AI Side Quests, Zaslav's Payday, SF Housing Market is Back | Shyam Sankar, Gili Raanan, Anna Patterson, Jake Loosararian, carried_no_interest

You're watching TVPN.

Today is Tuesday, March 17th. It's St. Patrick's Day. We are live from the TVPN UltraDome, the temple of technology.

The fortress of finance.

Let me tell you about ramp.com. Time is money, save both. Easy to use corporate cards, bill pay, accounting, and a whole lot more all in one place. Let me also tell you about Shopify because that's why I have this green suit. Shopify is the commerce platform that grows your business, lets you sell in seconds online, in store, on mobile, on social, on marketplaces.

Happy—

St. Patrick's Day. We are not drinking Guinness currently, but we hope that you are.

Yeah, enjoy it. Very exciting. I haven't seen a lot of St. Patrick's Day posts on the timeline. What's going on? We got to step it up. I feel like we should have planned much, much more. But it seems like Tyler is in the St. Patrick's Day mood and is wearing a fantastic St. Patrick's Day hat.

Let me take off my hat.

How did you wind up here? That's a nice hat. Where did that come from? Did we just have that?

No, I had this.

Oh, you just had it?

You had it handy.

You just daily drive that? Yeah. Okay.

Yeah.

It wasn't in your car?

Yeah, you wore it to the gym this morning.

Okay.

Oh yeah.

Cool.

Yeah. Apparently got to wear green. Uh, George is going with a little bit of a darker green today, but, uh, we will still be having fun. Uh, let's, uh, let's pull up the linear lineup, show you who we have on the show today. Carry No Interest, Anonymous poster is coming on the show at 11:45. Then Shams Sankar from Palantir Technologies is coming for a massive book launch. We're very excited about that. And then we have a lightning round going through cybersecurity, Ceramic AI. We have Gecko Robotics on the show. I don't know how familiar you are with Gecko Robotics, but it's a very, very interesting company where they crawl up oil and gas infrastructure, like literally like geckos, and will like inspect everything. There's a lot more to the business.

So like humans in gecko outfits. Crawling around?

No, robots.

Robots.

Yeah, they've been doing robots for a long time.

Okay.

Well, Linear, of course, is the system for modern software development. 70% of enterprise workspaces on Linear are using agents. So in the news, the Wall Street Journal has an article about— this is a scoop from Berber Jinn. We didn't have to break down the paywall because we get a copy of the Wall Street Journal every day. Although I don't know if this particular piece made it into the print edition today. It might be in the print edition tomorrow, but Berber Jinn has a scoop. He says, OpenAI's Fiji Simo told staff last week that the company could not afford to be distracted by side quests. Main quest only. Main quest only. What is the main quest? Probably just scaling compute, but we'll get into that. So the— and there's a whole bunch of takes back and forth, said the company execs are actively looking at areas to deprioritize. Of course, throughout last year, OpenAI launched a ton of different initiatives across consumer hardware. And I think a lot of people are starting to say like, okay, like, do you want to be fighting a battle on all these different fronts or do you want to just really nail consumer, nail enterprise? And now that is what they are signaling internally. So from the Wall Street Journal article, OpenAI's top executives are finalizing plans for a major strategy shift to refocus the company around coding and business users, recognizing that a do-everything-all-at-once strategy has put them on the on the defensive. Fiji Simo, OpenAI CEO of Applications, previewed the changes to employees at an all-hands meeting, telling them that top leaders, including Sam Altman and Mark Chen, were actively looking at which areas to deprioritize. They expect to notify staff about the changes in the coming weeks. We cannot miss this moment because we are distracted by side quests. Side quests is the key term there. We really have to nail productivity in general, and particularly productivity on the business front. I was listening to a podcast with the head of ChatGPT last night and there were a bunch of things. Nick Turley. Yeah, Nick Turley just talking about like it's such a weird product surface area because people come to it from all these different reasons. So how you prioritize things within there, like images in ChatGPT, clearly very important. That was a separate app at one point. It got rolled in to good success. Nanobanana was also sort of like a, okay, everyone should download this particular app because you're going to get the best. Frontier image model for a while. Sora.

Yeah, we had talked about this a while back and it seems to be confirmed now that Sora's functionality will just end up in the main ChatGPT app.

Yeah, yeah. So last year OpenAI announced a new, an array of new products including the video generator Sora, a web browser called Atlas, a new hardware device, and e-commerce features for ChatGPT. Some of those are like wildly different timelines for these things.

Yeah.

And remember during this time, that's when I started talking about viewing OpenAI's activities like they were a hyperscaler.

Yeah.

When Google launches a new product.

Yeah.

Not— you don't have to assume that, hey, it's going to work every single time. And in fact, it's the exact opposite. A lot of these experiments don't end up going anywhere. It's fine. And OpenAI, like, again, you have this like massively scaling, you know, core business. You start to say like, hey, let's experiment in a bunch of these different areas. Some of them are going to work, some aren't. It's okay. But then ultimately that creates a scenario where you're sort of like opening up maybe too many fronts, right, to the competition, right? You're competing with Meta in social, with Sora, you're competing with competing with Google in some ways. You're competing with even like the Microsofts of the world in other ways. And ultimately this just feels like, hey, let's like narrow the fronts. And I think a lot of people expected something like this for the last few months.

OTP in the chat is surfacing a very interesting old take from Ben Thompson about OpenAI should not be in the API business. And now based on the way things look, Originally that was like, oh wait, well maybe Microsoft will be able to handle and scale the API demand. But API demand has been so huge and it's allowed Anthropic to develop such a solid business there that it's undeniable that OpenAI should also be in the API business, especially—

Yeah, the other side of that is like they clearly need to be in the harness business too, which is why Codex is accelerating.

Yeah, yeah, yeah. And so the hyperscaler, Like your take of like, think about OpenAI as the new hyperscaler. I think that takes age extremely well, especially with recent interviews with Dylan Patel and Dwarkesh. Jensen just went on Stratechery. There's a whole bunch of new data points that show how compute constrained we are, how tricky it will be, how chips will be the key bottleneck. There's that interesting story that Tyler was recounting from, Dylan Patel about how the TPU was developed by the TPU team. DeepMind didn't realize how important it was gonna be, how compute constrained they were gonna be. So the TPU went and sold it to Anthropic and then DeepMind went back to them and was like, wait, wait, wait, we want those chips. Like we need the chips. And so there's this like the hyperscaler phrase is really, really important in that it designates not just a big consumer company or a big business, it's particularly about the ability to marshal compute at hyperscale, right? And so I was interested to look at like the history of side quests among Mag 7 companies, among trillion-dollar tech companies, because it's all over the place. And so it's very hard to paint with a broad brush. I think right now there's a huge narrative around like OpenAI was doing way too many side quests, they need to do zero side quests. And the reality is probably like, you know, we've seen this with Riley Walls joining the Labs team, like there will be small projects, there will be acquisitions, there will be, a continuum of bets. There's just a level of refocusing that's going on right now.

Yeah.

And reorganization. And this has happened at many, many—

Yeah, and I think the bet with the OpenAI Labs team is you have a really small group of people that are focused on creating or being quick to new product approaches.

Yeah.

But that can be, again, like a two-pizza team.

Yep.

And that type of experimentation is gonna make more sense if the rest of the company needs to, refocus on enterprise and the core business.

Yeah, Tyler, is that data point you shared earlier like public about the size of that particular team? I didn't want to mention it if it was—

The storage team?

Yeah. I think people know it.

Okay, people know it's storage team. At least at one point it was very small.

It was like 6 people or something.

Also, I was going to say, I think now that you have so many Neo Labs doing these kind of like weird, like kind of moonshot research projects, I think it also adds to this thing where like maybe you don't actually need people internally doing those same things, right? You can just acquire them if they work.

Yep. Yeah. Did you— we got to get into that conspiracy theory later, but there's this idea of like all the NeoLabs teaming up to marshal— we got to dig into that. Anyway, history of side quests in tech. Can we paint with a broad brush here? You know, spoiler alert, I think the answer is no. But there are some fascinating stories. So Google has a balloon internet Play. That's a serious company called Project Loon where they inflate balloons that go into the stratosphere, I believe. They go high up and then they deliver internet, sort of competitive Starlink. Some promising stuff there. They also have—

They have Fiber Play too.

Fiber Play. Oh yeah, they just sold that.

Yeah.

People are so upset about that because dealing with your ISP is one of the roughest things ever. I've had such a bad time with various ISPs. Oftentimes I would, running a small business, running a startup, I would put it on my personal account. And then I would have like 6 lines for like different people. And then one of them wouldn't get paid, and I'd get sent to collections for like—

That happened to me too. That happened to you? The only time I've ever had like some credit-related issue was I returned the router. And then even though I still had an active account, just another line with them, they were like, you didn't return it. We billed you.

Yeah.

And somehow—

Wait, Nick, are we still paying internet at the Jonathan Club or did we fix that? I sent you this. We canceled the plan. You did? But we haven't returned the equipment. Returned the equipment.

OK. We're definitely—

We're definitely getting fleeced for sure. But that is the status quo. And it was awesome when Google was just handling it with Google-level service.

But hard to sell ads against.

Hard to sell ads again, so they divested, I guess. But that's not even close to the weirdest thing. The weirdest Google side project that I found is they make a contact lens that will tell you how drunk you are. I'm not kidding about this. Google, they have a project. It does more than that. It's supposed to do biometrics, and it can do a lot of different things. But one of the things that it can measure— it can measure glucose in your tears. It can measure any sort of, like, biomarker that's in your tears, which apparently is like a rich source of data. But one of them is blood alcohol level. So you can put in this contact lens from Google and it will tell you how drunk you are.

Jarvis.

Am I drunk?

Am I good?

Am I good? Am I good to drive?

You're absolutely not.

He's like, "Sir, you should take a Waymo tonight. Would you like me to connect you?" See, this is all part of the plan. They're gonna stop people from drunk driving, give them Waymos, it works. So that was a weird one. They also briefly owned Boston Dynamics. But most importantly, they bought DeepMind, which was completely seen as like this wildcard side project. How does that fit in? You know, it's a bunch of researchers. And then it became like the most critical thing.

Amazon has taken tons of shots at drone-related delivery and home security stuff. They also own Twitch, but they never really linked the site to live shopping and closed that loop.

Have you thought that that's weird?

And Jassy doesn't do earnings calls on Twitch.

He doesn't do earnings calls on Twitch.

She should.

And a lot of the looks-maxing livestreamers have decamped to a different platform.

Let me get some, what do you, Ws? Let me get some Ws in the chat for this quarter.

For this CapEx guide. Let me get some Ws in the chat for the CapEx guide.

Of course, Apple tried to build a car, then pulled out.

Who knows where the Apple Vision Pro goes? Although of course I hope it continues. I'm hoping for another Apple Vision Pro, Apple Vision Air, just make it lighter, same screen. That's the trick. Make it like $1,000, I think it'll sell. And they have a bunch of health moonshots. They've been working on non-invasive glucose monitoring, which is— there's some book about it called like the white whale of biotech or something. It's something that people have been working on for so long and they've never been able to crack it. They've worked on this. The best thing I think they have is that you can get a continuous glucomonitor. Thank you for the W's in the chat, by the way. There's a lot of W's in the chat. You can get a continuous glucose monitor that is invasive, meaning it is pricking you. It is measuring your blood and then calculating calculating the amount of glucose in that. That's helpful for continuous glucose monitoring. There's a number of companies that do that in sort of a D2C realm levels. There's a number of biotech companies that offer that as like a healthcare product. And you can wirelessly connect that now to your Apple Watch, but they can't do it in the Apple Watch. There's always been the question about like, could you just shine a light through your skin and detect the blood, the glucose in the blood? Very, very difficult to do. Potentially always possible. Everyone thought, oh, it doesn't break the laws of physics. Lots of money spent with little to show for it, at least so far. Meta is probably the most egregious side quest.

They don't do side quests, they do full quests.

They create side quests in VR. There's actually one of the popular games, it's all about doing side quests. And yeah, I mean, they bought Oculus, they bought a bunch of VR studios, they rolled everything up, they renamed the company, spent tens of billions of dollars on What is starting to feel like less of a side quest, but still is so early. You know, the Meta Ray-Bans are like a success, but to the tune of millions of units, not billions of units or anything like that, or hundreds of millions of units. I think the iPhone has an install base over a billion now, and the ramp from here to there on Meta Ray-Bans is going to be long. Tesla launched a premium tequila in a lightning bolt-shaped bottle. So, you know, it comes from all different Mag7 companies. That one's more of a stunt. But truly, like, side quests come in all shapes and sizes. Some are just good for morale. Like, it's just fun. We do side quests here all day long. TVPN Simulator, that's a side quest. What was the other simulator? Jeremy Giffen Simulator. That was a side quest. You know, they're just fun. Some of them are good for marketing, good for attention, good for fun. Some are complete dumpster fires where you just pour money in and it just sucks all resources and you get nothing out. And then some reshape businesses entirely. DeepMind's a great example. And there's certainly others in the Mag 7 that have really, really changed the business.

And so— Yeah, Gabe said, seems like the entire Boring Company is a side quest.

Yeah, it's technically a separate company. But Elon's, yeah, king of side quests.

Although I think a lot of companies— Yeah, I mean, it feels like the company that gets the least amount of attention that's not on the critical path for many of the other projects.

Totally.

Totally.

Yeah, still a very cool idea. I mean, every time you're sitting in traffic, it's so tangible to know, oh, if we just had more space.

Creating a yearning for the mines, basically.

Yes, yes, yes. But very, very difficult. And I mean, it's been, what, over a decade? And there's really a very limited rollout of that technology. But it's still cooking. I think the business is still going and they're working on it. Some of OpenAI's teams come from very small teams with relatively tiny compute budgets, but they get a lot of attention sometimes because of the particular product category. So Sora is a great example where small team, not huge resource investment, probably a lot of inference and training cost, but relative to Codex 5.4, Sparrow, I don't know. I don't actually know like how we're looking on the order of magnitude there.

Yeah, I think a big part of it is like—

But super viral, right?

Yeah.

And what percentage of the team, the overall team's energy is going towards a specific product, right?

Yeah, yeah, yeah. And then, but so the idea of experimenting quickly and then consolidating efforts even faster makes a ton of sense. Nano Banana was a big deal for Gemini, bringing image, video, and audio generation together in a single ChatGPT flow is clearly the next step. And so some of this is not, it's probably not going to take the form of like, stop doing the side quests. It's like, let's fold that into the main quest because clearly the interaction pattern of ChatGPT is where people want to go. And so once we've done the experiment, it works, put it all together. Dylan Patel on DoorCash said the TAM for GPT-5.4 was north of $100 billion, which is crazy for a bunch of reasons. I mean, it's just a huge market that was created. In just a few years. The enterprise opportunity in general is crystal clear for anyone involved. At the same time, you still do need to experiment to make sure you are early to create the next breakout product experience. And so with that backdrop, OpenAI Labs and being more efficient with the shots on goal makes a lot of sense. I still think that the overall narrative around AI, just in AI broadly this year, will be about the main quest, which I see as like compute scaling. And so raise the money, do the deals, grow the capacity. I was thinking back to Travis on the show saying, if you're doing something and it's easy, it's not valuable. The key is if money matters, which I think we say, we would say it does, especially in certain categories. He was probably thinking of ridesharing, where there was a capital war, and AI compute, where there's also a capital war. You need to be the best in the world at it.

Yeah, and I think overall, there's real competition now. It's intense, right? You have OpenAI, Anthropic at the frontier pushing very, very hard. But the other side of this for OpenAI is like all the, yes, there was a ton of stuff that was announced last year. There was hardware, Sora, et cetera. But Sam was running around doing all these different mega deals for the compute side that now serve the main quest. And so the positioning is actually great.

Yeah, yeah. Let me tell you about Labelbox. RL environments, voice, robotics, evals, and expert human data. Labelbox is the data factory factory behind the world's leading AI teams. And let me also tell you about Gemini 3.1 Pro with a more capable baseline. It's great for super complex tasks like visualizing difficult concepts, synthesizing data into a single view, or bringing creative projects to life. And speaking of Google, uh, Boogle Capital Bloke, the longtime Google bull, uh, says, so Anthropic and OpenAI are going to just give this consumer market to Google? And I don't know how, how I mean, yes, they're like OpenAI reportedly planning to shift the strategy to refocus around business users and vibe coders. There's a lot of stuff going on there. Listening to some of the data around the retention curve of the various products, like I don't think that they're giving up on consumer at all. That seems like sort of an odd read. It will be interesting to see the next iteration of Google's consumer surface area because they have AI search overviews, AI mode, the Gemini app. It's in Gmail, but there's clearly some like UI fighting going on. I just— it's so funny when you're in Chrome and you have the ability to open one Ask Gemini panel in Chrome and then you can open a second Ask Gemini panel in Gmail, the app. And so both at the web layer and the browser layer, you have two, you have two chat boxes that look exactly the same and do the same thing. This feels like very much like a V1 of what they will do here. So there's certainly an opportunity for Google to reintroduce the AI features, make them more tightly aligned with what people are actually doing. And you have to imagine that a lot of the progress they're making on agent decoding can then come to the Gemini surface area in consumer overall.

Yeah, and we can pull up this chart that Sam showed of Codex usage. Again, this is like what I think is informing the battle with Anthropic as well as like this chart is gonna inform the strategy shift, which is like, hey, we can run, we can take revenue into the hundreds of billions of dollars with the current products that we have. Let's focus on them and make the best possible products.

Dean Ball is using Codex, is that what he's saying? Just hit a personal record. For single coding agent session of a little under 10 hours. GPT-5.4-X-High in the Codex app. Unsurprisingly, no flashy app, just really complicated economic research prompt. At this point, most of my prompts do not stress these agents all that much. To be clear, this is 10 hours of continuous work. I have meaningfully exceeded 10 hours if we include periods when the agent was waiting for jobs to complete. So what does this mean? Is he talking about like firing off one prompt and coming back 10 hours later or just or just going back and forth? Because he says, "Agent session." I think he's saying—

I think he's talking about firing off one 10-hour—

Yeah, what do you—

I thought he meant he's in Codex, like the terminal, and he's in there for 10 hours.

Locked in. That's amazing.

Oh, wow. Yeah.

Yeah, it is very interesting to see. I mean, Doug O'Loughlin was talking about how he moved a lot of his research. He wasn't building software.

No, he's saying it's not a flashy app that he's building, just a really complicated economic research prompt. To be clear, this is 10 hours of continuous work. I've meaningfully exceeded 10.

So, so I think what he's doing is he's, is he's asking like, okay, pull together all the census data about jobs. Now go and pull together all the economic indicators. Pull out, pull together all the inflation.

Somebody asked how much time was back, how much was back and forth was going on versus the agent spending time on its own. He said zero back and forth. So this is 10 hours of the agent just like Cooking.

Wow, that's really crazy. What is this prompt? It must have been pretty long, I imagine. That's crazy. It fit within his current rate limits on the $200 OpenAI tier in the Codex app, where rate limits are currently higher than usual. Did it outperform expectations? The prompt was primarily about the political economy of central government transfers to Indian states/union territories. I think honestly this matched my expectations of what 5.4 can do when it's properly prompted. What did it get stuck on? It encountered a bunch of problems along the way, but doesn't appear to have gotten stuck for too long on any one thing. It was just a steppy process, so it just kept working. That is absolutely crazy.

In other news, OpenAI is forming a joint venture with TPG Advent International, Bain Capital, and Brookfield Asset Management.

This is great news.

To form a joint venture—

AI is coming to start a new job.

Enterprise products across the firm's portfolio companies and beyond the proposed deal.

AI is coming to Fogo de Chão. Bain Capital owns Fogo de Chão, the Brazilian steakhouse.

Maybe now they can take Apple Pay.

Oh yeah, we got cooked on that. That would be a good— I don't know, maybe they deliberately don't. I wonder, I wonder if Apple Pay is expensive for them and this is actually a cost consideration. Fiji Simo said this news came out a little bit earlier than we planned. We're excited to be building a deployment arm and we'll share more details soon. Companies have a ton of urgency to deploy AI in their organizations, and we're sprinting to meet that demand. More than 1 million businesses run on OpenAI products. Codex is now at 2 million weekly active users, up nearly 4x since the start of the year. API usage jumped 20% in the week after GPT-5.4 launched. And Frontier, which launched last month to help enterprises build, deploy, and manage AI coworkers that can do real work, has way more demand than we can handle. That's why we launched Frontier Alliances, so we leverage our ecosystem of partners and scale and to scale. And that is also why we're launching a dedicated deployment arm tasked with embedding forward deployed engineers deeply inside enterprises. This project has been in the works with our investors and alliance partners since last December, and we are grateful for them and their partnership. We're still early, but the speed of adoption is a clear signal of where this is headed. We're excited to not just be building these technologies, but also building many ways for companies to deploy them and get impact. Interesting, because there was a big talk for a long time about like the next-gen private equity firm will buy businesses and like deploy AI inside them. And this feels like, well, maybe that works at the mid-market private equity level, but you don't necessarily need a new AI-native private equity firm because Bain Capital—

Yeah, this is what I've always said, like traditional private equity is not, It's not going to just be like, oh, well, we'll figure out AI in like 2030. It's not really on a roadmap right now. They're working as hard as they can to implement it. The question is like, will, will traditional private equity be able to basically like roll out and unlock the value of AI better than some of these like AI-native, more like venture-oriented rollups?

Okay, so 5.4 mini announced today, so So different sizes of models don't count as SideQuest. These are MainQuest, right? These are MainQuest-aligned. But what is the pitch for a mini or a nano model?

Yeah, I mean, so this is just like, I think someone ran some evals and the mini models are equivalent to GPT-5 when that first came out. But it's just way cheaper, right? So if you're cost-sensitive, if you're running a ton, a ton of queries, but they don't need to be the max, intelligence, then you just use this.

So cheaper but also faster?

Yes, probably, yes.

And potentially runs on older hardware?

Um, I think that's unclear.

This is bullish for our vintage NeoCloud. Yeah, the oldest GPUs.

Yeah, this was Gray's idea yesterday.

Yeah, we're going to run this on 1080 Ti's for my gaming.

People love, people love classic cars. Why not classic GPUs?

Yeah, I mean, it doesn't— how small the model is doesn't actually matter on what kind of—

Imagine running your vintage NeoCloud with a wood-fired hearth powering it all.

Yeah. But I mean, you have to imagine that— what's after Blackwell? What's the one that announced today?

Vera Rubin?

Rubin. Rubin. Rubin. Like, you have to imagine that— there will be models that only run on Rubin and need to be sort of re-architected to run on Hopper, I would imagine.

Sure. I mean, I think there's like small performance boosts that you can get.

Or it might just be slower.

But you can always— you can run any model on any hardware. It's just like it could be like way, way slower because you have to—

Way, way slower. The memory's not enough. A fleet of A100s when you could just be on like a smaller rack of Rubins. Interesting. Well, that's exciting.

CalSheet came out with the $1 billion.

Before we talk about this, let me tell you about Sentry. Sentry shows developers what's broken and helps them fix it fast. That's why 150,000 organizations use it to keep their apps working. And let me also tell you about Public.com, investing for those who take it seriously. Stocks, options, bonds, crypto, treasuries, and more with great customer service. So what did CalSheet launch?

CalSheet launched the $1 billion perfect bracket challenge. $1 billion for a perfect March Madness bracket.

So you can win $1 billion for entering this competition.

I haven't read the fine print, but it seems it is positioned that way. Vinny says it would be hilarious if Citadel built out a team to attempt to financially impair their competitors. So SIG Parametrics from Susquehanna. Yeah, yeah. Is, is basically like the financial backer for this promotion. So if Citadel can figure this out, they could leave their competitor with a $1 billion bill.

Yes.

Obviously, trying to nail the perfect bracket is functionally impossible, but never going to pass.

So there are 9 quintillion possible brackets. So the odds of a perfect bracket are 1 in 9 quintillion. And I believe that they are capped at— I like my odds. I like my odds. And I believe that they are capping entries at 10 million entries. So if you math that out and then you take some insurance out, like you should be able to, you know, hedge, hedge, hedge out any of the risk. If you are good and you don't suffer from any skill issues, you can basically with strong basketball knowledge, if you have ball knowledge, Apparently, you can get the odds down to like 1 in 10 billion. And at that, you know, not from 9 quintillion to 10 billion.

If you have 9 quintillion and you can sort them by how likely they are, take the top 10 million, that's, you know, maybe there's some kind of power law thing here where we actually—

Yes. You're getting your odds. You get close. But I think if there's 10 million, if there's 10 million, like assume that there's 10 billion reasonable brackets and out of the 9 quintillion, So you're discarding like all the craziest upsets to get to that $10 billion. You already ranked it and narrowed it down by like, what, 6 orders of magnitude or something. And then you're submitting the top 10 million of those 10 billion. Like, your odds of winning are still 1 in 1,000.

You can make it 1,000 accounts.

And that's a—

no, no, no, no.

I think that they are capping the entire campaign to the first 10 million entries. So you have to be on Kalshi and you have—

My open claw already submitted.

9 quintillion.

The SEC prepares proposal to eliminate quarterly reporting. If you liked earnings—

I was thinking like, is this bad for us? We like earnings days, but it's not like our whole shtick. So I think we'll be okay. We will continue soldiering on.

GoodAlexander says finally less transparency around financial results. Yeah, I don't know. I don't know if this is going to be enough to make it easier to be a public company. It's certainly a burden, but the kind of shareholder lawsuit risk and all that other stuff feels like a much bigger burden. What this is just going to do is create more volatility, right? If you're only getting updates twice a year, you can just see these like massive swings because it's like, hey, we haven't heard from this management team in a formal capacity for 6 months, the business could have changed wildly, growth rates fluctuating, et cetera. And so again, if you love volatility, you're probably going to love this.

So it's good for the VIX. That's bullish for VIX investors out there. There is a world where— The typical public versus private debate is that when you're private, you can think in multiple years, depending on who your investors are, maybe think in decades even, whereas when you're in the public markets, you need to think in quarters. And so advancing that from quarterly to 6 months, and you start putting management teams in, okay, what can we deliver that will show up in 6 months, as opposed to what can we deliver that shows up in 3 months? That's twice as long. It feels like you could wind up with better-run companies doing more ambitious things. I don't know. There is a bull case.

Yeah. No, I mean, the other side of it is you just get more complacency. There's less of that. There's less of that. You're not like on the daily march. You finish earnings and you're like, OK, 90 days, we got to do this again. Like there's no days off kind of thing.

So are you an advocate for monthly reporting?

Daily.

Daily earnings calls, potentially hourly.

Just put— I mean, this is the— this is the crypto folks. They're like, put it on the blockchain. I want to be able to see in real time the revenues of this asset stream and how things are moving around. I want full transparency at all times as an investor. I don't know. I think that there is a— there's a potential for a good outcome here, but we will have to keep an eye on it and see what actually happens. This is just a proposal. It's just the proposal isn't even— we are preparing for a proposal. So the proposal hasn't happened. We're preparing for the proposal.

It's an advanced idea.

No, it's a preliminary talk.

It's an advanced idea.

Yeah.

For early talks. For early talks. Around a proposal.

Exactly.

Around a potential proposal.

You know, the FDA actually works that way. ANPRMs. It's like Advanced Notice of Proposed Rulemaking. So they tell the industry, okay, we are thinking about making a rule.

Good luck.

Good luck. And then everyone's like, we're suing. Don't change the rules because I have set up my entire business for this particular set of rules. And if you change it, it's like completely over for me because every business is like set up like this, like narrowly thread whatever regulation is there. Really quickly, let me tell you about Vibe.co, where DTC brands, B2B startups, and AI companies advertise on streaming TV, pick channels, target audiences, and measure sales just like on Meta. And let me tell you about Cisco. Critical infrastructure for the AI era unlocks seamless real-time experiences and new value with Cisco.

So NVIDIA announcing NVIDIA DLSS 5 and AI-powered breakthrough in visual fidelity for games coming this fall.

Video games are going to get more realistic. Prepare over the next couple of decades. This is going to be— this is entirely possible. This happens. Continue.

Let's pull up the video. It's funny because when I see the video, I want to see it big. When I see the video, I— my, my thought was, weren't video games already this realistic? But I haven't played video games besides Tetris on the Matic.

They—

oh yeah, long time.

So let's play GeForce RTX DLSS 5. DLSS of course stands for Deep Learning Super Shading, something like that. DLSS, it uses deep learning. It's used AI for a long time. The whole pitch for DLSS supersampling has, has always been go from a 720p render to 4K, up-res, just interpolate the pixels. This is different. You can see that the— it's not just becoming higher resolution, it is becoming GenAI. Like, there is a world where you render that at higher res, the lighting is changing, the shading is changing, the makeup, the structure of that person's face is changing, but it's still driven by the underly— why are you posting the bubbles because I'm ranting.

Um, I want to see DLSS 5 do this.

Oh yeah, uh, but, but this is what we were talking about yesterday. We were like, what will it take to get TBPN Simulator photo real? I think we should try and pipe it through DLSS 5, although who knows what this is going to run on because NVIDIA is not shipping new graphics cards to gamers or something like that. They're delaying the next the next gaming graphics card. But I think that looks better. I don't know. People are upset. There's a community note that has yet to be accepted, but the community notes are not happy. Like one of the community notes is just AI slop. And it's like, yeah, like that's in the name. It's deep learning. Like, what do you think the DL stands for? And also, why are we on DLSS 5? Oh, because this is like a decade-old technology. I don't know. People are very upset. One of the community notes, AI technology has disastrous effects on the environment. The majority of gamers did not ask for this.

Is that the— are those the most photorealistic video games that we have today? And then they're showing how they're even getting more realistic.

No, no, no.

Yeah, because they don't look— they look like games that are kind of like animated characters that aren't trying to be photoreal.

So there's a range of games there. That some of them have pretty outdated engines, pretty outdated rendering pipelines. Some of them are on the latest and greatest and look very high fidelity. There's also a whole bunch of tradeoffs in game development around the scale of what you're building and how many, you know, poker chips you want to put into great graphics. If you're making some moody first-person shooter that's going to be dark and there's going to be lots of, you know, like blood splattering and aliens and stuff like, you could potentially make it ultra photoreal. But if you're trying to build like something like Minecraft, it's like this massive world that's interactive with a ton of different people, like by narrowing down the scope, like the beauty of Minecraft was that the whole thing's voxel-based, which is like these little blocks. And so the computer can store like an entire world very efficiently because all it needs to say is just there's a cube there and it's green. There's a cube there and it's black. There's a cube there and it's gray. What do you think?

I was going to say Minecraft is a good example of what this can improve, right?

Yes.

If you go to the next post.

Oh, yes.

If you pull that one up.

Yes.

This is what Minecraft could look like.

This is what Minecraft could look like. There are drawbacks, obviously, because you're driving the style transfer from the particular images, from the shape. You're not reconstructing the entire image. You're just applying this, like, you know, re-rendering on top of the underlying structure. And so you can get really weird outcomes like this. Of course, you can re-render the full scene. But if you're doing that on a per-frame basis— That's powerful. So right now, basically what we're seeing with DLSS is the best AI can do in 10 milliseconds, I think, or something like that, because it needs to do this. Basically, if you're running a game at 60 FPS, it needs to be able to do it 60 times per second. So you can't be waiting around for images in ChatGPT or Nanobanana to cook for a minute on each frame. It has to be real time.

But also on consumer hardware, right?

Yeah, on consumer hardware.

Just 'cause this is not gonna be like hosting the cloud.

Yeah.

The whole point is that it can run in your gaming PC.

And by that token, this feels like it's not just 4 years behind. This feels like it's like 1 or 2 years behind in terms of the actual quality of the output. I'm actually pretty impressed. If this was going back to like DALL-E 2 quality, I would have expected that because DALL-E 2 was running on a fleet of A100s, H100s, something like that.

Sure, but I mean, the underlying tech is very different, right? Because you have like a very clear ground truth, right? Where it's like, you know what's supposed to be there. There's this kind of upsampling.

Well, so I mean, the data pipeline here, the actual ground truth thing is a little bit harder for something like this. They probably have to use ImageGen to generate the high-res versions to actually generate all that. Because previously DLSS was really easy to train because you would just go and run the game at 4K on great hardware and then go and run the exact same. You would render it twice, like from the same exact game footage. So you'd have someone play the game and on one monitor it would be 720p, on one it would be 4K, and then they would record both and they would say, these are exactly the same thing. Every frame is the same. Learn how to take the 720p image and create the 4K image. Now they have to say, okay, well, we have a 4K image. Let's run that through a GenAI pipeline to make it like photo real. Do that a ton, probably on a fleet of GPUs that are not consumer hardware. Then you have your training data, then you go run the DLSS pipeline and get this style transfer algorithm that can run 60 FPS. Anyway.

But there have been open source image models that have done really good upsampling for a while now.

But at this level? Or wait, are you talking about up-resing? Or are you talking about like— I would consider this like re-rendering using something that feels like diffusion. This feels like— because if you look at any of these examples, it's very clear that they're not just increasing the pixel count, like the hair in that woman's image, like the hair has a different shape after you turn DLSS on. It's not just higher resolution.

Yeah, that's true.

So like there has to be something like that they're training on. But, and I think that there probably will be very small hallucinations still if you were like to zoom in Whereas that wasn't really true with like the 720p to 4K pipeline. Like you would get like, like whatever the cutout of that image was that you were upressing would remain the same. Whereas this is actually like re-rendering it basically.

Well, none of this will matter when we have a billion GPUs in space. Let's head over to space.

Let me tell you about Phantom Cash. Fund your wallet without exchanges or middlemen and spend with the Phantom card. And let me also tell you about Okta. Okta helps you assign every agent a trusted identity so you get the power of AI without the risk. Secure every agent, secure any agent with Okta.

Take us to space, Jensen.

Let's listen to Jensen talking about data centers in space, courtesy of Nick. I'll spend very little time on this at this time. However, we're going to space. We've already been out in space. Thor is radiation approved and we're in satellites. You do imaging from the, from satellites. In the future, we'll also build data centers in the, in the, in the in space. Obviously very complicated to do. So we have— we're working with our partners on a new computer called Vera Rubin Space One, and it's going to go out to space and start data centers out in space. Now, of course, in space there's no conduction, there's no convection, there's just radiation. And so we have to figure out how to cool these systems out in space. But we Got lots of great— I'll spend very little time on this one. Got lots of great engineers working on it. It was very funny listening to Dylan Patel talk about space data centers and saying that like, yeah, even if you solve everything, like it's still hard to make chips. He's just like, it still comes back down to the chip bottleneck above all else.

Yeah. I mean, this is why you see Elon doing the TeraFab thing, right?

Yeah. Yeah. I really want to know what his plan is for that because has he put in an order for a tool or is he going to do— Terra ASML, like, because there's like the supply chain is what, 10,000 companies?

I mean, historically they've done, they've gone super, super early in the supply chain, right? They've been like, you know, mining stuff and—

Yeah, yeah, yeah, yeah. But I don't know, toolmaking is a special, special industry. Well, speaking of Dylan Patel, Semi Analysis was featured at NVIDIA GTC. The inference king has been crowned. NVIDIA won a massive belt and it looks like Jensen's holding it up. He is in fact standing in front of an LED wall or projector, but a beautiful thing to see the Semi Analysis logo.

It's all WWE. The entire world. It's so cool.

NVIDIA Extreme Co-Design revolutionized token cost, the GBNVL. GPT-4.72 is the inference king with 50x higher performance per watt on InferenceX by SemiAnalysis, 35x lower cost. Very, very exciting. And congrats to Jensen for becoming the inference king and winning the Inference Max Award, or InferenceX as it's now known. Jensen is also confirming what we see in our GPU availability data. There is an epic scramble for pute, B200 basically unavailable, availability for GH, Grace Hopper, 200, H200, and A100 also collapsing. Low availability means high demand. And people were kind of going back joking about like he said like 1 trillion of demand or something over some period of time. And people were like, oh, so he's guiding down. I guess that's where we're at.

Yeah, it's hard when you're doing cumulative revenue.

Well, let's move over to a much smaller deal for Netflix and Warner Brothers and the final deal which went to Paramount, of course. And David Zaslav is in The Wall Street Journal. First, I will tell you about Vanta, automate compliance and security. Vanta is the leading AI trust management platform. So David Zaslav deal pay could top $800 million after last infinite tax benefit. So Warner Brothers chief executive David Zaslav could collect more than $800 million in severance and other payments after rival Paramount acquires the company. The sum of cash includes— sum includes cash and payments for options and restricted stock holdings, as well as a newly adopted tax reimbursement for Zaslav. In a securities disclosure filed late Monday, the total doesn't include the more than $20 million he's likely to get for shares he owns outright. So he's been investing as well. About $504 million is due to Zaslav if the deal closes, the company said, while $47 million would be triggered if he is fired or leaves under certain circumstances within a year of the close. $116 million in equity has already vested. Some $335 million would only kick in if other payments trigger a 20% federal excise tax on golden parachute severance payments he receives. The ultimate value of the payout is based on tax code rules that are expected to cause it to significantly decline with the passage of time, the company said in a filing. Without the tax payment, the total would be about $667 million. Companies typically try to keep severance at or below 3 times salary and target annual bonus to avoid or minimize the tax, and investors often criticize companies for reimbursing or grossing up executives for the tax. The company said Zaslav wouldn't have to pay the excise tax if the deal closes in 2027, which would save the company costs of the tax gross-up. So, there's been a lot of back and forth.

Yeah, a lot of people are—

Did he earn that?

Yeah, absolutely. And the reason that I think it is warranted from a business standpoint is when we were sitting here last year, talking with someone who won't be named, in the media space, somebody who has done deals not at this scale, but in the billions of dollars. He was sitting here telling us at this very table that he didn't think there would be any bidding process at all. And it was just gonna land with the Ellisons for a pretty predictable price. And Zaslav did his job as CEO to get the best possible price for shareholders.

And it's remarkable.

And that is why, you know, he was able to have outsized impact. He's getting outsized compensation. He increased the enterprise value through his dealmaking by tens of billions of dollars. And he's getting a significant, but I think warranted payment. And if you don't like this, then you probably don't like the game of business very much either.

Or you think Warner Brothers Discovery should have just used like Open Claw on a Mac Mini or something like that to negotiate the sale. Put the whole company on eBay and let everyone bid.

Just do an auction.

It's actually eBay's fees are pretty high. I think eBay takes like multiple percentage.

So they would take a higher fee.

I think they would take a higher fee actually. It's like a $100 billion deal. Don't they take like 3%? So by that, I mean, maybe we're saving money using a human in this case. But of course, some of what he did, you know, between the dinners and the photos and all of these different negotiations and somehow continuing to get Paramount to make offers, but then say that it wasn't their best and final. I don't know how you do that.

He's clearly multiple offers back to back that were not the best and final, but quickly.

Let me tell everyone about Lambda. Lambda is the superintelligence cloud building AI supercomputers for training and inference that scale from 1 GPU to hundreds of thousands. And without further ado, we have Carried No Interest, the anonymous X user with the most fantastic profile picture. How are you? As always, no interest.

No, I'm doing pretty well.

I'm going to repeat, Carried No Interest. As many times as possible so that I don't say your real name, Carey. But good to have you here, Carey. Thanks for taking the time to join the show. Why don't you tell us what's been on your mind lately?

Well, it feels like the chickens are coming home to roost in old private credit software land and going to take a little cheeky victory lap here and say, who could have predicted this?

Who could have predicted this? Except you, on December 12th, 2024, you wrote a post called The Bubble in Software Private Equity, Private Credit Edition, and you did a deep dive on a ton of lending data and discovered, you said, some hilarious stuff. You said for a while there have been rumors forming that a bubble in private credit related to private equity. I'm somewhat certain a mini bubble forming in software private equity, private credit. And then you get into it. What were you seeing back then that prompted the post?

What I was seeing, and it is wild that a nameless private markets investor with a PitchBook subscription, no intentional advertisement for them just there, was able to pull this data and do it themselves, but anybody with a PitchBook could have done it. What prompted this is that I was simply being shown deals by investment bankers where the financing structure, I think, would make the average American scratch their head. If you had told the average person on the street that you could take a business with $100 million in revenue and no profit and lever it up to buy it with no profit at all, and have to repay a good chunk of it in 4 years, your average American would go, no way, that is not a thing that you are allowed to do. Depending on how good your relationship was with a certain subset of private credit lenders, that is certainly a thing that was allowed to to do. What prompted the post was that I kept being shown deals with financing structures I could not believe, from very good sponsors historically who were excellent. Then I thought to myself, you know what, I'm just going to pull all this data from PitchBook, do this thing myself. What I found was this giant bar in the years 2028 and 2029, where a lot of my wonderful colleagues in software private equity would have to pay a good chunk of that debt back. It terrified me because something else was happening at the same time that was somewhat scary, which was you could see on a— there are a bunch of different investment banks who have published this. Carl Square, a bunch of the other ones that are software boutiques, you could see on a graph that the multiple on ARR was coming down in private market transactions. Let us not even talk about publics, which now become more relevant than ever. I looked at that graph and the graph of all the debt coming due, And, you know, my stomach churned a little bit and then I wrote it.

And so the sponsors, their idea was, hey, the debt's going to come due in 2028 or 2029, but we're going to roll it over and we'll just kick, kind of kick the can down the road. But then now with AI disruption and a lot of questions, people aren't going to want to lend in the same way to these businesses.

I mean, the irony is it all kicked off with interest rates back in like 2021 and 2022. And you can see in the post from Carl Square in 2021, an unprofitable growth software company in Carl Square's own investment banking data traded at 9.1 times ARR. In 2022— I'm so excited to hear what sound you use next. In 2022, it was 3.4. So you went from 9 times— yeah, there it is. That was very good. Very, very mature. You went from 9.1 times ARR down to 3.4 times ARR, and you can even— like, how much software M&A took place in 2020 and 2021 that was sponsored by— it was a lot. Who is in control?

I am. Sorry, I'll stop.

Oh wow, you're really good at that, man.

You got a lot of talent.

And so it's kind of scary. Now we have 30% of sponsor-backed transactions with software at one point. And everybody's been kind of kicking the can for a while. And now everyone is noticing.

OK. Take us back and explain— when I think about these software deals, take privates or buyouts of private companies, and there's debt involved, I go back to the LBO boom, the Milken era, Barbarians at the Gate, obviously those were big deals. Is that the correct structure that I'm thinking of? It's just LBO, but instead of like a cigarette company or a food company or some other industrial company, it's a software company now. And then after some of that history, I wanna know like what is fundamentally different about these software deals? Because that LBO boom seemed to last for decades. Yeah.

I think that it all comes down to the fact that they were borrowing against ARR and not EBITDA. There is very little profit to borrow against with a high-growth enterprise software company. What do you do? You need to juice your IRR somehow, so you need to borrow. You will go to a bunch of private credit funds and you say, look, I have never lost money. I am talking about a very specific firm right now. I have never lost money on a deal. You cannot lose with me. Just let me borrow against the ARR. Before it comes due, you know me, I'm going to sell it. We're all going to buy another Gulfstream, and we're going to laugh about this in 4 years. The problem was that the liquidity just left, both in terms of the most important person who could buy it, which was a strategic acquirer. When stock prices fell and interest rates went up, you're just not as acquisitive. Now, you lump in the fact that AI is creating all these existential questions for enterprise software companies.

Yeah.

And you have the depressed public valuations.

Yeah.

You know, and you have no profit. So your recovery rate with your lender might even be low. Sure.

Right.

Sure.

You can't even— you can't even take the keys to this business and milk it for dividends.

Yeah.

Yeah. You have to do a massive restructuring. That's a really scary proposition.

Did you— did you— did you see the— the— the AI disruption of enterprise software coming, or did that hit you like a flashbang?

I mean, I did, but I was also an AI or data scientist, as you know, for many years. So I thought it was inevitable.

That was good. Throwing flashbacks.

Yeah, no, to— almost said your name— Carey's credit, this time last year, you were talking about a specific product category and you were saying, 10 years ago to build a product in this space, I would've needed to raise $30 million to build something competitive. Now I can do it with $300 grand, and I'm gonna go after it. You were working on some deal at the time. And so I think you did, I don't know about on December 12th, 2024, but certainly at the very beginning of 2025, you were seeing that like, hey, a lot of these, companies that previously would have taken tens to hundreds of millions of dollars to rebuild and actually be competitive with, you can now do with much less capital.

Oh, 100%. And I mean, I think that that is kind of— I think that— and here's maybe even a spicier take that I swear I'm going to write a longer post on. I think a lot of the stuff around a really good enterprise software company is cooked because somebody is going to vibe code it. I don't think that's going to happen. I think that the scarier thing that's occurring now for private equity is actually that we're hearing rumors of customers that aren't signing 3-year deals. The entire basis of software PE being one of the best private market categories was that you had enterprise customers with 3-year contracts. If every other enterprise customer says actually AI is just so amazing now and we don't know where we're going to be and what it's going to look like, we're only doing 1-year deals. Okay, that has like much bigger implications than, you know, somebody vibe-coded a Notion clone on a Sunday. Sure, right? That's a big deal. And I think that the reason that, that they're doing that, and this is the bigger threat I think to a lot of these companies, is the adjacency within the market map. Like if you look like Rippling and you have a very talented like group of people like Rippling's employees, they can attack so many parts of the HR platform space now that they simply couldn't 3 years ago.

Sure.

I think that's a lot scarier than the vibe coding to my colleagues in software private equity than somebody just whipping one of these up and going out with a bunch of cold emails. It's the adjacency threats in the market map. Because now who's Rippling going to go after?

I don't know. Yeah. You have Previously, there was like, it's not an HR company, it's like HR for hospital or HR for golf courses or some sort of narrow niche in a niche. And that all of a sudden has been pretty easily added to Rippling's TAM, for example.

Yep. And another thing that I'm seeing just when I talk to people in the industry, I don't know that there's a scarier seat right now than at Toma or Vista. Because the liquidity is just evaporating. The appetite in the public markets, gone. The ability for these mega technology companies to do splashy 15x ARR acquisitions for $15 billion, gone. That's a really scary thing.

Yeah, all your— the hyperscalers want to put money into CapEx or various core AI bets. The public markets are going to put a massive discount because they don't know what you're going to be doing in 10 years or even 3 years. There's so much uncertainty. And who wants to buy these companies at at least the marks that they've had over the last few years because of that same uncertainty? So it's like, where do these companies go? The funds can continue to do continuation vehicles, but it's just like everyone is just kind of kicking the can down the road forever.

It's a scary time in the asset class. I don't think that there's, you know, I think that there's going to be a massive kind of resurgence of super ops-focused software funds that might even be far more AI-focused than we could imagine, right? Where you have PE funds where 50% of management fees is going to AI professionals. I could see that very near term.

Yeah. What's your read on, uh, partnerships between the labs and the private equity firms? We've seen a bunch of those happening versus like starting a new private equity firm that is like quote unquote AI native?

Oh, I mean, I think that, yeah, not to use the cliché, but like it has to be like founder mode from the top. Like you can't be trusting OpenAI to make your pork coast more money. I, I think that it needs to be very much from the top of the firm, really understanding what AI is capable of. Because I don't know how many times in history that it used to take 5 years to build a factory and now you can build it in 2 weeks.

Mm-hmm.

Right. That's the scenario. You probably want an expert in factory building at the top of the fund, right? Or near the top. And the rumors are that there's been a very slow adoption at some of these funds that were just used to XRP and they were rolling in it, right? For 15 years.

And I don't think— And when you say slow adoption, slow adoption for AI, like actually pushing AI-first initiatives across the portfolio companies? 100%. Like not the deal teams. But the actual operations teams. Yeah.

So, you know, you own, you know, 6 different businesses that are doing $30 million low end of ARR.

Yeah.

How involved is your deal team in— I'm not even talking about AI adoption on the ops level. I'm saying full-blown AI adoption across both product and ops. And I think that if you're not doing that, you're going to be in big trouble in like 3 or 4 years.

It's inevitable. What's the takeaway from Mark Leonard's, you know, kind of half-retirement last year, it felt like in many ways he, he like— perfect. I don't know if he, he didn't, he didn't like his exit, didn't top ticket, but he bailed right as there was going to be kind of like infinite questions around the sustainability of the model that, that got them to that point. Obviously built a fantastic collection of software companies?

Well, here's the trick.

Constellation never overpays. They didn't do what, in my opinion, firms like Vista and some others did, which was borrow and pay large ARR multiples for unprofitable businesses. Constellation does not pay frequently more than 6 times EBITDA for a software company. Price is such a good insulation. So if you just assume like a certain gross retention and you pay a certain price, Constellation is likely safe. I think that they face the same threat that others do in the adjacency of the market map, right? But their risk is much lower simply because they never overpay, right? So—

How are you seeing like diffusion amongst these like sort of deeper in the economy software companies where the small business that they sell to might not be tracking the latest frontier model development. Oh, 5.4 nano launched today. I can cut my cost. They're just like, yeah, my business is, is, you know, golf courses and I need to pay my employees. So I'm just on this particular platform and I don't even take calls from competitors because it's just so far down the stack. It's only 1% of my cost here. I don't think about it. Like, is there a world where because AI takes time to diffuse, there's a whole bunch of humans in the loop that some of these software companies are actually just stickier than we expect them to be, even though on principle they should be disrupted faster?

I 100% think that we are overestimating the rate of churn, right? Yeah. I think that that is pure timeline nonsense. As somebody who sees— I was just gonna write this, like, because I was actually getting frustrated. Yeah, I literally see these company financials, you know, I've seen them many times. I'm not seeing this alleged Vibe Code churn yet, right? It's just not showing up.

Horse carriage manufacturer says order flow looks great, we have a big backlog, there's no issues.

Sure. Like, and maybe at some point you're going to see, you know, every other, you know, middle market electronics distributor rolling their own ERP, but it's not happening right now. And so I try to keep like a check on my colleagues in other industries, right, that are non-tech. And I literally like once a month I just say, are you going to roll your own ERP this month? And the answer is always no, right? And at some point that answer will start to shift and then one or two will do it. It will go horribly. And then I will hear that.

But the, but But people have, you know, VCs are still funding AI-native ERPs, like they're funding more ERPs. Exactly. So I don't think the question has ever been, I haven't seen anyone saying all, you know, these like core system of record businesses are totally cooked. It's every other, it's the next, you know, 50 pieces of software that might integrate into said ERP that, there's a big question mark around.

So this actually circles back to— this actually circles back to, um, uh, what I was saying before, right? There's a reason I said like half of your— half of your like deal team, if you're in PE, should have some level of AI expertise because of that exact situation. That's actually what I'm talking about.

And, and sorry, sorry, to clarify there, deal team, you mean evaluating the transaction, setting price versus the operations team that's actually going to go and work with management and jump on the board?

Great question. Depending on the size of the fund, those could be the same group or they could be separate. But no matter what, there should be a good amount of either AI-first developers or just like AI-first product people that are helping your portfolio companies compete as the VCs subsidize the CAC of the next wave of quote unquote AI-first replacements. Sure, sure.

Yeah, because my default assumption would be Okay, yes, you need your deal team to be able to understand what moats are AI resistant, what aren't, like reevaluate the prices in the AI era, but then the real AI expertise of selecting products, advancing software development workflows, figuring out the right structure of the team mix, like that's gonna be much more on the operational side of the folks that you deploy into these companies, the management teams that you hire. Yes.

Like an investment banker, no offense to my pals in investment banking, is not going to help your portcos be like AI-first competitor-proof. That's not going to happen. That's your biggest existential threat. Right. So I think you're going to see a reshuffling for the best firms that can see what is so clearly the future of ops around that to deal with all of these AI-first YC and venture-backed competitors. I think that's inevitable, or you're going to be in trouble. I think another side of it is I think that you're going to see a lot of point solutions not trading hands at good multiples. I think that you will see a big flight to pure mission-critical software in PE, whereas I saw plenty of point solutions changing hands in 2023, 2024. Now, I don't see them trading, and I don't I think you can't underwrite that risk for your LPs. You simply can't. If I can swap it out in 12 hours, I don't think you can use your LP money on that in private equity.

How existential do you think everything that we're talking about is for like the big platform software PE funds? Like, do you think, do you think they can, you know, eventually figure out how to exit some of these businesses they have some funds that, that, you know, underperform or lose money, but then, but then they're able to kind of reposition? Or do you think it's, it's just, it's over?

I, I don't know, man. It, it's not looking good. Like, I mean, you get the public markets are, are saying, are telling us what the appetite is for these late-stage software companies without massive AI tailwinds. And it's a very scary proposition. I, I don't see a shift barring interest rates going down and everybody jumping back into the casino. I don't, I don't see it, but I've been wrong before.

So, well, the good news is interest rates are going up the last couple.

Perfect.

This is gonna be great for everybody. This is gonna be wonderful. Uh, but yeah, I think it's existential. I think there's no doubt about it. There's, there's no seat I would want less than at one of the bigger software PE funds right now. Scary stuff. And the rumor is that a lot of people are looking for the doors.

What about other stuff that's software-adjacent, AI-adjacent? We've seen a lot of the private credit funds do deals for large AI data centers. And that feels like a complete balance against— people are worried about software PE debt, but then if that firm has a some contract with Meta for some big data center, like Meta seems like they're going to be able to pay that bill.

Well, it's funny you say that. I think it all comes down to the recovery rate, right? Like, you know, if the recovery rates are poor enough, there will be nothing that Daddy Zuck can do to save them, right? There's nothing he can do, right?

If recovery rates are poor enough. Well, he can keep paying his data center bill on time.

Well, yes, there will be very little to hide from the recovery rates are really low. I don't know what those are going to look like. There is a whole new asset class coming, in my opinion, of software special situations that is imminent. You can think about a deal structure that that would look like. This is actually a really fun thought exercise. You have a software PE fund that was overleveraged. And the private credit fund takes the keys. And they're just trying to hope— they're hoping to get anything back. And you could see a special situation—

Yeah, this is John Zito saying, I think some of these deals, you're going to be lucky if you get $0.30, $0.40 back.

And is that technically a bankruptcy, or just the shareholders get wiped? Would you be declaring bankruptcy in that scenario?

Equity's taking a zero, almost undoubtedly. Depending on how the restructuring pans out, it certainly could end up that way. You could see where a very bending spoons-like entity might come in and say, hey, creditor, we're going to help you out. We're going to run this thing. We're going to shift it to the constellation model. Help me help you.

Yeah.

Help me help you. We're going to shift this to the constellation model. You know what? We're not going to focus a lot on growth. And we're just going to focus on dividends and we're going to try and get you out. Right. That's something that, that, that a special situations investor might look at. Right. And I think that's a much more reliable way to derive IRR than we're going to turn around this thing that went flat and isn't profitable at all today. And I think AI enables some of these special situations a little bit more. And that's kind of the whole bending spoons thesis on some level.

Right.

So I think there will be a new wave software special-sits people that come in and try to do something with these overleveraged software codes with the creditors in some form of agreement.

OK, so I mean, you said that software PE investors might be heading for the door. But isn't it possible that they spin up a new fund that does special situations, develop new expertise, just reorient their strategy around curing logjams?

Yeah, I mean, well, you're already seeing people leaving Constellation to pursue this. And I think that when LPs evaluate who's the better option, they're going to go with the guy from Constellation. Since then, sure, somebody from Telma or TA or—

I was thinking about, I was thinking about like internal to—

they're like, it was a special situation when I paid 50x ARR for this. It was quite special.

Look, you just didn't see my vision for this thing. Okay. What's your— Sorry.

Go ahead. What's your take on, on Zuck's move with Manus? You were an early Manus—

Oh, he got a steal!

When I was kind of— when I saw that acquisition, I was thinking, hey, obviously a talented team have built a product that can get to real revenue. But I think there were some questions of like, is Zuck just going to kind of wind down whatever they were doing and refocus on—

Yeah, was it talent acquisition?

Was it a talent acquisition or product acquisition? I was leaning more toward product acquisition. They seem to continue to be investing a lot in it, which is kind of interesting because it's a product category that Meta doesn't— they obviously serve a bunch of small businesses, but Manus is being positioned now as like an agent that can help you with your business or a school project. And it just feels like, you know, obviously hyper, hyper competitive. But what's your read on their strategy with Manus and their odds to actually compete in productivity against OpenAI, Anthropic, Microsoft, all these other players?

Zuck, if you're listening right now, you got a deal, man. That was good. That was a good deal. I love it. I think that Zuck looked and he saw an absolutely magical product that he could have and he bought it, right? It's magic. I have spent thousands and thousands and thousands of dollars on Manus, and it was worth every penny. There are simply things that you can do with Manus that no other inference provider or product can do. And I would reveal them, but they're—

it's Manus, by the way.

Manus.

Manus.

It's an American company now.

Yeah.

And like, there's, there's like literally things you can do with it.

I'm sure you saw the same thing. Like, they figured out context like LLM context in browser before anyone else in ways that no one else has yet to figure out. Right.

That's cool.

And I think that, I think that if Anthropic had figured it out, they would have pushed it. I think if OpenAI could figure it out, they would have pushed it and they simply haven't yet. And he saw a piece of magic and he said, what do we got to pay to get the magic? And I think it was right because it's generational. Like I said, there's a lot. I've spent a lot of money on Manus. There's things Manus can do that no other inference provider can do. And they benefit from open source. So there's another cheeky thing that Zuck realized, right? Like, we don't—

broadly an open source bull. I feel like a lot of people are saying like all the money flows to the frontier. You know, everyone just wants to use the best model, the Opus 4.6, the GPT 5.4. Codex Desktop can do a lot of these things. Have you, have you actually tried the other harnesses and found like a real big delta?

I am a Manus and Anthropic maxi. Those are my two favorites. I use Manus for certain tasks and Anthropic for others. But I'm certain that what Zuck realized is that I don't have to care about inference and who's giving it to me because these guys figured out some really special stuff around how does an LLM process data in and around a web browser. So he realized, okay, this is kind of inference provider proof. They figured out this piece of magic, this secret about LLMs using a browser. I'm going to pay whatever price I need to get it. And now if you've heard, he's already integrating it with like Ads Manager.

Yeah, that makes so much sense.

Manas, I want to grow my business. All right, give me every dollar in your bank account right now.

What is— Yeah.

No, credit to Zach for kind of, uh, either knowingly or accidentally realizing that harnesses would be quite important.

Yeah, yeah. And, and like I said, there's a lot of stuff Manus can do that no other, like, AI-related product can do for me.

Yeah.

And it's just, it's just really special. So yeah, I think you gotta steal. You pay whatever price you got to pay. Within, like, the re— like, the history of M&A, right? I think it's a good bet. It's a great bet.

Yeah, especially for their scale and the compute that they have. Like, there's a lot that they can deliver there that another buyer would not be able to.

I actually think Manus is like one of the most underrated AI products today.

Okay, I just downloaded it. I'm going to try it out. We had our intern Tyler download it and give it a try yesterday. We'll get a review from him later on the show.

Well, I'll read out one. Somebody in, John in the YouTube chat says, would be interesting to have Jeremy Giffen back on with Cary, no interest, do a sort of roundtable style back and forth on here. I like Cary's analysis so much, feels up Jeremy's alley.

Oh, that'd be cool.

I totally agree, we should do that sometime.

That'd be great.

I've heard of that guy, I've heard of that guy before.

Smart, smart cookie.

He's a good dude. You're both former TVPN guests. He's been on the show. Uh, well, we'd love to have you back. Let's, let's throw a smoke grenade and get him out of here.

Yeah, do it. It's great to see you, Kerry.

See you later.

I'll get behind this production.

We'll see you soon.

Goodbye.

It's been an honor. And let me tell you about MongoDB. What's the only thing faster than the AI market? Your business on MongoDB. Don't just build AI, own the data platform that powers it. And let me also tell you about Cognition. They're the makers of Devin the AI Software Engineer. Crush your backlog with your personal AI engineering team. Midjourney's financials are crazy and underrated. This is from Amrit. Did not raise anything, built things in-house. $200 million ARR as of late 2023, confirmed $500 million in ARR as of 2025, maybe. I don't know, that sounds reasonable to me.

Yeah, the question is like, how was the Meta deal structured, right? I assume that Meta was just like, we're gonna give you a ton of money Yeah, every year for some period of time. So I'm sure that, that, that factors in.

Yeah. Yeah. They must be making a fortune.

Okay, John, I said yesterday I wanted to order beef from a ranch with a livestream.

Yes.

To allow me to be present.

Yes.

Where, where the cattle are being raised. And this Indian startup lets you own a mango tree for $111 per season.

That's cool.

Allows you to rent a mango tree and enjoy the entire harvest. There are 3 types of trees you can rent. Base, you can get a base tree, standard tree, or max tree.

No way.

You can get 30 kilograms to 60 kilograms of mangoes.

That's a lot of mangoes. I love mangoes. We're going through them a lot.

This company is operating in 3 states in India. You select your favorite tree, pay the money. You will get a dashboard with all the information about the tree you rented, connecting farmers with direct customers who love chemical-free fruits. This is awesome. Very, very cool.

I love it. Yeah, we should definitely get a mango tree.

Tyler, can you get us one mango tree on the Max plan?

On the Mango Max.

I want the Mango Max plan, please.

So Tyler, have you had a chance to fire off a Manus prompt and see what it's cooking?

Yeah, I mean, I think so far, maybe I'm just not creative enough, but it's felt very similar to the cowork kind of thing. Essentially just like a wrapper around Cloud Code where it can interact with my desktop and interact with local files. I haven't tried using computer use yet. I'll do that now.

I have no local files anymore. Everything's either in the cloud, camera roll, or I get a new computer, I don't even transfer anything over. I just sign in to Gmail. I'm good. It's a very, very thin client these days. But if you do have a lot of files, I guess that makes sense. And I still do like that idea of like color grade every image off this thing. And I mean, just there's a ton of different research things that you can do. Anyway, let me tell you about Railway. Railway is the all-in-one intelligent cloud provider. Use your favorite agent to deploy web apps, servers, databases, and more while Railway automatically takes care of scaling, monitoring, and security. So the AI boom has exploded. The San Francisco housing market. I believe we will have an expert coming on the show soon to discuss.

Rohan Dar is going to be joining next week to talk about this.

But the Wall Street Journal has a report today that we will read through some of it. At a Pacific Heights open house in January, a line of people made their way up the steps of a 2-bedroom, 1-bath co-op. There were 85 of them— steps, not people. 8 flights, no elevator. The property received 14 offers and sold for over $1.62 million, more than $400,000 over the asking price. While much of the U.S. housing market has been stuck in a rut, slowly elevated by mortgage rates and slowed by elevated mortgage rates and home prices near record highs, pockets of San Francisco are rebounding in a big way. The AI boom, a new mayor, and other changes in municipal leadership have helped bring the city back, reversing a years-long slump that was compounded by the ripple effects of the pandemic, crime, and persistent struggles with homelessness. Rents citywide were up 14% year over year in February. And this is what we dug into when we looked at Daniel Gross's AGI bets. He was saying, is San Francisco the new Detroit? And it was unclear if he meant the current— is San Francisco going to be like the current Detroit or what Detroit was in its heyday as a boomtown? Well, it certainly feels like the latter. An uptick in demand with the city's notorious lack of housing supply, they gotta build the cube. They gotta build some cubes, they gotta build some skyscrapers. This is the way.

The Munger dorm. Yeah, the Munger dorm. Maybe expand that out to General Purpose.

Windowless dorms.

They should build one of those. They should tear down those houses. They call them like the Painted Ladies. You know, they should tear those down and build like the Munger dorm with no windows. That would be ideal. Maybe BlackRock should start investing in San Francisco. San Francisco real estate. I feel like they would have a good financial opportunity here. Wow, everything's booming.

Uh, condo prices, which have been sluggish for years, grew 12% year over year as of February, ahead of the spring peak, according to real estate brokerage Compass. The median sale price was $1.23 million. Single-family homes prices are up 23%, with the median price at $1.96 million.

It's just—

by comparison, the year-over-year median increase for existing home sales nationwide is just 3%. Last month, 16 homes in San Francisco sold for $5 million or more, a 220% year-over-year bump. It's just skyrocketed, said Kelsey Carlson, 34 years old, an attorney who is expecting her first child in June and has been house hunting with her husband since April. You're more likely to get outbid by an all-cash offer. Carlson was outside a packed open house for a 3-bedroom, 2-bath condo on Buchanan Street in Pacific Heights. The area, known for its breathtaking views and mix of mansions, Victorians, and pre-war apartment buildings, has long been sought after. Carlson and her husband have been outbid on 4 properties so far. Brutal. Even a house in nearby Presidio Heights that needs hundreds of thousands of dollars of work. With AI, everyone's coming in with these huge salaries, she says. We just can't keep up the pace.

There is some good news. There's a housing market. Market that I think is potentially better than San Francisco, growing a lot slower. It's only up 4% this year, year over year, as opposed to San Francisco, which is up 12%. It's Alaska. And I would highly recommend if you've made your money in an AI, pick up a house in Anchorage. Average house prices up there, $400,000, doing very well. And you can throw a Starlink up there.

San Francisco.

The locals will love Anchorage. The locals will.

John has a long history with—

I shouldn't even joke about it. I need to lock my account again. If you see me with a locked account, you know that I've pissed off the Hornets.

Nasty Alaskans. Anyway, so we're going to have Rohan, who is an SF realtor and has 100,000 followers on X, just talking about investing and advising clients.

Let me tell you about Plaid. Plaid powers the apps you use to spend, save, borrow, and invest, securely connecting bank accounts to move money, fight fraud, and improve lending now with AI. And let's see what Adam Neumann is up to because he's in Miami and was featured in an Instagram Reel, a TikTok collab with none other than Caleb Simpson. Caleb W. Simpson.

Adam Neumann, what's up, man?

Nice to see you. How much do you pay for rent in Miami?

In Miami?

I don't rent.

I give out loans.

Okay, developer.

I come into the main—

I want your review on this.

What's going on over here?

You should do one of these in Alaska. I don't know why he's doing it in Miami. He's getting after it.

Yeah, I've always wanted to do one of those classes.

You can go as low as $2,000.

Okay, what a view!

Oh my gosh, what does $2,000 get you in Alaska?

Safety, and clean.

After it's safe and clean, now we can talk about community. And if you can put those 3 things together, you're starting to have a great experience. There was a company once, there's a television show about it, it was called WeWork, I think it looked like this.

Huh.

I don't know what they're doing now.

Still got bits. Still got bits.

It smells so good.

That is so important.

It smells so good.

Yoga.

Yoga.

Is that a coffee shop?

Downstairs that's open and even the public can come into.

This is just for the residents.

All these products are all Flo products done in the Flo kitchen.

And there's a whole spread out.

Everything is fresh, and whatever one sells today will go to the residents at the end of the day. There are people working out in this gym all day long. I've lived here since it started, but I've lived in this building for about 4 years. So he was here before we bought the building. You got to see the building before before and after. When I moved here, it was a great building, but it's turned out way better. The same price in Alaska, you can get a whole house. There you go. But it's probably fly-in only, or there's limited road access.

Limited road access.

Limited road.

How many square feet?

Uh, at least like 2,000 square feet.

Okay, okay. Yeah, I mean, so, so, uh, I think Adam will create a great apartment living experience. The question is how different is it from every other kind of apartment complex in Miami? How much do residents care about a cafe differentiator, workout classes? A lot of apartments have gyms, they have spas, things like that. So I just think, yeah, ultimately I think Adam will be successful, with this purely because he has high energy. He cares a lot about customer experience. And then the question, well, like the question coming out of WeWork is like, can he keep the costs in control? Because it's if you don't want to make it, if you don't care about making money, it's really easy to deliver the best apartment living experience in a given area. The question is like, can he deliver services and keep the margin because I think ultimately this time around the business will be valued. Maybe there's some brand equity to the Flow apartment platform, which is, which is a brand that he's doing this around. But ultimately it will be valued based on earnings given, given enough time. So I think it'll just be a balancing act. But looks, looks nice.

Yeah. Well, Florida man, Florida man sold his house in just 5 days after letting ChatGPT handle the entire process. Instead of a real estate agent. There's some community notes on this. He, he didn't go to ChatGPT and say, sell my home. He like you. It's the same thing as the dog cancer story. He used ChatGPT to conduct market research. Everybody wants to move the goalposts and yeah, but you can just sell your house. That's, that's good news. I was listening to Ryan Serhant talk about the opposite scenario, which He said he was working on, I think it was like a $50 million house sale. And he, and they'd spent, he was like, our agents had spent like months getting these two buyers very educated about the value. They'd agreed to a price. And then the seller went to ChatGPT and said like, this, should I sell or is the price too low? And ChatGPT said, you're absolutely right. It is too low. You should pull out. And then the buyer went like, is this a good deal or is the price too high? And ChatGPT said, you're absolutely right, the price is too high. And so they both walked away and Ryan Serhant was like, what the— like, now I got to get these people to come back. It's just like a funny thing.

Yeah. Who knows?

I think they need to be in a shared chat window going back and forth arguing with the AI, the same AI in the same—

Yeah, yeah, yeah. A group.

You can do that.

Yeah.

Yeah.

Yeah. That would be the correct way because then they We can fire back and forth. Anyway, let me tell you about Figma. No matter where your idea starts, Figma may cloud code, codex, or a sketch. The Figma canvas is where ideas connect and products take shape. Build in the right direction with Figma. Uh, reorg in the, uh, Copilot division of Microsoft. So company scraps divide between consumer and business app teams. AI chief Mustafa Suleyman will focus on AI models. He's freed up, apparently. So Microsoft is reorganizing the teams that work on the different versions of its flagship Copilot AI product. This is from the Wall Street Journal, altering a strategy that some employees said created a disjointed user experience and consumer confusion. The software giant is unifying the teams that work on its Microsoft 365 Copilot productivity offerings and the consumer version of Copilot, according to a memo from Satya Nadella. Adela, Jacob Andreu, who leads product and growth for Microsoft AI.

The man who has sat right here.

Yeah, he, he, he brought back Clippy. He's the man that brought back Clippy.

Yeah.

He will become the executive vice president of Copilot, of just Copilot itself. That's the brand. And it's a great, it's a great brand name. I'm a big fan of Copilot.

Yeah.

These are prosumer tools.

Prosumer tools.

There's, they kind of, they're They're going to be used in and out of people's work life. I think businesses everywhere are kind of realizing that individual employees are bringing their own AI tools to work just because they help them do a better job. Yeah.

And this is what happened with Apple. I mean, Apple was like a consumer company and then all of a sudden everyone was using Macs at work. And I think that's also a little bit of what's happening with the OpenAI story about like refocusing on business enterprise API. It's like, no, like there, Thibault is also saying he's going to put Codex in ChatGPT. So you're just going to be like, build me an app that does the thing that I want. There's a lot of times when you go to ChatGPT and you ask something to, hey, go, go transcribe this YouTube video. And like within ChatGPT, it might not have that tool, but Codex could absolutely one-shot it. And so bringing that ability, whether it's cloud-hosted or running on your computer, So bringing that over makes a ton of sense, and Microsoft should probably follow a similar strategy. So exciting news there. Let me tell you about the New York Stock Exchange. Want to change the world? Raise capital at the New York Stock Exchange. Just do it. And we have our next guest, Shyam Shankar from Palantir. He is the Chief Technology Officer in the Restream waiting room. Let's bring him in. The TV's going to— I'm Shyam.

Good to see you.

Good to see you guys. How are you, Jordy? Welcome back.

It's been far too long. It's been like almost a year. You were one of our earliest guests, a huge moment for the show. Thank you so much for joining then, and thank you so much for joining today. How are you doing?

It's a big day.

I'm doing great. I'm pumped to be back on St. Patrick's Day, but more importantly, on the book launch day on Mobilize.

Yeah, what's the Straussian read of—

You sent us a green package.

Should we open this? Can we open this?

You got to open it. Oh yeah.

You sent us a green package because you knew It was going to be on St. Patrick's Day.

Look at this thing.

This is a real piece of hardware.

It's an actual US Army ammo box. It held 20mm electrically primed munitions. And now it holds Mobilize. It's got the book, it's got some swag.

You've outdone yourself. So cool.

I'm going to step back.

You can move that.

This might get in the way of the camera. Anyway, so I mean, you've been working at Palantir for a long time. Has it been almost 2 decades now?

It'll be 20 years this Friday.

Oh, this Friday. Congratulations.

Amazing.

So when did you feel like, OK, now is the time to write the book? What was the thesis? What was the moment where you were like, OK, it makes sense to actually take all that experience and distill it down into something that can be instantiated in an actual book?

Well, the book is the long-form version of the 18 theses. So I put the Defense Reformation out there in October of 2024, and that was kind of the welling up of all these feelings I had working more or less in the bowels of the Department of War for 18 years, watching deterrence slowly erode, seeing that we had a problem. You go back in time and you say we had the annexation of Crimea in '14, we had the militarization of the Spratly Islands in '15, breakout capability for Iran to get the bomb in '17. We've had a pogrom in Israel, the invasion of Ukraine, conflict with the Houthis. What's going on here? Like, we're spending $1 trillion a year.

Where's our deterrence?

And that's one piece of it. And you look at what's looming with China. Then you look at the other side of this and you really look at history and you say, like, what gave us deterrence in World War II, in the Cold War? How did this stuff work? You recognize how much the industrial base that won World War II and the Cold War is not the industrial base we have today. You know, the astonishing statistic is that only 6% of major weapons systems were built by defense specialists in '89 when the Berlin Wall still stood. So 94% of spending went to companies that were what I call dual purpose. Chrysler was the prime on the Minuteman intercontinental ballistic missile. So Chrysler made missiles and minivans. Ford made satellites. General Mills, the cereal company, made torpedoes. We had an economy that was equally invested in freedom and prosperity. But the corporate story is not enough. So that's a precondition. You could kind of say, well, what were these companies like? You know, because we think about it today as Northrop Grumman and Lockheed Martin, but it was Glenn Martin, it was Jack Northrop, it was Leroy Grumman. It's something that your audience would understand foundationally. It was founders. You know, our entire industrial base was made up of founders. What's happened, you know, at the end of the Cold War, we had this enormous financialization of defense. These companies became run by the third, fourth generation, buybacks, dividends, financial engineering over real engineering. And by the way, that's not specific to defense. The same thing happened to Intel. The same thing happened to a lot of great American companies. The rejuvenation of our economy comes from the heretics. And then I started researching a lot about the history of innovation in defense. It's like nothing worked because of the system. Everything that worked, worked despite the system. You look at Andrew Higgins, this Scots-Irishman in Louisiana who built the boat that won the war. 92% of all boats in World War II were Higgins boats.

Wow.

But the Navy didn't let him compete. Then when they finally let them compete and he won, they stole his designs and failed to copy it successfully. In the end, after all these things, they're like, fine, we'll buy the freaking boat. And the boys in Normandy landed on Higgins boats.

That's great.

You know, and you go story after story.

Yeah.

Hyman Rickover, who built the nuclear Navy, his first office was a women's restroom. And I think part of documenting these stories of the heresy because of course the Navy wanted to humiliate him into quitting. And then think about the chutzpah, like Oppenheimer told him, this is not gonna work. You're not gonna be able to build a nuclear-powered submarine. And in the face of Oppie telling you this, you're like, no, you're wrong, I'm gonna do it.

That's bold.

You know?

Yeah.

Bold. Have you heard the story of Ball Corporation? They make Mason jars.

Yeah.

You know those Mason jars? Also a defense contractor.

Yeah.

I think they eventually sold it to BAE Systems in 2024. But for a long time, the company that made the Mason jars that are in every, like, hipster millennial burger joint in America was also making like satellites and sensors and all sorts of stuff. And so there's just endless stories about that of, of reindustrialization. I'm wondering, like, there's a huge boom in new startups that are saying we're going to build boats from scratch, we're going to build missile systems from scratch, we're going to build satellites from scratch. But is there some underrated industrial capacity in America where we haven't actually gone to Chrysler and said, what can you do these days? I know Chrysler might be a bad example because it's an older company, but is there still some latent industrial capacity where in the worst-case scenario, like, America can actually adapt?

Well, that's the important part of the book is we're telling the story that it's not the facile version of the story where we flipped a switch in World War II and then bam, the automotive industry started making munitions. It was actually a journey. It took 18 months to retool and rebuild factories to produce war material.

Yeah.

And so you really want to get moving early.

Yeah.

Now we have a lot of this latent capacity. You think about GM produces a new Escalade every 90 seconds. You know, right about now we need some SM-6s, SM-3s, and Tomahawks rolling off the line every 90 seconds too.

Yeah.

And so how do we take the kind of exquisite artisanal approach to a low number of munitions we built and start scaling that out? And you're going to need a breadth of approaches. You're going to need the new entrants building entirely new classes of things. And you're going to need to make the exquisite things much more quickly. What makes this stark is we have 8 days of weapons on hand for a major fight with China. Nobody thinks that's deterrence. Nobody thinks that's enough, right? We need 800 days. How do we really fire up the arsenal of freedom here and get serious about building? And we're going to have to build those things in new ways. And a lot of that skill exists in El Segundo. It exists in the modern American manufacturing economy.

Yeah, yeah. I'm always reminded of this this like Palmer Luckey take about like the younger generation throughout the 2000s got obsessed with building consumer software ad platforms. We love ads, but I take the point. Is there something similar going on right now with AI? Because AI can be important for the military, but also you see there's only a few Caterpillars or electricians and they're working on data centers. If that doesn't become critical to the defense and deterrence of the nation, it actually just winds up being more just juice for the economy, which is probably good. But at the same time, it might be sucking capital, sucking human talent out of true industrialization efforts.

Well, I think so. Like, I think with AI, we have to remember that we have huge human agency. AI doesn't do X or Y. Humans use AI to do X or Y. Let's pick intelligently. Let's pick things that are in the national interest, that give the American people prosperity, that actually propel civilization forward and aren't AI slop and, you know, AI slot machines.

Yeah.

And I think that the promise in front of us is that AI is an opportunity to give the American worker superpowers. How do you make the American worker 50 times more productive than any other worker anywhere in the world? And that solves the math equation of like, how do we reindustrialize economically? This is actually—

yeah. How do you, how do you, uh, how do you rate the current reindustrialization process. There's a lot of founders you mentioned in places like El Segundo that are working as hard as they possibly can. But, but is, are we 10% of the way there for what you want to see? Are we 20% of the way there? Are we 5%? Like where, where do we stand right now given all the effort that has already started? But we're still, you know, early in this process.

Well, I'd say 3 years ago at the first Reindustrialize, there was an aspirational aspect to it. Now I think we're closer to 5%. Like this is happening. We're in the early innings of it, but it's happening.

Yeah.

And I think people are starting, you know, one of the amazing things about the American spirit is people just roll up their sleeves and get busy trying things, you know, and I'm working with people on the factory floor every day who are using AI to change how they do production. One of our submarine parts manufacturers actually added a third shift. They were able to use AI to automate the planning process, which meant instead of having to have tools down while they did planning and quoting, they were actually able to get that done in 10 minutes. They needed to hire a third shift because there was more work to do. And these are the sort of narrative violations that aren't being reported. And I think the underlying phenomenon is that we are listening. So these revolutions are always tools-driven revolutions, not concept revolutions. And the impact of the revolution is determined by the people who wield the technology, not the people who invented it. Galileo did not invent the telescope. He used the telescope to discover the planets and planetary motion. The microscope, the power loom, the personal computer, thing after thing, it's the wielder of the technology that determined its impact on society. Today, we are way over-indexed on listening to the inventors of AI. They're very smart, but just like their creations, they have their own jagged intelligence. And the future of AI is going to be written by the American worker.

Yeah. Yeah. How do you think AI interfaces with the reindustrialization effort? There's also like, yes, use AI in the factory, but I imagine that retraining is a really underrated opportunity. I've already heard just years ago, I was talking to Chris Power about, from Hadrian, about how he was able to hire someone and just get them forklift certified and teach them how to use these things. And reskilling has always been happening, but it feels like we're going through an acceleratory phase of reskilling. But what are you seeing on the reskilling side related to AI?

Well, enormous thing. So I mean, I think Chris has really led the way with Hadrian here. He's gonna have a huge factory opening on Friday. Hopefully you guys get a chance to cover that, Factory 4. The Panasonic Energy, I work with them. So they're located in Sparks, Nevada, inside the Gigafactory. They make every cell for the Gigafactory for Tesla.

Interesting.

The population, your employee base there are prior casino workers. And this is high-end exquisite Japanese equipment. It used to take 3 years of apprenticeship to learn how to be a battery technician for this equipment. With AI, it now takes 3 months.

Wow.

So that's a very concrete example of the reskilling. More profoundly, I'd say, you know, one of the things we cover in the book is the story of Colonel Kukur, the father of MAVEN. And it's the newest heretic. He's a contemporary. He's alive today, and, you know, obviously. And what I think is really compelling about that story is that this is actually the most consequential AI system in the world today. But because it exists in the Department of War, it's not something that broadly the Valley interacts with or thinks about. And I think one of the reasons it's so consequential is the stakes are existential. People are not— they don't have lame goals like, how do I get 10% more efficient or reduce headcount by X or Y? It's really like, how do I have complete dominance and overmatch? And as a result of this, the people who are building it are not just formally trained computer scientists over here, but it's become a platform that the vocational expert, the intel warrant officer, the fires officer is able to really encode their knowledge, build agents that are their kind of team. Working with that to get things done. And so the efficiency, the speed, the scale of what's going on, like really we're learning more from those users today than we are anywhere else.

Yeah. How are you— how are you thinking about the role of the forward deployed engineer in the AI boom? It feels like there's the capability overhang, an incredible amount of, you know, genius intelligence from the machines. And yet there's so many processes that are still manual. I went to the doctor recently and I had to fill out a paper form. And so in many ways, like there's still a capability overhang just from like HTML web forms. And it feels like as amazing as the AI is getting and we're seeing so much progress there, there's still something that needs to fall into place to actually get systems deployed.

I think that's right. There's a huge, I mean, in air quotes mockingly, I'll call it the last mile problem. You know, everyone's, the first 80% of the problem was building the genius technology. The second 80%, it turns out, is actually how do you implement it for economic value? And that's where we're— that's kind of our jam. That's what we do for a living. And I think it's never been more fun to be a forward-deployed engineer than right now, because the speed at which you can take new product ideas that you're learning, systematize them, generalize them— it's crazy. And I, I know Ted's talked about this, but like, we have to reinvent forward-deployed engineering as we go right now. Things that we used to think would take 4 weeks take 4 hours. And so the amount that you can get done, how you go to market with that, It's like, let's just sit in a room with the customer that there's no sales meeting. Sit in the room. Let's go build agents that are automating actual workflows today. By tomorrow, you decide.

How do you think about advice for young people? I imagine that you'd say, you know, come work with you. But also, it seems like there's some potential alpha in being a young person that goes to a Chrysler and says, I'm going to be the AI czar, I'm going to be the Ford fully deployed engineer, fully FOIL deployed, because I'm just going to work at this company. Where are the opportunities for young people during this tumultuous AI revolution?

So I'll give you two answers to that. The first is, what would I tell my children? Like, what should they learn right now?

Yeah.

And really what I would want to cultivate in them is agency, extreme agency. Like, I think all the other skills you'll be able to figure out as you go. But, you know, do you really believe that your human effort can make a dent on the planet? And how do you experience that and live that? Then where would you spend time? I think Palantir is an amazing platform to have impact on the world. The things that you do in the commercial sector impact the government and vice versa. And you have access to the problems. You're in it. But second to that, it's like when you're thinking about being inside of a company, I think AI is going to be the antidote to the managerial revolution of the 20th century. All this power that was sucked away from the frontline worker who actually knew what they were doing to an amorphous blob of middle managers. And even actually they sucked power away from the senior leadership. That's being reversed because all of the bureaucracy is getting cut. The agency that someone has— I was thinking about this because I, in the military, I'm seeing incredible AI application developers who are not formally trained computer scientists. And I was like, what happened? I've been doing this for 20 years. This feels like a discontinuity. Where do these people come from? And I realized like, oh, they've always been there. The thing is like, what would this guy have done 10 years ago? Make a PowerPoint, try to convince some program manager that his ideas were good, only to be told they weren't, knowing full, you know, he's too smart for that. He's not gonna waste his time. Now he just goes away in a corner for 2 weeks and builds it, and he's arguing about something that's empirical, and the commander's like, shit, this works, let's go, you know?

And I think that's like the most underappreciated part of this moment. I mean, we've been covering, we covered a story yesterday of a guy in Australia who's gone on like a year journey to try to cure his dog's cancer, cancer, and he had experience building in AI and ML but didn't have any experience in biology or pharma or any of these things. And he— and just by leveraging the models, he was able to kind of figure out the right path to go down, figure out like even he took a recommendation from ChatGPT of like which professor to go to in their sort of local university system to get help with the problem that he was working on. And he ultimately has been able to show like real results on this sort of experimental vaccine. And you apply that to that type of, that sort of nationwide realization that like, you don't need to be an expert, you don't need to have gone to school for a certain thing, you don't need to be a software engineer to build software, or you don't need to be an electrician to start figuring out how this stuff works. And I think that that unlock across the entire world where just like bringing down the kind of like knowledge boundary around so many different tasks is going to dramatically transform huge parts of the economy.

Yeah, I couldn't agree more. I mean, and talk about an example of agency. Like, he couldn't have started that unless he thought, I could do this.

Yeah, it's going to work. Yeah. Yeah. And like you said, it wasn't like he just one-shotted anything. And that's the point. That's what people need to realize. It's like if you just want AI to one-shot everything, that's like saying like, I want results in life and I don't want to have to work. And it's like anybody throughout all human history, if you want results and you're not willing to put in the work, you're going to have a bad time. Yeah, but like now there's never been a better moment in history to want to build something, to do something and have a better shot at actually achieving that or learning how to do that than right now. And so again, it's all— it all comes down to agency. I guess the question is like, Can you teach agency, or is it innate? I find if I'm talking with somebody about their career, sometimes it's like, you know in your head all the different moves that you should make in order to achieve the outcome. And yet some people just think, OK, I'm just going to go back to submitting resumes that never get answered, because that's the straightforward path. So anyways, we're going to find out if—

He wanted to go back in, remember? You know, he wanted to eat his steak.

Yeah.

And I think, look, obviously there's a part of it that's innate, but there's a part of it that you can cultivate.

Yeah, I agree. Yeah, I was talking to my wife about my 5-year-old last night and was talking about AI and sort of like what he might do for a job and how it is nervous. It's nerve-wracking. It's like, okay, what? Like, what if I try to predict and set him up for success in some particular career? Like, how is that tractable at all? And then I was reflecting on my own career, and I was like, well, for like 10 years, I sold things online. And when I was born, e-commerce literally didn't exist, because I was born before amazon.com and before Webvan. So it was not fathomable to click a button—

You would have been a door-to-door protein salesman.

Exactly. And now I have a live stream which was not a thing before the internet. And so like every career opportunity I've had has been adjacent to other things. People sold things before, but, or, and they talked on, you know, TV before, but the actual shape of the career has been wildly different. And so I felt very relaxed at the end of this conversation, but it is nerve-wracking if you really do want to just think, okay, yes, like doctor, lawyer, merchant chief forever, and it will never change. Even the lawyer's role has changed a ton with the electronic revolution, the information revolution. But yeah, it's just a fascinating time. Where should we go next, Jordy? Do you have—

Any other stories that stand out? And you can give kind of like a trailer so that we want people to go and buy the book. So just give us another trailer. I know there's so many great stories.

Yeah, well, we can talk about Colonel Cukor a little bit more. You think about— so you have this Marine colonel who's just I call them heroes and heretics, you know, because it is somehow their rebellion that gets all these things started. So he had this seminal experience where he was on a helicopter waiting to land on Mount Sinjar to evacuate the Yazidi who were fleeing ISIS. And there was a young Marine who thought he saw a rocket-propelled grenade. And because of this misidentification, this human error, they waved off the landing. It was obviously unsafe. And you have order of 1,000 people who are raped, tortured, and enslaved from this small little decision. And so this is the sort of thing that he was like, there has to be a better way of doing this. And that kind of set him up on this crusade to go figure out how to bring AI to the department. What I think is interesting we document is how everyone tried to kill him in doing this. I mean, up to the point, you know, the services were threatened by it, the bureaucracy was threatened by it. People filed IG investigations. People claimed he was housing Iranians in basement. Here you have this devout Mormon with 4 daughters living in a 1,400-square-foot home that has no basement. They actually sent investigators out to go do this, but just through that, never giving up. And that's what you see consistently when you see Rickover. You know, you see in his memoirs that the humiliation of the women's bathroom, all these slights, it's not that they didn't get to him. You know, he documents how much they did hurt, but despite that, he would push through and get all these things done. And perhaps one of the greatest heretics, Billy Mitchell, who's the father of the Air Force, He didn't even live to see the creation of the Air Force, but his little rogue act of rebellion was the Navy was trying to sink a ship at this— they call it a sink-ex, an exercise to do it— and they were failing to sink this stationary ship. And he's like, you know what, I'm just going to drop a bomb from a plane. And there was no permission, there was no rules. And, you know, you get this like feisty— he sunk the ship. But before then, people thought air power was about sending messages back and forth across the front line. Nobody thought about actually using these as weapons of war. It's totally crazy. And so I think hearing these stories, what I really hope is both the heretics inside and outside the building are inspired because your country needs your heresy right now. In every one of these wars, it really comes down to Churchill and the tank. As the head of the Royal Navy, Winston Churchill funded and built a landship. He could only build ships, of course. That's the tank, because the British Army was like, we got horses. We're good, dude. No thanks. You start discovering these stories, and you get emboldened to say, like, I got to pursue what I think is right here. And you go back to World War II, we built 154 different airframes, different types of aircraft. I think 10 really mattered. But just, you know, in the sense of, like, the American free market system, like, obviously, you can't know. You need a market for competition. And that's part of what makes defense really hard. It's a monopsony. There's a single buyer. People have a penchant for control. I like to quip that that, you know, everyone's given up on communism, including Russia and China, except for Cuba and the DOD. You know, we have this deeply centralized planning approach that we thought would provide for what we needed to win wars, and it's just not the case.

That's a good hot take. Do you think that the next batch of defense tech companies should go public earlier? Like, what advice do you have for the current crop of private defense tech companies? Companies. It seems like there is appetite in the capital markets. Palantir's obviously done very well in the capital markets. At the same time, the private markets seem to be able to find any amount of money in the couch cushions, especially if there's AI attached to the narrative. But how, how are you talking to leaders of private defense tech companies right now about the markets broadly?

My advice to them is all the same. I think one of the hardest things about doing the defense tech thing is is you need to hold two contradictory ideas in your head. One is like you need to run towards the pain. Like, you know, you need to run towards proving real results operationally. But that's not your buyer. And so, you know, you could, you could say that there ought to be a mark-to-market moment right now. Who's in the fight today in Iran?

Sure.

Where have these companies started bending the curve? What are the opportunities to prove this capability? You know, I'm hearing about incredible things Shield's doing in Ukraine right now.

Sure.

That's really important. You're not gonna get paid for that, but that's the validation you need. And then you have to figure out how to get programs of record, all those sorts of things. But if you just focus on the programs of record, if you just focus on treating the Defense Department as a buyer, you'll lose the magic. You'll lose your own heresy that leads to the innovation.

Yeah, there was a little bit of that in the space economy where we saw there was pretty quickly a bifurcation between space companies that were doing doing a lot of interesting work and signing deals. And then other space companies that were like, they got on the rocket and they went to space. And I feel like that was like an important binary. I don't know how the binary is sort of coming down now as more companies get to space, but it felt like actual deployment where the rubber meets the road, it's always important in startups, but it feels especially important in this scenario.

What about stories from history around copying what works? Does anything stand out? And the reason I ask is because one, it was somewhat surprising seeing that we have American-made versions of the Shahed drone. And I feel like America has always prided itself on being inventors, you know, these sort of like zero-to-one, zero-to-one projects, right? You said 100-something airplanes created or airframes created during World War II. And yet here it felt like the smart decision was like, hey, this is like a battle-tested, form factor, like we can just make the thing that is delivering results on the battlefield. But are there any other kind of stories that stood out around America kind of swallowing its pride and saying like, hey, this thing has worked, let's fast follow?

The best one that also speaks to the importance of founders and people, the primacy of people, is Operation Paperclip. You know, as the Nazis were losing, we started, we actually had two competing programs. We had FIAT and Paperclip. Fiat's theory of the case was, we don't need these people, these dirty Nazis. We're going to go steal all the technical papers. And we're just going to be able to figure it out just by having all the papers. That was an abominable failure. It did not work at all. Instead, what worked was you go get the founders. You get Werner von Braun. You get the people who actually know. Because there's something more three-dimensional to the knowledge than what's on just the 2D paper. And I think that's a huge swallowing of pride. We had to really hold our nose to these Nazis. Recognized that actually it delivered ICBMs, it delivered the space program, it delivered a fundamental offset against the Soviets. But we did that other times as well. I think probably the most famous one is there was a North Korean defector. I don't have the dates exactly right. He flew out of North Korea on his MiG. The MiGs were killing—

Oh, his personal MiG.

He flew out on his MiG. They tried to shoot him down. He escaped. Fortunately, we didn't shoot him down. We figured out he was defecting. So we reverse engineered the MiG and built our own. I'm forgetting, was it the F-86 or the something like that? And that actually helped us restore air dominance once again. So, you know, it is the primacy of winning, whatever it takes to win. It's, you know, you don't want to be like, hey, I didn't steal anything from the adversary, but I died nobly.

No. Yeah, that doesn't make any sense.

Yeah.

Well, thank you so much for taking the time. Congratulations on launching the book.

Yeah, I can't wait. I can't wait to get into it. We're going to have to— we'll share our copy.

We will.

We'll fight over it. I think there's two in here. Maybe you got two.

But thank you so much.

Congratulations on the launch. We'll talk to you soon. Great to catch up.

Goodbye.

Cheers.

Let me tell you about Graphite, code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster. And let me also tell you about Restream. One live stream, 30+ destinations. If you want to multi-stream, go to restream.com. The timeline is in turmoil. Doug O'Loughlin at Fabricated Knowledge is going back and forth with bucco capital bloke over Google. So Fabricated Knowledge says, "People keep thinking that their distribution," think Microsoft or CRM, "is bigger than the technology, AI. But the fact that people are buying Macs to mess around with Claude Code and other coding agents, they're also good machines," well, tells me that this technology is much bigger than any distribution this time. Like whatever distribution is there, if you're willing to go to the Apple Store, buy a MacBook Pro or a Mac Mini, wire it up, install OpenClaw on it, verify everything, like you're willing to jump over the default of the distribution.

Yeah, I think another question here is how many small businesses in America don't pay for any AI products directly? And yet their employees are bringing the product to work.

Yeah.

And it's just not appearing on the, on the actual—

but Google Capital bloke fires back. He says these people are dweebs. It's like 0 point— it's like 0.000001% of the market that's buying Mac Minis. In his opinion, distribution does matter. It will always matter. It doesn't mean the incumbents will win, but it certainly gives them an advantage. Just look at the sentiment shift on Google in one year. That's a good point. Doug fires back. He says, "I am going to just put that distribution friction is the only thing that matters if we are going to be software brained. I do not— I do understand and appreciate it, but I do think this is a big thing and making it all about distribution is such a technology loser way to look at it. Lots of great tech companies with awesome distribution lost. Like IBM had CTO relationships with every company. LMAO. Buco says, yup, I am not pounding the table for Salesforce, but they're not dead yet. The long game. Also, IBM beat QQQ over the last 5 years by a lot. I had no idea.

Let's give it up for International Business Machine.

And they keep going back and forth, but let's give it up for IBM and let's give it up for our guest from Cybersecurity starts. How do I pronounce his name? Let's bring him into the TVP and AltruDome.

What's going on? Great to meet you.

Help me pronounce your name. I don't want to get it wrong.

Gili. Hey, John. Hey, Joji.

Thank you so much for taking the time to join us. How are you doing?

I'm doing great. I'm speaking with you from Miami. We just finished the first day of our annual conference, CyberSparks. Yeah, we had a lot of fun, lots of terrific guests and speakers. We just had Nikesh Arora from Palo Alto Networks, Joe Kurtz, CEO of CrowdStrike.

We love Joe Kurtz.

Doug Leone, my friend from Sequoia Capital.

Royal flush.

The founders of Wiz, a small company that you might heard of. Completed a tiny acquisition last week.

Yeah, congratulations to you on that one. Yeah, great, great outcome. Uh, yeah, what is— what are people talking about? What is top of mind for everybody at the event?

Well, the whole idea of CyberSparks is to bring together the top 300 leaders in cybersecurity. So you've got here about 100 executives, you know, CEOs and founders of the top cybersecurity companies in the world and the top practitioners, you know, chief information security officers of Fortune 500 companies. And the whole idea is to work together and talk about what's next and how we can work together in order to deal with what's upcoming. And there's lots of risk and you know, an expanding threat vector. You probably talk a lot about artificial intelligence and the new capabilities that it brings in your show. But with every technology wave, you know, think about internet, think about cloud, there are new risks introduced. So we kind of spend a couple of days together to talk about you know, the future of cybersecurity.

Is there, do you think there's more fear or excitement around the technology shift? Because there's obviously this new kind of new threat vector, but at the same time that creates opportunity for the industry for product expansion. And so it's go time from a business standpoint to kind of meet the threat. But also I'm sure some people in the room are a little bit scared of what's coming?

I think there is mostly uncertainty about where this thing is going. You know, what we've seen in the past 12 months is accelerating. And there's— there are so many things that we simply don't know. You know, I wish I could stand here and give you all the answers. But in the room, we had a lot of conversation around the uncertainty. On one hand, and the pressing need to make some decisions about safeguards to make sure that, you know, the Arnold Schwarzenegger Terminator movie, you know, doesn't become a mild, a mild story relative to the reality in 10 years. And I think that what we are going through right now, we call it or I call it technological doubling. You know, if you think about technology used by us as a race, as human, you know, and you think about, you know, what we have today as the 100%, and let's say the zero is the invention of writing, like 4,000 years ago when we lived in caves and started to write in order to accumulate knowledge. We can build tools, then the 50% is probably at the point in time of the invention of the steam engine, which allowed us to build machines that would perform tasks too difficult for man. So the last doubling took 170 years. The next doubling would take place in the next 25 years. That means that all of us, all three of us and everybody who listened to or watched the show, are the first people that would witness technology doubling within their lifespan. That never happened.

That's an exaggeration.

No living person has seen doubling of technology. And you know what? When you just follow the curve, you know, the curve I just described to you, That means that in, you know, 25 years we'll be at 200%, but in 100 years we'll be at 2,000%. This is just math. It's not a prediction. This is just following the math, meaning that people that would live 100 years from now would look at our technology in the very same perspective we look at the technology used by our ancestors living in a cave drawing something on the wall. That's exactly the same. So we are going through a radical change in technology that we have never experienced, and that would require dramatically different type of thinking about, you know, safeguarding the new capabilities. So because this is enormous opportunity ahead of us, you know, AI would change everything. About healthcare. It would change everything about education. So there's an enormous opportunity, conditioning under the condition that we can control it and that we are not losing control. So that was the core of the conversation.

Yeah, very cool. What is, uh, what is the shape of the cybersecurity threat landscape right now? It feels like there's entirely new capabilities when I think about like like a phishing attack with a deepfaked voice on the phone. That's like, I guess you could get an impersonator, but it's sort of a new threat area. But then there's also just like hammering a coding model to spam a whole bunch of like SQL queries or SQL injections, like all the old stuff, but just multiplied. Like where are you seeing the biggest new threats emerge or what are people discussing on that front? Front?

It's all over the map just because of what you said, because it becomes very asymmetric, even more asymmetric than we used to see, because the guys on the offense, you know, threat actors, they have a huge advantage because they can simply take a cloud and just apply it for new attacks while the pace of adoption of AI tools and machines within, for defenders within large organizations, within the enterprise is significantly slower. So that makes, you know, that puts AI in the hands of the bad guys much faster than it puts it in the hands of the good guys. And if that would be the case for long, then the bad guys would win. And the bad guys today are really bad. You know, you look at state-sponsored attack, then, you know, everything that's going on in the world, then that's a real risk.

Is there a bit of like a white pill here in the sense that because there's a lag between open-source capabilities and proprietary systems, and then you also have the big frontier labs, DeepMind, OpenAI, Anthropic, like they're definitely running agents over there, user bases and their APIs to know, hey, this person just spent $5 million on our API and it's all cyber attack related prompts. Like, let's maybe turn them off or figure out what's going on over there. They have a huge incentive to sort of, you know, control their customer base so that their customer base is not using these tools maliciously. The hackers sort of wind up on the lagging edge, not on the front frontier, but all of the cybersecurity companies like Palo Alto Networks, like CrowdStrike, like the folks that you've had at your conference, they maintain access to the frontier. And so they're always fighting with a bigger weapon. Is that sort of the equilibrium we should expect here?

That's a great question. And by the way, we did have today the two top cybersecurity experts at Anthropic, you know, the head of security and the head of product security, sharing the roadmap and thoughts about the upcoming capabilities of Anthropic and other platforms. I think the answer to that is our continuous investment in innovation in the space. It's not just about Wiz that I mentioned or Sierra. Those are large established startups. But we did have one company going out of stealth last week, Onyx Security. All they do is agent security. We had today a major launch out of stealth for another company, Surf AI, which takes care of, leverages AI to complete tasks tasks much faster for organizations. So there's definitely, there'll be a battle between machines, machines that are used by threat actors and machines that are used by the good guys. And we obviously, our job in this world is to identify the best teams, the best talent and help them build and utilize AI. So we live in a safer, better world. And that's So that's the plan.

What's the general sentiment from guests around the actual competitive threat of the labs releasing various agents and security-focused products? Obviously they feel threatened when they announce new products 'cause they tend to send the stock price down, or they have over the last month, but is there a real sense that this is a threat the labs could threaten their business models, or is it more just kind of frustration with, with how the market perceives the threat?

We always overestimate the impact of disruption in the short term, and we always underestimate it on the long term. So I think that what we would see here is that in the next 2 years, the frontier guys, you know, they have different focus and different priorities. So their impact on cybersecurity companies would relatively be low. But in 5, 7 years, I think that they, they are, they have a chance to take over. And I think that we would see 3 cohorts, 3 groups of competitors. You would see the traditional cybersecurity platforms, you know, the Palo Alto, the Wizz, the CrowdStrike of the world. You'd see the cloud providers, you know, Google, Amazon, Microsoft. And there'll be a third group, which is the AI platforms. And, you know, we shall see who would hold keys for cybersecurity. And, you know, that's, you know, if you, if you ask me the same question 12 months ago, I would not even mention the third the third group. So it's really hard when technology moves at this pace. It's really hard to make predictions. And I'm not afraid of making a fool of myself. I do that almost every day. But, but making predictions right now is guaranteed to make a fool of yourself.

I was just going to ask you for another prediction, but I guess I'll table it for next time. Next time. Thank you so much for joining us.

I can share with you the predictions I made today in a closed room.

Please.

But maybe I'll start with the predictions I made 12 months ago. 12 months ago, I told the room that my prediction— and keep in mind, it was really pre the big wave of AI— I told them that I expect to see a million-dollar ARR company, a million dollars in revenues company with a single employee within 2 years. And I told them that I expect to see a $100 million revenue company with less than 100 employees within 2 years. I was wrong twice because those two things materialized in less than 12 months. You know, we've seen, we've seen Base44 with We have seen one founder who reached $3.5 million ARR before acquired by Wix. And we have seen Cursor reaching $100 million of ARR with less than 20 employees. So things are accelerating. And therefore, the predictions I made today were that I anticipate that we would see Fortune 100 companies with a security, a cybersecurity group of less than 10 employees. Keep in mind that today those companies employ thousands of employees. So I believe that we would see AI making huge impact on cybersecurity and would turn cybersecurity from a profession into function. That would be a major shift in the market.

Yeah. Makes sense. Well, thank you so much for taking the time to come chat with us.

And we're looking forward to seeing how these predictions play out.

My pleasure.

We make predictions all day long.

Risking it all.

We're in the random prediction, off-the-cuff business. And I completely agree with you. It's extremely hard to make predictions that hold for any amount of time right now. Right now, but it's an exciting time.

I love it. It's entertaining. Thank you so much. Really enjoyed it.

Yeah, give our, give our best to everybody at the event. Great to hang out.

Have a good one.

Cheers.

Goodbye. Let me tell you about ElevenLabs. Build intelligent real-time conversational agents. Reimagine human technology interaction with ElevenLabs. Let me also tell you about Gusto, the unified platform for payroll, benefits, and HR built to evolve with modern small and medium-sized businesses.

All right, someone in the chat shared this earlier, but it is important to the Manus deal. Oh yeah, in the New York Times: China ramps up scrutiny of a Meta deal. The country appears to be cracking down on people linked to the acquisition of Manus, well, Singapore company with Chinese roots, as President Trump prepares to visit Beijing. Chinese government is taking actions to penalize people linked to Meta's $2 billion acquisition of Manus in an apparent effort to discourage Chinese AI executives from moving businesses offshore. It sounds like they want to make an example of them. Officials at China's National Development and Reform Commission, a high-level ministry that oversees economic planning including AI, called in Meta and Manus executives for a meeting late last week to express concerns about the deal, which was announced in December, said the people who declined to be named publicly. The scope of the Chinese government actions remain unclear but appears to include an effort to restrict Manus executives from departing China for Singapore. Beijing has issued exit bans in the past for corporate executives who are under scrutiny. The transaction complied fully with applicable law, said Andy Stone, a Meta spokesman, said in a statement. The outstanding team at Manus is now deeply integrated into Meta. He added, we anticipate an appropriate resolution to the inquiry. Manus did not respond, but they are owned by Meta and Meta did. So anyways, not surprised to see China frustrated that one of their great AI teams just kind of bounced. Poached. Poached by the Zuck. But who hasn't been poached? If you haven't had somebody poached from Zuck—

If you haven't had Zuck come to town, you're not doing something good.

Work harder.

Exactly. Chamath Palihapitiya chimed in on the timeline. He said, what if AI doesn't need to show an not immediate ROI, but instead is the plausible deniability companies use to riff 50% of the workforce they already knew did nothing. And expatnon says, the real question here is why is an allegedly cutthroat hyper-capitalist economy with every large white-collar firm maintained by a workforce 3x the size it actually needs to run its operations Why would they stop now and not before? And that is a good question. Like, the private equity has been, you know, trying to find the right size for every company for a long time. I think there's like natural bloat that happens. We were talking about the question of like, you are an AI company if your revenues are accelerating. It would be very interesting to see what companies ramp up hiring this year. Obviously, like the startups and high-growth companies are. But of the older school economy, what management teams are guiding towards more human capital needs, that should tell you a lot about where that management team sees the business and economy going in the AI era.

Great story here. What happened? Formula 1 chief Bernie Ecclestone, the 80-year-old billionaire, was badly beaten up in a brutal mugging outside of his night bridge office last month. I believe this is a— this is certainly a historical piece. But undeterred, he allowed his bruised face, complete with an impressive black eye, to be used in an ad for an exclusive Hublot brand of Swiss watches last week with the slogan, "See what people will do for a Hublot." That's great. Gnarly.

Bernie Ecclestone got mugged and his reaction was, How can I turn this unfortunate event into money? That's remarkable. Yeah, there's a number of crazy Bernie Ecclestone stories. The F1 acquired episode goes into a lot of Bernie Ecclestone's history and stories there. It's a remarkable episode you should go take a listen to. Let me tell you about Finn.ai, the number one AI agent for customer service. If you want AI to handle your customer support, Go to finn.ai.

Noah Smith sharing some unfortunate news. We certainly hit the gong when the GDP numbers came out in Q4 of 2025, but the AI productivity boom story is gone, at least for now, according to Noah. Instead, it's all just AI CapEx. Data centers are the only thing keeping our economy afloat. Of course, the other thing keeping our economy afloat is all the economic activity that is still there despite a lack of excessive growth. Yeah, but certainly data centers are making the overall picture a little bit rosier.

Makes sense. Let me tell you about CrowdStrike. Your business is AI, their business is securing it. CrowdStrike secures AI and stops breaches. Tyler, have you had a chance to fire up Manus again? Take it for another spin?

I mean, I had it organize my desktop.

Okay, was it effective?

Yeah, it did a good job.

Because you look like a man with an organized desktop now.

With a suit like that, can't afford not to have an organized desktop.

What is a good benchmark these days? We need a new benchmark like our comedy bench, like our What is it?

The—

Shrimp fried rice?

Shrimp fried rice bench. We need a benchmark for desktop, something with a very disorganized desktop that would be difficult.

I mean, I feel like that's not a good benchmark because that's like not—

Yeah, arbitrary.

You don't need to use that very often, but it's something like interacting with actual applications. Like, can you open Premiere and edit a file? That's a great computer use benchmark.

Yeah, yeah.

We're very far away from that.

Yeah. When I was thinking I was thinking like I would like to be able to pick a song and then have it go and find stock footage, AI footage, movie clips, and cut together a vibe reel to the beat like I see on Instagram from just a prompt. If I just have an idea of, oh, I like this song, I would imagine that this song with this footage would go really well together. That's like several hours of work. I could make a lot more of those videos, have a lot of fun with those. MetaVibes is a little bit of that because you pick a song and then you can generate one Midjourney image and do some light animation on top. But it's not truly finding iconic footage from around the internet. And that feels like something you could do in OpenClaw. You could do with a Cowork product or Manus.

Yeah, because I mean, you can just use like FFmpeg. FFmpeg, and you can cut videos down from the terminal. They could do it without even using computer use technically.

Well, the models could do that. Is that what you're saying?

Well, I'm saying like Cloud Code could do that, right? You can tell Cloud Code to edit a file and edit a video from this type of—

And it should download FFmpeg, do it, but who knows how good it is yet. Okay, that's our new benchmark. Music video driven live reels. Let me tell you about console.com. Constable builds AI agents that automate 70% of IT, HR, and finance support, giving employees instant resolution for access requests and password resets. And without further ado, we have Anna Patterson from Ceramic AI here.

What's going on?

How are you doing, Anna?

Good to meet you.

Great to meet you.

Yeah, you too.

Thanks so much for joining us.

Are you guys having a good St. Patrick's Day?

I see the green. Yes, we're very—

We're greened up.

Yes, we have these bright green suits from a show we did on Black Friday about Shopify. And I thought, certainly I will not be using that until next Black Friday, but here I am.

Here we are.

Well, happy St. Paddy's Day to you. Since it is the first time on the show, please introduce yourself and the company.

Hi, I'm Anna Patterson. I'm a founder of Ceramic AI. I was a longtime Google engineer. I started in 2000 and 4, and I'm best known for building large search engines.

That's amazing. So, yeah, give me the pitch for Ceramic AI.

So Ceramic brings the cost of search in line with the cost of inference. As you know, inference costs have been going down and down, and actually inference is faster and faster. But search is $5 to $15 per 1,000 searches. But inference is maybe $0.50 per 1,000 searches. So the kind of analogy I like to use is tacos and salsa, right? Tacos is kind of the meat, the meal, and that's kind of what inference is. It's the thing that's really delivering intelligence to your application. But salsa kind of makes it better, right? But search is now the dominant cost. In tacos and salsa, you're adding search, but it's $5 to $15 per 1,000 queries. So it's really kind of time to bring salsa in line with tacos. So ceramic is $0.05 per 1,000 queries.

Amazing.

Right. So you're saying this historically, it was like you were getting a taco and it costs you $5, but then they were like, well, if you want salsa, it's going to cost you an extra, like, you know, $500.

Yeah.

And you're like, well, I don't— not sure I really want salsa.

Weird, weird.

Exactly.

Exactly.

Okay.

So help me define define— help me understand what search means, because search can mean over the internet that's already sort of baked into LLMs. It can mean web, active web search, like in a proprietary database or just the open web. How are you thinking about the surface area of search?

Yes. So we do have a 40 billion page web search, and that is the open web. And then we've built proprietary systems as well. So one of the things that we're announcing is this idea of supervised generation, that as the model is generating, it's double-checking what it's saying with search. And it's also double-checking with search, what else should I say to make a comprehensive topic? And so that way, you can really enable new applications with 10x fewer hallucinations. But also make the whole product affordable and fast.

Okay, so who do you sell? Sounds super valuable. Who do you sell this to? Is this going to be maybe similar to kind of like the data labeling market where there's like 5 customers that like really matter and you want to get all of them? Or are there a bunch of other applications that you want to actually sell to vertical-specific AI applications that can— vend you in, in like kind of in line with their LLM products?

Yeah, I think there are, um, kind of two strategies there. Um, you mentioned one of getting all the big players. Uh, that definitely would be nice, but we also have a self-serve, every agentic workflow. We have one startup, their, uh, agentic workflows do 1,100 searches So this, um, our search engine responds in 50 milliseconds.

Wow.

So it's both affordable and, um, more real time than the other search engines. So, um, we see a number of agentic workflows happening. But to your idea of, um, a custom index or a custom application, we see that as well because let's say you're a pharmaceutical company or some banks are very privacy-centric they don't want their searches going out and they don't even want their searches to models going out. So they kind of host their own copy of a model. And here they'd host their own copy of search as well so that they get their own proprietary environment for all their agentic flows inside their enterprise.

Makes sense.

Talk about just how the open web is changing in the age of AI. I've seen some crazy stats about how much more of the internet Googlebot sees because everyone has been indexing and been very friendly to Google for a very long time. Other publishers are getting more closed off. How easy is it to actually search the internet broadly these days?

So the 40 billion pages that we have are available on the open web. We do not disobey, or I should say, we obey robots.txt. So we don't actually crawl news sources that have blocked us, but we are in active talks to make deals to them and to have a proprietary API that costs more but also reflects back revenue to those proprietary sources.

Cool.

Are you— is there any value in having— like, I've always been interested in the flip side of Google Search, which is Google Alerts, where the search is happening internally and then the information's actually getting pushed to you. That product is like probably like 0.0001% as important to Google as Search, but it's always been interesting to me. Is that interesting to you? Is that relevant in the age of AI? Does anything change about that ratio going forward?

Well, one of the interesting things is that Inference, a lot of times with these MOE models, inference has a lot of spare compute because it's memory bound. And so with that, it means that it could be thinking. So as it's inferencing, it could be getting a stream of search results, searching all the time, and actually bringing you sort of like alerts, only the most interesting information or information that it doesn't think that you already know. By looking at your history. So I think it's going to enable a lot of new applications.

Amazing. Jordan?

Wildcard question.

Hit me.

How long until we see ads in Gemini? There's been some reporting this week. Obviously, Demis had come out and said, you know, why would we put ads in Gemini?

For the record, on this show, we are extremely pro-ads.

We're extremely pro-ads.

We love it.

Exactly.

Yeah.

And I, and I, We both expect ads to be in Gemini. I would say this year is my guess, but—

That would be my guess as well. Before the end of '26. Yeah.

Let's go.

Cool.

That gives me a lot of hope.

We're going to be very excited.

And excitement.

Nice.

Faith in humanity restored.

What's the story of the pink guitar on the wall there?

Yeah. Well, if you want to do something really humbling, learn an instrument from your children.

Oh, okay.

They're absolutely brutal with telling you to practice, everything you're doing wrong. And so I kept borrowing my daughter's electric guitar after she taught me acoustic guitar. And so she decided to get me my own because she saw her guitar, like, laying on my couch, and she said a guitar should be hanging on the wall. And I said, I think I've been told to clean up my room by my child. So my work here is done.

That's amazing.

Well, thank you so much for taking the time to come chat with us. Have a great rest of your day.

Great to meet you. Excited to follow Ceramic and we'll talk to you soon, Anna.

We'll talk to you soon.

Thank you.

Goodbye. Speaking of ads, let me tell you about AppLovin. Profitable advertising made easy with Axon.AI. Get access to over 1 billion daily active users and grow your business today and without Without further ado, we have our final guest of the lightning round, Jake from Gecko Robotics. Jake, how you doing?

What's going on?

Doing great.

Thanks so much for taking the time. I don't believe we've met, but I've heard about Gecko for a long time. I think Trey introduced me to it when I was at Founders Fund. But since this is the first time on the show, I'd love a little bit backstory. Like how'd you get into the industry? How'd you start the company? And then we can kind of get up to speed on what's happening today.

Yeah, well, Trey's awesome and I'm so glad he, who originally spoke so highly of us, I'm sure. So I founded the company about 13 years ago out of a college dorm in western Pennsylvania, about an hour north of Pittsburgh, Pennsylvania. And I was studying electrical engineering, really wanted to figure out how things like energy was created. And so I went to a power plant in Franklin City, Pennsylvania. For those history nerds out there, that was where the first commercial oil rig was drilled and got to see a power plant, how it was made. And I dove in headfirst through this little manhole that I could barely fit through. And I got to see what a boiler was. And so that was like a football field-sized room that was completely covered wall-to-wall with these steel tubes. And this whole job of this boiler was just to get really, really hot. And so, you know, shot water through it, got really hot. Anyway, this boiler kept on having failures. 30% of the year it was shut down because of catastrophic failure that was occurring because the pressure vessels would keep exploding. And the only way to stop it from exploding was a guy in a robe 100 feet up in the air, you know, trying to figure out where the next explosion was going to happen, and that just wasn't working. And that guy actually died that year before doing the job of gathering datasets by hand in the real world. And so I was like, my gosh, like, where is, where is the tech innovation like for these guys that are making, making sure that our homes stay heated? And, you know, just began to look more and more at this, like how we understand the health of the built world and, and also what kind of technology exists for these sorts of heroes that are, you know, that are hidden behind the scenes, if you will, whether they're a port engineer or they're a boiler engineer and folks that are just like helping us do all this stuff. And Silicon Valley, for the most part, forgot about them. And I decided to build a company that was specifically dedicated to helping these guys out.

Yeah. So I'm imagining a big vertical tank, sort of like what you might see at a brewery, but instead of filled with beer, it's water that's boiling. Why not just fly a drone up and use video camera and and just be close. Why did you choose this? What decisions did you make technically, and why did you make those decisions?

You basically, if you're diagnosing the health of built structures, whether it's a boiler, a pipeline, a bridge, whatever it is, inside of a ship, you got to actually get close to it just like you would for a sonogram. So you use jelly, and then you use ultrasound to see inside of a belly for the pregnancy example. Same kind of idea. Use ultrasonics as one way of gathering information datasets. You know, what you're seeing is some of the electromagnetics that we developed, the sensors, and the robots are just the vehicles by which we get sensors around places that are typically hard to reach. And then also localizing and seeing like where to track that year to year to year to understand how do you predict into the future. I mean, this idea of creating the Minority Report for the built world was kind of this idea of predicting, you know, what death or catastrophe was going to happen for like built structures before it did. And the precogs in this case are these robots. Yeah. You know, but all the data is being collected. Collected and then like fed into the central source of truth, which is, it was a cantilever. And then we sell cantilever, you know, as our way of predicting and preventing catastrophic failures for the built world.

Yeah. So what's the shape of the business? It feels incredibly dual use. I mean, we just saw a video of the robot crawling along what looked like an aircraft carrier or battleship, but I imagine that the oil and gas industry, the industrials industry, like there's huge demand for this. What's the mix of the business and what's What are the best practices for working with both the government and private businesses?

Yes, we started in the energy sector, in those power plants. I actually bootstrapped for 3 years, and then was down to $100 to the name. I know, it was bad. I was like homeless, living with a best friend, sleeping on the floors, really roughing it. And, you know, I was out in Pittsburgh, Pennsylvania. There's no VCs out here. I got an acquisition offer and then the folks over at YC in 2016 were just like, you're going to build a huge trillion-dollar company and you're having incredible success. Come out and do that. That's awesome. And so I was like, I'm already poor.

This is an amazing reason.

I already know what the worst feels like. You know what I mean? So it's like, what could be worse than this? And so just decided to do that. Got one up to California and yeah, we're in this YC batch. It's like really just like black sheep of the batch working in energy and robotics space in Pittsburgh. For a first-time founder. So, you know what I mean? It was just like, it was kind of wild and crazy. Then we came out like one of the top companies, but we started in the energy sector because that was just what we knew.

What was the actual first company? Because like energy sector could be, you know, Power. Oh, meeting with the CEO of Exxon, or it could be like the local guy who just wants to buy these as like a prosumer tool almost or something like, what was that actual deal like?

It was this group of power producers called IPPs. So independent power producers. So these guys were just like, they actually don't have these massive contracts utilities have that they can just rest and just pass all their losses down to me and you to pay more bills if things just blow up and don't work. And so these folks actually have to be making sure that they make stuff. And so in PA, there's a lot of them actually in the tri-state area. And so we actually get a lot of access. So the first 3 years I was bootstrapping, I was every single day at a power plant for the most part, trying to to build the robot actually in the environment. And that was a really core thesis and core principle for the company. And we kind of still hold that today of build technology in the real world, not in the labs. Actually, one of the most prominent investors in Silicon Valley said, don't leave, build this in a lab, make it autonomous, and then launch. I was like, that's fucking stupid. That's so bad. Dumb idea. And today he's like, you know, top 3 investor in the world, you know, saying the company's worth like $10+ billion. I was like, me and my co-founders—

Just one shot it.

Don't iterate.

Don't do—

don't, don't, don't talk to your customers. Trial and error. Just one shot it in the lab.

So anyway, so power, man, it was like we— I moved back to Pittsburgh because it was close to these customers. And so that was just like the core ethos of the company was build tech for the people in the environments. And so that led us to oil and gas next. We could apply the same kind of technology to diagnose the health of structures that make the oil and gas assets go. And then it was mining and metal manufacturing. Then I began to get into things like building and manufacturing for building ships or submarines, injecting the same kind of tech and robotics and software there because we have the most data about the health of and the material science of the world. And so applying it for actually building and welding. And then the Navy side, what we're doing there is helping to achieve readiness. 80% readiness was Secretary Phelan's objective. You know, right now it's about, you know, 2 out of every 5 ships are stuck in dry dock somewhere, and that's a global issue. And so our technology allows for us to be able to make up that difference by getting so much information, you know, in some cases 2 to, you know, 2, 3, 4 months faster to be able to get these ships out of dry dock in time and then begin to plan it for the future. It's just like idea of if you always had a living, breathing understanding of the health of these sorts of assets, my goodness, Maybe you'd never have to shut down. That's the thing that I'm trying to build. And you can see now some really large energy companies, like the ones that we work with, are beginning to adopt this. The data doesn't exist. In order for the data to be able to drive AI models to actually be impactful, we actually have to go out and gather information and datasets. And oh, by the way, maybe one day, maybe not into the distant future, you can actually begin to augment these very commoditized sectors and industries that are very capital intensive with robotics-native, AI-native approaches and operating systems that make commoditized industries less and less commoditized. And so that's the kind of vision that we're trying to build, being the company that's very pragmatic in its results and aren't just promising in 5 or 10 years all this impact, but actually delivering the thing today.

Where are you—

I'm assuming the same kind of VC that told you to just one-shot about the product in the lab, make it autonomous, and then go to market would also ask you, are you doing anything in data centers? Can you, can you get any, are there any tailwinds there? I can imagine, I can imagine one of your robots just crawling around a data center, but what are you seeing on that front?

Yeah, well, it's a good question. You know, what ended up happening was we ended up really putting a lot of effort and energy into customers where when things weren't working, it was extremely painful expensive. So think of oil and gas. When you're down for a day, it could be $30 million you're losing. Or if you're down, if your ship isn't patrolling certain places in the Pacific or Atlantic, that's really expensive and harmful too. So that's where we focus. But what's happening now is all this attention on AI infrastructure has actually put a large attention on how efficient can you run your power plants, in this case, a lot of natural gas, and then also how reliable are those assets. And also the assets that were depreciated, not invested in for 10, 20, 30 years because banks wouldn't fund continuation of putting capital into coal facilities or even natural gas facilities because of these carbon strangleholds on these companies. It ended up creating this really interesting opportunity for companies that are really good at understanding the health and the value of assets and can convert them into something really much more valuable, something that we're taking a very close look at. So anyway, power and the ability to make a power plant run more efficiently and also more reliably is core competency of GECCO for a long time. And so we're putting a lot of efforts into that. You'll have a big launch in the middle of— for the 250th anniversary and coming out on July 4th on the power production side. So it's, you know, we've got something that's, you know, if we pull off what our, you know, 6 or so trials have proved, we might be able to in the thermal fleet be able to increase by 15 to 20 gigawatts the amount of power in the US without even building a new power plant.

Whoa.

It's this kind of efficiency, man, that is possible.

Yeah, it's all right.

We said 30 to 50 gigawatts.

We have a $71 million—

I know, let's hit the gong.

Let's freaking go.

Where's the mallet? That's a big number.

Hit it hard.

Oh my God.

I feel the vibration.

Vibrations over here.

I know, we're feeling it. I wish you were here.

For a much dumber question, how do you actually stick a robot to a vertical surface? I mean, gecko, I imagine suckers, but are magnets involved? I imagine that there's some surfaces where a suction cup won't work because maybe it's stuck out or something. Do you have multiple tools in your toolbox? Talk me through that.

Yeah, great question. Well, I really was desiring to use nanofibers just like a gecko would use. Unfortunately, we can't live up to the biology. And so, we actually use neodymium magnets.

Why not just get hundreds of actual geckos?

Yeah, yeah, yeah.

A chariot of geckos.

It's a biotech company actually.

That's a great idea.

So, you use rare earth magnets, is that right?

Yeah. So, we put them in a Hallbach array instead of wheels to optimize pull force, so magnetic force into the Typically, like most of our infrastructure is carbon steel, so it's magnetic. If it's not, if it's concrete or if it's stainless, then we'll use— the best adhesion is actually suction for non-ferrous materials. Like, you were right. Okay, so it's actually just like creating a really great vacuum. Vacuum, yeah. And so that can actually—

so you have an air pump that's sucking in air and creating that, like, little—

yeah, literally just a Dyson vacuum, like, you know, head and And then you create like a nice chamber, you know, for, you know, for you to be able to stick. Unfortunately, it's like, it's not great for dusty environments, but like, you know, you can— so it's like, it's less applicable, but also there's less cases. Like, so for the food and beverage space, a lot of stainless steel. And so there's not much dust there either. And so you can actually apply tools there.

Okay, that makes sense.

But it's the same kind of concept too, which is like, you don't want to crack, you don't want a failure to occur.

And you don't want to damage the material you're inspecting with some crazy, you know, rock climbing shoe with spikes on it. You're not going to be ice climbing up the side of something that's going to—

Totally. So actually, funny, a really, a really like geeky cool like thing that we built for the Navy was you typically have like, you know, when, when Maverick lands his like fighter jet onto the plane, onto the, onto the, um, the aircraft carrier, he's landing on this like, um, um, this, this coating, this non-skid is what it's called. It's like really rough and grainy. You have to actually remove all that stuff every like 3 years or so to be able to evaluate how healthy, you know, is the platform, is the flight deck So there might be a crack underneath that and you got to take all that off.

Got it.

Totally. And so just to even know, like, you know, how healthy the structure is underneath it.

Yeah.

So we've developed technology that actually, you know, uses electromagnetics to excite the surface. You can actually measure how healthy things are underneath the non-skid.

Sure.

You know, it's like things like this where you just remove this, like, you know, this orthodox way of thinking about how do you figure out, like, what's wrong, what to fix, how much budget to put to use. And where does the supply chain, you know, how's the supply chain meet my need in terms of what things I need to fix and get the ship, you know, back patrolling and deterring conflict. It's like this kind of stuff that we, you know, have just like built such an expertise and now models around, you know, ensuring we get all this information and data set. So it really is like, you know, it's going from like a 3 to 4 month process to now click a button and you know exactly like where to make all the repairs and the next time you do it, you know, you spend a lot less time in the dry dock and you should do this stuff more and more just like as it's, as it's like, you know, patrolling and on duty. So like robots, you know, specialized robots that can be like on these ships, that's like, that's the, that's what we want to work towards.

In the, in the basic example of like there's a boiler, the gecko robot climbs up, finds the crack, let's say. I don't know exactly what happens, but erosion, corrosion, corrosion. Are you thinking about an act 2 where you're actually doing repairs well? Are you already doing repair or is that something where it's like, oh, that's a big technological challenge?

It's a great question and it's something we're working on right now on the manufacturing side. But I think the key was the first idea was like, what if you could own the health data of the world's most critical piece of infrastructure? Yeah, that'd be cool. That seems like a pretty interesting thing to do. And then you begin to get into, if you can understand the health of things, then you begin, then like you can just pull in existing information, time series datasets. Like these are the kinds you pull in now. And so if you're a power plant or you're a refinery, and you want to really capitalize on the fact that oil barrels per day is $100 right now, and so you want to maybe increase production, well, you don't know if you can. This is where the consultant with software and AI companies really have a hard time is because you don't know if you push the asset, will it break down and not work and explode? So that's where the missing data set that never existed for that atoms to bits side of the robots allows for us to be able to make these operational changes now, and which is a lot more valuable than actually fixing the problem right then and there. So we're going after these big value propositions. And if you think about the ability to run functional finite element analysis, so these ANSYS type of models, these are the sorts of things that Gecko is working on because we have this unfair data advantage that we've been collecting for 13 years on half a million assets now inside of cantilever. And so, what we're getting to—

Overrated success.

So, what we're trying to get to, man, is like, you're totally right in like where to go. Like, you want to find the problem, fix it right then and there using these tools. And so, you want to build towards that, but you want to be really smart on like the best techniques, the best kinds of approaches and ways to fix things and solve the most valuable thing first, then keep on solving more problems that your customers offer you.

Awesome.

But man, I think the future is really going to be interesting because these are capital-intensive industries. That can adopt technology like we're talking about here really, really quickly are going to be so unfairly advantaged. And this is where we're in those rooms creating the strategies for the top 10, top 15 large oil and gas companies in the world, helping them with how do we make and optimize this future because the future is going to really belong to folks that can figure out how to run this stuff unfairly.

Well, the chat absolutely loves you.

Yeah.

So come back soon. This was really fun. Congrats to the whole team on the new contract. And I'm so glad you started this company 14 years ago. You're not coming on just pitching the concept, but you've actually done the heavy lifting to figure this stuff out. And yeah, the opportunity—

Yeah, this is amazing.

—scale of it is just insane. Yeah.

Have a good one.

Thanks.

I appreciate it. Nice to see you, brother.

Thanks for hanging, Jake.

Thank you. Happy St.

Patrick's Day.

We'll talk to you soon, and I will tell everyone about Turbo Puffer, serverless vector and full-text search built from first principles on object storage, fast, 10x cheaper, and extremely scalable. We have some good news. For just $3.9 million, you can fly private to every race, every, every F1 race on the calendar this year. It's called the ultimate experience. All 24 races, $3,911,100 per person. So don't think, oh, they're flying me. I'll just have my buddies tag along. You're going to have to put your buddies in a really large suitcase.

It's so funny because for some people, they would have to pay them $3.9 million to go to all 24 races. Because it's not just— if you're really doing this on a fixed schedule, it's like at least 4 days a week, 24 times a year. And this feels like a full-time job. Also, there's some AI researchers out there that would just say, like, I can make— Yeah.

So there's a wrinkle here. They're saying, hey, flat price for private jet to all the races. And they're like, well, like, even if you're based in Paris, like, there's a race over there. You know, you're based in Miami, there's a flight over there. They haven't considered the fact that if I live off-grid in Alaska, it's going to cost them $10 million to fly the private jet to some remote landing strip that I've constructed outside of Nome, Alaska.

Yeah, is there, like, a SaaS company that could take advantage of this, basically being like, we know we've got a bunch of buyers here. We're going to take our top rep, and we're going to have him jet-set around the world, 24 strike missions. Yes. Make it math out. Yes. If you're selling bigger deals.

Skip the race, spend it in a conference room, pitch it, do a demo. The whole week.

The whole time. Yeah. Demos and the helmet.

And then hop back on the jet. Somehow, I think that if you price these flights individually, it would be cheaper. But it is very funny. Not by much. You're talking about going—

it depends on the jet. Yeah.

I mean, it's a lot of travel, a lot of miles, a lot of hours. But $4 million is a lot of money. And I think it gets you a lot of flights.

Going to all 24 races, if you're not— a driver working does not sound fun.

Also, they didn't— they just said private jet travel between every race on the calendar. They didn't tell you how many other people are on this private jet. That's what I'm saying. Imagine you hop on and then there's 17 stops while they pick up people.

Unfortunately, standing room only.

Yeah, it's kind of a Southwest vibe. Yeah, yeah, yeah, yeah. Anything could happen. Anything could happen. But we'll let you decide whether or not you want to spend $3.9 million to fly private to every F1 race on the calendar. If you do, let us know. Give us a review. Send us a message. Subscribe to our newsletter at tbpn.com, then email us when we email you the newsletter and tell us how the private jet experience was. The ultimate experience.

Last but not least, the new Roadster is apparently going to be unveiled next month.

Hopefully. Hopefully. If Elon puts hopefully in a tweet, he says it will be—

it will be a banger next level. I'm very excited.

And Travis Kalanick said, "When I've run into people who are in the know, I inquire. They tell me nothing, but their eyebrows raise and their eyes widen in a way that can only mean something of sorcery and magic is coming. Hopefully it's a flying car." And I hope that it is revealed in April. He should unveil it on April 1st. Everyone would be so confused. It'd be very funny. If not, I can wait till May. I can wait till May. I'm very excited for the next roadster. Expectations are incredibly high. There have been so many electric supercars, a lot of depreciation, not a lot that have filled that special territory. Even the electric sports cars have not done well.

The Boxster. My sense is it's not going to be a true supercar. It's going to be a Turbo S, the kind of thing that people are going to just daily. Daily. And it's super fast.

Yeah. It's fun. But it probably— I mean, I would expect that it's—

But it's not the kind of thing you keep with low miles because you expect it to appreciate.

Yeah. I would basically expect a Rimac Nevera, which is a $2 million car, but at a $200,000 price point. Everyone's like, it goes 0 to 60 in 1.6 seconds or something like that. It's going to have some headline stat that everyone debates and is like, oh, well, technically it wasn't this, blah, blah, blah, blah, blah. But it will be— be like shocking in its own way, and I'm sure people will have a lot of fun with it. So excited to track that story. Well, thank you for tuning in to TVPN today. We will see you tomorrow.

Go have the best day. Maybe give us 5 stars on Apple Podcasts and Spotify. It's been an honor.

Why don't you throw that flashbang? Throwing flashbang. We'll see you tomorrow. Goodbye.