4 Tech Arguments (Number 3 Will Shock You), Mansion Section, John Arnold Joins | Viswa Colluru, Jake Adler, Adam Korsunsky

You're watching TVPN. Today is Friday, September 25th, 2026. We are live from the TVPN UltraDome, the temple of technology, the fortress of finance, the capital of capital. Let me tell you about vpn.com. Time is money. Save both. Easy to use corporate cards, bill pay, accounting, and a whole lot more all in one place. Be the king in the castle. withramp.com. Uh, we got some debates on the timeline. I identified 4 debates for you. You, you don't want to call them debates though, you want to call them arguments, fights?

I thought arguments.

Okay, we're throwing down, we're throwing down.

Wouldn't, wouldn't someone rather listen to an argument than a debate?

Okay, yeah, I don't know.

I don't want it to be— I don't want it to be a health— uh, healthy, mature conversation.

I want—

okay, I want to be reaching over.

Debate fight. First one, Microsoft. Is Microsoft about to have their Muse moment? They've launched Copilot. Now they're launching a new product today. Is it going to be a breakout success? Is it going to compete directly with Meta and Muse, or will it be a different thing entirely because it lives in the enterprise, it lives in the Microsoft ecosystem? Let's go to Tae Kim. He says shots fired. Microsoft will go after Meta's Muse from Alex Heath's interview of Satya Nadella. He's doing the rounds. Talk to Diadra Bosa. Talk to Alex Heath over on Sources.News. Still cooking. Quote from Satya Nadella or from Alex Heath: He wants to bring Autopilot's AI chief of staff to your personal life too. It's not staying in Excel. It's not staying in Outlook. It's going to be cooking dinner.

Big move going from a copilot to autopilot.

Autopilot.

Yeah, it says a lot about About their intentions with the product.

Yeah, yeah. It is actually sort of the like, you know, is this AGI? Is this not AGI? Like if you're in the Copilot mode, you're sort of signaling we're not really in the AGI era. Autopilot feels more like, although it's a term, it's a vague definition, it is significant in the brand term. So it's a hardened version of OpenCLAW. The agent gets its own computer workspace and memory To work continuously. There was a very funny post from somebody in Meta who was like, you know, a lot of these agents, like the value prop is that it has its own computer, it has its own virtual machine with some real hardware behind it. So you can have a file system. It has, I think, 8 gigs of storage. So you can say, oh, I need you to download some images, some files, like keep them in your cloud so you can access them at all times. It can obviously build a database, save records, your contact information, your calendar, it can mirror that. So it doesn't need to be pinging every service all the time, doing everything in the context window. In the Ben Thompson parlance, it can write things down. This is very good. This is what everyone was doing with OpenCLaw on their Mac Minis. This is what people are doing with Codex, Claude Code on their computers. And Codex of course has like cloud threads that do something similar. The funny thing is that Meta was like, it's a real computer, this is a real computer. And then they're literally using a virtual machine. Which is still a real computer, like you have access to hardware, but it is a virtual machine. Like when you're not using it, it goes dark and someone else uses it. And I think there's still some confusion because people were like, this isn't a computer at all. And it is.

Well, yeah, it sounds a little bit too good to be true that— They're not just in a rack for you and just let you—

if you sign up once and never use it, that's just inefficient compute resources. But it's just funny that we live in such a virtual time that someone can very seriously and basically correctly be like, it is real, it's a virtual machine. It can't get any more real than that. It's like, yeah, it could, but why would you? So no one cares. Anyway, Microsoft's doing the same thing, virtual machine for everybody. It'll have its own computer workspace and memory to work continuously. When I ask about Microsoft's consumer strategy, he points to its 100 million plus consumer subscribers. People that subscribe to consumer Microsoft services, Call of Duty, Xbox Live. These are the consumers. Now, Xbox is still like pretty anti-AI. They've been signaling like, hey, yeah, we're not going to lean in. We're not going to stuff Copilot all over Xbox Live. I'm personally, I hate that because I think—

Yeah, I want to basically like a company-built aimbot.

No, I mean, yes, of course, that's funny. But no, what I want is I want to be able to text my Xbox and say, hey, like Grand Theft Auto 6 is coming out. I want you to preorder it. I want you to get it installed and I want you to open it up and do the little config. Because every time you open a game now, it's like downloading an update, accept the terms and conditions, accept the privacy policy, adjust the HDR, adjust where the HUD is on the screen. Do you want any accessibility features? Do you want to do the tutorial? It's going to blow your mind, but if you press forward on the stick, you'll move forward. Lock in. We got to do this tutorial. Here's how to jump. It's going to blow your mind. It's the X button. Push it now. Push it again. Okay, you aced the tutorial. It's like, I want to be able to jump into the game, and I do think an AI agent could help me with that. It can certainly do it on desktop. You can just say, hey, go play the whole game for me. Level me up until I'm ready to really get in the fight. Because this is my first rodeo. But Xbox been pushing back on it. Microsoft broadly, they're going all in with Autopilot. Autopilot should also go to the consumer side. He sees consumer agents potentially cutting out today's middlemen, making that market more zero-sum. In the enterprise, he thinks agents will make the market bigger than cloud by orders of magnitude. We got room to run. Satya said it himself. He's going to 300T, $300 trillion company.

I like it.

What do you think?

So, I mean, the first thought is that again, everyone is building the exact same thing.

Yep.

So we went through this for the last few years where everyone was building—

Everyone had a chat app.

Everyone's building their chat app.

The funny thing is the lore on GPT-4, that was first available in Bing. Like GPT-4, the thing that, like GPT-3, GPT-3.5 powered the original ChatGPT. GPT-4 was only available through Bing. And then everyone was like, oh, I guess like Microsoft's gonna make a play and Bing's gonna be doing AI and like this is gonna be a challenger to Google. And then they sort of like pulled back, but—

Well, and at that time when Microsoft was backing up the Brinks truck for OpenAI, they certainly, I don't think it was— it certainly was non-obvious at that time that they would be creating a company that they would ultimately compete with as much as they are partners. Yeah.

So Peter Steinberger, the founder of OpenCLAW, chimed in. He said, Microsoft shipped a really compelling product on top of OpenCLAW today. We worked with them since March to make the codebase ready for large-scale deployments. And of course, there might be some IP sharing agreement, right? Because Didn't— oh no, because OpenAI hired Peter, but they didn't acquire OpenClaw.

OpenClaw is open source. Yeah.

So anyone can build on top of it.

Yeah.

But what's odd is that I think Nat Friedman stated that they didn't fork OpenClaw. They built something from scratch, but there are just conventions. And so you wind up having similar looking .md files. It's just like, you know, if you build a web page, like you're gonna have a product— if you build an e-commerce site, you're gonna have a product detail page. You might have a similar sitemap and architecture. It doesn't mean that you forked Shopify to do that or whatever. Let me actually tell you about Shopify right now since I mentioned it. Shopify is the commerce platform that grows with your business and lets you sell in seconds online, in-store, on mobile, on social, on marketplaces, and now with AI agents. So the debate point is how big will this be? What is the actual impact to Microsoft? And like, there is the cool macro tech story, which is just that the models have advanced to a point where there is a new use case. Like, they're useful in a new way.

Yeah.

Not just knowledge retrieval, not just go and write a bunch of code. They are useful, but that requires a new type of harness. Just like when the models got good at code, the harness became very important. We saw Cloud Code and Codex and Cursor and a bunch of other companies. Cognition really defined that era. Now, the models are pretty good with personal agent stuff, sending text messages, talking to people, integrating with Slack, et cetera. So, the harness and the integrations and the partnerships matter. So, we're talking about partners with Amazon. Who's going to be Amazon's dance partner? Who's going to be Walmart's dance partner? Who's going to partner with Shopify? How big is the open ecosystem? You need a lot of— You need a different harness, you need different integrations, and you need a lot of new user education because I think there's a lot of people— I mean, there's people that are still online being like, these things hallucinate. Like, they have a knowledge cutoff because they're talking to, like, the voice model.

Yeah.

The main thing with this launch is that I don't think there's anyone on X that's seeing Microsoft launch an OpenCLAW-powered agent and thinking, I got to use that. I'm so excited. I really want it. I Okay. That's my view and I think it's correct and you're free to steelman it.

Okay.

That being said, they have such incredible distribution, they can still roll this out to a bunch of people who are gonna try it and be like, wow, that's really cool and useful and helpful.

Yeah.

And be the first way that they experience this new paradigm of agents, right?

Yeah.

And so as much as like, I just think there's basically zero excitement from like the core insider agent maximalists. There's going to be just natural excitement from everyday Windows and Microsoft users.

Yeah.

Yeah.

It is uncommon in tech and maybe it's uncommon in X to run your entire organization, your enterprise on Microsoft. The default tech stack for most startups is Google Enterprise, Gmail. You set up the Gmail account for your employees and then You add Slack, so then you're in the Salesforce ecosystem. And Microsoft is there, but usually with one-off Excel licenses here and there for the finance team. Does that resonate with you in terms of how these products fit together? There's usually some Microsoft, but you have to go to a financial company or a real economy company to see, okay, these guys are all in on Microsoft.

Yeah.

Yeah.

As a tech stack, like Outlook down to Bing down to, you know, the databases and like fully integrated in Microsoft. But there are hundreds of millions of people who their life at work is just pure Microsoft and they're like, yeah, I don't use Gmail at work. I don't use Google Calendar. I use Outlook. And for those people, they're like, okay, I get this thing like probably for free added to the subscription. Maybe we already have some Copilot stuff. So this just rolls in and I get You know, near frontier capabilities, good enough. And there's such an overhang in terms of what I can do that just going from, you know, zero to taking advantage of some decent automations is going to be potentially popular. It feels like it could be more successful than the old Copilot. And a lot of that's driven just by capabilities. The models are better, the harnesses are better, the integrations are better. And so. you can do more things. I think there'll still be a lot of point solutions and a lot of competition. The big question is, like, when is Google going to fire back? Because are they going to wait until the next I/O? Like, that's a long way away. But they do love releasing on, you know, those annual schedules. And so, it's like we might get a Google— like a crazy Google OpenCLAW agent, personal agent, in mid-2027 and then Apple's close behind in like 2035, something like that? Would they launch something?

I mean, Google does have— there is Gemini Spark, which is like a 24/7 always-on agent. But like no one—

Does it have a VM? Does it have a real virtual machine?

I mean, it says even if your phone and laptops are turned off, it'll still keep working. But I've seen like very little— like no one's really talking about this.

They aren't really pushing that because they are pushing Gemini as like the chat interface. But man, that thing falls short sometimes.

It's so funny and frustrating that you have all these companies building personal agents, but none actually has 100% coverage over the services that you need.

Oh, monopoly guy over here. You're going to advocate for monopolies?

Not advocating it, but it would be helpful right now. Because even Google, right? Everything around the full Google set of products would be absolutely amazing, but then you don't have native iOS integration.

No, this is the irony of monopoly is that it's bad for pricing in some ways.

But it's great for user experience.

It's great for user experience and it's often great for customers. Like how many times are you advertising on Facebook and you're like, oh, I wish like Ad Manager would just like also run Reddit ads and also run X ads and also do my YouTube ads, like just do it all. And the same thing is true even when you fire up like a Netflix app. You're like, oh, is the movie I want on Prime or Apple TV or Netflix or Paramount Plus or HBO Max or Peacock? Like there's 25 different things. And as a consumer, you're like, I'd love just one, but then they would have pricing power over you. And so you gotta look ahead and know that there be dragons if you allow the monopoly to flourish.

Yeah.

Potentially. At the same time, We've been in this odd regime where the monopolies have not actually damaged consumers. Like, this is how Amazon got out of antitrust again, was like they didn't raise prices perniciously against their consumers. And so, the current antitrust law is written such that you have to prove customer harm. It's not enough just to have concentration in a market. You also have to be extracting monopoly rents. You also have to be raising prices. increasing profits and hurting the consumer financially. And so for something like Google, it has a search monopoly. It's still a free product, so it's very hard to prove consumer harm because you're just like, yeah.

G Suite, they are steadily hiking prices.

Okay.

But there's viable alternatives.

That's, yeah, that's much more oligopolistic for the technology suite. Anyway, it'll be interesting to see if there's like, you know, a niche community of like, yeah, I'm a I'm a Microsoft Autopilot guy. I mean, the Claude integration could change things. It just depends on how you're positioning it. Like, if the brand— if you talk to somebody and they're like, oh, are you a Claude guy? Are you a Codex guy? A Cursor guy? Like, all those things say certain things about you aesthetically. You're Muse. OK, we get it. You're like the boo-boo guy. But if you come out and you're like, yeah, I'm actually a Microsoft Autopilot guy, That could say two things. It could say Excel Jockey, king of the financial markets, apex predator of the economy. Or it could say, you know, dropping a nuke on COD.

Prestige Max.

Prestige Max. What's the—

He has Diamond Camo.

Diamond Camo? No, it's Dark Matter.

Dark Matter. Yeah, that's when you get Diamond on all of the— Yeah.

So if it helps you grind to Dark Matter, I think they got a winning product there. Don't you think?

I agree.

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. OK, we spent way too long on that. We got 3 more, and our first guest is joining at noon, so we're going to breeze through this.

And we've got to figure out a way to make the tensions go up and actually get an argument going here, because that was—

We were fighting a little bit.

That was too friendly.

That was too friendly. OK, well—

You wanted to argue that everyone All the tech insiders were couldn't wait to get their hands on autopilot. That was your stance, right?

Yeah, it's a tough one. Tough to steelman that. Anyway, Doge Designer shared a photo from a dinner that happened at the White House featuring important tech people and none other than Xi Jinping, the Chinese president. So Elon Musk, President Trump, First Lady Melania Trump. China's First Lady, AMD CEO Lisa Su, Nvidia CEO Jensen Huang, and Apple's Tim Cook got the prime seating right in front of Donald Trump, and they were at the prime table together. No translator that I can see in this image, although there is an empty chair, so it's possible that there was a translator at this.

The empty chair would almost certainly be for Donald. Right? Oh yeah, that would be—

yeah, that's right.

Okay, so we're gonna, we're gonna put the translator just between Xi and Melania.

Yeah, because, because Xi Jinping is— whenever he speaks, he, he, uh, he gives his speeches even in America in Chinese. Um, but, uh, he seems to be able to hang with, uh, Tim Cook, Elon, Lisa Su, Jensen. Although maybe there's some, uh, maybe, maybe those tech leaders at least know That one important phrase. If you want to really ingratiate yourself with a Chinese speaker, what do you say to them? 我很喜欢啤酒.

Right?

That's what you got.

Always worked for me.

If you're ever seated with Xi Jinping at a dinner at the White House, just throw that down and he'll be like, one of us.

One of us.

Now it's time to do a deal.

Notable that Elon and Tim, no plus one. They're locked in.

Okay.

No time for— no time for guests for them.

Yeah, no John Ternus. I feel like Tim Cook should get the John Ternus plus one. Elon should be able to bring like Nikita Beer along or something. That would be fun. But it was a pretty tight guest list, although some people did get plus ones. Jensen Browne's wife, Lisa Su, brought her husband. What's interesting is that Dario wasn't there. He's sparred with the administration before. We've seen this. He's also attended plenty of meetings in DC. And Tom Brown, the co-founder of Anthropic, has been on a reset with the administration and even went as far to praise Trump's Let Data Rain post on Truth Social. This was September 1st. He said to Howard Lutnick at the G20 Innovation Ministerial in Chapel Hill, quote, I really loved President Trump's post from earlier this week. And so there's clearly olive branches going out, but they haven't been fully received because the dinner invite was— maybe didn't materialize, or maybe he turned it down. We don't actually know. But I was offered 15 plus ones, maybe 10 plus ones, maybe 5 plus ones, and he decided to lock in and grind for the Grind for the ASI.

Well, you know who was there?

Who was?

Lynn Martin.

Lynn Martin.

From the president of the New York Stock Exchange, our friend.

To see it.

Holding it down.

Let me tell you about the New York Stock Exchange. Want to change the world? Raise capital at the New York Stock Exchange. There's Lynn. She's been on our show multiple times. Dear friend. And also, you know, an important person to get around the table when the, when the capital markets, when the future of the global economy is at stake. So very, very exciting to see her there. As well as many other guests who made the list. The list is wild. Signal was really having some serious FOMO. This was hilarious.

I'll read it. Okay.

Would have paid my entire net worth just to be a fly on the wall at this gathering. I can only imagine the conversations, the subtext, and the tension in that room. Also, where are the venture capitalists?

It's such a good point. Actually, wait.

Also, where are the venture capitalists? VC congratulating themselves. Won't someone please think of the venture capitalists?

Yes. Yeah, it really showed that there was levels.

There's levels. It is. This is like a— is this a copypasta of the like $15 million or dinner with Drake meme?

Yeah, dinner with Jay-Z. $500K or dinner with Jay-Z.

$500K or dinner with Jay-Z.

Or the Grant Cardone like, yeah, I slept in my car last night to come here.

And then it is very funny because like There truly are levels. And if you, as the humble poster— I obviously have a ton of respect for Signal, but if he or I were at this dinner, we would just be seated way in the back and we'd be like, wow, look, it's Jensen. Wow, that's so cool. I wonder what he's talking about.

Cool.

And then you get to go up to him for 2 seconds and be like, dude, I love the GB.

Well, to be clear, he didn't want to be— he didn't—

dude, that's so cool.

John, he didn't want to be a guest. He wanted to be a literal fly.

Oh yeah, that's true. That is different.

And you can see in that original image, I hope there's no flies around here.

We got some flies on the wall today.

Sorry.

But yeah, it was a stacked, stacked list. There's a lot of folks— Sam Altman, Greg Brockman, Anna Brockman, Sergey Brin, Satya Nadella, Sundar Pichai, Cristiano Amon from Qualcomm, our guest this week, made it to the dinner, came back from Hawaii, David Solomon from Goldman Goldman Sachs, tons of other folks, ExxonMobil. It was really a who's who of the global economy. Son of Bernard Arnault, also Bernard Arnault is there.

A lot of people are saying Tai Lopez was snubbed.

Yeah, I would have liked to see a Jocko Willink. I would have liked to see a Joe Rogan on this list. Sort of disappointing that they didn't include podcasters, except for David Sacks. He did make it. And also—

Yeah, he's a podcaster and a venture capitalist.

Secretary Kennedy is also has a podcast. There's actually a fair amount of podcasts.

Brad Gerstner podcast.

Sam Altman ran the Y Combinator podcast years ago, startup school, so former podcaster. There's a few other people on here. But lock in, get your— I think if you put your entire net worth into potentially becoming the CEO of a really important company, maybe you do wind up as a fly on the wall. Never doubt yourself, Signal. I think you can make the next cut. I believe in you. I also believe in Railway. Railway is the all-in-one intelligent cloud provider. Use your favorite agent to deploy web app servers, databases, and more while Railway automatically takes care of scaling, monitoring, and security. My take on this is that it's— the question is like, should Dario have been invited? I think even though they're having beef with the administration, like, it's such an important company. It's important that he's in the conversation. He also has like a very, very hard line on China, and I Would— I would feel better about his positions on China if he's like, yeah, I've met Xi Jinping, I really do hate him, or something like that. I don't know. You know, it's informed from like a 30,000-foot view.

Wait, Dario?

Yeah, Dario's like—

He worked at Baidu.

Yeah, but since then, his positioning has been like, we gotta win the race against China, we have to dismantle authoritarian regimes, et cetera, maybe. he looks Xi Jinping in the eyes and gets the stare and is like, oh, wow, I had you all wrong. You never know. But I want to see the sit-down happen. I want to hear the downstream. I want to be the fly on the wall. Anyway, let me tell you about Figma. Agents meet the canvas. Your AI agents can now create and modify your Figma files with design system context. Has the AI slop gone too far this time? That's our 3rd debate. The 3rd debate that will shock you. So, We can pull up this Calci ad. It's their latest direct-to-consumer, direct response ad.

For everyone at home, this is a karaoke moment.

Yeah.

You're sitting in the office. This is going to be on the open floor plan. Don't be afraid to start singing.

Yeah.

Add it to the Spotify playlist. So Brian Pempus says this Calci ad is among the most dystopian ones I've seen from any betting app. And you can sort of hear The AI song over the Pixar-inspired AI-generated video. Uh, the lyrics are also odd. Not to nitpick, but it says like he sends his girlfriend money and then he's handing her money. Every month I send her money from groceries. So, uh, there's a lot of like— it lacks the polish that removes something from slop. Every once in a while you see a video that is AI-generated, but you're like, okay, there was enough Attention to detail there. This was let's just rip it. This is actually a popular format of AI swap that's been going very viral. Starts with a really dramatic hook. In this case, the man thinks his partner is having an affair, but actually she's just hedging food prices on cow cheese.

Yeah, because she has inside knowledge on the market from working at a bakery in the morning.

Yeah, so that even though egg prices are. Going up, she's able to net out and remain even because she is profiting on trading the egg futures market on Kashi. I don't know how deep the market is. I don't know how real those markets are. Do they really have egg futures markets on Kashi now? That's kind of interesting. I don't know if you can actually hedge your groceries. I want to see a version of this ad from the Chicago Board of Trade where she goes a lot further, gets a Bloomberg terminal. Establishes a relationship with JP Morgan, Morgan Stanley. She's visiting the Chicago Board of Trade regularly, trading in size. She's talking to counterparties, Millennium, Citadel. She's on the phone constantly moving huge volumes of commodity futures at scale, elevating herself past just the prediction market small ball and getting into the big leagues. That's the future.

Kane says, would you rather your girlfriend insider trade egg futures or cheat on you?

No. So, this is not insider trading. Like, you can actually do this. You can be a baker and realize that prices are going up and trade the commodities markets. In fact, the commodities markets are designed specifically to give that signal, and that is the purpose. The purpose is to Is to understand, okay, the weather is bad, people can trade that. Fundamentally, there's no law against insider trading in commodity markets effectively, but there's a question of like, is this a consumer-ready product? Like, is this something that someone who is— maybe doesn't have, you know, a whole desk and a research team and a Bloomberg terminal can trade?

One interesting thing is I found Ag prices up in September, the market right now, there's a 72% chance.

Okay.

But there's only $417 of volume.

Yeah, but if you're only buying $20 of eggs, that's enough depth. So it is a viable use case. It's not fraudulent. I think most people are responding to just like the sloppiness and like the clickbaitiness of like the weird hook. But somebody posted one of these that was done by a much less prestigious brand in, I think, like weight loss. It was some sort of weight loss pill, but it starts with like this insane hook about a secret relationship, very soap opera script. And I feel bad, but I actually watched the whole thing because I couldn't look away because it was really Pandora's box. It was infinite jest for that 3 minutes and all the comments were like, Why did I watch this? And someone had put it on the timeline as like, this is the worst thing I've ever seen.

Yeah, the funny thing about this ad is it seems like no one proofread it because there's a lot of elements of it that just are not even phrased properly.

Yeah.

But sometimes putting in those details gets people to talk about it, gets people to go in the comments, actually amplifies things. It draws your eye in. And so you never know with The algorithmic ad generation, where all this stuff goes. What about the? The interesting thing is I think the consensus, and we're going to talk to John Arnold about sports betting and prediction markets at noon today. But there is another interesting side of this, which is is people seem to be coalescing. At least the vibes and the timeline are like this is not good. Like like this is like. Either you should personally stay away from it, like I don't advise you to be sports betting, and I think that's John Arnold's position. There is a second debate point in question, which is what about the dopamine sites? How do you feel about dopamine sites? So in South Korea—

The thing that I think is—

Consumerism, but continue.

My current concern about the state of the hyper casino that our world has turned into is that, like, as an adult, I feel, like, confident in my own decision-making ability and the ability to see ads for gambling all the time everywhere and still make the call that it's not for me. I'm not interested.

You've never lost a dollar?

I've tried.

One weekend.

John got to witness me one weekend. I was like, look, I think I should I think I should try this gambling thing. It seems super popular. I gotta get to the bottom of it. So I tried it. I'm running in the red. And this is generally true for anything. But that being said, like, you know—

You're stuck in your ways.

As a 30-year-old, I've been advertising gambling products for a decade now, right? Yeah, yeah, yeah. And my concern is that while adults have the ability to just make the decision, like, hey, this isn't for me, and you have like More—

yeah. If you start young, you're more likely to stick with it for a while.

Yeah. But young people today are going to go through— that are growing up on the internet are going to go through this just like— they will probably be served like a million gambling ads by the time they're 18.

It's a lot. Yeah. I guess I didn't get served.

It's so normalized. It's so a part of society now. It's instantly accessible. If you're getting advertised stuff like this, that's basically saying like, Hey, you're sending this message that like this is your way out, right? Like this is—

Yeah, the deep irony is like the way out is to not participate and be one of the people that makes it through without forming the addiction.

Yeah.

But maybe we need to run a slop ad about that. We wanna do a slop ad like this where it's the same premise. Oh, I think my partner's having an affair. And lo and behold, they're just watching TBPN for 3 hours every day, getting informed about technology and business news. What do you think about dopamine sites though? Because dopamine sites, which is where—

Explain this.

Dopamine sites are fake shopping experiences. It's an e-commerce site where you can buy luxury goods, but you don't actually have to pay them and they don't arrive. So this is from an article shared by Adonis on X. The final frontier of consumerism has arrived in South Korea. Quote, this week I placed orders for a $44,860 Patek Philippe hand-engraved watch, a $12,500 Hermès bag, a $9,800 Tiffany diamond ring. Cha-ching! Oh, I like— I didn't realize that was the analog one. That's great. And a $7,350 Cartier love bracelet in yellow gold. I don't need any of them. I certainly can't afford them. Thankfully, they'll never arrive. Instead of Amazon, this individual spent the past week browsing a new breed of website known as dopamine sites, a trend that emerged in South Korea. These sites— these websites like Dopamine Shop and Food Never Comes recreate the entire ritual of online shopping. You search for products, compare reviews, add items to your cart, Enter a shipping address, place an order, and even track your delivery, but it's all imaginary. Is this not as bad as sports betting? Is this not bad at all? What is this?

It's sort of strange and sad and a reflection of our society that is so infatuated with pure consumption instead of creation.

Yeah.

And it's dystopian, but—

We should just make a dopamine site for TPPI merch.

Well—

You just order them, create a cart, and—

Well, I was thinking you should make like Dopamine ARR where you log in.

Oh, build a fake startup.

It looks like a Stripe dashboard, and then you just see like—

There's a game that I played that is basically that. It's like a cookie clicker game. What are those mindless games? I don't know.

Like Adventure Capitalist?

Yeah, I played Adventure Capitalist, and you just like click a button, and it's like open lemonade stand, and then by the end you're like farming robots on the moon, and you just watch the numbers go up.

Yeah, we made a game called Burn the Runway back in the day.

Oh yeah, yeah, yeah.

I put it in the chat. You can pull it up.

It's still up.

Wow.

This is pre-vibe coding. You did it the old-fashioned way, by hand. It's crazy.

Do you think this counts as LARPing? Because it's not actually live action, right?

But it's like the spirit of LARPing. You will develop like an encyclopedic knowledge of luxury goods if you spend enough time on here. I don't know. It seems like it's nowhere near as bad as something where real money is at stake because you are wasting time, but you're not wasting any money and there's no risk of downside. So why would you buy bananas?

Well, that's the cheapest item.

Okay.

Like if you need them for— but you can buy the max amount of bananas and then—

Oh, okay, okay. So you buy different things for your startup and then you want You wind up running out of runway at the end?

Yeah.

Interesting.

Game over.

Wow.

To put this into perspective, there were no AI companies even on this list.

Wow.

Back then it was WeWork. You had early Anduril. Anduril had only raised, I think, a couple hundred million at the time. Oh.

Good to get in.

You had Quibi on there.

OK. OK.

The whole—

Well, it's not all bad and AI slop. Video that Ates shared by popular demand, his magnum opus.

No, it is all bad and AI slop.

You think this one's bad?

Rewind. There's this guy Henry says, "I am so beyond obsessed with this account." For me, this is the moment that I am now very, very, very bullish on AI video. Okay. There's an account called Jean Philanthrope, who's like a French guy who's a boxer.

Yep.

It is just an Somebody using, basically making videos and then—

Translating this guy with a funny haircut.

Yeah, he looks very silly.

And he got 230,000 followers on Instagram.

Yeah. And million, you know, when he launched a coin, hundreds of millions of views. And then, and so the new, the new meta, I guess, is, you know, making an AI character.

Sure.

That's silly, strange, run up the views and then pump the coin. This was one of the top tokens.

Okay.

A few days ago.

Yeah. Well, sounds like a mess.

Anyways, over to Eights.

To what?

This is the video you were talking about.

Oh, yes, yes, yes. By popular demand, his magnum opus. He has repurposed the famous scene from American Psycho, the business card scene with pickles.

New pickle.

What do you think?

I can't really hear it, so we're doing IEMs.

Very nice.

Look at that.

Picked them up from the deli yesterday.

Good coloring.

That's dill, and the brine is something called mustard seed.

It's very cool, Bateman, but that's nothing.

I remember when you had to do this. You had to go into After Effects, track the actual business card, roto out the hand so that the card— whatever you put on the card was underneath, and then you needed to blend the new card in if you Wanted to make a joke. It was very time intensive. Took hours. Now, you can even change the voices. It's pretty crazy. Anyway, this is a nut pickle. There's something funny about pickle coming out of a whole peppercorn. Extra crisp.

Impressive.

Very nice.

Let's see Paul Allen's pickle.

See, there's still like a spark of genius here, you know, mixing pickles with—

Let's see Paul Allen's pickle.

With, uh, and the fidelity is just— There's a little olive coloring, a tasteful thickness of it.

Oh my God.

All right, we can move on.

We can move on. Let me tell you about Cisco critical infrastructure for the AI era. Unlock real-time seamless experiences and new value with Cisco. Anyway, there's a white pill on the slop narrative. So AI video is clearly in the depths of the slopification, but code was there a year ago. People were saying that the slop code would never get good enough and that if it did, it would destroy everything. DHH is out with a new speech at RailsWorld 2026, his opening keynote. He says, it's pencils down, people. Writing code by hand is no longer an economically valuable, viable skill for most programmers at most companies. Of course, if you're in some highly secure environment, it might make, might make sense to still write it by hand. And so it sounds like a black pill. But then he says, but the future of making software has never been brighter. Don't you dare black pill this beautiful moment. So you can go watch DHH's full presentation.

Looks like a megachurch preacher.

Yeah, he's having a good time. So our 4th debate, our 4th and final debate. To what degree is the AI buildout driving interest rates upwards? People are going back and forth about this. Interest rates are Yes. Specifically, the 10-year yield has jumped 12%. So, it's actually— what is that, 1,200 basis points or something? Because you're trying to explain the move in the rates. So, it's not that we're at 18%, we're at 5.18%, but we were down much lower in the 4s earlier. And it's causing nervousness. Last time interest rates went up, the venture capital world collapsed, and nobody likes high mortgage rates. Can you afford a home? There's a lot of knock-on effects. We've covered this a bunch. The vanilla explanation is just the Iran war was unexpected. It caused an energy crisis that caused inflation that caused interest rate hikes and expectations for future interest rate hikes, and thus, interest rates are higher. AI is important in driving the economy, but mainly a sideshow in this story around interest rates specifically. But in tech, there's a view that the data center debt is simply too good to resist. And so, the capital markets are flooding over there. But there's a lot of debate around this. So, Rune said Iran oil prices is short-term driver, but broadly, This is due to demand pressure from extremely attractive AI CapEx opportunities that are borrowing at nation scale. They're certainly hoovering up trillions of dollars. I think the total data center CapEx hyperscaler debt is now almost 20% of what the US government is buying, something like that.

No, so it's actually higher.

I looked this up.

So basically hyperscaler and Nvidia like debt issuance as a percentage of like total Treasury bond issuance. through 2026 has been like 70%.

Whoa.

So, and I think 2025 it was like around 30%.

Yes. And so Astrid Wild says, this is correct, global yields will continue to go higher because why would you invest in literally anything other than data centers when they have such a short payback period and you can throw tens of billions at it? I mean, we saw ClusterMax 1, I think, had a few dozen NeoClouds. ClusterMax 3, which we talked to Jordan Nanos about yesterday, had 326 NeoClouds. And they reviewed, I think, or they rated 200-something of them. And then that's not to mention that there are also data center construction projects from the actual hyperscalers. The labs weren't on there. And then the chip companies Jordan called out. He said like Cerebras will have a NeoCloud. and Positron and Etched will have NeoClouds. And so, there's going to be a lot more of this. At the same time, there's the flip side of the argument. Deep Dish Enjoyer says this is definitely currently not true. Those opportunities are not riskless. And importantly, you can't use private illiquid AI CapEx as collateral. So, if you're a bond investor and you're looking at 2 different markets. Are you really shifting out of government treasuries that are historically the risk-free rate, that they have a monopoly on violence, they have the ability to tax their citizens to pay back the debt? A NeoCloud that's doing a data center, they go bankrupt, you might actually suffer a capital impairment and get back 70 cents on the dollar, maybe 90 cents on the dollar. So there, it stands to reason that there is more risk, associated with the AI buildout. And so, they're not completely apples to apples, but I still think that there is probably some sort of effect here. He said, it's true that as real GDP growth increases substantially, rates should rise though, but we aren't there yet.

Alan Roche says, lots of scary talk in bonds these days, but it's mostly recency bias. Since 1960, the 10-year has averaged 5.8%. Where We're at 5% below average. If you'd fallen asleep 20 years ago and woke up today, you'd think nothing happened in the bond market the entire time. Ignore all the sovereign debt crisis talk. Inflation expectations are adjusting to something more historically normalized. Carry on.

Carry on. So, which do you think it is?

I'll defer to John Arnold on that.

Okay, we'll talk to him about it. I think that the recent spike is 80% Iran war because I think that the AI buildout has been going on for a while and has been maybe driving rates upwards broadly, but the AI buildout has not been the thing that's been driving this most recent spike. Which we've seen over the last 30 days, and it's not like we've we've been seeing a flurry of like oh wow like like we didn't know that Oracle was going to back up the truck. We didn't know that Amazon was so there's no like the expectations in the last month on like like we've been covering this every day. We haven't done that many shows. We're like whoa! We didn't realize that this was going to happen. We didn't know that this was going to be a thing. Now maybe you could say Jensen. Securitization of GPUs initiative with all the big banks, maybe that's having knock-on effects now. But in general, it hasn't felt like there's acceleration that's a departure from the trend. It's like, yeah, we're going to 3x compute every year. We're on trend. We've been talking about this. Some people don't buy it. Some people are skeptical. But in general, there hasn't been like a massive revision to the upside. It's hard to say like, yeah, rates are spiking because Muse is taking off and personal agents are going to drive token demand, which is going to drive the buildout further. Like, we're not quite there yet. There's a couple million.

Yeah.

The other thing is when you look at these large pools of capital, it's not like they're sitting there saying, I was just going to go all in on Treasuries, but instead I like— this is still put it— they're still putting capital at risk.

Yeah.

Right? And instead of make— there's a number of other investments that they might make outside of AI CapEx, but it wasn't like it's just binary thing, like one or the other.

Well, whatever you want to do in the age of higher interest rates, do it on public.com. Investing for those who take it seriously. They got stocks, options, bonds, crypto, treasuries, and more with great customer service. So The Wall Street Journal is sort of putting the higher bond yields in context, saying the robust US economy powers through rate hikes and rising bond yields. The US economy keeps powering through inflation, tariffs, and higher borrowing costs, defying a run-up in Treasury yields and a Fed rate increase that has the bond market spooked. The usual economic brakes aren't slowing growth, hiring, or an AI investment boom that looks to be unstoppable. To some, AI's potential returns seem so bright that even steep interest rates won't slow down tech companies' investments. And it does seem like the math on the AI buildout is wildly different. I was talking to somebody who was saying, like, he knows a lot of people that are falling in love with the spreadsheet. And so, they want to start neo-clouds and data centers because the payback period is shorter. I also saw on Instagram sort of like Course Hustler talking about how he bought a bunch of GPUs and renting them to a data center and saw that as an attractive opportunity, à la like, oh, I bought a multifamily apartment complex.

You must have watched the Dylan Patel interview, the recent Dwarkesh one.

Oh yeah, yeah, yeah, yeah. I mean, this is all related to that for sure. And also a lot of the AI buildout, a lot of the AI companies were very much born in the era of higher interest rates. So, I do believe that if you're a software company trading at 100x revenue multiple and you weren't forecasting real earnings for 10 years or something like that, like the traditional SaaS playbook, and interest rates go from 0% to 4%, that's a lot more damaging than, okay, you were already born in the era of higher interest rates and interest rates go from 4% to 6% or something. It's just not going to move the needle in the the DCF that much because like you were already discounting cash flows in the future 4%, and so as opposed to zero, you're basically counting future cash flows as today's dollars. So anyway, that's the take from the Wall Street Journal. The Wall Street Journal also had an interesting article about real estate bargains in a higher interest rate environment. They say, looking for real estate bargains? Watch these apartment loans. And it's an interesting article. So, one pool of multifamily apartment mortgages issued in 2021 is already 53% delinquent. And so, what this means is that whoever bought the building had a floating rate interest rate and is now— the interest rate rise has put so much pressure under them that they might have to sell units, they might have to sell the building, and this is public information. So, you can actually— I actually went to Codex and was like, find a bunch of apartments that could potentially be on the market for way below market rates because there are public filings of the loan delinquency, and because those might be interesting things. I found some interesting stuff.

Yeah. I mean, one— One indicator of just how brutal the whole multifamily investing world is, is like 5 years ago, the multifamily guys were extremely active on X, right? Posting multiple times a day, clearly incredibly fired up, doing tons of deals. And then most of them have gone completely silent since then.

Yeah.

And it's in large part because a lot of their bets are completely underwater. Basically, there's been a number of funds that have basically not just lost money, but gone to zero.

Yeah. No, totally. There's also some interesting local knock-on effects. Here in LA, Los Angeles implemented a mansion tax for houses, in theory, houses above $4.5 million, $5 million, something like that. But it applies to commercial properties as well. So apartment buildings that change hands have to pay something like a 4 or 5% tax. And so that's obviously going to hurt the amount of building because when you go to sell and transact the building, you're taking this 5% fee off the top. But it's also putting pressure on existing housing stock. And sort of is like an odd knock-on effect because voters obviously go and say like, yeah, if you can afford a $10 million house, you can afford a $500,000 transaction fee, go for it. And it was a very popular bill or change to the law, popular tax. A lot of people would be like, wait, my apartment is going to be taxed too because I just live in a building that's over $5 million in aggregate? Like, this is maybe not what they wanted. And so, who knows where that goes, but people are disappointed. Anyway, moving on to CrowdStrike. Your business is AI. Their business is securing it. CrowdStrike secures AI and stops breaches.

America's newest trophy asset is a country home in the Cotswolds. Cotswolds. Can never pronounce that properly. Harry and Meghan's return to the UK has put the region in the spotlight. A bevy of US buyers are looking for homes in the area. Not very American to buy or invest out here. Strange to see that it's becoming sort of an American investment trend.

Well, everyone saw what Paul Graham did, and they said, I want to be like that.

That's right.

I want a putter.

Are you interested?

Absolutely not.

Yeah.

Next story. Moving on. A modern home in Silicon Valley, though, lists for $44 million. Sort of the Cotswolds of America.

That's right.

Silicon Valley, 5-bedroom house which features 4 distinct courtyards, a pool, and is the most expensive listing in Palo Alto, California. Now I want to do a little tier list. As I read you the various features, I want you to put them in buckets from F tier to S tier. So I'll read through the article and I want you to quickly assign each feature of this $44 million Palo Alto home, Atteer. So, for years, tech entrepreneur Asher Waldvogel and his wife Helen McLean dreamed of building a modern house in Palo Alto. The couple, however, worried about clashing with the Mediterranean-style architecture typically associated with their neighborhood of Old Palo Alto. So, they tapped an architect to design a home that paid homage to its surroundings with stucco, mahogany, and titanium Zinc plating. Where does titanium zinc plating go for you? Kick it off. You like it? A tier?

A tier.

A tier. Okay, we're leaving some room for an even better feature. They spent $20 million over several years building a house completed in 2005. So now looking to be closer to their adult daughter on the East Coast, they are putting the 5-bedroom home on the market for $44 million. Uh, Wilgefold is an angel investor who co-founded Redback Networks, a telecommunications equipment company. McLean previously had a career in fundraising. The couple purchased roughly—

Career in fundraising, let's go.

Generational run in fundraising. So, the couple purchased the roughly 0.4-acre site. In terms of lot size, where are you putting 0.4 acres? Is that enough?

F tier.

F tier. Wow, okay. You need at least 3 acres to really make a splash and earn S tier. They demolished the circa 1930s Spanish Colonial home. The house they built pinwheels around a central staircase. What do you think about houses that pinwheel around central staircases?

I like classic California one-story ranch-style homes.

You don't want a staircase at all?

No.

Because the alternative is 2 staircases. You have a front staircase, a main entryway, and then a back staircase. This is a central staircase, but you're a no staircase guy. So where is in-wheeling around—

Once the kids are older, grown up, then you introduce the staircase. I could explore it, but for now I'm putting it— staircases in general going in D tier.

D tier. Okay, staircases are out. It has 7,900 square feet of livable space. Where does that go?

Split across multiple floors, uh, I'm going, uh, B tier.

B tier, okay. 4 distinct courtyards. Are you a courtyard guy?

Problem with courtyards is like they're, they're very cool in theory, but how often are you actually like— it's really just like a, it's really just, it's like a feature that goes unused.

Sure.

I feel like the average courtyard gets like 5 minutes of someone's time annually.

So I'm hearing like a C tier?

D tier.

D tier. Brutal for the 4 courtyards. What about the pool? It's got a pool. Where are pools for you?

Pools are still underrated.

Okay. So what's your—

A lot of people would say pools properly rated, everyone thinks they're great. I think they're better than the average person thinks a pool is.

I'm hearing S tier.

S tier.

Pool is S tier. Okay.

We need to— I want to see what the pool looks like.

Okay. Key feature of the home is a cast-in-place concrete wall or spine. What do you like? Do you like a spine in your house? Concrete wall, 2 stories high, 80 feet long. There's a little bit of—

I like it. I like it. Some sort of a continuous theme through the home. Yeah, I'm into it.

B tier.

B tier.

B tier. Okay. He said his wife and him are thinking of the next phase of life. They also have a home in Sun Valley, Idaho.

S-tier.

Second home in Sun Valley is S-tier. Okay. Right now they're trying to emotionally let go and decide what to do next. Palo Alto is the center of venture capital and tech startups in Silicon Valley, home to some of the country's biggest tech titans. Sales volumes and prices are rising with a median sale price of $3.5 million for the 3 months ending in August, up 5.8%.

Yeah, it's tough to get excited about this home in LA. This is like a probably $12 million home, you know, depending on where it's located.

But over there, it's—

But over there, like, apparently it's S-tier. Yeah, apparently it's S-tier.

Well, someone will pick it up, and hopefully they enjoy it. Hopefully they're into courtyards, they're into concrete spines in their homes, and—

That's right.

They, they make it a wonderful I think it's that time. It is that time. Let's bring in John Arnold, co-founder, co-chair of Arnold Ventures. Welcome to the show. How are you doing?

What's going on?

Doing great. Thanks for the invite.

Yeah, thanks so much for hopping on. I would love to start with just a little overview of your career since it's been a fantastic journey following it from afar and hearing you describe it. But for our audience, How do you tell the story these days?

Yeah.

So, my first job out of college was at Enron. So, going there as a 21-year-old, it was kind of— got there in 1995. If you remember, the company went bankrupt in December of 2001. And I was there to and through bankruptcy. And it was kind of a remarkable ride up. And a very educational and sad ride down, and probably learned more on the latter than I did on the former. But, um, came out of there— we can talk a little bit more about that if you want to— but came out of there and tried to decide what I wanted to do next. Um, you know, kind of really wanted to run my own division of something. I was running a trading desk at Enron. I was, you know, in very, very quick order had become the head trader at Enron, which was the largest natural gas trading firm in the country. I had that top seat in the industry, and I wanted to take the next step and run a division of something. I looked at, should I do this at a bank? Should I do it at an energy company, at another hedge fund? But I got the chance to start my own hedge fund. I got approached by a few people and said, we'll back you if you want to go do your own thing.

And well, yeah, why were you— why weren't— why wasn't your immediate impulse— most of the people that come on the show, for better or worse, are like, I couldn't imagine working for anyone but myself. Why were you more excited initially about running a division at a bigger sort of platform?

Yeah, so I think the question was How much of the information flow that I had at Enron that kind of helped create this kind of very profitable guess there could I replicate on my own? And if you could replicate 95% of the information flow, did that translate into 80% or 90% or 95% of the profitability? Or was that—

Was it the 5% that—

Was that 5%?

Was that all of it?

Did 95% of the information flow mean that you got 0% of the profitability?

Yeah. What was the information flow at Enron in natural gas in 1995? Did you know that you were stepping into an organization with a strong footing there? I know the later years pretty well, but I actually don't know what the brand positioning was to attract a 21-year-old at that time?

Yeah. So, if you look back at the history of the natural gas industry, and like many industries in American history, at one point, it was very highly regulated, to the point where the United States set the price of natural gas. And you had producers who would go produce it, and they would say, if you produce gas, you get X dollars for it. It was viewed as as an industry having some monopolistic characteristics. And those characteristics were mostly around the pipelines. So there's not necessarily 3 pipelines between point A and point B. And it's more of a monopoly type of industry. And so over the years, that kind of government price setting, as you can imagine, has a lot of downsides. And so sometimes you would end up with shortages. Sometimes you would end up with up with surpluses. And so, starting in the '70s under Carter, and then continuing with Reagan in the '80s, you had a deregulation of a number of sectors, including natural gas.

Yeah.

And the way that gas was deregulated was said that the production of natural gas is a competitive field. The end use of natural gas is competitive. But interstate pipelines are monopolistic at times and should be federally regulated. And so that happened. Enron at the time was this integrated natural gas company. And all of a sudden, you have a pipeline division that has Chinese wall between the production side and the end-use side. And much like a bank, now all of a sudden, somebody has to be that intermediary between the two. And so you have a producer producing gas in South Texas, a certain amount wants a a certain pricing mechanism over a certain time period, and an end user someplace else who wants different characteristics about how their gas is priced. And somebody had to get in the middle of that. And in fact, Enron's first business coming out of this deregulation was Enron Gas Bank. Again, very similar characteristics to what a bank does. And so this industry was really getting started in the In the late '80s, early '90s, and when I got started in '95, I had the benefit of this was not a highly mature industry. In fact, it was still trying to find its legs. And so, as a young guy coming in, I wasn't that far behind, because everybody else was starting afresh as well.

Yeah. How did you process the famous Enron pitch to commoditize and trade bandwidth? Were there other people that were like, ah, it might be too homogeneous of a market, I don't know if it can be commoditized, or was there at least some glimmer of hope there?

Yeah, maybe luckily, this was not my area. But I think there was this notion that by pricing and managing risk, and connecting the producer of a commodity with the end users that you could create value.

Mm-hmm.

And so, it just started for Enron in natural gas, although the commodity markets had existed long before that. Oil had been a traded commodity for a long time. But electricity started to get deregulated. Enron moved horizontally into that. And then, it started thinking about what other areas. What about water? What about trucking? What about bandwidth? And I think many of those had characteristics where you could tell an argument about how they could be an actively traded commodity, but many of which I think when Enron got into them was a little ahead of their time and didn't work out very well.

Sure. Can you tell us the story of the All the options that were open to you post-Enron, there's the famous story of Ken Griffin and Citadel coming in. And I'm wondering, going back to that pitch of the thing that a platform can offer, the actual profit pool that's available, how was that articulated? And how did you make a decision there ultimately?

Yeah, so I think around this time, there had been a number of hedge funds that were doing these arbitrage-type strategies. And so, Ken Griffin famously got started with convertible bond arbitrage. And there were kind of a number of those strategies. But they were limited in number. As they were successful, more capital came into them. And they started to get arb'd away.

Yeah.

And so, the returns for the hedge funds that were doing this were starting to decline. And their assets were increasing. And I think a number of those hedge funds had realized that they needed to start moving a little bit further out on the risk spectrum and take more risk rather than doing pure arbitrage-type trades. And the most aggressive of that was Citadel.

Mm-hmm.

And to the extent where whenever Enron went bankrupt, I think Citadel was already interested in the commodity markets, was already trading some of them. But they viewed the Enron collapse as a great catalyst in order to get into it. And they sent a whole team down to Houston, set up offices in the hotel across the street from the Enron building, and just started interviewing almost everybody on the floor. And they were just recreating— what was the edge that Enron had? What was the org chart? Who was doing what? doing what, who were the real kind of value creators of the field. And, you know, I, I could kind of see what was happening. I didn't really want any part of it. And so I kept kind of— you know, the recruiter was calling me, uh, will you come talk to us? Uh, and I saw people around me were doing that, and I just said no. Uh, and then there was, um, kind of a weekend trip with, uh, kind of industry trip. I was headed to Aspen. Ken Griffin called me at the airport and says, can I talk to you? I said, I'm actually headed to Colorado right now. He calls back a few minutes later and says, if I come to Colorado tomorrow, will you meet with me? I said, of course. If you're going to put that effort out, then you're serious about it, and I'll meet with you. I had a great conversation with him. I have great respect for him, great respect for the Citadel organization. But by that point, I realized that the economics for me would be better if I did my own thing, and that the real value that they could provide me is capital on day 1. And if I could raise that money, then I should do it on my own.

How did you think about risk?

Yeah.

When you were getting your firm, maybe in the first 3 years of building your own firm, having gone through, been able to experience, at least participate in, be involved with a firm that experienced total collapse, I feel like that probably does something to your psychology. At the same time, you were quickly able to put up the kind of returns that typically will would take a manager taking on extreme levels of risk. But I imagine that was probably priority number one, don't blow up.

Yeah, I think that's priority number one for anybody managing money is don't blow up. And especially if you think you have an edge, and then there's real value in getting to come back tomorrow and play again. And so the biggest challenge was coming from a corporation, using a corporate balance sheet, and not really thinking about capital usage.

Mm-hmm.

So now all of a sudden, I have just an investment account with X amount of dollars in it, and I got to stay under that in terms of capital usage, as well as thinking about what's the right the value at risk that I should be running. And these markets, the natural gas market, especially in that time, could be exceptionally volatile. And so, there was huge risk in— there was huge tail risk. And that created a lot of the opportunity, was that there was— you could get paid if you had capital to provide some insurance at times, but to do it smartly, to size it correctly so that whenever you had to pay out on that, you could, again, show up the next day. But I think that the biggest challenge for me was this transition from corporate balance sheet to running my own money.

I think most people have heard the story of how you made your first dollar. I'd love to have you tell it again, but I'm wondering if there's actually like a true lineage between your trading baseball cards to running arbitrage at Enron in '95, or are these sort of just coincidences that these 2 things happened? Or do you think that there is a clear line between like what it takes to make it as a trader showing up early in your life?

Yeah.

I actually think that there were a number of similarities between these 2 things of trading baseball cards and trading natural gas. And to some extent, sports cards are a commodity. Maybe a little bit less so today, as there's so much specialization. There's the 1-of-1 card or 1-of-10 cards, trying to make them not commodities. But back then, they really did feel like commodities. And I made money by knowing what every product was worth at every moment. And so, I connected when I was 14, 15 years old through this dealer network that was on an early bulletin board that connected dealers around the country and into Canada. And I was one of the few guys in Texas. And so, there was a lot of geographic arbitrage that I was doing, but also just that the prices were volatile. They would change week to week. And if you knew what a product was worth in New York that week, you could go around and buy things cheaper and send them up to New York and make that money. And so, that notion of know what every part— what every product within the industry is worth at all times certainly stayed with me and was a skill I used throughout my trading career. And I think the second part was just this entrepreneurial aspect of that I wanted to do my own thing.

Yeah.

It took some courage at age 27 to go out and start my own firm, hire people, put most of my capital at risk, and just say, let's give it a shot. If it doesn't work out, I'm still young, I can go do other things, but let's give it a shot.

It feels like there's a fork in the road that's been intentionally designed to Potentially take advantage of young people where you can get into sports, which are fun. I mean, every kid plays sports as a kid, peewee football and Little League and AYSO soccer or whatever. And then you can—

You can get into cards.

You get into cards and then you can get into natural gas trading or—

Not even getting into cards, but the sort of digital representation of cards that function as slot machines and sort of fully disconnect the sort of passion for the thing from the just sort of finance. It just becomes purely sort of a financial activity. And we were talking earlier on the show how I feel very lucky to be fully— You didn't get hit with the gambling bug. Yeah. As a 30-year-old, By the time we entered this era of hyper-gambling and everything becoming a casino, my prefrontal cortex was developed and I could say, maybe that would be fun, or maybe there's some money to be made there or lost, but it's not for me either way. But it feels like there was maybe a beautiful moment where you could just be into trading cards and then establish a business. arbitraging, but not go down this crazy road.

The industry was much less of the lottery system that I sometimes feel it is today. There was a card that was just pulled a couple months ago that got auctioned today or yesterday for $8 million, this Cooper Flagg card. And so someone buys a pack of cards, gets an $8 million lottery ticket in there.

Yeah.

And I think there is this lotterization of much of society. You see this in a lot of financial markets today. It's not sexy to buy the index fund and just sit on it. As the markets have progressed, you've gotten to shorter and shorter timeframe trades that are more on the binary nature of I either lose 1 or I make 10, and that have a lot of characteristics of a lottery.

Yeah.

And so, back 20 years ago, this was day trading. And there was only so much risk out there. That was more of 1 for 1. You can make 1, lose 1. But then financial, quote unquote, innovation happens, and people come up with products that allow you to take a lot of leverage, either through options or through just straight leverage. And so, you can put $1 up and try to make $10 or try to make $100. And there's this whole conflation between investing and trading and gambling. And there's always been some gray area. cross flows. But you open up a Robinhood app today, and it's, do you want to day trade? Do you want to buy zero-day options? Do you want to gamble on the sports game tonight? Or do you want to buy the index fund? And it's kind of all presented as, this is all investing. And I think that's a really, really scary message that we're teaching. You know, the 16, 18, 20-year-olds today.

Are you—

so I think there's a ton of agreement there. I'm interested in where this goes forward. Obviously, there are things individually you can do, just like, you know, stay away from the aggressive gambling products. There are things that you can do in your community encouraging people not to take on unnecessary risk and stay with the safer investments. On a national level, there is a conversation about regulation, restricting, banning more aggressive financial products. There's another side, which is maybe fund more education to teach people about this. It's sort of what we did with cigarettes, where there was a lot of regulation for who could buy them, but also a lot of education where pretty much everyone knows that they cause cancer, there's labels on there. And I'm wondering if you're more optimistic that we get out of this with an education strategy or through a regulation strategy, or is there no hope and it's every man for themselves and they have to resist the temptation individually?

Yeah, education strategies take a long time. And you also kind of have some vices that kind of catch on like wildfire and will eventually burn themselves out. And you can look at a lot of illicit drugs will have that type of characteristic.

Yeah.

And regulation's always going to be slow, but I think it's faster than the education side. I think the scary thing today is that there's a number of products that have been, again, innovated to be of a higher intensity and easier access than what they've ever been in the past. And the challenge is regulation lags on those. And so, you can think about things like marijuana. And the marijuana of 20 years ago is very, very different than it is today. And the access is much more available today. And you can think about pornography. The intensity of porn today is very different. The access today is very different. And gambling is the same way. Historically, you either had things that were high friction and high speed of play. So think about a slot machine where you have to go to a casino. Oftentimes, that meant Getting on a plane and going someplace, right? And you sit down. You could pull the arm of the slot machine pretty quickly. In fact, it was too slow, so they made it a button so you could just sit there and boom, boom, boom, boom, boom, right?

Right.

But, there was friction on high speed of play, or you had something that was low friction, like I could buy a lottery ticket at the supermarket when I'm buying my meat, but it's low speed of play. I buy my ticket, I wait 8 hours or 3 days, or I find out whether I won or not.

Yeah, trading cards now, you can buy a pack of cards, but you don't have to wait for them to arrive with you to know what you got. And you can decide, I don't want any of them. I'll just roll the dice again.

Roll the dice again, yeah. And so, what gambling is today, with the phone, access through the phone, and being able to— I don't even want to wait 3 hours and figure out whether my team won or not. I can bet on the next pitch. I can bet on every play.

Yeah, there's 1-minute markets now for, do you think Bitcoin will go up or down over the next 1 minute? And so, that is effectively a binary option on 1 minute, which is fairly close to just a pure slot machine.

Yeah, it's just flipping a coin, right?

Yeah, flipping a coin. Exactly, because no one has information on a minute-to-minute basis. I mean, maybe somebody does, but certainly not the average gambler.

Yeah, and so you were mentioning earlier just about, luckily, this is happening when we're older. But I have a teenage boy, and teenage boys are kind of the most susceptible to these things. And there's tens of millions of teenage boys out there.

Yeah.

Who are all getting drawn into these products. And I think there's real risk to society. And so the question is, what's the right way to allow access to products? I have libertarian bones in my body.

Sure.

I think we need to be careful about banning access to products. But what are the right and smart guardrails to put on them? Yeah.

Yeah, in some ways, In some ways, multifaceted approach to regulation would help a lot of companies that are in positions where the right thing to do for shareholders is actually to make the gambling product, right? And we've seen this because there's investing apps now that make more money from gambling products, whatever they want to position them as, than stock trading.

Mm-hmm.

And we even saw 4 months ago, there was a series of headlines around how Meta was exploring building prediction markets, which I always felt was completely at odds with their stated goal and mission. But at the same time, you can make the argument as a Meta shareholder that they have a lot of attention. There's people that are making a bunch of trades based on information they get on Meta, and that product should be integrated into the product because it's generally legal, right? And it just feels like companies in many categories will get punished by not rolling out these products because their competitors will, and they will onboard more users and generate more revenue and be able to spend more money on marketing. And it's sort of this like really, really vicious cycle. And so I've gotten to the point where I pretty strongly believe that a lot of this stuff needs to be pulled back or banned. Of course, that sort of human desire and energy to speculate will just naturally flow to other places, right? You saw in some ways, the AI trade over the last year sucked a lot of the wind out of the sails of crypto, right? There was way less appetite for Bitcoin when you can make 10x on some bottleneck trade. Micro- I did want to get to as much as I want, as we could spend hours talking about our casino world. I did want to get your point of view on how you've processed the overall sort of AI supercycle. There's so many different elements of it that I'm sure you've drawn— you have experience in through the early 2000s in that period, whether it be obviously, energy markets coming into play in tech for the first time in a really meaningful way, massive CapEx and debt cycles, and seeing some of these companies that have historically had pristine balance sheets figure out all these super complex structures. And then I actually want to get into compute markets as well—

Sure.

And how we should be thinking about that from your point of view. But maybe start how you've processed the the sort of the last 5 years and maybe your mental model for where we are today?

Yeah, you know, I probably started talking to people in this space around the 2020, 2021 timeframe. And then kind of like everybody in '22 was kind of surrounded by it. What I'm amazed at is The number of predictions that came true.

Yeah.

And when I heard it, as a guy that's kind of outside of the tech world, that it just seemed farcical, the growth that they were projecting, that they were saying that things would keep scaling forever.

Yeah. And tech people have been wrong so many times, VR and flying cars and self-driving and NFTs and all this stuff. And then this was the one where it was like, oh, again, with the crazy predictions. And then it's been right on track.

Very different. And that you could keep scaling, and that you would never hit that plateau. And again, like in every other field, you hit plateaus, you start getting significant decreasing marginal returns. inputs. And that was my expectation of how this would play out. And it certainly hasn't. And so, I think I've been open to the possibility, but skeptical. And now, whenever the big money was coming into data centers a couple years ago, and you had the calls for the bubble, I think my view was just because a lot of money is coming in doesn't make it a bubble.

Yeah.

Now, it is true that you always say it's different this time. Occasionally, it actually is different this time. The real test for the industry is, can you find useful products and services that people are going to pay for? And if so, it works out great. And if not, it's going to be overinvestment. I do think that every commodity market goes through these booms and busts. And that's because you have the producers see the same price signal, producers and end users see the same price signal at the same time, and typically react in the same way. And so, that creates Whenever prices are high, people increase investment. There's a delay before that comes on. Oftentimes, the end users have figured out ways to optimize their demand or to find substitutes. And then you get that bust that happens because too much capacity shows up in the future. And I think that's the big risk now with data centers is that Every megawatt that's been built thus far keeps being worth more. And it's kind of a highly backwardated curve where if you can provide compute tomorrow, that's worth a lot of money. Compute in 2 years is worth less. Compute in 5 years is worth a lot less.

Right.

But everybody is seeing that, that if I can build today and supply today, I can make a lot of money by even renting it out in the short term. And I think the question is, is everybody overbuilding a little bit because they look at the same price signal and say, even if we overbuild, we're okay, because we can sell it to somebody else?

Yeah, or even you can imagine some of these models saying, well, yeah, well, you can discount the rate by 60%, and we're still We're still making— we're still going to be making money here. So of course, there's— that should be plenty of margin of safety when in reality you could see a much larger drop depending on how much of an oversupply there is.

Yeah. Have you been able to tease out the impact of AI on markets from other factors? I mean, We were looking at 10-year interest rates are at, you know, very high. There's been a big spike over the last 30 days. All the tech people are sort of saying, look, it's evidence that data centers are just too good, and so everyone's investing there. At the same time, there's a war in Iran, and so that might be a bigger factor. But even in the GDP numbers, it's very hard to actually get to a ground truth of the impact on productivity. The internet famously never really showed up in the productivity statistics. An incredible amount of wealth was created. So how do you think about teasing out the different effects? Does it even matter? Or do you have a process? Do you have a view right now on how much growth or, or effect on markets AI is actually driving?

Yeah, I think it's really hard. I think you did a great job Kind of setting the stage for the difficulties the field is having. I think the financial world and economists are notoriously just very bad at predicting interest rates and inflation. And it's even hard to describe why have rates spiked so much just in the past month. Yeah. And part of this is, there were times when everybody was scratching their head about why are rates staying so low for so long. And now, it's the opposite, that inflation is manageable, at least today, at that 2.5%, 3% levels. Having the 10-year, what, 5.20% or so today? It seems like those real interest rates are at a level that I think people, many economists, were doubtful that they would see again. And so, I think that the question is this is how much of this is being driven by the AI super investment cycle? And is there just too much demand for capital because there is both all the sovereigns are issuing enormous amounts of debt, as well as everybody associated with the AI industry is trying to do the same. I think it's always easy to assign causation after the fact. But trying to predict where this is going in the future, it's ended the careers of many a trader.

Yeah. How do you think about the bullwhip effect? One thing that a lot of tech insiders have been pounding the table about is why isn't TSMC building new fabs fast enough? We finally got AI working. It's going to continue to increase and scale, scale, scale forever. Nvidia is ramping up. Everyone's fighting for chips. And yet we can't seem to get TSMC to mention AI on an earnings call. They're They're starting to, but they've been burned before. And I'm wondering if there's parallels in other markets that you've seen. And how do you actually think about the nature of that bullwhip when it actually comes around?

Yeah, and I think that's exactly it, right? That if you're a commodity producer, like the memory chip companies kind of are, after you've kind of fallen for this The boom-bust, that you expanded capacity at the top of the market. And again, all your competitors saw the same price signals and did the same thing. And you have that bust. And that bust just causes so much financial pain that at some point, you're like, I'm not doing this unless you, my customers, are going to sign a 3 or 5-year offtake agreement. at prices that pay for the whole factory and then some, such that if we've overbuilt, that's on the customer, that's not on me. And you can certainly understand why a couple of years ago, all the hyperscalers were asking memory companies and the chip companies to increase production, increase capacity, and because it would be beneficial to the hyperscalers who were buying this stuff. And the producers were just saying, we're just much more hesitant to do so. Because again, like a couple of years ago, AI could have gone a number of different ways. And, you know, it's gone the bull case to date. But, you know, where it goes from now, I don't think these companies want to bet their future on it. And you've seen the same thing kind of in the oil markets where the reaction to high prices, the industry Interest and investor interest in funding, recycling all the money back into drilling is much lower than it used to be. And investors are saying, we fell for this in the past. Whenever you're making the big returns, pay them out to investors.

Can you reality check me on my history around the shale revolution? I'm interested in, could that have happened anywhere other than the United States? How much of it is a technology story? How unpredictable was it? Because right now in tech, everyone's talking about energy. People are talking about nuclear capacity that might come online in 2030, 2035. If it comes, I don't think it will be a surprise to the technologists who have been raising money and proselytizing and evangelizing for nuclear. But from my perspective, as like an energy outsider, the shale revolution looked like something that was an unexpected boon to energy in America. But what was it like on the inside?

Yeah, so the shale volumes have really been over the past 20 years. But this is You know, this industry kind of got started maybe 50 years ago.

Okay.

So in the '70s, uh, federal government started investing in the technology through some basic science grants. Um, in the '80s, George Mitchell, who was really the kind of the grandfather of this industry, um, started investing and doing some, um, you know, using his, you know, his company's capital to go test wells. It was really in the '90s, whenever George Mitchell kind of came up with the slickwater fraction— sorry, fracking.

Yeah.

Um, that, uh, that you could see wells start to become economic. And the industry kind of had the bonus of the commodity boom that happened in the 2000s, um, kind of culminating in 2008 when Natural gas got to $13.50 versus the $3 it's at now. Oil got to $147, where a lot of those wells, even though they were expensive to drill, were economic. That's when that 2005 to 2008 period was really where the shale started to work.

Yeah.

All triggers had happened, and you had high prices. And then the 2010s, you had kind of the industry really start to figure out how to do this at scale and get all the benefits of coming down the learning curve and the benefits of going from drilling a well here and a well there to really kind of turning this much more into a manufacturing process.

Do you think—

And in 2020, it's— a mature industry. So, it was 50 years in the making. And I think the question is, un-nuke, what happens? In many ways—

Oh, yeah.

There are aspects in which the United States is very conducive to the nuclear industry. We have very robust capital markets. We have a lot of great technologists here. We have a lot of land in the United States. very deep electricity markets. So, if you come on with a gigawatt of power, you can place it very easily. And you have bipartisan government support for the industry.

Sure.

So, all those things are working for it. And the downside is that the US actually has fairly cheap electricity. We have fairly high cost of labor, especially the craft labor. And And to build a nuclear plant requires a lot of labor, at least today. Siting anything is hard in this country. And you have this kind of very bifurcated utility system. We don't have a national utility.

Yeah.

It's kind of state by state. It's utility by utility. And they're investor-owned. And does the investor get the benefit of this? They certainly get a lot of the risks. And then last, I think we just don't have the workforce that's trained in this field today. And so in some ways, maybe the optimal thing is we develop the technologies here, and they get built somewhere else where it's just easier to build and cheaper to build.

Yeah, we've seen that with a couple nuclear companies that have done their first plants in Southeast Asia and internationally and seen opportunities elsewhere, but I'm certainly hoping for it to hit America. I have one more, but please, Jordi.

How have you processed the growth of the private markets and the subsequent disconnect between the private markets and the public markets? I've been personally having made at this point 70-odd private market, early-stage investments over the last number of years, I've been really feeling like the inflation, the feeling of inflation in the private markets where you have companies that are a $10 billion company, but only because there are a bunch of venture capitalists with $10 billion, right?

Mm-hmm.

And so, you just have these prices pushing up, not based on the quality of the underlying asset or or the category. And I just— so my head is really thinking about how long can you have this disconnect? The disconnect can't continue forever with how extreme it is right now. And so, I would expect it to correct sometime in the next 2 years, maybe, who knows? But how are you processing it?

Yeah, we're certainly seeing that question Right now in the energy tech industry, where a year ago, it was probably the peak of— almost exactly a year ago, peak of a lot of the SMR, small modular reactor nuclear companies. And they've been on a steady decline since. And then a few months ago, you had Fervo go public. That had a great first day, first couple of weeks. It went public at a big number, and that has gradually sold off. When Fervo went public, a lot of companies in the energy transition space thought that the IPO door was open. Everybody called their banker, the bankers call all these companies and say, okay, get ready, we are going. Then everything is sold off from there. And now, they're all getting the same message, which is, OK, the door closed really quickly. And now, what do you do? And the highs are higher in the public markets, but the lows are lower, too.

Yeah.

So, if you have these companies that still need to raise a lot of capital, going public, if you catch it right, can work out really well. it can also be handcuffs on you, that if your stock trades poorly from the IPO on, and you have to go do secondaries in the market for a stock that's down 50%, 60%, 70%, it becomes enormously difficult. And so, there's a lot of CEOs of those energy transition companies who are Now, trying to decide, do we force this, even take a lower valuation and try to get public, or do we stay private where we can have more nuanced conversations? We only have to convince a few investors to come in rather than the whole market.

Yeah, it will be more stable.

Yeah.

The other dynamic that is endlessly entertaining to me is companies repricing based on the category leader in their category. So you saw this with SpaceX, right? Like SpaceX is far—

Oh yeah.

Is probably like 1,000 times more important than the next space company, right? But every company right after the IPO starts to like reprice in the private markets based on that. And you can see the logic from the private equity side private market investors. But again, there's those disconnects that have to be righted in the fullness of time. Like, you can't— no, it doesn't matter how big your venture fund is, eventually your businesses are gonna price based on market realities. Last question I had for you. I know we're over time. How did you process the compute markets ban? We saw Recently, we saw a couple prediction market players try to put up compute markets, and apparently, they got a tap on the shoulder from Washington and were asked to shut down the markets. It seemed somewhat curious, but how would you read into that?

Yeah, I haven't followed this too closely. I've always been somewhat— the big question is, is that compute is not a commodity where one data center is providing the same product as another data center. And in order to have a real tradable market, you either have to deliver into a certain product that needs to have a lot of buyers and a lot of sellers that are able to take that specific product to create and have an arbitrage-free type of ecosystem. Or you need some type of index that the industry trusts that might look at across a number of different types of compute and say—

Yeah.

Compute that has characteristics in this range The cost for that for the month of September is X.

Mm-hmm.

And so, yeah, either a robust index or a kind of physical delivery product. And I think both are really challenging in this field of compute. There just aren't that many short-term deals. Transparency into those is difficult. The distinctions between the chips and the design of the data center and all the specific needs is just different. And so trying to either come up with the physical delivery mechanism or the index that everybody trusts to be right is really hard.

Yeah, the other dynamic is just how much more valuable compute is to a couple companies. Relative to the broader market, which feels quite different than some of these other commodity markets. I did have one, my second final question.

Please.

And then we'll actually let you go. How much have you opened your mind to various sci-fi predictions, given how correct many of them have been in the last, let's say, from 2020 until now? 2015 until 2025. So many of them sounded crazy at the time and almost impossible. Things happened that were almost impossible to predict. And given how correct many of them were, it's hard not to place way more weight on them going forward, except that they actually only get more insane from here.

Right.

And so, you have to take another Yeah. Leap of faith?

Yeah, it's a really interesting question. I'm a natural skeptic. I've always kind of been a bear trader. I make more money when the markets are going down than when they're going up. I would be a terrible VC because of that. I would tell you all the reasons why this company is going to fail rather than why it's going to work.

Sure.

I think getting to an entrance into the tech sector that I've had of late has helped open my eyes. And rather than being naturally skeptical, I've tried to be naturally open-minded to this. But in terms of specifics, I think the next 10 years could be crazy, it could be a normal technology. And I think that's this huge question here. And I try to keep my mind open to both of those.

We're going to find out one day at a time.

Yeah, just one day at a time. Well, thank you so much for taking the time to come chat.

Yeah, I really enjoyed it.

Really informative conversation.

Yeah, it's been great.

Hope we can do it again soon.

Yeah, great to hang out.

Have a great weekend. Have a great rest of your day. We'll talk to you soon. Goodbye. Let's move right into our next guest. We have Viswa Pollaroo, the founder and CEO of Inveta, with a huge fundraising announcement. Get that gong ready. This one.

How are you doing?

Welcome to the show.

Thank you for having me. Long time listener, first time caller.

Sorry for the delay. We had to shift things around. I'm so happy that we get to talk to you. Let's kick it off. Jordy's warming up the gong already. Tell us about the most recent fundraising round.

The most recent round was a lot of capital to get a lot of medicines to people.

There we go. And what's important for a lot of A lot of catalyzers. 311 million.

I'm guessing I'll be back.

Okay. So yes, what you're building.

We're building, the easiest way to describe it is think of it as a sequencer, but for life's chemical code instead of life's genetic code.

Sure.

About 400,000 compounds have been discovered by the collective human endeavor from throughout history to now. And it's expected that there's about 1 to 10 billion. So 99% of what makes up you, a tomato in your garden, or a random sample in the Amazon rainforest is still a mystery to science.

Mm-hmm.

And that's 'cause we've never built anything that can take a biological sample and answer 2 really important questions. What are the molecules and what do they do? So we've built the tech to do just that.

Talk about the positioning of the company as actually developing drugs going through FDA processes versus selling a software product to existing pharmaceutical companies or versus focusing on more general use artificial intelligence technologies. How'd you land where you did and What are the benefits of that, of that approach?

As far as we can tell throughout the history of the industry, there's only been one way to build a big company in this space, and that is to make and own drugs. Right. And even the biggest companies in this space are largely underpinned by cash flows from a single blockbuster franchise. Yeah, something like 70% of Lilly's EV is forward sales multiples of the GLP-1+ family. 80% of Novo's is the same. And 30% of Sanofi's is half of one, a drug called Dupixent for eczema.

Wow.

So if you want to build a large and impactful drug company, the rule is super simple. Make drugs and make drugs that matter.

Gimme some timelines. AI is focused on curing cancer, and I feel like we gotta cure peanut allergies before we cure cancer? It just seems easy to me. Am I misinformed? What are your peanut allergy timelines? You know, it's just something that seems a little bit quicker to test than cancer, which could take, you know, months to actually fight and verify that you're in remission. If you have a GLP-1 for peanut allergies, how do you test that?

This weekend, John, figure out how to give yourself a peanut allergy.

I need a million dollars to compute, I think.

And then fix it on Sunday.

But the more anodyne diseases, are those actually more tractable or is it a straight shot to cancer and like the big ones?

I think they're both difficult in different ways. I'll start with the simple fact that cancer isn't one disease.

Yeah.

And it's thousands of diseases.

Yeah.

And we just call all of those diseases cancer for something that grows without control and invades other organs. So I think that's going to have to be disease by disease. And we've made a lot of progress for a lot of these diseases, but it's going to help, I think, for us to be very specific about what cancers have been most recalcitrant and why. And I think I'd love to see AI come up with completely new ways, whether it's molecules or mechanisms, to be able to hit cancers that we struggle the most with.

Yeah.

Common diseases, are difficult for 2 reasons. They're complex and multifactorial. So in other words, obesity is something you understand quite simply as excess fat in your body. But we now know that at least 2 organs, likely more, are involved in obesity. So your fat tissue, your brain, your muscle, your pancreas, all of those coordinate to ultimately give you that. The second reason they're hard is the way each of these factors collude to give you obesity can be massively different across the population. I was just reading something as I was waiting for you guys since you went over that 10 to 15% of patients don't respond to GLP-1s.

Whoa.

Right? And it has probably something to do with the fact that hormones affect GLP-1 massively. So women tend to respond better. And there's lots of other things that are just beginning to come out from the data. And these are the 2 things that make common diseases more complex for AI. But I think there's no reason to stage gate. We should absolutely attack all of them at once and continue pushing the frontier for better medicines so each of us can live a better life on a day-to-day basis. And that's the test we want each of our molecules to pass.

So yeah, I mean, it seems like you're taking a pretty broad approach, but is there a decision criteria where you're looking at the economic opportunity of a particular drug and treatment versus the total harm caused? Like, obviously, cancer is at the top of the list, peanut allergy is a little bit lower. Or are you more just like, let's generate a bunch of stuff and see what the impacts are? Oh, turns out we cured, you know, we cured cancer. day and then the next day we cure peanut allergies and it's sort of a random downstream effect. How much of it is you picking the target versus just seeing what you're capable of?

Yeah, I'm glad you asked. So at Enveda, we haven't started efforts on trying to treat cancer yet.

Okay.

We're focused on really large diseases that still have big, big swaths of what we call unmet medical need. For example, if you suffer from asthma, there hasn't been an oral medicine that isn't a steroid that has been approved for over 25 years.

Interesting.

Wow. Right? If you are one of the many, many Americans and the world citizens that has trouble managing your weight, you can get on GLP-1. In 1 or 2 years, you're likely going to be off of it for one reason or another, and then you have no idea what to do about your weight.

Interesting.

Our first 2 drugs are hitting exactly those problems. We want to make a safe nonsteroidal oral for diseases like atopic dermatitis or eczema and asthma, for which there's millions of people in America alone, and there isn't an ideal solution. And our second molecule discovered a new hormone that's produced after you go for a sprint. So it's essentially the chemistry of exercise. And we did to it what semaglutide did to GLP-1.

Hmm.

We put it in a pill that you can take once daily, and we think it'll allow you to maintain your body weight and metabolic health over the long term, giving 55 million people in America an off-ramp over the next 7 years that will have taken a GLP-1 and then not know what to do next.

Wow.

A lot of discussion over should we give superintelligence a wet lab? What is the actual human in the loop? I mean, it feels like you're hiring feels like this is a very rigorous process where you're coming up with candidate molecules, taking them through the full FDA process. But how do you make sure that your product is safe? Are you confident in the current system, or do we need to create new rules or slow down? Or what's your view on the future of actually getting new AI-aided, AI-designed products in the hands of Americans?

For better or for worse. 3 quick things. The first one is NVIDIA's core AI models allow us to understand what evolution's already made. So they're not optimized for creating new things, either benevolent or malevolent or anywhere in between.

Okay.

And so this allows us to, for example, push forward our frontier very, very quickly without being particularly bottlenecked by the safety question. Second is, for people whose models are designed to make molecules, I think that most, the most important aperture or filter that those models have to pass through is the simple fact that most things that a model can design are not something human chemists can make. So, and this doesn't get talked about often at all, but—

Interesting.

99% or more of the chemistry that any particular model comes up with is constrained by the physics and energetics of putting it together.

Oh, interesting.

So I'd say just because AI can come up with an incredibly powerful toxin doesn't mean that it'll be produced right away. And that's the bottleneck even when you're trying to make medicines produced by AI. Third, I think the big bottleneck overall that encompasses issues 1 and 2 really is the fact that AI's really bad at moving things in the physical world. And I think a lot of a lot of biological surgery, whether it's of cells or genes or assembly, involves things that require custom workflows and automating and roboticizing that still is an endeavor that we're not very good at. So unless there's a rogue agent in the loop with the AI, I think it'd be hard. Now, could then a rogue agent be significantly more damaging? Yes.

Yeah.

But I think it's a more human-centric problem than it gets written about.

Okay, I like that. Well, thank you for breaking it down. Congratulations on the progress.

Yeah, great to meet you. Wish we had more time, but come back on.

Yeah, we'll have to have you back on soon. This is really great. Congratulations.

Sign me up.

Thank you, Chris.

We'll talk to you soon. Goodbye. Let me tell you about Console.com. Console builds AI agents that automate 70% of IT, HR, and finance support, giving employees instant resolution for Access requests and password resets. Coming back on the show, we got Jake Adler, founder and CEO of Pilgrim. Jake, how you doing? Sorry for keeping you waiting.

No problem. Good to, uh, good to be back on. Thanks, guys.

What's the latest in your world? You raised some money. Anything gong-worthy? Anything that's going to make Jordi stand up and smash that while you're giving us the important details, deafening us all? What happened?

Yeah, so we, uh, we closed the $25 million round to build America's first biology prime.

Fantastic.

There you go.

It's so funny. Like, 2 weeks ago, I was— it's been probably a year since you've been on the show, but 2 weeks ago, I was thinking about you and Pilgrim because of all the biosecurity talk, and I'm like, he's got to be raising a new round, just because it feels like you've been building to sort of meet many of the current and future problems of our world.

Yeah. How are you pitching the company right now, and then the specific products? Because there is sort of a differentiation there between being America's biological prime contractor, Kingswell, the actual product, correct?

Yeah, there's certainly been a bit of an evolution, but I think the core, the core thesis has remained the same.

Yeah.

Which is really this idea that we're just trying to deploy biotechnology more efficiently.

Sure.

Right. So there's like an abundance of like really compelling technologies that it's like language in academia. And the bet at Pilgrim is really on, on the mechanism to take these like really critical technologies and get them deployed. So, you know, at the core, when I think about what we're building today with like a system like Argus— so this is one of our early prototypes that we deployed— the idea really is to build up what is a really substantial wedge and get technologies like this into the hands of the warfighter today. But on the basis of building what is a bioprime, it really does come back to that core mechanism of being able to actually get technology into the hands of warfighters sooner and ensuring that more you know, warfighters can go into conflict and actually be able to return home. So it's been a very exciting mission, but it's all really revolved around just the idea of getting bio into the hands of the people who need it the most.

The sign behind you says, smallpox, keep out. Is that because there's smallpox in your facility and you don't want people coming into the facility, or is that because you don't like smallpox and you want to keep it out of your office?

You know, you'd be shocked, you know, how easy it is to procure some like really gnarly biological agents.

Black Belt.

But yeah, exactly. But, you know, we don't have smallpox in the office. We have a bit of monkeypox here. And you'll be able to procure, you know, like a couple of chemical weapons as well. But like, you know, the amount of validation that's required for these things is bewildering. You know, they'll call me up on the phone and be like, like, Do you know what you're doing with these things? And I'm like, yeah, we work with the government. And they're like, that's all we need to hear. Which is, again, incredibly alarming, you know, considering a lot of the infrastructure we have to deter a bioweapon attack or to stop the proliferation. None of this has changed in like 20 years. It's been a really cool mission. It's a bet that we made 9 months ago on Argus. And candidly, you look to things this year, like, you know, for example, the Pentagon shutting down 4 months ago due to like a false anthrax alert. Yeah, there are really serious events that are transpiring. And I think that the, the, the rate and the incidents around bio have grown dramatically, even to a point now where I think the sheer lack of that infrastructure has become incredibly alarming.

Okay. That ARGUS system looks portable. It's on a tripod. Where do you deploy that? I've heard about In terms of our biodefenses, sometimes airports, wastewater treatment systems, there's a whole bunch of other places to plug in. How did you land on that portable of a system as opposed to tapping into basically like a stream of biological data somewhere else in the American economy or American industrial system?

I think the core bet is really on the platform itself. We really want to move beyond what is like the current precedent, which are like these very manual workflows. So right now, for example, like if I was to, you know, if I was to pull Jordi in and have him go work the line at TSA, for example, and some guy comes through and I gave you like the best and most portable sensor in the world, it's able to tell you whether or not some person has Ebola. The question quickly becomes like, what are you going to do with that information? Right. So we definitely knew that the the system itself had to be portable, but it had to go just beyond, you know, from a platform standpoint, just doing the detection to actually enabling identification and characterization of the threat as well. The reason that we opted for this form factor first was very simple. It was really on the basis, especially in the airborne environment, that like that is the primary mode of transmission for viruses and for pathogens. Right. So when we're thinking about how these viruses spread between populations, It's very important that we're actually monitoring the air. There's a lot of information that we don't typically see. So we wanted this system to look scary and really to stand out in an airport environment because we have to make a statement to our adversaries that if they attempt to come into our country and attempt to proliferate a biological weapon, that we actually have the defenses in place to be able to deter and identify that proliferation very quickly. Yeah. So that alone is really what contributed. But the core platform being able to detect viruses autonomously, that's just going to be deployed across all of these different types of media, wastewater, clinical, everything.

Last question for me on the actual sensor fidelity. Is this going to be something where I'm going through the airport, I take off my shoes, they take a picture of me, they scan my thumbprint, I give them my ID, I give them my boarding pass, and then I also have to like breathalyze into this thing? Or is this passive? Like, are they taking my saliva? Are they going to take a pint of blood from me? Like, how far does this go?

Yeah, you know, I think if we make the TSA line any longer, people aren't gonna, gonna be very satisfied. Uh, really, the, the goal here is simple, is we just want to monitor the environment.

Okay.

A lot of, a lot of tools in the past, like, uh, you would call like contact tracing, and like even just thinking about like just how crazy of a time COVID was, I remember my sister like Lysol wiping bananas You know, this is sort of like the world that we are living in. A lot of the technology that has been built tends to be hyper-personalized. And what we recognized more than anything is like the current timeline to even detect a threat in the air is like a week. And, you know, that means by the time you find out about a pandemic, it's typically in the newspaper before one of our sensors are even picking up on it.

So we're like Google search trends for I lost my sense of Taste or whatever. Yeah.

Yeah, like literally, right? It's like these very small ideas. So we wanted something that'd be quick and monitor the air. And it's like a— like the joke more recently is like a Shazam for the air.

Like, how do we just like— Shazam for the air. I like that. Any plans to put it on a robotic dog? You got the drug-sniffing dog, the bomb-sniffing dog. I like the idea of a pandemic-sniffing robotic dog. Could probably carry that payload, maybe? Who knows? What are you thinking?

Yeah, I mean, it could certainly be interesting. I think that there's also a possibility of putting this thing on drones.

Or wheels. Yeah, I mean, you can drive around. Yeah, on wheels.

Yeah, just like have it run after people at TSA, just like coming up to them and staring at them. Yeah, it would certainly be very interesting.

Well, good luck. Congratulations. I'm glad you have fresh funding to help keep the—

Yeah, very important work and fantastic Yeah, we love to see it. Great to see you, Jake.

Thanks so much for coming on, Jake. We'll talk to you soon.

Have a good one.

Let me tell you about Codex. Codex is a powerful workspace for getting work done with AI agents, whether you're writing code, analyzing data, creating content, or automating business workflows. Codex helps you move projects forward from start to finish. Our final guest of the show is the co-founder of Freebean. We have Adam here. We're the first parents, but But there's been a little Easter egg this whole show. We have his product here with us in the TBN UltraDome on the desk.

Cracking open a can of Ramp.

How you doing?

That's how we started the show.

This was a dream of Jordi's. He had this idea, uh, a year ago, 2 years ago.

I was searching for more ad inventory. We had, we had more ad inventory than any show on earth, but I wasn't gonna settle there. I wanted more ad inventory, and I thought, we're drinking, we're not getting paid paid to drink on the show. How do we fix that?

You're not, but now you have some new ad inventory courtesy of Ram.

Yes, and Freebean. So yeah, tell us about you, your business, your career. Get us up to speed on all this.

Absolutely. I'm Adam Krasinski. I'm the founder of Freebean. Freebean is a targeted, trackable, and tangible out-of-home advertising medium built on free coffee.

Okay.

We started the business while I was in college, a senior at the University of Michigan, and I've never had a full-time job. So, yeah, go blue, baby.

Wow.

Very, very, very great story.

Yeah.

Thank you.

How much do you know about the sort of like white space in advertising inventory? I'm obsessed with, you know, AppLovin finding that people are playing hyper-casual mobile games, they went and found this pool that wasn't on Instagram, wasn't on Facebook, and they blew it up to a huge business. At the same time, on the other end, I know someone who did free napkins and fortune cookies for Chinese restaurants, but put ads on them. And it was pure margin, because the napkins are pretty cheap. And so they would go and say, hey, restaurant, do you want some ad-supported napkins? And the advertisers would say, yeah, we'd love to put our logo all over that bar. What is the shape of these opportunities? Why did you pick beverage? And then, you know, how has it actually been going distributing this to various places to get in the hands of consumers?

Distribution has been going incredible. We just crossed the 100,000-can mark across a bunch of different channels, accelerators and incubators in partnership with Entrepreneurs First, Antler, Techstars. Conferences and trade shows are a huge channel for us as well. And then on the consumer-facing side, college campuses and things like that.

Cool.

The reason for beverage mainly is because it's tangible and consumable. We like to say that this is an ad you sip, not skip. When talking to advertisers, it's, it's really one of the only formats that's actually consumed. It's, it's super unique.

Yeah.

And the reason we chose coffee, 66% of US adults drink it every single day.

Yeah.

an extremely popular beverage.

Yeah.

And it's— it allows us to target virtually any type of demographic regardless of what brand is on the cans. It could be Ramp or it could be Red Lobster, another customer of ours.

Sure.

Their customers all drink coffee.

Yeah.

Or at least 2/3 of them do. So it allows us to target those people effectively and drive high return for, for our advertisers and brands.

Yeah. What's the sweet—

oh, sorry, fellow Michigan student just texted us, and, and, uh, or former student, uh, just texted us, uh, didn't graduate, but, uh, uh, just texted us if you plan to expand into beers. Is that, uh, is that, is that a potential category?

I have been asked that a lot. Alcohol is tough, so we're, we're sticking with coffee for now. Maybe down the line we'll expand to beer, but we'll have to—

I say straight beer, straight shot to the original Four Loko recipe. Bring it back, ramp it, wrap it in a brand.

Four Loko might be, might be the move. I know you, the, uh, on YC Demo Day when the guy had the Humanoid and you were like, can you shotgun that Ramp Cold Brew? And then somebody was like, oh, Four Loko. So yeah, maybe, maybe these are shotgunnable though.

Maybe in the future. What, what is the, what is the key, uh, to success as a brand partnering with you? Is it I go you, I buy a bunch of cans. I mean, the brand looks great on the actual can.

Thank you.

But then should I mail these to my customers? Should I have this at my conference? Like, how should I think about actually getting the most out of the pallet when it's delivered? Do you handle logistics if I want to send it to my customers to remind them that they should be talking to their sales— to my sales reps, for example? Like, what is the right way to actually get this in the hands of potential customers?

Sure.

So when it comes to distribution, if we're looking at gifting, something that Ramp does a lot of actually, there's usually third-party fulfillment partners who will handle the packing of those cans and then shipping them out to, you know, maybe a prospect or something like that. For the right distribution channel for the brand, it's, it's dependent on a couple of factors. Some brands take the approach of, hey, this is a brand campaign, or some look at it as an acquisition channel. And conferences work extremely well. We've been outside some major ones. Figma Config with Paper went super viral on Twitter and they basically called it the highest converting booth at the conference without a booth.

Wow.

Because Paper was everywhere at their direct competitor's conference. And so again, you think about like putting something that's tangible in somebody's hands where traditional advertising formats don't really deliver anything to the consumer. Billboard companies, the consumer just sees a billboard. But when you allow a brand to put something in someone's hands for free, especially something that's expensive and increasing in price, you create this reciprocity effect where a person is opting in to seeing an ad and They're saying, okay, I'm gonna take an ad, I'm gonna take a free coffee as well. And it just drives higher return for the brand.

Yeah. Last question for me, talk about what it takes to work with you. Do you have hard minimums? Do you have maximum capacity right now? What's the sweet spot of a brand coming to you? Is it 1,000 cans for a conference, 10,000 cans? What's a, what's a really down the line pitch for you these days?

For a conference, it's primarily dependent on the size of the conference. If we're talking about Dreamforce, yeah, a brand is gonna want a lot more cans. If it's a smaller conference, they'll want a lot less cans. We do have a hard minimum specifically for our storefront, which we just launched our grand opening on Wednesday in New York City between Bryant Park and Grand Central. It is It's the world's first free coffee shop.

Cool.

And as you can imagine, the coffee is literally free.

You walk in, you ask for it. Ad-supported coffee.

You ask for it.

Yeah, wait, so what are the different— yeah, what are the different— what brands can I enjoy free coffee from?

So at our storefront, we have Ramp, Novig, Warp, Outset, Paper, and Bland.

Wow.

That's so good.

So you can get AI voice-supported coffee, you can get gambling-supported coffee, you can get— Enterprise payment supported coffee. It's really like a choose your own adventure kind of thing.

It is. What I think is also really interesting, there's a negative sentiment around AI advertising, more so in New York than San Francisco, of course, but people don't seem to care. And as I alluded to, where you provide value to the consumer, yeah, they're like, hey, I'm opting in to get this advertisement. And yeah, it's okay if You know, a voice AI company is saying, hey, call this phone number to learn how customer calls can help.

Yeah, nobody's vandalizing these.

Any plans to add a speaker to the can, the bottom of the can, so it can chirp out, you know, while you're enjoying this sip, you know, you might be interested in learning that, you know, Ramp customers save an average of 5% on their expenses.

Probably no speaker, but we are We're exploring thermochromic ink, similar to like a Coors can, where if it's cold, it displays something different than when it warms up a little bit.

I like that. I like it. I love your clarity of thought on this opportunity.

Thank you.

It's clearly very sharp. And yeah, I'm excited for you to scale to billions of cans annually.

Thank you, guys.

Thank you, guys.

I appreciate it. Come back on the show soon.

Yeah, let's do it. And we have a bunch more advertisers, so let's get them all on board.

We got more cans. We need more cans.

We have a solid amount of crossover, more than you think.

MongoDB. Yeah, fantastic. Ramp, obviously. We'd love to do more. Thank you so much for coming on the show.

Great to meet you, Adam.

Have a great weekend. We'll talk to you soon.

Sounds great.

Goodbye. That's our show, folks. We'll be back on Monday at 11:00 AM Pacific.

Leave us 5 A lot of excitement heading into the weekend.

Yeah, a lot of people.

I mean, you can just— it's— you can feel the energy in the UltraDome. Okay, uh, people are just getting here. We got 3 more sleeps till Monday. Yeah, people are counting it down, counting it down, counting.

Have people been counting down the days to Christmas? I feel like we missed the road to Christmas today. What is it? I think we're at 90 days, a nice round number. That's a huge deal.

We got 90 days until Yeah, and again, this is the weekend to get on it, get your tree set up, and get after it.

It's also only 455 days until the next Christmas, so you could start thinking about that as well. But yeah, you know, you want to focus on the next 90 days.

We'll see you Monday, folks.

See ya. Goodbye.