I went to Yale. I studied computer science. My junior year, I realized what a good programmer was and that I wasn't one of them. Because what would take a good programmer an hour would take me took me 6 days. So that was kind of sobering. And I thought, OK, I've got to think about something else. And I love movies. I always watched a ton of movies. It was kind of the peak for, I think, kind of the movie business. We were about to roll into the DVD era, which, if you think about it now as an analyst, was essentially the studios monetizing a library again at virtually no cost. So everybody's building DVDs. Media's at the peak. You had Mike Ovitz on the COVID of Newsweek. So I learned as much about that industry kind of reading the New York Times on Mondays, which was kind of the digital edition, reading Vanity Fair and Premiere magazine and anything I could get my hands on. I said, OK, well, Hollywood sounds like a lot of fun. And CAA has this training program. Well, I don't know anything. I don't know anybody. I've never really been there. So getting trained sounded pretty good. I went to my alumni house. And we had these binders. And you could look up industries. And there was one that said entertainment. And I saw there was one agent who had gone to Yale who was at CAA. So we're now in my senior year. I called her every day for 6 months at the same time. And I got to actually know her assistant pretty well, because agents have to pick up the phone, right? Because that's how their business is built.

Wait, what time of day? Like early morning? Late? No. How did you settle on one particular time? Because— 3:00 PM Pacific. By default, I would mix it up and try to press it.

I knew first thing in the morning. No, it's kind of an interesting thing.

If you just get a random call here or there, you're not really paying attention to it. But if every single day at 3:00 you get the same call, by the third or fourth day, you're like—

—I knew she wouldn't take my call first thing in the morning. —Sure, too busy. Yeah, exactly. So I'm like, right about right after lunch, right? It's like the mid-afternoon nap. Maybe I'll hit them in a weak spot. So I got to know the assistant, and we would kind of joke around. But finally, one day I got a callback from her. Her name was Sally Wilcox. She was a book agent. And she said, well, what's it going to take for you you to stop calling me? I said, well, I want an interview. And she actually gave me, I thought, an amazing answer to that. She said, well, if you agree to move out and you call me after you've moved out to LA, I'll get you the interview.

She wanted you to move. Move first. Move first.

Before you even get the interview. That is a crazy answer.

But I realized what she was doing was she was kind of testing my conviction. Sure. And she probably got a lot of calls from people who said, oh, I want to do this, I want to do that. And she's like, well, if you have the conviction in yourself to move out with no job, that shows you really kind of want it. It was probably a filter on her part. So, I did. And it was my second time, I think, in LA. And I called her. And I only had one egg in the basket. There was no other egg in the basket. This was the only egg in the basket. Basket. So I called her. I said, well, remember, you told me if I moved, you would give me the interview. And she did. I interviewed, I think, on— I moved out May 22, I think, after I graduated. I got here first week of June. And by July 7 was my first day of 1997 in the CA mailroom.

In the mailroom.

And I think you guys were just maybe there, right? It was just this iconic— place, and a lot of people I kind of looked up to had started in the mailroom. Ron Meyers started in the mailroom, David Geffen, Barry Diller. So I'd read about all these legends of the business kind of starting.

There was a dress code back then, too?

What's that?

There was a dress code back then, too? Oh, absolutely. Yeah, there's still a dress code today. And I think that's the thing that sticks out the most about the mailroom is just the number of— talented young people that you have in a pretty small space, and they're all dressed perfectly, which is just—

And every experience you have in that mailroom is kind of unique. So I'll tell you guys this story. I don't think I've ever shared this one publicly, but we did a lot of work for Ralph Lauren. And this is back kind of in the Friends era. And if you remember, Jennifer Aniston's character on Friends worked at Ralph Lauren. So he's coming in, and he's doing a taping. And so I'm asked to go pick him up at the Beverly Hills Hotel and just escort him for the day while he's shooting the scene. And if you remember the scene, it's in an elevator. And Jennifer's character, Rachel, bumps into Ralph. And it's like a 12-second scene. You can look it up on YouTube. So it took like 3 minutes to shoot. So we get on set at 9:00. I think he's in denim on denim, just classic Ralph. And right before he goes on, he's like, call her up or call her down? I'm like, call her up. Let's go. We're rolling with this. And so it takes like 3 seconds. And so we arrived at 9:00, and like 9:08, we're kind of done. And Ralph says, well, I didn't think this would be that quick, and my next meeting is not till 4:00, so why don't we go and spent some time together, and I'm really kind of interested to go where you shop. Oh, interesting.

Oh my gosh. He wanted to jump right into market research. He did. Yeah.

So we spent the whole day going together, shopping around. One memorable one was going to Fred Siegel, back when Fred Siegel on Melrose was— every new brand was kind of getting broken into there. And I just watched him walking around, and the way he engaged with every single salesperson, never talked down to them, —to your point—did market research. Why are people buying this style, not that one? It was just amazing to watch, and how just curious he was, and how much he wanted to learn. So, we finish, and I'll get to the end of the story, but at the end of the day, he's like, look, I gotta ask you something. It's really bothered me the whole day. I'm like, yeah. He's like, why aren't you wearing Ralph Lauren? And I said, look, honestly, the cut's not great, and the store on Beverly Drive's a little old, and then there's a lot of wood. It doesn't really feel kind of new. And he's like, man, you're so right. But he's like, you really should be wearing Ralph's. So go there, and just tell him I sent you, and you can get anything you want. Wow. So we kind of parted ways, and he was really lovely. But I debated whether I should go or not. And the next week, I go. And I go, and I introduce myself to the very pretty lady at the counter. And I said, look, I'm sure you get this all the time. You're going to think I'm a crank. I said, I know, but Ralph sent me, and he said I could get whatever I want. And she paused for a minute, and she looked me up and down. She said, oh, we've been waiting for you. No way. And I kind of spent an hour kind of going through the store. But the exposure that you got to people through that job was really amazing, watching actors, watching directors, watching entrepreneurs like Ralph. So, I was really grateful for that opportunity. Talk about— you had a—

we were catching up yesterday off air, and you had a story about— you can just talk generally about your— what maybe entrepreneurs could learn from actors and actresses that are breaking in through Hollywood and what they have to go through from a competitive standpoint to actually break out. I'm sure you got to see a bunch of different stars over the years, but the, you know, everything from, you know, the rejection to the constantly having to, you know, constant hustle because, you know, as soon as a project ends, it's like, okay, what's the next thing?

Yeah, I think people underestimate, first of all, how hard it is to be an actor and how competitive it is. First of all, especially in the movie industry, right? You have to kind of find a new job every you know, 6 or 7 weeks. So, you're constantly unemployed. You're constantly having to search the new job. Not only that, when you go and you actually interview for a job, you might run into 10 other people that look exactly like you, right? All of your competition— like, imagine if I was pitching an entrepreneur, and I'm like, "Oh, there's Sequoia and Andreessen and Benchmark, and we're all kind of sitting lined up next to each other waiting." That's kind of what actors go through when they audition. And then, by the way, when we give you feedback, it's not even feedback on your business. It's on you. We don't like how you talk, how you look, you're tall, too skinny.

Yeah, or your ears.

Your ears are— I represented Jordana Brewster, who's a friend now and is married to a friend of mine. But one day on one of the casting sheets, it said, hey, a Jordana Brewster-like actress for this role. So I called, and I said, well, what about Jordana Brewster? This is literally who you said. They're like, no, no, no. No, I want someone like her. So, well, what about her? Right? But that is the kind of stuff. And obviously, the competition is intense, right? And it actually reminds me a little bit of what it was like to be an entrepreneur in China, right? Very similar. Like, you would have— if you had a ridesharing company, there were 150 other ridesharing companies that you had to kind of get through, just to then win your city, to then go and compete with all the other cities, to then at the end compete against Uber in China. as an example. And Dee Dee won that. So it's an extremely competitive industry. And I relate to that as someone who worked with actors. You would sometimes talk to an actor and say, how'd the audition go? They would tell you, I nailed it. It was just so good. And then you'd call the casting director or the director, and they would say, that was the most unprepared, bad audition. And so you have to figure out how to communicate that feedback in a way that's constructive to the client, right? Because just telling them, "Oh, you did great, but you didn't get it," isn't necessarily useful. But in a way that doesn't also, you know, is detrimental, right, to their mental health and things like that. So, it's very much a people, a reputation kind of business. And I really enjoyed it for the 7 years I did it.

Yeah. What was the process of getting out of the mailroom? I imagine that you have a few really iconic stories from the mailroom, but there were plenty of days that were just photocopying. Is that roughly correct?

Yeah, I mean, we did shopping for groceries. Oh, sure. We copied scripts, we delivered scripts. So there was a whole set of kind of stories on that. And honestly, we could fill a lot of podcasts with just the different things that we kind of did. But look, it rewarded hard work. It rewarded attention to detail. Ultimately— and you got invited to the retreats, right? And CA did these retreats every year. And after one of them, I had some thoughts. I wrote them down. And I hand-delivered a letter to the managing partner saying, I was at the retreat, and here are my thoughts, right? I wish I still had it. I don't know what I said. And then I didn't hear anything for a long time, but when I came— Off what was called the runs, where you delivered scripts, you then waited to be picked for a desk. And only the top 5 got to interview for desks to try and not keep people jammed there too long. But the desk of Brian Lord opened up, and he was the co-chairman and is now the CEO. And he asked to interview me, 'cause he had read my memo, and he said, "I wanna interview the kid from the memo. Even though he's not in the top 5, I don't care." So, I went up and I'd never really met him or spent any time with him before, but we sat down and somehow we got talking about John Steinbeck, which was my favorite writer, and I was reading a biography of him at the time. And we spent 30 minutes talking about John Steinbeck. Nothing about business or how do you answer the phone or what he's looking for in an assistant. So, I came back down and all the guys were like, "Ah, how'd it go?" I'm like, "Well, I don't think it went well, because he didn't ask me a single question other than I was talking about John Steinbeck. So I'm like, there's no way I got this job. And then he ultimately kind of gave me the job. And I worked with him for almost 3 years. Unbelievable experience coming off of Mike Ovitz having gone to Disney, then having flamed out and starting his own company. So much kind of turbulence in the industry, and he was an amazing guy, and it was, it was a great formative experience for me.

Uh, since we are in Hollywood, running a show that covers venture, was there anyone in that era of Hollywood that was, uh, from the talent side that was leaning in and investing in startups of any kind?

No. I mean, what I remember from that era is no one really thought about the investing side of it, per se. People thought that, wow, the business is going to be digitized, so let's think of creating content for these new digital channels. And so it was more on the monetization side of, here's a new distribution channel called the internet, and how can we adapt our businesses to that new channel? I had always loved investing, so I was trading public market stocks on the side. And my brother and I kind of started kind of doing that together. And I just bought companies that I liked and was kind of doing it as a hobby. Eventually, when I was promoted to an agent, I kind of realized two things. One is I preferred being an assistant to an agent. I preferred being an assistant rather than an agent. So that was one. And then two, I kind of really like this stock thing on the side, right? And someone said, well— always invest in someone's hobbies, because that's what they choose to do in their spare time. And so Philippe called me, and this was a couple of years into CO2, and said, well, we're kind of trading this account together anyways. Why don't you come over and do this? And that's how I got started.

And what year did you get started at CO2?

What was the strategy then? 2003.

Walk us through post-dot-com crash. How are you feeling? What's the strategy?

2003, so we're just coming off of 2000, 2001, 2002, and down 80% on the market. '03 kind of had a snapback, so the market was up 50%. We were well positioned for the downturn, not as well for the snapback. But now it's about, OK, what does the next kind of decade kind of look like, right? The easy money's been made on the quick rebound, right? NASDAQ, I think, was up 50% year to date or something like that. Like that. And we were really looking for a semiconductor analyst and couldn't find one. And this is something I'll forever be grateful to my brother for. But eventually he said, look, since we can't find one, he dropped a copy of the universe. And this universe was a printed memo of essentially all the stocks and key metrics about each name— PE, volume, sector, the semiconductor universe. And he said, you know, why don't you go ahead and do it? Can't find anybody anyways. And I had no real training as an analyst. And I would sit in our bullpen, and there were analysts from Morgan Stanley and Goldman Sachs. And I thought, man, my training was in a mailroom. I don't know anything. But what I started to realize is the downside is I didn't know anything. The upside is I didn't have any bad habits either. In terms of how— so I really learned from my brother directly, and I started to define how we want to invest versus how maybe you learned it at Goldman or a mutual fund or something like that. And so I kind of relearned from first principles, Philippe and I kind of working hand in hand. And I started learning semis, and pretty soon, if you started in semis at the time, And all the roads led to one company. And that was in Cupertino, and it was Apple. Why? Well, because the iPod was starting to really gain traction, and an iPod was a semiconductor product. It had NAND flash. It had a processor. So that was the gateway into Apple, which eventually led to the iPhone. And it was kind of an iconic investment for us. And we got to know that management team really well.

Yeah, I wanted to ask you about that. How much of your investing philosophy at the time was quantitative, pulling metrics, building models, versus doing expert calls, talking to management teams, listening to earnings calls?

The model was a really sacred place for me. And because I hadn't trained as an analyst, I felt I needed to build every model myself, because I didn't trust myself with someone else's work, because I wasn't good enough. So I'm like, well, I'm going to build every single cell myself. That's going to mean I'm going to understand it if I built it myself, right? Versus if I take someone else's complex model, I'm not going to understand. I'm going to rip a DC since I'm number 10 for the day. So let's go. I knew when you had coffee, by the way, after dinner last night, I was like, OK.

My god.

Get ready for the caffeine. Yeah. So I felt like I had to rebuild every model myself because I couldn't understand. Someone else's model. And what happened in the process of building that model is, as I would go through a sell-side model, and there were lines that I thought were irrelevant, I said, well, why should I add that to my model? Like, this is driving no value. But it was kind of a verboten thing at the time. It was like, well, hold on, the company reports it this way, or this is a revenue line, and for the sake of accuracy, you kind of need it in there. But I didn't know enough to basically say, well, to me, it doesn't drive any value to my investment thesis, so I'm just gonna lump it into this other bucket, yeah, called other. And I'm gonna rename things the way I understand them, not the way the company chooses to report it, or— Yeah, except basically just what made sense to me and what my thesis was about, so that when I pitched my thesis, my model actually reflected what my thesis was, right? "Let me pitch you my thesis, but now I have a 1,000-line model, and I have to go from row 2 to row 7 to, like, the other tab, then back to line 250 to kind of explain it to you," like, that made no sense to me. Like, the narrative was, "No, let me show you from the top line all the way to the bottom how it flows and what the key functions are." Yeah, what's—

That started developing, by the way, our philosophy. Yeah, doesn't that kind of still define —like a partner meeting? Like, aren't you guys— like, let's say you meet an entrepreneur, you're excited about them, you do the work, and then from what I've heard, you guys will spend like hours and hours and hours still just in the model, like ignoring—

ignoring— And we will, but a lot of times you might be getting a model from the sell side, right? Because it's faster and it's more convenient, but it might not be exactly how your thesis is being laid out, right? So one of the things I try and talk to all of our analysts and say, well, let's have a model that really reflects our simple view of what the thesis is and what the drivers are. So I think a model to me is kind of a sacred place. And in fact, our Apple model, which I then passed on to Jaymin Rangwala, who's now our CIO, was kind of this sacred— I didn't let anyone edit a cell on that model. I knew every single cell. I knew the color. I wanted a very specific kind of color for the background of certain cells, right? And eventually that kinda got passed on. But to me, at the end of the day, I learned this in— actually, back in Hollywood, we had all the trainees one day got brought into a meeting with Steven Spielberg, and Steven said, "Every great story can be pitched in 3 sentences," no matter what the story was. And he said, "So pitch me a story or a movie and I'll pitch it to you. And no matter how complex the movie was, he understood the essence of it, and in 3 sentences you got the whole movie. And what I realized is it takes a true understanding of story to be able to crystallize it in 3 sentences, right? If you don't understand something, you'll say, "Okay, well, TVPN is a podcast, and it's about these 2 guys, and they do this." Well, do you really understand what it's about? Because you just gave me 10 minutes of rambling stuff, right? And all the great investors that I've met, like Stan Druckenmiller, or my brother Philippe, or Dan Loeb, or some of these kind of legends of the hedge fund world, right? They have an ability to take any kind of story and just drill it down into its essence, to what the key pivot points are that are going to make or break that stock at that particular moment. And so we really try and say our thesis should be simple. We should be able to explain them very in, you know, few sentences, right? And you should have a model that reflects that thesis. So on the public side in particular, then we dive into, okay, let's go into your model and let's say, okay, can Apple sell 50 million phones? Can the ARPU be in the out year? Is it Is it gonna increase? Is it gonna decrease? You know, when I look back at our old Apple model, I actually think we did a pretty good job on units. Where we were way off is we had the price of the phone declining 5 to 10% a year, 'cause that's what every consumer electronics product did. TVs. And in fact, the ARPU doubled, right? I think the first iPhone was like $600, right? Unsubsidized. Now people spend $1,200 easily. So never would have kind of forecast that. Interesting. So yeah, so models are quite important to us.

What was the mood in the hedge fund industry broadly during that time around different strategic expansion opportunities? There's obviously a high-frequency trading boom that's happening. There's more quantitative strategies, there's debt strategies, there's so many hedge fund can can mean so many things. How were you thinking about defining what you would do best and where you would expand to or decline to expand to?

Look, I think for tech, it was an amazing environment to be in because almost every company was getting swallowed up. So if you were in TMT, you really felt like you were the center of the universe.

So there wasn't much of a pull for distraction. Correct.

That makes sense. And then also, we had companies going public pretty early on. So you could do a lot of differentiated work in companies that the market cap was a couple billion, right? I think what really changed for us in the early 2010s was Meta, then Facebook, and Alibaba staying private for longer. We'd just never seen companies of that scale who were that important to our research and our market not go public.

It was Meta, or Facebook, went out at like $60 billion, I believe, something like that.

Yeah, I think around there, right? And right around 2012, 2013.

So if you're used to buying a company potentially at $2 or $3 or $6 billion, that's a big miss.

It's like a 10x difference. But not only that, we actually were an investor in Google from pretty shortly after the IPO. Sure. And Google had this stretch post-financial crisis where it kind of traded sideways for a long time. And the reason was, here comes Meta, or Facebook, and they're going to replicate a private internet that Google won't be able to search. And so Google is going to be under pressure. And we weren't able to talk to Meta, so we didn't know what they were thinking. As soon as the company went public, ironically, Google stock started working. Because they didn't come out and tell you, we want to kill Google, or we're replicating the internet. We're doing kind of something different. Both stocks ended up working. So that was kind of a big eye-opener to us. It felt both offensively minded and defensively minded.

How could you tell at that moment? Obviously, it was correct that companies would stay private longer, but I'm sure you were debating the question, is Meta the outlier? Are they the exception that will eventually prove the rule versus there's a structural shift in private equity venture capital that will propel many more companies to stay private well into the tens of billions of dollars in market cap?

We would not have foreseen what ended up happening. We had an instinct that it might happen. And remember that Spotify is kind of getting built at the same time, right? And it's redefining music. And we really, you know, as I mentioned, Apple was a core thesis for us. And now wait, here comes subscription music, which goes directly against Apple's model of selling you an album problem. And so what's going on? They just did a round at $3 billion. Felt like a lot. And then Uber, right? So it was just kind of— you felt— and Airbnb, right? So I would say, like, those two companies, the mobile internet coming out, these companies getting big in private markets, it felt like undeniable momentum, right? And so— and China, by the way, right? The same. So we felt like we had to participate. So what was the first private investment you made? Evernote. Evernote. I think was one of the first. And we said, look, we're going to do later-stage deals, $100 million-plus in revenue. So we did deals like Evernote and Box. And ironically, our most successful deal is one that broke all of the rules that I just laid out, which was Snapchat, where I think We led the Series C in that at a valuation of about $1.5 billion.

Yeah. And what was the reaction from other more traditional venture investors when you guys started leading around?

I think that, look, venture felt very different back then. There was fewer firms. There was more atrophied thinking. Andreessen is just kind of starting. We actually shared a building with them. So we were kind of starting at the same time about as they were. And it felt like, wow, we're going to bring a bit of a different competitive energy to this market. It felt very clubby. You hadn't seen these new firms like Founders, that obviously you know, and a bunch of others kind of really make their mark. So it was sharp-elbowed for sure, and still is in many respects, which is probably what I like the least about that market, because I love talking about ideas. I love trying to be a positive-sum thinker, which the public markets— Easier to do in the public markets.

Correct. Yeah.

Did you bring the models to private markets? Were you building financial models, or the team? We did.

I think we brought that. We brought analytical thinking. We brought kind of— deep research. Like, I remember reverse pitching Aaron Levy at Box a big deck that we had done, and we had just done what we thought was kind of public market-like research, but we brought that to a private entrepreneur, and he hadn't seen that kind of work before. So, that was a differentiator for a minute until other firms realized, "Wait, we can do that, or even better, we can outsource it to Bain," right? And so, we had to to quickly adapt to that. But in the beginning, it was novel, and that was an industry that was really done in Word, and we were Excel thinkers. So that was kind of a very different kind of mindset that we were bringing to the table.

Was there any shift required in the messaging to LPs, the fund structuring, anything that you had to work through in order to actually set up the fund for success in the private markets?

I think our LPs— first of all, LPs are not talked a lot about in venture, which is kind of interesting. Like, we talk a lot about the founder. We talk a lot about the companies. But I would say, for us, we are pretty clear that our customer at the end of the day is our LPs. And so the trust that they give us means a lot. I have virtually no outside investments. Right? I do some as favors and things like that, but almost everything I have and own is in our funds. So, we kind of act as entrepreneurs and as owners ourselves, and we ask for trust from our LPs, and I think at the end of the day, they were willing to give us a chance in our first fund. I don't think they held us specifically to exactly what we said we were going to do, but they're like, these guys are pretty disciplined, and they're pretty smart, and they're entrepreneurial and aggressive, and let's see what they can do. And so I think over the years, what's helped us most in our business, and I think why we're still in business 25 almost years, or 27-plus years later, when a lot of our peers have disappeared over time, is We never lose sight of our investors, and hopefully we've made some good decisions, but we've also made some bad ones. But I think our investors learn more about us on how we deal with our bad decisions, right? And so I think we've earned, hopefully, some trust from them over the years. So, I remember distinctly an investor, when I called about the Snap deal, and saying, "Look, I know this is a bit off-brand." This was one of the largest investors. In the fund, but I just have a lot of conviction in this deal. And he said, then do it. That's what ultimately why we're investing in you. And so I think the trust that you build, the relationships that you build with your own investors over those periods of time are really important.

So relationships with entrepreneurs, building models for Apple ARPU, projecting units, that feels all very micro. How have you processed macro statistics and factors like interest rates? Everyone talks about when interest rates rise, all the DCFs change, there's a pullback in the private markets. We lived through this with like the end of ZIRP. But how much are you tracking the labor market, the GDP numbers, the interest rates, and how much of a factor is that on the strategy day to day, month to month, year to year, or even like broader terms?

Yeah, so data science has really become a much larger part of our business than it was back then. So we now have, I don't know, maybe 20 or 20-ish, 20 to 30 people, something like that, in data science that are just processing different types of data and alternative data. So think it's not just macro data, it's App Store data, it's clickstream stream data, it's credit card data. So we use a lot of that for our investment research, right? And some of that we even make accessible to our portfolio companies, right? So that made us smart about a trend. We were early customers of Databricks and Snowflake, as an example, that let us invest in those companies. So that is way more of a presence in today's world than it was 20 years ago. So we're constantly looking at data. As an example, I think OpenAI is probably the most important company in the world today in the sense that it's the driver of AI, both consumption and spending. So I look almost every day at the chart of ChatGPT users, download share, how it's weathering the storm versus other competitors. That's really something that wasn't available 10, 15 years ago. Sure. We had this amazing data set called Onavo, which actually gave you engagement data from, from users on phone. And then Zuck bought it and turned it off. And I remember thinking, like, damn, that guy, that's a great move. It was the only data set that really gave you engagement data. Yeah. So we're always looking for kind of new data sets. Right? And then obviously, so that felt like a major shift, right? Going from data science-enabled research, and now obviously, we're getting to AI and agentic research and—

How are you thinking about AI as a category from an investor perspective versus the Databricks and Snowflakes, which to me feel it's easier for me to maybe understand the financials, the model that I would build, how I think about value accrual and competition in Databricks and Snowflake, fantastic businesses, but feels like easier to pattern match against previous eras of software and tech innovation versus AI where you have infrastructure and CapEx and training costs and inference budgets and all sorts of different, your entire product's getting copied by open source every 3 months, and it just feels like a different puzzle to solve when you're thinking about underwriting those businesses. Like, how have you grappled with that? Do you see it as an extension of the tech investing, or is it an entirely new motion?

Well, for me, it was almost coming back home to what I knew, because the infrastructure layer was really semiconductor-driven. Sure. Right, so I think our knowledge of, semis and our team's knowledge of semis was a great head start because a lot of people just hadn't done semis. So we're kind of new to semis, and I had a lot of relationships in the industry, and that led us to lead the series being Cerebras. Yeah. Um, that I think will be kind of a generational kind of company.

Yeah. I love Andrew. He's been on multiple times. He's great.

Um, so that felt very natural to us and pretty quickly when we saw what Jensen was building and the momentum that NVIDIA was building in the data center. So that was kind of our first telltale that, wow, something big is going on here. So we had seen semis before in the mobile era. So we felt very equipped when AI first came around to look at it from a semiconductor, GPUs, memory, TSMC. I've personally been to Taiwan many times, visited TSMC. Great anecdote. I'm driving back to Taipei City with my host from TSMC, and we're on the highway, and there's a golf course, and it's nighttime, so there's lights and people playing, and I just very innocently turned to him and said, "Wow, you guys play golf at night here?" And he very innocently looked back at me and said, "Well, when do you play?" And that's when I realized, like, that's— we're in a different level of— of work here. So we saw it at the infrastructure layer first, right? Where it just became obvious, you didn't have to necessarily worry about who was gonna win, like the whole infrastructure layer will win. So I think that was kind of layer 1. I think layer 2 then came kind of the models, and obviously we're investors in a number of them. I'd say the most complex element of AI today is you can almost talk yourself into a bull case and a bear case. For almost any name in tech. And I think software is kind of seeing that right now, right? So is software going to win because of AI or get displaced, right? So you've got that. In infrastructure, you've got the, well, when is the peak? And what multiple peak should things kind of trade at, right? So it's both an exhilarating dynamic, but also very complicated in environment in the sense that, for example, when Databricks came around, no one thought, well, gee, Databricks is going to put Salesforce out of business. Yeah. Right? It just felt like a new architecture.

Yep. Yeah.

There's something about AI that feels a lot more disruptive. And what if your model, not today, but in 2 years, can just build you a Workday right off the bat? Yeah. What does that mean for Workday? Does that mean their data is more valuable on the left-hand side, or no, does that mean they get fully wiped out? So I think that battle, and it's being played out kind of in the public market today, right? You kind of see it in these names, is—

How can a public company CEO actually communicate a vision for that case, the bull and bear case? Case around AI, what effect AI will have on their company?

I mean, you're seeing it. I read something that we're kind of moving into a selection market, right? So now, like, some companies are gonna do well, some companies are not. I mean, look at Square, right? Jack came out and just said, no, I'm pivoting the entire infrastructure of this company for an AI era, we think you're going to need to be remote, right? Because you're going to move faster if you're remote. And so he's kind of laid out a whole vision about how he wants to run the company.

And that's because you're sort of generating the necessary context because you're not getting the in-person interaction. And small teams. Yeah, small teams. You'll more easily be able to be AI-native if you're basically explaining process.

Yeah, very small teams moving really quickly without a central organization, or, you know, kind of like the Borg, right? No central organizing force, the model drives everything. And then you have other companies that are saying no product— well, the labs themselves, right, are all in person, and they believe, right, that product development needs to be done kind of in person. So you're kind of seeing a lot of these different ways of— you have different models. Some people are going to charge for tokens, some people are going to charge for data access, or ingress and egress. So I think what we're seeing right now play out is a true Darwin-like survival of the fittest, where software companies are like the mailroom. Exactly, right? Um, you saw Aneel come back to Workday, right? He probably thought that— I think Carl is an amazing executive and, um, is a good friend of mine, but maybe he thought in order to make the changes, I need to have that kind of founder mindset, right? Founder mode, as Brian calls it, right? So we're seeing now a lot of these different approaches kind of compete with each other. I think it's too early to tell who's gonna win, but eventually, I think we should see kind of separation between winners and losers, right? Which should be good for our business. But I think right now it's still so early that it's not clear who will win.

Is it enough to look at reaccelerating top lines? I imagine we've talked to a lot of founder CEOs, maybe they're unicorn status decade in, completely reinvigorated by AI. They come on, they see that the growth has returned. It feels like a new startup even though they're maybe coming back from a sabbatical. Maybe they're coming back maybe they're just coming back in with a new vigor, but what are you seeing more at the earlier stage or mid-market stage around companies that are starting to show signs of being winners in the AI age?

I think, look, the good news is no one is head in the sand about this. So I do think in prior cycles, you had more of a head in the sand. Head-in-the-sand mentality, right? So, for example, if I remember when cloud got started, there was— do you remember virtual cloud, right? That was gonna be the big thing, right? I can't just have my stuff in Amazon. I'm gonna have this virtual cloud. And, you know, obviously that hybrid cloud, right? That's another one. No, it's gonna be half— all the big data, right? All those things basically just got torched and went by the wayside, right? So there was a lot of head-in-the-sand. Similarly with the iPhone. You need a keyboard. It doesn't have 3G. It doesn't support Adobe. It doesn't have copy and paste. Exactly. The battery life, right? Uh, BenGate, you may remember that one. That was a whole weekend, whole weekend wasted on BenGate. Um, so there was a lot more to me head in the sand in prior, uh, investments, uh, cycles, right? In prior tech themes where people were just pushing back against the idea that this was gonna work. I think AI is one where the consensus view is, it is gonna work. It's an extinction-level event, and so the sense of urgency is high. So that does feel a little bit different to me than maybe prior cycles where I think that took time for people, right? The carriers are like, well, I'm not gonna allow Apple to have an App Store, and I don't wanna be a dump, pipe, and, right? So they— all these things that they fought. Yeah. And over time, tech won. Yeah. The difference to me with this specific cycle is everybody agrees it's going to happen, and it's happening quicker, and the stakes are higher than any other. Yeah. So I think every board is ultra motivated. Every founder is focused on this. Now, they're bringing different approaches. and we'll see which ones win out, but I would say they're tracking token consumption, right? So how much of my revenue is token-based, right? How much of my COGS is token-based? How much of my G&A is token-based? How much of my spend per developer on Cursor, OpenAI, and Anthropic is happening, right? So that in, I do sit on a bunch of boards as an advisor observer most, most of the time, so I kind of see a lot of that kind of happening. So the awareness is absolutely there. People are doing different approaches. There's some that are, no, I'm gonna whitebox a totally new product, right, where I think I'm uniquely positioned to build it. And then there's others who are like, no, my data set is so valuable, I'm not gonna allow my customers to build apps directly using my data. So you're seeing a lot of kind of different approaches. Yeah. Right? But everyone's awareness is at a 12 out of 10. So there's no convincing needed, right? Everybody's aligned, every board member's aligned, every investor, and now it's about, okay, what does that sense of urgency mean for this company? What are the things that we need to track and the things that we really need to go and execute on?

What does it take to make it as a new hire at Coatue?

So we do— perform is number one, like seeking big themes and big ideas. That's a big one. And then the kind of the risk management piece, right? That's kind of kept us in business for a long period of time.

How is AI changing the role of early analysts or career or analysts who are earlier in their career on the investment side?

Too early to tell, I think. Too early to tell. Obviously, look, we use AI every single day. I use it a lot to test my thinking, to clarify my thinking. I've always had a weird dichotomy personally where I love reading, but I'm a terrible writer. And one of the things I like about AI and ChatGPT specifically is it's helped me actually write in a way that I can be proud of. Sometimes I write something, an email to somebody, I'm like, this is 'cause it's just so badly written, you know? And it's just, I don't know how to make it better. I know it's not good. I don't know how to improve it. And it's so frustrating 'cause I know what great writing is from my reading, but I just can't do it. And at least now—

Yeah, I've had a family member send me something that was very obviously, obvious to me, AI-generated. And I think people have an aversion to AI-generated text.

I totally don't get that.

But the thing is, like, I was reading through it and I was like, this is very cool because I know this person would not have been able to articulate their thoughts in this way, but they went line by line and I know they mean it, right? And so they were able to communicate something that they never would have been able to communicate with text. Maybe if we sat down and spent, you know, a couple hours talking through it, I would have been able to get the gist. Exactly.

But some people take out the em dash because they write it. I don't want it. I'm like, what do I care? Yeah, like, I don't leave it in. Yes, a lot of emails that I write are helped by ChatGPT.

The Arnold Schwarzenegger line, you know, he's, he's, I smoke my stogies. Oh yeah, everywhere. Why? The em dash.

Judge me on what I've said. Yeah. And my idea and whether it's well written. Yeah. Who cares whether it was written by AI or not? That I totally don't get that.

Yeah, or polished by AI.

In fact, I hope that more people are able to communicate things that maybe they couldn't before, right? Because they didn't know how, or they only knew granular things, you know? And now more people can write, more people can communicate, more people can express themselves. Like, to me, that's an incredibly empowering right vision.

Talk about how you guys have approached investing in multiple companies in the same category, or the same general category. How has that evolved over time? Were you mocked early for doing that from maybe some of the more traditional funds? And then how have you managed to make it work in practice? Maintain the trust of entrepreneurs and management teams.

I think conflicts, which is kind of what you're— has definitely changed a lot in the Valley as companies have stayed private longer, right? And I think we have to be kind of precise by what we mean by conflict, right? So as an example, funding two Series A companies at the same time that are pursuing the same opportunity is an obvious conflict that I think no firm including us, would ever do. Yeah. Right? So let's just kind of be very clear on that. I think it's quite different when now you're talking about these very late-stage companies, right, that are kind of competing with each other. But look, every company is kind of competing, cooperating. You know, Apple and Google are great examples. You know, they compete, but they're partners. So I think the distinction and the conflicts distinction has to be— kind of changes as companies and markets kind of mature, right? So I think you're seeing that become much less of an issue in mid to later stages, even by firms that would typically view conflict as core to what they do, like traditional venture firms, right? Have not kind of moved in that direction. And I think it's just the nature of the market. So that would be my first point. I think the second point is the execution of it, I think, really matters. So, you know, if I view a perceived conflict between companies, even if they're later stage, I will always let the founders know directly. And I'm not asking for their permission, so I think you also have to be clear with the founder, because what if they say no and you still want to do it? Then you're in trouble. Now you've just— broken your word, and I won't do that. But I will inform them. I'll be very direct. I won't let them hear about it from somebody else or something like that, and I'll explain kind of the rationale, right? So I think communicating directly, both good news and bad news, that is something that I learned as an agent. I'm not afraid to have difficult conversations because I think we can grow from them, and I ask the same of founders or employees that I work with to both come to me and say, if you have an issue, let's just kind of talk about it. And look, I've hired a lot of people. I've fired a lot of people over the years. I've asked a lot of people to go and look for different career paths. So I'm comfortable having those conversations. So that's not something that— I think your reputation and trust, then, is kind of the second point on the X-axis. Execution, right? We take information security incredibly strongly. That's when, you know, we're SEC registered, and, you know, even before coming here, I got like a 4-page memo from my lawyer about how you can talk about this but not that, and SEC this. And so, of course, you know, I have to read it.

You can tell the Ralph Lauren story.

Yes, yes, I hope so. I mean, specifically, by the way, I— the way it— well, you know, so So side note, I believe that meetings should be recorded, as an example. Now, my compliance will say, shit, we can't have meetings be recorded because it creates a paper trail.

Insane discovery.

And let's just not even talk CodeTwo specifically. Let's just talk an enterprise. Yeah, yeah. But then I say, OK, let me posit enterprise CIO says, no, I can't have my meetings recorded. I'm too afraid. I'm like, OK, let me posit two scenarios to you. To you, OK? Scenario 1 is— and in each, you have a bad actor that's doing bad things, right? Whatever that is, like belligerent, talking down to people, whatever. So scenario 1, nothing gets recorded. You know nothing. And 10 years later, someone comes out of the woodwork and says, by the way, X, Y, and Z, 10-year pattern of deception, nothing happened. OK, so that's scenario 1, right? Scenario 2— is every meeting is recorded. The first time said person does something that's not right, the compliance system, right, which is always listening, sends that person an email and says, "Hey, by the way, better you didn't do this, talk to that person that way, disclose this piece of information," depending on the severity, right? Second time person does it again, says, "Hey, I now have to flag this to HR." I would much rather live in world number 2, right? Because you know what the problem is. There's a system. There's flags that have been raised. And eventually, someone kind of gets involved, and you either remediate, or you terminate the person, or whatever, right? So to me—

You avoid the back and forth of he said, she said, all of that.

Exactly. And who knows? Maybe that person, if they had gotten that first warning, might have realized, oh wait, yeah, you're right. I'm being abusive, or whatever the case may be of whatever they were violating, right? So I think, to me, that's a better world, right, than kind of the ignorance of while getting no feedback, and then you just learn much later that kind of you had a problem.

Yeah, I mean, it certainly seems like a trend. I mean, Bridgewater's written about it. —a lot, Ray Dalio, but then also—

And he did that with no analytics, right? So that's different, but yeah, correct. I think to me what will happen is the analytics are going to get so much better. Interesting, right? And these systems are going to know, and they're going to be able to look at your WhatsApp and your messages and your emails and all of your calls, and they're going to be able to just say, hey, by the way, just don't say this, or did you think about this, or maybe you could have they're going to coach you. They're going to say, hey, you told this customer to fuck off. Be like, well, maybe you shouldn't do that. Here's like 2 other ways you might have mentioned, you know, like whatever this is. Yeah, your frustration with the situation, right? Um, yeah, no, very interesting.

Instead of the token maxing dashboard, you're going to have the social credit score. Yeah, it's the bottom 20%. No, no, there's a lot of environments that it can make.

Yeah, maybe there's a divide on like the in-person versus remote work crowd.

Well, let me be clear, I'm talking about work context. Yeah, yeah, I'm not talking about after work.

Yeah, yeah, I just wonder if that would become something that employees select into or out of for various reasons, just like some people are huge fans of remote work, some people can't stand it, and there's a variety of—

I also think there's a clear distinction between transcription and recording. OK. Right? They don't necessarily go hand in hand to me. OK. Right? So I don't necessarily need a system that transcribes every word that was said and keeps it in some database somewhere. Yeah. I'm not sure that's necessary. But key takeaways from the meeting, what was said, what was agreed upon, like, that's useful. That's a good corpus of data. To me, just because someone's, quote, recording doesn't necessarily mean that it's transcribed. Yeah, it could be compressed already. I like the option of deleting. In an ideal world, I would recommend, well, you can delete the transcript. The transcript's not that relevant because maybe you batted an idea back and forth 3 times and you said something that turned out not to be true, but you figured that out later. So you don't need any of that. What you do need is what was said, what was agreed upon, was anything done out of compliance or not, right? So it doesn't mean, it doesn't imply a world where everything you say is recorded.

No. Right. No, it's funny because there's so many opinions on this, but we record everything all day and livestream it on the internet. Right. So, it's—

What, uh, how have you processed a number of these venture funds that have become publicly listed, that are taking a lot of the different names that CodeTwo is in, you know, trying to find basically the most in-demand secondaries, putting them into funds, As I've watched, I think what I've seen is like, yes, it's very obvious there's an incredible amount of demand from the public to invest in these names. But the big issue is that as soon as that demand floods in, you know, supply and demand, price shoots up, and then you have a bunch of people investing at effectively, you know, 10 times what the private valuation or the underlying asset value. But how have you processed Do you think there's a more elegant solution over time?

So we do have a fund called CTEK that I'm allowed to mention that addresses some of this, and people can kind of go online and research more about it. What I would say generally is people want access to these companies, and I think there's a lot of arguments for going public. One of, I think, the most powerful most powerful, in my opinion, is to democratize access to companies. Let's take an Anthropic, let's take an OpenAI, right? And enabling the retail investor, enabling the Trump accounts, which I think is a marvelous idea that my friend Brad Gerstner really spearheaded. And I give him a lot of credit for this idea of, wow, why can't we have every single new child already have— be invested in the market and participate in the value creation of these companies. So I love that democratized nature of it. So what I think it speaks to is there is incredible demand, right? Let's say you were sitting— you're not a VC investor, you're maybe a dentist, and you're seeing OpenAI and Anthropic, and you're like, wow, why am I not able to participate in that? You know, why is it just like an elite group of funds and accredited investors and so forth and so on? That to me will have to change. And I think it should be bipartisan, frankly, right? And I think there'll need to be guidelines and stuff put into place, and we don't want bad behavior and all that kind of stuff. But I think there is incredible demand from the retail investor base to participate in the value creation. And I think what we're learning as a society is the cost of not having broad participation is incredibly high. Completely agree. Right? And will be. Yeah. Right? If you have a whole generation of young people that don't own their house and have student debt and don't feel like they're economically levered to OpenAI or Anthropic, or even more so are directly threatened by their technologies, I don't think that's a great future for any of us. Totally. So I'm not saying that OpenAI or Anthropic going public is the solution. but it's like the one thing, right? Obviously, there's going to need to be a lot of things that are done. But I do think the transparency that comes with it, the democratized nature of it, will make a huge difference.

Yeah, makes a ton of sense. Jordy, anything else? Last question.

What is— what are— give us a scenario.

Can I make also one point on in-person? I'm so glad I got to do this in-person. I did not want to do this remote because is the tactile feedback you pick up in person. As an example, for the viewers at home who have not been to this office, the open jar of creatine, just what— I mean, what a pull. Open jar. Just for everybody to describe it, there's a goody bag. And there's just a jar of creatinine. It's just wide open. The scoop's right in there.

It's so inviting.

If you want to power up before you get on, it's just right there.

It helps if you're sleep deprived.

Right. That's just the kind of tactile feedback you want.

Yeah, you don't get that on Zoom. On Zoom, no way.

I'm really glad you could be here.

I'll save the last question for your next appearance. Okay. Let's do it again soon. Well, thank you so much. We're going way over.

Yeah, we went way over.

Delighted to be partners in OpenAI now with you guys. I'm really proud of your success. Honestly, I love hustle and people that break into industries. And so congratulations.

Yeah, we bring a very mailroom approach to podcasting. We do.

Yeah, I didn't think about this. There's a lot of mailroom here, a lot of suits here too. Well, thank you for watching. Tune in tomorrow, 11:00 AM Pacific. Leave us 5 stars on Apple Podcasts and Spotify. Sign up for our newsletter at tbpn.com, and we will see you tomorrow. Cheers, man. Goodbye. See ya. We love you. Goodbye.