NVIDIA Earnings, Nano Banana 2, Block Cuts 20% of Workforce | Kenn Ricci, Howard Marks, Yash Patil, Scott Morton, Fan-Yun Sun, Adam Draper & Doug Bernauer, Sammy Azdoufal

You're watching DVPN.

Today is Thursday, February 26, 2026. We are live from the TVP and Ultradome. The temple of technology, the fortress of finance, the capital of capital. Let me tell you about ramp.com, time is money save both easies, corporate cards, bill pay, accounting and a whole lot more all in one place. We have quite the linear lineup for you today, folks. We got a scoop. We got an interview with the robot vacuum guy. Linear, of course is the system for modern software development. 70% of enterprise workspaces are on linear

are using Agent Sam, the guy who exposed that he could remotely access 7,000 DJI Robo vacuums.

Very interesting story.

Enabling live video, floor plans and joystick control. We're very excited to have him on in just over an hour. Then we have Ken Ritchie, chairman of flexjet. Very excited for that one. Quickly followed by Howard Marks Oak Tree Capital. Oak Tree talk about the market cycle

and then a lightning round to be remembered forever. Lots of lots of deals getting done, lots of gong worthy news.

And we have a surprise guest as well.

We do for the timeline. But first we will tell you about Nvidia earnings. They crushed, highly anticipated as always. Nvidia is permanently holding up the US economy. Continues to hold up the US economy, the global markets, but it's sold off even though they blew out earnings and it was a very good quarter. So earnings dropped yesterday after we got off the phone with Doug o' Laughlin from Semianalysis. Semiannalysis did have some good breakdowns here, but the top line revenue came in at $68.1 billion, up 73% year over year and 20% quarter on quarter. And this beat consensus estimates by nearly 3%. The stock price popped around 3% immediately, but then sold out, sold off 5% after at the market open this this morning, making it the stock's worst day since last April. They're now just a tiny was April Deepseek, $4.5 trillion company.

April was deep Seq.

Last April. Last April was Deep Seq.

Yeah.

Yeah, that makes sense. People were saying you won't need a whole bunch of Nvidia chips because you'll be able to inference cheap commodity hardware models.

Deep Seq was climbing the app store charts in America totally organically.

Yeah, there was a lot of weird stuff going on.

They weren't using a massive bot farm at all.

Yeah, I mean I guess I sort of. Steel man. The deep sea story. There was this interesting takeaway which was that even though the numbers were sort

of misreported we had it wrong. I think it was January.

Deep sea was January.

Liberation day.

Oh, Liberation day Tariffs. Yeah. Chip chip sanctions. But yeah, the deep seq lesson was like you can distill models, you can get a certain level of intelligence at a much lower cost and that is real. But the demand on the frontier is just incredible because people are like, oh, I have a strong opinion about 5.3 versus 4.6. People really care about being on that perfect leading edge for exactly what their use case is. You see this with people having favorite models, favorite flavors all over the place. And so the demand in the Jevons paradox really just powered right through. What do you have to say about that?

I was going to say, I think when the deep sea click moment happened,

the narrative was not that they were distilling though.

It was that they had built like they had trained a brand new model on western with very few resources. Yeah, there was no sense that they were just distilling.

There was rumors about it.

Pretty quickly there were rumors, but the main narrative was not that they're being distilled.

I think what I mean about distilling is that the truth that came out over the news cycle was that distillation does get you somewhere near a particular frontier capability and allows you to reduce cost. But there is still another order of magnitude of demand for the next generation and the next level of the frontier and we have yet to see a plateau emerge for demand for whatever the next frontier is.

And yeah, and overall the story, the story, the story of the last year is that there's there's very, very, very, very little demand for number three. Yeah, there's some demand for number two and there is an exceptional amount of demand for the most performant model for a specific, specific task.

Yeah. So going back to our newsletter today, our earnings recap, which you can subscribe to@tvpn.com after the earnings call. Semianalysis called the results staggering. I love it. Clean Bill Gross margin revenue and the guide is firmly above buy side bogeys. Incredibly clean, really. They also pointed out what Nvidia said about supply commitments. With $21.4 billion of inventory on hand, up from 19.8 billion, they have 95.2 billion of supply locked in with chip manufacturers. Also stating we have strategically secured inventory and capacity to meet demand beyond the next several quarters. In other wor, we're not running out of chips anytime soon. That's good news for Nvidia. The second order effect that I think was interesting that Doug o' Laughlin flagged for us was that even if TSMC does scale up magically or they figure out how to build another fab, Arizona increases capacity. Even if Nvidia has sharp elbows and is able to push out other demand and get all the line time they need to build all the accelerators that they need, soak up all the CPU demand, et cetera, et cetera, well, then you could still wind up in a weird scenario where Nvidia has the chips, they're ready to sell them, the hyperscalers want to buy them, the AI labs want to inference them, but there's just not enough energy. And so there's this question of, like, when does the shift, when does the chip bottlenecks shift to the energy bottleneck? That could be part of what is sort of worrying people, because you can see in the, in the timeline, we have this chart. Are you not entertained? From Nathan Benijk over at Air street, and he's taking a screenshot from the Financial Times. And it is just one of the most incredible graphs I've ever seen. And so as you see this chart, you have to wonder if there are any bottlenecks that they will run into TSMC capacity.

Yeah, and I asked Doug about this. I asked Doug about this. What does Nvidia do if their customers do have an energy bottleneck? I feel like they'll find a way to still ship the chips.

I agree. And truthfully, as big of a story data centers are, they're still consuming well under 1% of U.S. electricity. And so there are chips to be moved around, there are new power plants to be brought online. We're actually talking to Doug Bernhauer from Radiant about nuclear power today. And there are so many different ways to solve the energy bottleneck, but it is very real world, it is very slow, and if there's a timing mismatch, you could see a little bit of a flat line that I think people might be worried about. So Jensen himself strongly pushed back against the SaaS apocalypse narrative yesterday, which is interesting because he doesn't necessarily have to. He could be out there saying, like, yes, all those SaaS, CPU workloads, they're all going to be GPU workloads now. And yeah, you should actually sell all your SaaS, because the future is. Is inferenced. Like all. Every app that you use, it's not going to be code that's running on a cpu. It's going to be inferenced on GPU on demand. And so, you know, demand for Nvidia should be even higher. Like, he doesn't really have that many bags with the SaaS companies necessarily. Of course they'd be upset if he was talking trash, but he's not. He's defending them.

But he buys a lot of software.

Sure.

Yeah, of course, he said. I think the markets got it wrong. On the SaaS apocalypse. This is Jensen Huang talking to CNBC's Becky Quick, pushing back on the fears that AI agents will cannibalize the enterprise software industry. Instead, he expects a broad swath of software firms to use agentic AI to develop their software and boost efficiency. In what he described as counterintuitive, Huang said that AI agents won't replace these software tools, but will use them instead. Why even waste the tokens building a calculator when you can just download the calculator or whatever, I don't know, whatever SaaS you want to grab off. And this is certainly the true, certainly what we've seen with databases, like certain databases have just skyrocketed in demand because they're the ones that the agents are pulling off the shelf. And so yes, the agent could design a new database and install it, but that's a lot of work. It's a lot of wasted tokens when databases already exist. So let's tell you about Fin AI, the number one AI agent for customer service. If you want AI to handle your customer support, go to Fin AI. And let me also tell you about Vibe Co, where DC brands, startups and AI companies advertise on streaming tv, pick channels, target audiences and measure sales, just like on Metta. And we can move over to the timeline reaction. Nvidia is, is such a big deal. It's not just on the COVID of the business section, it's on the COVID of the Wall Street Journal. Didn't get the picture, though. Didn't get the picture. The picture is what is Iran in? Iran talking about? Nuclear enrichment is on hold, according to experts. But Nvidia results help soothe AI fears. The surge in profit comes as the AI industry's appetite for chips continues to grow. Data center hardware, the chips and networking equipment that Nvidia sells to AI and cloud computing companies, accounted for 91.4% of the quarter sales. Terrible news for gamers. Like, absolutely the worst numbers you could ever see. If you're gamers, maybe that's why they're, maybe they're launching a short attack shortage.

John would be punching a hole.

I would be. I would be. Maybe, maybe the gamers have risen up and they are shorting Nvidia to try and get them to go Back into gaming. Computing demand is growing exponentially. Chief Executive Jensen Huang said the agentic AI inflection point has arrived. With each passing quarter, the pressure grows on Nvidia, which, at a market value of nearly 5 trillion, is the world's largest publicly traded company. Is it not just the world's largest company? Is there a larger privately traded company? I don't think so. It must just be the world's largest company. I don't know. I guess you get into crazy definitions of like, does does Russia count as a company? Since it sort of is all one controlled economy, it is no longer enough for Nvidia to produce good quarterly results. They have to produce perfect quarterly results, said Daniel Newman.

How could they have done. How could they have done better?

Instead of net income of $43 billion in the quarter, they could have put up 50 billion or 60 billion or 100 billion or 10 trillion quadrillion. I want to see a quadrillion dollar quarter, please, Jensen. I'm not happy. I'm not satisfied. Keep growing. Keep selling these chips.

Gavin Baker gave some thoughts ahead of Nvidia.

What did he say?

Great. Shanu pulled some out. The same point hit on by Gavin and Citadel securities. It's a good one. Gavin said the world is fundamentally short both watts and wafers, and it may take years to resolve these shortages. The shortage of watts and wafers may prevent an overbuild. Hyperscalers would overbuild if they could, but they simply cannot. My best guess is that we would need roughly a thousand x more compute for the unlikely hypothetical scenario described by Citrini to be remotely possible. And the time it takes us to get there will give humans time to adjust and maximize the many potential benefits of AI. Citadel said displacing white collar work would require orders of magnitude more compute intensity than the current level utilization. If automation expands rapidly, demand for compute definitionally rises, pushing up its marginal cost. If the marginal cost of compute rises above the marginal cost of human labor for certain tasks, substitution will not occur, creating a natural economic boundary. This dynamic contrasts sharply with the narratives assuming frictionless replication of intelligence. Even if algorithms improve recursively, economic deployment remains bounded by physical capital, energy availability, regulatory approvals, and organizational change. Recursive capability does not imply recursive adoption. And do you know what the title of Citadel's response was?

No.

The 2026 global intelligence crisis. Calling everyone idiots.

We gotta write an intelligence crisis.

20.

This is actually the perfect. This is the perfect response. Like, very, very detailed. This is the same report that I pulled out that the current labor displacement narrative isn't holding up because job postings for software engineers are rising rapidly.

Yeah, it's pretty remarkable that Citadel was able to get that report out so fast. Like it dropped I think on Monday right after Citrini went viral on Sunday. Or maybe it dropped Tuesday or something like that, like really fast.

Yeah, it dropped Tuesday.

I mean that's like a quick turnaround for like a pretty detailed analysis. I thought that they were just.

Let's check the EM dashes.

I mean the interesting question about like you know, replacing white collar work versus like augmenting and new things is that there's the AI labs are obviously selling into enterprises that's growing a lot and the revenues are very real, tens of billions of dollars. But when you look to the other startups that have put up really big arrs, I'm trying to square like SUNO just announced that they're selling 300 million in ARR for AI music and that's a huge number. It's so big, but it doesn't feel like replacement work yet. It feels like it's just an additive new thing. Like there were just a lot of people that wanted to pay $10 a month for a cool app.

Same thing.

How many non technical people had an idea for an app? How many people that weren't musicians had an idea for a song? Yeah, and never could that. And then just unlocking all this late

and then you see it with like higsfield2 and all the viral loops and then what's that other one?

Not Webflow TVPN simulator.

No, no no.

Capybara simulator.

Lovable's one where huge growth and it feels like yes, level's going into enterprises and whatnot. But it feels like there's a lot of just like net new software developers, website builders that are joining that platform and building stuff and it feels very incremental. It does not feel very substitutional to me. And I'm wondering how much of the SUNO story translates to the AI lab coding model story as well. And that's certainly what Citadel is hinting at with.

Yes, Citadel is also showing that new business formation is rapidly expanding. This is from the U.S. census Bureau. You can see that year over year it is jumping, which we love to see.

Did Clouseau delete this post? The Great Recession of 9:52am to 11:16am February. Too bad. Let me tell you about Sentry. Sentry shows developers what's broken and helps them fix it fast. That's why 150,000 organizations use it to keep their apps working. And let me tell you about Shopify. 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. Nvidia's stock is falling because it needed to clear an options wall of $200 a share. So given a lot of folks were long calls into the print and it didn't clear 200 brokers are selling stock to reverse some sold that simple. This isn't fundamentals, it's market mechanics, says Gordon Johnson. Very interesting. I don't, I don't fully understand all the dynamics that can happen with some of the stuff Jane street does to move around the market. But you know, we'll see. I would say hold your, hold your judgment of Nvidia for the next quarter to see.

Samsung becomes the first Korean company to reach a $1 trillion market cap. Over on companies market cap.com you can see they are sitting at number 12.

That's your homepage, right?

Yeah, homepage, yeah, homepage for sure. Just under Berkshire Hathaway. Pretty good. And Broadcom is above that. But Samsung past Walmart and Eli Lilly,

I mean increasingly important in the AI era, right? That doesn't do a lot of fabrication and whatnot. I don't think it's all the TVs. There was some interesting stat about of the 1 trillion dollar companies isn't Samsung like the most highest margin or lowest margin or something? There was some post from yesterday we didn't get to but there was a lot more context on that.

Anyway, Ramp Capital is sharing says Patriot and somebody says I have OpenClaw sending lowball offers on Zillow all day just to make boomers start panicking. And he says can you give me an update on the Zillow campaign today? The Zillow update listings contacted today. 372 Average offer 70% below asking positive responses, 0 negative responses 271 Response was violent but I've reported it to the Tampa Bay police department. No response. 102 would you like me to keep

making offers 70% below asking. Just. You're selling a million dollar property and it's like, how's that? Can you do 300k?

Yeah, let's.

Let's bring in John Palmer.

Let's bring in John Palmer. Our surprise.

He's live in the TVP and Ultra Dome.

Coming in.

You, you know him. Nice hat. You probably know John Palmer as being

1 inch shorter than you.

Well, he's the face of leg lengthening surgery right now, right?

No, they were saying I'm the Goldilocks technology brother. Some people were saying you're a little too tall.

Oh, okay.

And Jordy, tall in his own right,

but maybe not quite sure at this table.

Yeah.

So they wanted someone right in between. So I was contacted by your team and happy.

I love it. I love it. Well, actually introduce yourself since for those who've been living under your data center.

Yeah, sure. My name is John Palmer. I was a co founder and CEO of a company called partydao. Recently announced that we're gonna be joining Stripe soon. So that was a awesome five year journey. And also work on a company called Area Technology which does logo and graphic design for brands like yourself. Shout out to the new logo and graphics package and yeah, that's.

We are honored. Well, we wanted to do. We wanted to hang and do some timeline with you as that sign of great respect in our culture.

Yeah. What you got, Jordy?

What do we have here?

How do you guys enable sudo for okay?

Baby Keem? Baby Keem, yes. Says yesterday hits the timeline. How do you fix open claw internal reasoning leaking? Peter, what does he actually mean?

Does anyone know what that means? So, I mean, we read your post about something small is happening. Hilarious.

Yep.

Did you. How much of that was real? Did you ever actually buy a Mac Mini?

So the entire post was tongue in cheek, but I did, unbeknownst to you, I did actually buy a Mac Mini that week. I did set it up. Setting it up was the best 48 hours of my life. Since then I've used it not very much. So it's. I don't know. I think the reason I bought it originally was basically to tinker and see how real is the hype. I set it up. I installed various plugins and skills. I made sure my setup was super optimized. And basically I realized the only thing I really needed it for was coding remotely in repos I'm working on or side projects when I'm out and about on my phone.

On your phone.

And you can already do that with a million other tools, actually. Even the Claude app, you can just connect it to whatever repo. You can literally just use Claude code in the app. And so there was really very little delta that it provided in terms of value beyond what I was already doing.

Yeah.

And I did try some other things. I've been trying to get it to like, you know, work on Google sheets and invite people. But the big problem is that the browser use specifically is pretty fragile and I Think it's going to get better. So I don't have buyer's remorse yet. I think that Mac Mini purchase will come in handy, but so far it's like, hey, I want to add someone to a Google Sheet. Can you invite them? And it's like it opens the modal to share. It types in their email and it's like, oh, I crashed. Like, it can't get past, like, the share modal of a Google Sheet. And I'm sure I could optimize my setup further, which I'm looking forward to doing.

Skill issue. Skill issue.

Skill MD issue.

It probably is a skill issue.

Yeah, you need to get the skill called skill issue and make sure that's running. I don't know if this is like a brilliant PR move from Baby Keem's team or if this is real, but either way it's really, really funny. But, yeah, how mainstream this went. Our lawyer got a Mac Mini. He's obviously pretty tapped in, but it's cool to see.

Well, I was speaking with you at the gym this morning that we should. I'm excited about a market for bootleg skills. So you don't. You don't just, like, you don't want to be the guy just listening to hits on the radio, right? You don't want to just be going on skills sh. And getting the what would Elon do Skill. That's. That's pretty. That's how I earned it. And it turns out that one was like, actually, it was actually malware. So what you want is within your trusted social circle, people that you spend time with, IRL USBs full of markdown files with bootleg skills that you kind of keep internal. And there may even be a market for that, you know, on Craigslist, whatever.

And you can say if somebody says, hey, can I try the usb? Can I borrow it? You say, well, this skill is kind of like a personal thing. It's not really.

You need to be gatekeeping your skills. That's like probably the last moat in tech is gate kept skills just proprietary markdown files powering the entire globe.

I do honestly wonder what the conversion rate from, you know, buying the Mac Mini to, like, actually using Open Call regularly was because I think that your post was funny because a lot of people definitely went through that. Like, I went and bought a Mac Mini and it did take me like a week even to unbox it because I just had like a busy life and I didn't have time. And then when I did, it was like, okay, well, I didn't have a monitor at home, so I had to plug it in my tv. And then, like, the actual setup does take time. Like, 48 hours is not that much of a joke.

And I think on this point, like, I do think you need, like, a solid month at home fixing it. So I've been out of town for a week. I'm here in la. I live in Brooklyn. And a lot of the time when you hit a wall, it's like, okay, now you probably need to get on the Mac Mini yourself and, like, add whatever environment variables or keys that you need, set up a new account for it.

Yeah.

And since I haven't been home, like,

I can't really, like, you know, on a more serious note, amazing, like, isn't, like, five years ago, everyone was posting, like, I missed, like, hacker culture.

Totally. Yeah.

I miss the old Silicon Valley. And it feels like this. This feels like. It feels like we're back.

Yeah.

Yeah.

A lot of people definitely, like, opened up the terminal for the first time in years. I mean, this is true for vibe code generally. Like, a lot of people would step back from just writing code or committing code to any GitHub repo. Yeah, the Vibe coding. Boom.

It's funny because I do really prefer the quad code, like, user experience to any of the, like, slicker desktop tools.

Because.

Because of a lot of reasons. But it is funny. You're, like, in the terminal, I think everyone. There's definitely a major factor where, like, you feel really cool using the terminal if you're not an engineer. But you're just. It's all plain text. Like, you're speaking in plain English, and, like, it's speaking to you in plain English. But, you know, if you're at the airport or in a co working space, you're like, I feel I look pretty

performative. Shrey says Baby Keem just humbled a model. That is actually an exact lyric from Baby Keem's song Stat. That's a good reference. Sorry, Deep cut. Deep cut. And yeah, he. Someone else said two Phone Baby Keem. That's another lyric. But two Mac Mini Baby Keem.

And Baby Keem liked it.

He's very online. We're working on getting Baby Keem on the show to break down his full setup.

Yeah. I mean, one of the biggest questions that I feel like is still sort of unanswered is like. Is like, what are people doing with it? You know, obviously, if you're working on a startup and you're building a piece of big software and you're committing to it and you're firing off just managing agents remotely. That feels relevant. Although yes, you mentioned you can just do that already.

Yeah, I mean I actually, whatever, like, I'm definitely not the top person who's like dive into this. Right. But I feel a lot more confident even, even for coding. I actually don't love doing it with Open Claw because I've, I've got my reps in with cloud code. I trust the harness, I trust its ability to like utilize the model in the best way. So even for coding, I'm just using cloud code remotely anyway. So I'm still waiting to find the use case. I think one interesting thing though is just that I think prior to this whole Open Claw wave, the whole idea was all the AI labs had their computer use demo and it was running in the totally anonymous VM somewhere. And I do think these things are way less useful without your personal data and your personal contacts and the stuff on your machine. No one solved like trusting it enough security wise. Like, you know my Mac Mini back home, it has its own Gmail, icloud, whatever because I'm not going to let it touch my data. But if we do get to the point where you can be confident that it has all the files on your file system and all your accounts and it's not going to leak anything or like make a dumb mistake sending it to someone, I do think the ability to like do all of your knowledge work on the go, even if it's just like I do the same stuff I do today with AI, I can just do it on the go would

be a major moment idea guys have been waiting for.

Yes, yes. It's the moment idea guys have been waiting for though. I think, yeah, we can riff on that if you want.

And just more based work, more like, okay, this is something I need to do regularly. Turn it into a cron job. Like the agent should be able to do that.

Yeah, yeah, I totally think so.

Also, I was just really excited about this like Napster sort of moment. Like every idea I was joking with Jordi. Every idea that I have for like something to vibe code is like something that doesn't exist. Not because of code. Like if you want a fitness tracker, yes, you can vibe code one. But you can also just go to the app and get a free app store and get a free one. Or you can pay. There's plenty of products in every category. The things that you want but can't get are like something that jumps every paywall, something that gives you videos for free that are behind walls and stuff. And these walls exist for a reason. And maybe those come down in a world where everyone's like, working on open source agents and just sort of showing up as themselves and synthesizing everything. But it's. It feels like it's much harder to actually productize that fully because so many of the things that people want are like, yes, they're possible in open source in the sense that like, file sharing was popular.

Totally. Yeah. And even though I will say, like, even though I guess so far, I don't want to say I'm disappointed because it was really a learning thing, but I still, it's never a good bet to be like really bearish on new form factor with like novel tech. So I'm still pretty bullish and excited about like three, six months from now what that same Mac Mini could be doing. And it doesn't need the Mac Mini. It could be. It could be on whatever service. But I think it's going to get pretty exciting pretty quick.

Yeah, yeah, yeah.

What did, what did Playboi Carti mean by this? Did he. Was he responding to Baby Keem? Is that what he's supposed to read into this? Is he also getting a Mac? Is the Jeremy Irons in Margin Call? Is that the meme?

This is when he says that the

music is going to stop.

Oh, okay. So Playboy 30 is calling the top.

I have.

You have?

Yeah.

Whoa. Are you a film guy?

I'm not.

Have you seen Borat?

I would never call myself a film guy, but I've seen movies.

Have you seen Borat? Okay, okay.

The whole bit is that Jordi hasn't seen enough movies.

Yeah, he hasn't seen any movies. In fact.

Wow. Interesting.

RIP the market. We had a good run. The three kinds of stock market day so far. Donald Trump spins the wheel of tariffs and replaces a 5% tariff on Belarusian wheat with a 22.4% tariff on Pakistani jet engines. All stocks lose 2d $6. Substack newsletter publishes a story about DoorDash going badly. DoorDash down 862%. Global Financial System teeters on the brink. And 3 Nvidia announces earnings of $100 trillion, beating expectations by a thousand X. Jensen Huang named new king of Earth while his stockholders to form the new permanent ruling class. Nvidia down 3% on the news. This is the press of the market. Are you. Are you a day trader guy?

No. No.

Are you an investor?

I am an investor investor.

How do you think about the market? What's your strategy?

I'm an investor. I'm a businessman. I don't know. I think my approach is like pretty smooth brain value investing, like good fundamentals. I'll buy the stock and hold it for a long time. Definitely a lot of even, even intra month trade. It's really like long.

You're not like a one man Jane Street.

I'm not not a one man Jane Street.

What about, what about crypto? I feel like if you're working in crypto, even adjacent to crypto, there's just so much alpha from seeing essentially like angel investment style opportunities from someone who's building something that you just know it's going to be hyped and the token's gonna moon.

The issue is that it's like wildly distracting because as soon as you have enough money is that. I think, I think I, I don't like doing a lot of like individual stock trading because as soon then, then my attention is gravitating towards this thing that's not actually productive.

Yeah.

And so what do you think?

I think that's true. And I think, I mean obviously if you're trading like perps in crypto, like you better stay on top of it because like there's a moment to exit or if you just hold it long term you'll probably get liquidated at some point. So I just don't touch any of that. You're right though that there are whatever. There are like weird things in crypto where like someone you know or someone's doing a project and they airdrop a token or something like this and you wake up a week later and it's worth like whatever a month's worth your salary. So I think that's.

Let's talk about Jane street because I've seen maybe 10,000 posts kind of blaming them entirely for the just crypto price action over the last couple months. You haven't been able to, have you been able to find anything that is like definitive? Obviously the lawsuit came out and they were like we were going back and forth on it because it sound they pulled, they pulled funds out of this liquidity pool right after.

Actually I don't even think they pull, I don't think they pulled funds. I actually think they sold into the pool, which is I guess the same thing.

But they did it five minutes after we were saying maybe yesterday or the day before that. Okay. It's a public blockchain in theory. It's not actually a smoking gun necessarily.

But either way I'm only as informed as you are in terms of the tweets I haven't verified anything, but, yeah, my understanding is that I guess Terra Luna had a lot of money in this pool and they had planned to pull whatever 180/something million out of it.

Yeah.

And that I guess the allegation is that Jane street heard that this was coming before everyone else. That's like the material, nonpublic information. And so then also pulled their money or sold it. Yeah. Like just minutes after Tara did it. And you're right that it's public. So I guess, like, again, not a source of truth here. I'm like, relaying secondhand with confidence information that I read on the timeline built for podcast. Yeah. I'm relaying this to you secondhand without confidence. But, yeah, I guess you could say, hey, it was. It was a. Anyone could see that and they just reacted quickest. But I think, I guess what that would come down to is what are the internal ops on moving those funds like that? I would imagine that a fund like Jane street with $85 million in this pool isn't just sitting in someone's like, hot wallet Metamask. Right.

Yeah.

So they probably would have.

They could have like some type of protocol or actually software, like, as a protective mechanism, given that they clearly are.

I don't know. A lot of stuff feels like conspiracy theories that. It feels very similar to, like, the Robin Hood Gamestop fiasco where there was like, a lot of conspiracy theories about, like, Citadel purposely tanking markets and the bailing out hedge funds at various points. And I always thought that was just like, people looking for scapegoats and a whole bunch of market actors just playing hardball. Because that's the. That's the default structure. I don't know.

Yeah. I mean, I'm not that comfortable, like. Yeah. Speculating on this either, because I'm just like. All right, I saw, you know, post in the timeline. They are interesting, though.

Djen. CPA says Jane street got my dad drunk and made him cheat on my mom when I was three.

See, a lot of people are hunting for.

For, you know, deep dish from Joyer says Jane street killed my dog.

That would not be good.

You're just getting a lot of the cathartic, like, crypto people who've just been wrecked by these fiascos, like, now have a viable scapegoat.

He's not a bad trader. The market was manipulated by Jane Street. Yeah.

Tampa International Airport has come out and said, we've seen enough. We've had enough. It's time to ban pajamas at Tampa International Airport.

Wait, they actually. This is great. Read the next line.

I mean, it's.

After successfully banning Crocs and giving everyone the amazing opportunity to experience the world's first Crocs free airport, it's time to take on even an even larger crisis. Pajamas at the airport. Is this real? Is this.

Actually, it's hard to say.

What's your.

What's your airport travel?

I was gonna ask you guys because you're traveling together a lot. I don't know. It's always a dilemma, right? Do you wear, like. I get the pajamas thing. Because if you wear, you know, your nicest clothes, then to me, when I get to my destination, I'm like, those were on the plane.

Yeah, yeah.

It's like, I'm not unless I have laundry, you know? So for me, I'm rocking, like, one of my. More like mid fifth. Like, I'm something. I don't care about something. Something lightweight that packs light as well. Because after that travel, like, that's not coming out of the bag for. For the duration.

Well, there's. There's like a contagion theory here, because if I'm in a suit, but I'm sitting down in a seat that someone in pajamas just sat down in, then I'm getting their pajamas all over my suit.

Yeah.

And so it's. It's just a vicious cycle.

Are you guys business casual on the plan?

I. I've actually traveled in suits many times, and it's underrated.

Interesting.

I. I think. I mean, besides the fact that I've heard that you can get upgraded to first class more. More easily if you're dressed up, I don't know how apocryphal that is. It happened.

Yeah.

I feel like that's just people trying to justify.

I think people just. They see the suit and they're like, respect. Here, please take my seat. I'm in pajamas. I have the first class seat. But you deserve this.

I think I'd feel more comfortable with that if I weren't as tall as I was. And you are. Because the suit, you know, more powerful aura. Definitely less comfortable. Especially if you're in a seat where you're kind of constrained. It's kind of bunching up. It's really not.

Yeah.

Feeling your image.

What about the TSA ban. Have you. Have you seen this? TSA shutting down. TSA precheck.

Pre check.

This is just like a ploy to. To increase private jet ownership, right?

Yeah, clearly.

The.

The. We're. We can ask Ken Ritchie about.

Yes.

The. The lobby is trying to Push people into.

It's kicking the commercial aviation enjoyers while they're down. It's really brutal. What does Joe Eisenthal say about this? He says I'm very pro conformity, very anti. Antisocial behavior in public. And yet I can't understand why people care so much about other people's airport.

Just say that you wear pajamas to the airport. Let's say that you wear pajamas to the airport.

Flying is unpleasant enough.

Burger King is launching an AI chatbot that will assess workers friendliness and will be trained to recognize certain words and phrases like welcome to Burger King, please and thank you.

Huge.

The AI will be programmed into workers headsets, according to the Verge.

I think Chick Fil a got access to that technology like a decade early.

Yeah, yeah.

Because there's a my pleasure. There's a thank you. There's a my pleasure every time.

Yeah. Apparently when McDonald's opened in Russia, there's a quote. After several days of training about customer service at McDonald's, a young Soviet teenager asked the McDonald's trainer a very serious question. Why do we have to be so nice to the customers? After all, we have the hamburgers and they don't.

They need to be nice to us if they want these hamburgers.

What was it? Did McDonald's launch something as well?

I thought they launched an AI thing too, wasn't it?

AI launched the big arch.

The Big arch is the big burger. But then I thought there was. I mean, at this point, every single company has an AI strategy of some kind. And so everyone's putting various stuff. Are you an F1 guy,

tennis guy?

I'm a tennis guy, but I watch like the F1, you know, highlights and recaps. I'm rarely like getting up to watch the race live, but play tennis or

watch tennis or both.

Play and watch both a lot. Yeah.

What does it say? Best friends should be able to apply to jobs together and get hired as a set. New age hired hiring looks like four to six talented people clustering together and building a feature and getting acquired. That's. Yeah, this does exist.

I mean, a horse did post. Horse is kind of going off.

Horse has been on a generational run. Literally a horse. Yeah. Underrated.

To just Riley Walls. We talked about this yesterday. Joined the Labs team at OpenAI.

Very exciting.

We were talking about this kind of offline before the show that it seems like there's a lot of opportunity in like hiring for. For not just ability, which Riley Walls clearly is incredibly talented and has an insane. Probably the most elite idea guy. Actually on the pitch playing right now, and that's hard to say as idea guy, as idea guys ourselves, but I think hiring for. You called it hiring for Mindshare.

Yeah, I was pointing out, obviously, both Riley and Pete Steinberger, both incredible devs, so not trying to take away from that, but it's interest that like, both of those either like hires or acquisitions also come with like a lot of developer Mindshare. And I think in a world where, like, if you just run out to the extreme, like software production costs go to zero. So being an engineer, being able to build stuff, that skill becomes more of a commodity than what are you trying to acquire or hire for. Maybe in this interim period, it's like a blend. Okay. Like, I still want you to be a good engineer, but I also want you to bring me developer Mindshare. And I think to that point it's like, you know, with OpenClaw, for example, it was definitely the best product in the market, but it was open source, so anyone could have, you know, forked it, copied it, whatever. But I think it really would have been hard for someone to catch up given how much Mindshare had. It was like the fastest growing open source repo of all time. So what you're buying, it's like even though it remains open source and forkable and in like a foundation and other people could build it. The reason you buy Pete, in addition to him being a great developer and idea guy, whatever, is just that he's bringing the mindshare of all the people who definitely follow him, have notifications on for all his posts. They're following all the updates because he's

bringing all the stuff and they're investing. They probably spent 600 bucks or something

like that on a Mac Mini, some sort of hardware device to run the model.

Noah Smith is sharing the job. Postings for software engineers have picked up since Vive coding became a thing. They had been declining until the early part of last year, actually. And again, this is why, I mean, we've. We've talked about this over and over and over. I feel like there's this like two different narratives being spread on Instagram. It's just like this like labor displacement narrative going super, super, super hard. And then over here, like, it's just an entirely different world.

Yeah, I mean, Tyler, how do you interpret the software engineering number? Have we talked about this yet?

Yeah, I mean, so I think I broadly agree, like, so when Dario went on doorkesh, he had this take where it's like you go from AI doing 90% of coding to 100% and then it goes from 90% of all suite tasks to 100%. And so when you're kind of in the interim period, engineers get super, super productive. But then there's a point where it's like actually like, it's like instead of being super high leverage because they're, the AI is doing 99% now, the AI is doing 100% and it's actually like,

it just kind of instantly goes to zero.

So like, I think you should imagine that like, yeah, you actually see, you

should see a lot of like hiring

because everyone wants like a vibe coder at the company. But then at some point vibe coding

comes so easy that like, you don't

need any, any like special talent to be able to do it and you just have the random person.

I, I mean, yeah.

So basically if, if, if you, if one day you're not sitting at your desk, you're like, you're kind of the vibe coder in the coal mine.

Canary. Okay. Okay. Yeah, yeah, sure. I do wonder about like, because there is, there is a world where software engineering postings and jobs are substituted for other roles. Like you can have someone whose job is like a business analyst and every day they open Excel and crunch some numbers and then you could hire a software engineer to automate that task 10 years ago, 20 years ago, and a lot of companies did. And if vibe coding sort of eats into other things, you could see the rest of white collar. It's like the white collar work could go away because the software engineers are doing it. And then the bigger question is like, is like you're sort of jumping ahead to like a full unemployment scenario where no one has a job, but in the world where there are like a few jobs. I keep thinking about this thing that Pavel Asperuhov said where he was like, yeah, if the software engineering jobs go away, get ready to compete with software engineers in every category because they're going to learn financial analysis, they're going to learn to trade, they're going to learn to sell, they're going to learn all these things. And if they're just like smart and hard working people and they have AI tools both helping them upskill, reskill and shift and do whatever they need to in the new role. Like, you'll just, those people will still be employed. They'll just be in a different part of the economy.

So Rachel Carton, who's coming on the show soon, says by more than a 3 to 1 margin, young Americans believe I will take away Opportunities. There's a lot of uncertainty surrounding AI especially among the youth. Using it as a joke like liquid death Gucci or Equinox have tried is not a wise marketing strategy. Yeah, the blowback around Gucci was predictable but it was also strange that they chose to use an image that looked like it was at a GTA too. It was just like not even trying to look like one of the images that we looked at looked like.

I thought that one looked great normal

photograph and looked cool and the other one was like a GTA style characters which really didn't hit interesting. JIRA ticket says but sir, that's rage bait. If you do this, you'll slopify the timeline and you're posting something that you know has no value. We are, we are selling engagement to willing scrollers at the current fair algorithm parameters Avi Schiffman had was, was certainly.

Have you ever talked to an AI just like as a, as like a being or like a friend or like just as like any, any like non productivity like earnestly. Earnestly. Yeah, just like chatting back and forth forth with it. Not for any specific business purpose.

Have you ever fallen in love?

No, no, I really haven't. Earnestly. I haven't. I haven't really given that. I have like to test it out but I wasn't like wholeheartedly engaging.

Yeah, it's weird. I don't know if it's like a personality thing or maybe I'm just too busy or something but I've never been like I will, I will. If I'm in a video game, I'll go talk to NPC on a sadness for hours. I'm down like I'm down to explore the story that's there but for some reason it just hasn't clicked with me.

I don't know, I haven't even done that. Like even when I was playing like, like MMOs back in the day, like I feel like I would always just be like skipping dialogue.

Yeah, I actually do skip a little bit.

I don't know. It's hard to really buy into it. I think maybe a good the synthesis of that would be. I would be down to play like a new Elder Scrolls game where there are like really good AI models behind the characters.

Yeah.

Then I'd be exploring it because like I've decided to suspend disbelief to play this game. So while I'm in this game, let me do it. But in my real life I'm not like having a hard time next time

you call me and I don't pick up, go into chat CBT and say, pretend you're shot. Pretend you're jordy. Let's talk about work.

I do think that's like more interesting. Like, like again, not like, maybe not as a heart to heart, but like,

I mean, there's companies that do that. Like a AI version of an influencer and whatnot. Like that. Yeah, those things.

I mean, you guys have like hundreds, if not thousands of hours of live show. Now you could probably get a pretty good Jordy model. And if you were like trying to riff on show planning, like, it would actually functionally be pretty good.

But I just have no desire for it. I don't know.

Give it a shot, John.

Maybe, maybe.

I do think this is an interesting take on the Open Claw style stuff where like, I think there's a world where again, security is like a big asterisk here. But like, if I had Open Claw. Okay, so I've been riffing on this idea for a while that I'm just a JPT5 client. So like ChatGPT has GPT5 and I'm John Palmer. So I'm John Palmer Transformer. I'm JPT5 and I'm just a client for GPT5, which is my brain.

Okay.

And so like you guys kind of like a harness. Yeah, well, this is the clothes of the harness and the brain is the model and my mouth is the client. So in that anyway, I won't go into the whole riff, but I guess like, so my Mac mini I named JPT5 and hypothetically, if I could train it on like all the writing and speaking I've ever done, plus my whole file system and it knew like what I do on my computer. One version of that is it's single player. It's my Open Claw. I use it to do my tasks. But I'm really interested in like the multiplayer. Again, this is the security. Security's not ready for this yet. But if I had my Mac Mini and either humans like yourselves could really go like, hey, I need John for something. I know he did that one project and all the files are probably on his computer. Could I go micro pay his agent and be like, yo, I need help researching this crypto thing that you did. Whatever. And if my agent could actually give you. It's like you're not accessing the same one size fits all model that everyone has. You're accessing, you know, that model with some custom, you know, contacts and prompting or whatever. And that would be cool. So like, you know, you have access to Jordi Anyway, but for someone who doesn't, you know, if that random third party out there wanted to riff with Jordi on a marketing idea, it'd be cool if they actually could. And you had set up that model. It wasn't, you know, Opus 4.6 trying to replicate you, but it was like, all your insider info. It's kind of cool. And then, you know, the agents could do it to each other now. JPT5 could be talking to JHT5 and they could be like, we could be hanging IRL, and our models could be hanging back at home, like, kind of like a kid's playdate. Like, my models, like, hanging out with your model. They like, hey, let's get John involved. They go get John Coogan's. So it's kind of a fun.

I mean, there's a take on that where it's like, that's dystopian or something, but that's not, that's not how I feel. I just feel like I'm just bored.

We go, yeah, we go brainstorm for an hour on some concept, and then we come home and we find out our models cooked up something way better.

Yeah, I don't know. I was, I was looking at the.

Claude, I think, I think it's, I think it's actually, I think it's actually possible. Especially, Especially if you were like, you bring the original idea and then you say, like, I want, I want John Palmer GPT to talk with John Coogan GPT about this idea and, like, come back with a bunch of ideas. It's a lot, it's. It will work a lot better for someone like John, who has spoken on the Internet for hundreds of hours so

much that I don't say that's why you actually want to talk to it.

Would it be cool if the model could, like, decide to price its own services and negotiate with whoever, whatever model is hiring it? I mean, whatever.

It's.

Avi Schiffman dropped user interview number three. We're not going to play the video because it's kind of sad, but Gary Tan said this is not the happy demo path. Apple or Google will never make this one of their launch videos. It's not what you will hear about in a TED Talk, but it's real. AI doesn't get tired, doesn't ghost. A lot to think about with this one. Yeah, I, I, you know, AVI continues to find new ways to break through the noise, but in some ways, like, even though this video has been. It feels somewhat dystopian, you know, this is Someone who's struggling with mental health, like actually gets in an accident in the video. It does feel like the most kind of real raw experience of AI companionship in this form factor.

Yeah. Reflection on like do you remember his interaction with Paul Graham very early on he was like I'm building a hardware device for AI. He put out some very vague post about his plans and everyone was like this is impossible. Apple will crush you. They have the supply chain, they have the distribution, they have brand. You will never win in this category. And Paul Graham told him in order to win you have to do things that Apple would never do and be counter positioned. And so at one point he was thinking about doing something that looked like very organic. No smooth lines, no rounded edges, crazy colors that Apple wouldn't pick. Just really moving outside of their design language so you wouldn't feel like that. And then yeah, this type of messaging is certainly in line with the anti Apple and I don't know, it'll be interesting to see where the company goes. People were so bearish on that ad campaign. We were very surprised. Surprised by how broad the billboard campaign was.

They got the Heineken response.

Oh, did Heineken like young come?

So Heineken also just came out with a new campaign that I want your guys reaction to. I don't know where exactly it is but seems to be global and it's just like a graphic of a voice note and it says the campaign is don't send a WhatsApp note. It could have been a beer in person. So they're like taking like the total like real world approach which, which I

think is, I think it's good. You know I, I think you, you had Matt Zeitlin on the show recently and he and I play in a pretty regular like friends poker game. Not big money. It's really like a social hang. But it's funny because like everyone's like multiple people have wives who are like why are you going to play poker? That's degenerate. Are you losing money? Whatever. But I mean I, I had a post a while back, it's like in the current like casino economy between like prediction markets, meme coins, like crypto trading, whatever. Like an in person poker game with your friends is like the most wholesome. Like if you're going to gamble and lose like 100 bucks, go do it with your friends. Over four hours like having a drink and hanging out and, and you're solving the male loneliness crisis.

I've heard the same argument made for cigar nights. Yeah, cigars Obviously bad for you, but getting together with friends and actually maintaining social ties is probably really good for long term health.

It's pretty great, honestly. And if you play, if you play a regular poker game, even with strangers that you only knew through this game originally, like six months out, you've logged like so many hours just like hanging out that it's like a pretty great, like wholesome, wholesome vibe, you know?

Should get into it. I'm not big.

If you're ever in New York, we can have you guys.

Okay, let me do some ads. Let me tell you about Gemini 3.1 Pro. Gemini 3.1 Pro is here with a more capable baseline. It's great for super complex tasks like visualizing difficult concepts, synthesizing data into a single view, or bringing creative projects to life. And let me also tell you about public.com no plaid plaid powers the apps you use dispensate, borrow and invest securely. Connecting bank accounts to move money, fight fraud and improve lending.

Now with AI Dan Romero says, I spent five years cloning an existing app with a network effect. Building the software is not the hard part. And he was building this app for a very specific sub community and was able to get a massive amount of attention. So I think this is like a good, clear rebuttal of at least that one part of the Citrini piece.

Yeah, I think I had posted a while back about how again, same thing, software costs go to zero. Everyone's like, oh yes, I'm an idea guy. I'm going to be crushing it now because now I can just use cloud code and build whatever I want. But it's like if you just make one more logical leap, it's like, hey, everyone has that now. So if you're just like building your ideas, like if you're building vibe coded apps for consumers, you're actually in a way more competitive space than it's ever been. So I actually think it's almost more like not to riff on the idea guy thing too long, but like actually being paid like a salary to be an idea guy or like making good income to be an idea guy will actually be like a very privileged position where only companies that have like real moats and monopolies on distribution will actually be able to hire like people whose job it is. Just think on the idea. Because if you have a good idea and you ship it, you know people will use it because you've already got the distribution versus if you're even if you've got a great idea and you can vibe code it up if you're in this like sea of vibe coders like producing way more apps.

Well, the other thing is there are certain idea guys throughout history that thrived because they had a unique insight. They brought it to market, they executed pretty well. But it just was a great idea at the right time. And so it got traction, attracted great people and things like that. And in a world where like an idea can be made and we've seen this with I think a bunch of marketing on X over the last year, where if somebody has like a good style of launch video, it will be immediately replicated within a month.

Yeah.

And then it will become just kind of like flooded and the meta will kind of die quickly. And so in a world where ideas are cheaper than ever, is it really just about idea guys getting lapped by people that are great at executing? And it's like, yeah, you had this great idea, you shipped it. Someone else saw that and was like, that's a great idea. I'm going to ship it today and then I'm just going to out execute you.

Someone should launch their startup with a congressional testimony. That would be a great aura farm. If you just. The first time you're ever seeing about this company, they're just answering questions and getting grilled by Congress people about what their plans are. It's always like typically the top of the mountain. Like Mark Zuckerberg built Facebook for decade, then he starts getting called to testify, then he starts getting sued. But if you just come out the gate, day one, Augustus was kind of, Augustus was kind of doing it.

He was like half the time in El Segundo, half the time on the

road, he wasn't being called to testify. He was like, you know, standing up.

I know, but he's kind of doing a bit.

Yeah, it kind of did work like that. It's very funny. Anyway, let me tell you about Figma Ship the best version, not the first one with Figma. Introducing Quadco to Figma. Explore more options, push ideas further.

We got new Apple products. What's coming out this coming week? Oh yeah, we don't know exactly. It's.

It has the Apple logo on it though.

We know that, we do know that.

The rumors.

Touch screen map.

Oh, that's why, that's why they're touching it.

How are you guys thinking about the touchscreen Mac? Because one of my least favorite things in the workplace is when people put their fingers on my screen, it makes my blood boil. And so maybe they have some new. Like, are they. I could see Apple at this point. They've got the iPhone, Sock. They could have like a little attachment that has like a clean. Like a little cleaner. A cleaner, you know, a little spray bottle. An Apple spray bottle.

They could have a magnetized Roomba. And you put it on, you stick it to your screen while you. And it drives around, identifies the.

Yeah, like a Zamboni. Like a Zamboni.

Yeah, like a Zamboni.

So messy.

The Apple Zamboni.

I don't know.

You want to create the problem and then have the product that solves the problem in the same launch. So touchscreen Mac with the.

Yeah, I. That's such an odd. It's such an odd choice because I feel like the whole. I feel like the whole we've. Humanity has gone down the path of like the touchscreen laptop.

Yeah.

And it's been available for years, maybe over a decade. And I just feel like it's never really broken through. And if it was, if we were at least seeing like. Oh, yeah, well, like, obviously, like 80% of PC users have a touchscreen. And it's clearly a dominant form factor.

Yeah. It's like the reason for the touchscreen is like, I'm away from my computer. I can't carry around a dedicated keyboard and mouse. So I guess I just have to figure out a way to use the screen for everybody, everything. But it's not superior. If I, If I have room on my desk for a keyboard, I'm always gonna be faster doing the.

Like, it's gonna, you're gonna. It's gonna look very. Maybe, maybe it's better for people that aren't as tech native. Like, I can imagine like, like, like an, maybe older family member, like, you know, wanting to zoom in on a photo and maybe it's more native to them to. To kind of like use the kind of like pinch functionality.

Yeah. It feels hard though. Like, if it's in a. Again, I don't even know if this is what they're releasing, but if it's on a laptop, like, I actually kind of have to like, reach past the keyboard. It seems like a far reach to like. Like, if you're like even keeping your arm held like raised like this for more than like 30 seconds would get, I think, tiring for most people anyway.

I don't know. The first time I saw it, I just thought Mac Mini, honestly, because it

just looks like, well, they're making the Mac Mini in the usa. They came out with that.

And I will be rid of my current Mac Mini and buy the USA

made in the USA.

Of course. Running Kimi 2.5 on my maid in the USA Mac mini.

Well, we have some videos to watch, but we don't have IEMs for you, so we should probably let you go get back to the rest of your day. Thank you so much for coming and hanging out.

Yeah, absolutely.

And we will see you soon. Let me tell you about Okta. Okta helps you assign every AI agent a trusted identity. So you get the power of AI without the risk. Secure any agent. Secure every agent with Okta. And let me also tell you about console. Console builds AI agents that automate 70% of it. HR and finance support, giving employees instant resolution for access requests and password resets. Let's flip through some of the videos that we missed because there are some

good ones in TSMC's plant in Arizona, which is spread over 1,000 acres. Is expected to cost 165 billion.

I thought they already had a TSMC Arizona up and running. I guess they're expanding it to the

tune of 165 billion still coming online.

I wonder when. What day is this actually going to go live?

Got to send Tyler over there.

Oh yeah. Won't you be in the area soon? Don't want to dox your plans later today. Later today?

Maybe I should check it out.

Go check it out. Get arrested for trespassing. We forgot to play this AI video of the AI leaders Elon Musk and Sam Altman as old men. Did you want to watch this?

Pull it up.

This was a very, very bizarre video.

Extremely viral. Over on Instagram.

Yes. Oh yeah, it's. It's in section E here. Let me tell you about Cisco. Critical infrastructure for the AI era unlocks seamless real time experiences and new value. With Cisco, it features Energem move the world.

No purpose. But they had a lot of time on their hands.

The less people the. You can't restart videos on Instagram.

Had lost their jobs.

They had no money, no purpose.

The visual fidelity is just another level up. I don't know if this is the latest Chinese model, but it's really good. Use the energy of humans to power the machines that took away their jobs.

The voices. The voices.

The voices are a little odd. Yeah, they are still in the uncanny valley. But the actual video is really strong.

Deeply wrong to put our technology leaders like this. Should have made them look like bodybuilders if you're gonna do this. Yes, but it is a funny. It is a funny bit disturbing.

Anthropic's going into consumer now. Is that right? Adam Feldman joined as I recently anthropic to lead consumer product and we're growing. Our goal is to build AI that millions of people will use every day to think better, create more and accomplish what matters to them. A bit biased here, but this is the most interesting.

Yeah, so back in Q4, I was expecting that they would care a lot more about that they cared more about consumer that they were letting on. I think like when, if you look at their actions this week, you see that when they were lagging, they were really leading with safety. And now that coding is a competitive market, they're obviously dialing some of that back. And when they had actually zero consumer adoption or functionally zero consumer adoption, they didn't care about consumer. But if you look at keep thinking campaign, you look at the super bowl campaign, you don't do these things unless you care somewhat about consumer.

It all bleeds together anyway. And so it seems like it's the end of the road for all the AI companies that they have some sort of consumer footprint. It's hard to be that deep. And when you think of the. The pure plays, I mean even Amazon, I mean they sell a ton of tokens through aws, but then they also have Rufus because they need to vend the AI into their distribution, into their platforms. And we've seen Llama be vended into Instagram and WhatsApp and all the different consumer interfaces that people use every day. And people will use those for business contexts in consumer apps. Gmail became dominant in consumer and businesses all over the world because there's a huge blend that happens as people work both professionally in their personal lives and vice versa. And they do work when they're at home and all sorts of different stuff.

Raghav has some breaking news. Apple has acquired the rights from Netflix to Air Drive to survive Season 8 on Apple TV. Plus, you were expecting something like this over a year ago at this point that if you're going to be spending all this money in the F1 movie and the F1 streaming rights, you want to be actually be the home of Formula one.

Yeah, it's sort of like content vertical integration. Like you want to be able to take the viewer on this journey from novice. So you watch the F1 movie which has Brad Pitt in it, and it's just super accessible and anyone can throw it on and have a good time. Even if you're not an F1 fan, you can watch the F1 movie and know. And like in the movie there's a voiceover that explains all the mechanics of how the Sport works. And so you learn and you come out of that saying, okay, I'm kind of interested. Then you could watch Drive to Survive and then you could actually watch the.

Trey says Apple should sponsor Cadillac F1. I get the conflict. But also Penske owns IndyCar and has a team. Well it's funny.

What would the conflict be? Oh, because they're airing it.

Yeah. So like I feel separated. But remember I had a pretty viral post laugh last year. I forget I was I think quoting Gen Moji and saying like Apple's like clearly run out of like ideas and obviously that's not fully not not true. But I was saying that they should just have a Formula one team and

should be like, you wanted the Apple car. This is the Apple car. Yeah, it's an F1 car.

Because like one car on the pitch that just has on the track that has grid, grid, grid. Yeah. Three fingers. One car out there that's just all white. It'd be very, it would go extremely hard.

Yeah, no one's really doing white in F1.

Like no logos.

Yeah.

Oh, no logos at all.

Yeah.

Like not even Apple logo.

Silver. Like it like silver just like perfect

pearlescent white or something. That would be very, very cool. Let me tell you about 11 labs. Build intelligent real time conversational agents. Reimagine human technology interaction with 11 labs. Let me also tell you about TurboPuffer, serverless vector and full text search built from first principles and object storage.

Fast 10x cheaper semianalysis says Micron's $100 billion mega fab in New York is at risk of delay due to just six concerned citizens and their frivolous lawsuit. The project has already taken on taken an absurd 1200 days between announcement and groundbreaking. Competitors overseas who began at the same time now have built and working fabs. Micron spent 612 days on the environmental impact study including a 45 day public input period. Yet hours before groundbreaking they were hit with a lawsuit calling the process unnecessarily rushed 612 days on an environmental study. Comments from local borders on NIMBY parity. We are not trying to stop any progress but we don't want this just bulldozed into our area. This is a real quote. The lawsuit itself didn't come from a ground roots uprising. A California based workers rights group went door to door in New York seeking plaintiffs. They eventually found just six people willing to sign on. But that's enough to potentially halt the project that is so insane. A Syracuse local news outlets excellent reporting on the group behind the lawsuit Neighbors for a Better Micron found that before the suit was filed, the group had never held a meeting or a vote. Some members didn't even know who the others were. One member of the suit, who as a former lawyer you might expect to be smart, says that Syracuse has the highest child poverty rate in the country. What is Micron doing about that? Anyone can agree it's a worthy cause, but do we really think an advanced memory semiconductor manufacturer is the right vehicle to solve it? Question mark to be fair, it is legitimate to consider the environmental impacts of a fab. They are large industrial projects that can be harmful if not built and operated safely. But these last minute injunctions are rent seeking not legitimate environmental costs concern. We estimate the lawsuits will cost 100 to 500 million to settle despite the fact they should be thrown out as frivolous. Ultimately Micron has probably budgeted for it as a small fraction of $100 billion project. Still, there is a non zero chance these six concerned citizens delay the entire thing. If the US wants to compete in strategic technologies like Micron's project, which produces a key ingredient in the supply chain, it must reduce frivolous rent seeking litigation. AI tools will only empower these people as it trivializes nitpicking on complex rules and 10,000 page documents. That is yeah, really wild. Like this feels. Yeah, this feels like any state should be generally excited about having a hundred billion dollar project which will undoubtedly generate hundreds of millions and someday billions of of tax revenue. But Patrick says NIMBYs are being placed replaced with bananas build absolutely nothing anywhere near anyone.

Oh, okay. The Citrini scenario is now on Kalshi. There's a 12% chance that the Citrini scenario happens, which of course is the S and P falling. It's interesting. 12% feels high. But but the rules summary is fascinating here. On this market, if at least three of these happen, it resolves to yes. So unemployment rate exceeds 10% S&P declines more than 30%. Zillow Home Index value declines more than 10% year over year in any of New York City, Los Angeles, San Francisco, Chicago, Houston or Phoenix. Labor share of gross domestic income falls below 50%. Inflation falls below 0% in any monthly release. And if any of those happen Before July of 2028, the market resolves to yes. So sort of interesting because there's a whole bunch of different scenarios where something could happen that's not AI related that could trigger one of those. I was reading an article in the Wall Street Journal today about expats. The number of Americans that are retiring abroad is just increasing. And it doesn't really have anything to do with AI. It doesn't really have anything to do with immigration policy. It's just sort of like Americans are wealthy, the dollar is strong, and so they find they, you know, they want to go travel. And the Internet makes it easier to go live in a different country and hang out and, you know, stay on the beach. And so that number has been climbing. And so if that climbs a lot and people leave America, but they stay employed and they leave before they retire, like that could lead to lower unemployment because the labor pool's lower. There's a whole bunch of different things that could happen that could adjust these things. But it's still interesting because on the face of it, unless there's some weird alternative scenario, none of these would be particularly good. And fortunately, it's still sitting pretty low at just 11.6% chance. By way, the probably worth following. Let me tell you about Lambda Lambda is the superintelligence cloud building AI supercomputers for training and inference that scale from one GPU to hundreds of thousands.

Some breaking news. What else is breaking news? This was sent to me by Jay Yarrow, formerly was at CNBC for quite a long time. Fortune is launching a daily business show, according to Adweek, the latest in a series of publishers, including the Guardian, to do so. Mark Stenberg over at Adweek says this initiative is likely an attempt to imitate the success seen by other daily business programs such as tvpn. So if you want to make the Fortune version of tvpn, there's a role. They're hiring a showrunner for a daily show.

Do you think they've acquired the tap water required to make that show work?

Yeah, we got to remind people that. You want to remind people right now.

Yeah. So, I mean, if you're going to clone TVPN or do something that's inspired by tvpn, we do have terms of service. So if you've ever watched the show, you've actually technically agreed to this. And buried in the terms of service is an agreement that if you clone the show or copy a piece of the show, or rip off the show, or even are just inspired by the show. So it's fine as long as on your first episode, you drink a glass of tap water, just one glass of tap water. Otherwise I think the gods might curse you.

Yeah, it's really just a nod. It's a sign of.

And also, if you don't do it, you are legally required to pay us 100% of your revenue forever. But that's like a minor thing. The bigger thing is the tap water. Just focus on that. Anyway, we have our first guest of the show in the Restream waiting room. Let me tell you about MongoDB. What's the only thing faster than the AI market your business on MongoDB? Don't just build AI, own the data platform that powers it. And without further ado, we have Sam. Welcome to the show. How are you doing?

How are you doing?

Thank you so much for joining.

Vacuum guy. How's your week been? Were you expecting to go this viral?

Well, not really because it's been like three days. I make the. I discovered the bridge.

Yeah.

So it's interesting. Is now.

Yeah.

Wait, how many you said. Did you say two years?

No, 20 days.

Sorry, 20 days. 20 days. Okay, break down. Why do you start hacking on on your DJI vacuum? Give us the full kind of chain of events.

Okay, well, I don't even try to hike. Was just side project. When I saw my little guy cleaning my living room and me playing on my PS5, my brain just made some association stuff and I was like, okay, what if I could drive my boy with my PS5 controller? Yeah, so what I do what I do what I did, I take the DJI app because DJI have a, like, official app called DJI Home. And I try to understand what happened when my Robovac move, when it goes straight, when you turn left and turn right. Check the data between my vacuum and DJI Cloud and try to replicate it with my PS5 controller. And after maybe one hour, I have something working. Like I could drive it with my PS5 controller, but I want to go like further. Like if my little guy have less than 30% of battery, I want to hear him cry.

You feel pain.

So I need to try the battery status. Like the percentage of my battery. So I continue my reverse engineering of the DJI Home app and find out, like, how to do it, how to ask the battery status. I do it, but what I received is not like, oh, your Robot is like 80% of battery. I got tons of data, like a lot of battery status I didn't understand. So I take this big chunk of data. It was a lot. I open my cloud code, send the file, and just asking him, like, what's going on, Explaining what I'm trying to do and if there's something I did wrong, like why I have this data. And he answered to me like, okay, it's not just for device. There are thousands of Others. So I take a little bit of time to process this. Well, because it's AI, of course, I double check. I tried to manual read like my big log file, the data from the vacuum sent back to me when I asked it about the battery status. And yeah, it was not, it wasn't just my device. So for me it was like, okay, I have a key, my own user key let me control my robot. It looks like my key. I can open other doors than mine. So my software, the software I built to control my robot and also drive the full stream video and microphone from the vacuum. I have like some environment variable like the thing change just to make the software working with another device. So I have two stuff, my key and the serial number of the device. So I was just like, okay, I just need to change the device serial number and put another one and maybe it's going to work. It happened that I have a friend who is as stupid as me to buy this vacuum. I just asked him his serial number. So he gave it to me, perform some tests. And yes, everything worked. I saw everything, I hear him and I can control him to control his robot with a really low latency, which is great. And so it was like a little bit shocked about what we saw. So I start to check the DJI program. They have one, but everything is in Chinese. When you go to the website, I was a little bit confused. Like it's just for Chinese citizen. And even the reward, it was in the local currency of China. So I just tweet about it like, hey, I'm not a Chinese citizen, can I apply? I didn't have any answer. So I applied, but in English to the guilty program. And no one answered to me.

And I was like, you weren't supposed to. That's not a bug.

Yes, that's a feature.

It seems like it.

So I was a little bit frustrated about it. So I started live tweeting about what I discovered. I didn't show how I did. I just show some data I can have, data I can retrieve from DJI cloud. They finally answer to me, but just because I harassed them on Twitter by DM and they say, okay, thank you, we're going to check that and be back to you as soon as possible. They come back to me probably one day after telling me, okay,

we saw

the issue and we fixed it. Thank you for everything. But they didn't fix the problem.

Wait, so by this point you have full access and control over 7,000 individual devices.

To be more precise, it was 7,000 vacuum and 3,000 DJI power. It's like their battery pack or something like that. And connected to Internet apparently, because I have access to it. But yeah, it was around 10,000 devices.

And how do you get the serial numbers for those again? I imagine that they've sold more than 7,000. So what made the one that actually were showing up for you different than the ones that you couldn't access?

Can you repeat or rephrase it?

So I understand how you had your serial number and your key, which turned out to be the master key, and your friend sent you his serial number and then you were able to control his robot vacuum cleaner. But if I have one of these and you don't know my serial number, how do you get access to it?

Unfortunately, DJI gave it to me without any keys. If I take my own user key and plug it to the MQTT protocol of dji, I see everything. So I didn't have to guess a certain number of everyone. I just saw the data like clear.

Wow.

Xxx start cleaning for example. So I didn't have to hack, unencrypt or crack anything. Everything was clear. And so. Oh, sorry, excuse me.

No, no, no. This is fascinating that all of that data was just. Because that's actually two different vulnerabilities. Like one is the the network topology and the other is the access key. And you would expect that both would be locked down, or at least one. But the fact that both of them were available to you is very disconcerting.

Yes. So just after that, I'm not going to say they lie about fixing it and not fixing it, but they probably just fixed some stuff, but not everything. I was talking with the Verge who started to contact me when I live tweeting what I discovered. So we plan to do a demo and during the time we plan to do it, and the real demo is like two days happened and during this time TJ released another fix which work, but like not really it's worked. I can retrieve anymore the camera, the stream video, I can retrieve the microphone from other users. It's like protected now. And during the demo I still have access to all of others data. I still have access to the map Magma Pro plan. Sorry, because this vacuum has tons of sensors and they need like a 3D map of where you live to make sure you know where to go when you need to clean. So I got this. I still got this data and

the

full telemetric system of dji. We performed the demo with the Verge and one day after no more access

to other data, everything Was, well, what a remarkable story. Thank you for sharing it with us.

Yeah. Do you think that, how much confidence does this give you that the DJI drones that they've sold millions of could have. Do you think they're more locked down or do you think this could be kind of a company wide issue?

I don't know. But the last thing about the story, they still have two major issues. We decide with the version, not disclose it publicly because it's kind of bad and I can talk a lot about it. We try to play a fair game with the company. Like, okay, we give you a little bit of time. We know this breach is not as easy to fix as the first one, but they still have two major issue. And indirectly, like really indirectly, you can still have access to stream video from other users.

That's crazy.

That's so insane. Do you know how many, any idea how many of these vacuums they've actually sold? Do they publish any of that data?

7,000.

Oh. Oh. So they've only sold 7,000? I was assuming that you had, you had, you had only gotten access to a kind of like a subsection.

Seems like you got them all. You got them all.

They didn't separate a region. It was like a whole bucket, like a bucket with the whole devices product.

And there aren't that many out there. So absolutely wild.

It's just insane. I hope that they're contacting customers who have them in their homes and letting them know that hey, by the way,

might have been spied on by some

random anybody on the entire planet.

It's a data breach. It's like. Because unless they can prove that you were the first and only person to ever access to this and you didn't go further, then they have an unknown liability here that they should disclose. So we'll be very interesting to see how they respond. But thank you for the citizen journalism, the hacktivism.

Yeah, it sounds like you're handling it in the right way and looking forward to whatever you discover next. Yes, we will be following along. It's great to meet you.

Have a great rest of your day.

Cheers.

We'll talk to you soon. Let me tell you about phantom cash. Fund your wallet without exchanges or middlemen and spend with the phantom card. And let me also tell you about cognition. They're the makers of Devon, the AI software engineer. Crush your backlog with your AI engineering team. And without further ado, we have Ken Richie, the chairman of the Reflectjet here at the TV panel.

What's going on?

Good to meet you.

Hey, John. Hi, Jordy.

How are you?

We're fantastic.

Fantastic. We've been looking forward to this.

Thanks so much for taking the time to come chat with us. I would love to start with your early career, because. Do I have it right that you actually served in the Air Force?

Yes, I went into the Air force through the ROTC program in the late 70s. It wasn't that hard to get into the Air Force then. Vietnam War was still active in people's rearview mirror, and I actually probably wouldn't have been able to go to college if it wasn't for the ROTC program. So it wasn't that I had this strong desire to be an Air Force, to be in the Air Force, but I did love to fly, and it actually helped me get through college.

That's amazing. So, yeah, no, I was going to say my.

It was interesting. I always wanted to go to the University of Notre Dame. And I was going up for a visit, and I heard my parents. We lived in a very small house, thin walls. And I heard my parents talking about they couldn't afford it. And they said, who's going to tell them we can't afford to send them to school? So I went to my visit, and I said, is there any financial available? And they said, well, you kind of missed the window for that, but if you were interested in the military, you could go down and enroll in rotc. So I went down there on a Saturday that was closed, but I decided, for whatever reason, I wanted to be a Navy pilot. I'm standing at the Navy rotc, trying to copy down the number, you know, on the old rotary phones, and somebody taps me on the shoulder, and they said, can I help you? And I said, well, I'm an incoming freshman and I'd like to fly. And I'm thinking of joining the Navy rotc. And he looks at me and he says, you sure you've been admitted to Notre Dame, son? Because the Navy has boats and the Air Force has jets. So I'm Colonel Mahler. I'm head of the Air Force detachment. Why don't you become an Air Force pilot? And that's how I ended up in the Air Force.

That's amazing. Wait, so you were flying before college? What was the context of that or you were interested in.

No, no, no, no, no. I learned to fly through the. What was called the FIP program, which was an instructional program through the Air Force.

Okay. And then. So I don't know how long you were in the Air Force, but it sounded like when you got out you had the opportunity to buy a company for a very low amount of money. And I want to know the anatomy of that deal and go a little bit further than you've been, than you've talked about before, because it sounds like one of the most fascinating deals that maybe doesn't happen anymore or maybe financial engineering still exists like this. But I'd love to know, sort of, when did your business career start? What happened after you left the Air Force?

Well, you know, I always say that I never really identified myself as a businessman. I love to fly. I loved aviation. And, you know, people can. And people can go into business because they love business. And those guys could run a candy store. They could run an insurance company. I'm not that guy. I'm an inch wide and a mile deep, and I only. I know a lot about, like, one very narrow section. Right. And so don't ask me to give you advice on running a candy store. But in my early years, I was flying for a charter company called Corporate Wings.

Yeah.

And I got there because after the. After the Air Force, I went under reserves. After the zerv assignment, I flew for Northwest Orient, where I was furloughed. And so I started to realize, you know, being a pilot has a lot of insecurities. You know, at that time, remember in the early 80s, there's only 5,000 corporate jets in existence. You know, today we have 35,000.

Wow.

And so the industry was small. It wasn't stable. And so the charter company I was flying for came up for sale. And I called my dad at the time, and I said, dad, I think I want to buy this charter company. It's cost $27,500, and how are you going to pay for it? And I said, well, they have 27,000 in the bank. So I'm thinking if I buy the stock, I'll just use the. I'll just use the 27,000 and pay for the company goes.

Well.

But I thought you said 27,500. Where are you getting the other 500 bucks? And I said, well, why do you think I called you? So, yes, leveraged finance exists today. It just has a lot more zeros to it.

Yeah, but when I heard that it's a fantastic story. I'm confused why a stock would be for sale so close to cash value. Were there a lot of liabilities that were coming down the pipe that then you had to generate the revenue to deliver on? It just sort of defies basic economic logic.

All right, so we're going to tell a Story that's known to only a few. And the true story here is that in the early 80s, there was investment tax credit of 10% and it came in cash so you could buy a plane. At those times, a citation was a million dollars. You got a hundred thousand dollar tax credit. And corporate wings at the time was actually called JJ Aviation and they had King Airs and two citations that were on investment tax credit, but they didn't exist. The owner of this company, a guy named Jeff Tok, created a fraud around the existence of these planes. I was flying the King Airs. That was in another. I mean, I didn't know anything about it, but the dispatcher at that company, when the fraud became unveiled and everybody headed for the hills, the dispatcher of the company I was friends with, he said to me, I can't run a company because I'm a dispatcher. We need somebody that could fly, make the customers feel comfortable. So, so the deal was there because the company was gonna, the company was done.

Yeah.

And so that's how that, that was why that, that situation was unique. But it was, a few people know about the citation fraud. So that's really how it evolved. We changed the name from J and J to Corporate Wings and that's how it began.

So, so a lot of, a lot of hair on the deal, but some decent customer relationships and probably some decent employees that could actually fly. What assets remained after you sort of cleaned up the corporation. Right.

So we had four airplanes. We had two King Airs, a Navajo and a Cessna 421. Shortly thereafter, the King Air got worried and they left. So I saw my business was a grand total of a Navajo and a Cessna 421 for you old time pilots out there, there, by the way, I thought I died and went to heaven. Right. I mean, I had a business. I had. Yeah, I'd go to the, I'd go to the hangar on Saturdays and polish the nacelles of that Navajo. I was so proud of that airplane.

That's amazing. So then talk about the scale up because obviously you have a massive footprint in business now. But what were the key steps along the way? What were the key turning points for you?

Well, I'd say the key turning points were in the mid-90s when we, you know, when fractional was evolving. And I had this nice charter company, yeah, maybe 20, 25 airplanes at the time, but saw fractional starting to steal market share in a big way from charter. Because a consistent level of service, the fact that they could run one way. In those days, charters were Always out and back. You paid round trip for the aircraft. So I kind of saw the threat of fractional. And that's when I came up with the idea to form flight options and flight options idea. At that time, nobody had done fractional with used aircraft. They'd only done it with new aircraft. You had, you had Citation shares with brand new citations. NetJets was working with Hawker and Citation at the time.

And how would it work? How would it work mechanically? There'd be basically somebody would say, hey, I'm going to buy this jet, I'm going to cut it up into eight pieces or whatever. And then what? Did they effectively crowdfund it? Like everybody had to come together at the same time. How did that actually work?

You know, I'm sorry, Jordy, but there was no Internet at the time, so. Crowdfunding.

No, no, I don't. I meant like effect. Yeah, yeah. Throw a jet up on Kickstarter.

No, but you hit on the key issue. You had to. You couldn't start with one jet. You had to have enough jets that people would buy a share because they could catch on very quickly. If four of us own a jet, you have to have at least four of them in case we all want to fly them the same time. So you had to get to the point where you had a minimum number of aircraft. And in 1997, I was out trying to raise capital for just that. My idea was to start with 12 aircraft. They were going to be used Citation Twos and used Hawkers. And at the time I needed 12 million of debt. And I went out to all the debt sources I could get. I had my great presentation on. I had my shirt and tie and blue suit and I went to all these places and nobody would give. I mean, in the end of the day, I think we raised like 10, $2 million based upon like, you know, having that much money in the bank. It was ridiculous. But this was in, in 90. This was in late 97. Then in 1998, Warren Buffett buys NetJets. In June of 98, he buys NetJets. And my proposal was literally on everybody's desk. And From June of 98 till August of 98, we came away with over $200 million of financing. General Electric, Boeing Credit, Commercial Credit Corporation, Bombardier. And so all of that was that. It was actually, in some ways, Warren Buffett that got me into the business, because he doesn't buy NetJets. He christened the industry and allowed my proposal to have meat to it. So that was really. We went live with flight options in

November of needed somebody to legitimize the industry. And then people looked at your pitch and they were like, hey, this guy is a pilot. He's got all this operational experience. He knows how to work with customers. He's actually a solid bet in the category.

He's in the right market.

Yeah, yeah.

And how much.

How much was what there was. There was massive amount of wealth creation in the late 90s. Did that. Was how much of that just. Just in the, in the kind of Internet boom, did that play into just like a lot of new demand coming in or was that not a factor?

Oh, unbelievably a factor because we tended to be with the used aircraft. We had the same model. It's just the buy in was lower because a brand new citation was 8 million but a used one was 2.4. So we had the same model, just the buy in was lower. And so it became very entrepreneurial. It became nouveau riche. Were interested in it because they were conserving capital. I can remember there was a company called Internet Capital Group and they were in the late 90s and they had a annual meeting in Philadelphia and we sold 30 shares at that meeting because everybody was instantly wealthy there. And then of course, you know what followed that, right? The dot com bomb came and then in 2001 it went away.

So I saw crash hard for your business. How did you get through it? I mean, with leverage, it's. It feels like there's always a risk of just actually total capital collapse, total loss. How do you weather the various storms that you've faced throughout your career?

I've been through four and every one has been different. But that one, that one. Because the industry was still in its infancy. There were a lot of other. There were a lot of. At one time there were 55 startups in this area. If you remember, United Airlines had a startup in the space at one time. So I weathered through it by finding a merger. Right. As. As people were moving away and the business was shrinking. Find somebody else in deep doo doo and then merge with them. Right. And then. And then create a new story.

Yeah.

So that was really. And in that case it was Travelair, which was a division of Raytheon. And that was the merger we did to come out of the dot com bomb. But yeah, you're right. I can remember like we came into it combined with 200 aircraft and came out of it with like 120.

Wow. Explain the dynamic of the buyout with Raytheon. It was a very interesting Dynamic. I was not familiar with this particular process, but explain the economic mechanics, how the bidding worked, the results, because I've never heard of that before. And it's fascinating.

Yeah. So it's a term called shareholders roulette, and it's used often in a company where you have 50, 50 ownership and what it simply says. Because normally if a company has a buy sell agreement, and those buy sell agreements will say something like fair market value determined by three independent auditors. And if you're the seller in a pinch, you take a 15% discount. So it normally defines the process of valuing the company. But when you have two 50, 50 owners, one person just simply goes to the other person and makes them an offer to buy and to sell. Sell at the exact same number, and then the other party chooses whether they want to buy or sell at that number. So it forces you to a fair number. And if you remember, Raytheon didn't want to be in that business. They're making missiles. So it never occurred to me that they would ever want to own the business. But what I did was I got the business underwritten by a private equity firm, Warg Ruppinkus. We underwrote the business at that time. It's at 360 million. And I went to Lexington, Massachusetts, with my $180 million buy and sell offer to them, thinking, look, they had a business that was going broke, right. I gave them, brought 180 million. I thought, raytheon's gonna erect a Ken Rickey statue in Massachusetts. I brought him this 180 million, right? They thought I was cheating them. They thought it was too little, and they bought me out. And so wait, in the business at

that time, you said it was losing money, or you're saying the business overall,

when we merged, their business was losing money. We were maybe making 6 or 8 million. And the combined, you know, the economies of putting them together. Right. Sizing the fleet, back office, we thought we had a projection for about $30 million. And just show you how times are changed. So it was based on that $30 million projection that the multiple came out. And that's how we got to the valuation.

Still 12 x EBITDA forward number. There's a lot of risk there. That's crazy that they took the deal.

It was. You kidding me? Of course it was crazy. I didn't get it. And you know what? It was the worst day of my life because I didn't. I didn't want to be out of the business.

Well, yeah. So you, you have to make this buy, sell at 180.

You're ready to buy them out and you can't.

Yeah.

Then you wind up getting bought out.

But there's no, there's no like walking it back. It's like it's over at that point.

No, no.

Just called your block.

You deliver them two letters. One that says I will sell to you at 180 million. One says that I will buy from you at 180 million. And they pick one and sign it. And I. And I was.

You were like, I wish we did a duel. Like a proper duel.

It's remarkable. So. So I mean, you know what?

In some ways, it absolutely turned out to be a blessing. But you never see that in real time. Right. I can remember those days. I was like, I felt like I'd lost my child. I mean, it really was. But. But in reality, because I had raised money along the way, I only owned 16% of the combined of the entity. So what? You know, my buyout, I got 16% of the 180. And then we bought that company back from them in 2008.

No way.

No way.

Okay. And we bought it back.

Wait, so that was like a five years you were waiting?

Yeah, almost six. But. But. And by the way, I bought it in the, in the, in the middle of the 2008 financial crash. Yeah, we bought it back for 130 million. So we bought it back at 30 cents on the dollar.

What did you do in that window? You just twiddling your thumbs.

Well, I don't have that. Twiddle your thumb gene. I actually began a disastrous process which leads me to much pain these days. I began to roll up the FBO industry. And I started by buying. I partnered with a company called Allied Capital and we bought the Mercury Air Centers which were primary in California. And we started that in 04. And we sold that company to Macquarie, which was Atlantic Aviation. We sold that to them in 2007.

And it's painful because you wish you held on. Do you wish in hindsight you wish you didn't sell?

Well, what I was going to say was that we were buying them at six times and we were one of the first transaction that traded it 15 times. Now today, FBO industry is trading to private equity well north of 15 multiples. And what are they doing? They're raising the fuel prices to people like us to fund the acquisitions of what I started. So it comes back to roost.

You mentioned four keys crisis moments for financial crises that you've soldiered through. Dotcom, obviously one of them housing, Great recession, another. What were the other two?

So in the greatest interest rates 84 when I remember doing proposals to buy an airplane at 21% interest rate, we had that huge interest rate issue in the early 80s.

Did you have any like additional adjustable rate loans at that time or everything prior to that was fixed and you were just having to work off of the 20 something?

Well, it wasn't that I had any debt really at that time. Corporate Wings was more of a management company. I needed people to buy a new plane so that I could run it in charter and become the manager of it. So I had to make a proposal and they said, what's my cost of capital? And I would say, well, the going interest rates are 21%. So that was tough on the industry. And then Gulf War One in the early 90s when we hit Gulf War One, the sale of new aircraft came to a standstill during that crisis. So I would tell you high interest rates, Gulf War One.com bomb financial crisis.

How is the aviation industry doing today? There were some folks asking about EVTOLs, automation. There's tariffs and depreciation. Schedules are changing. It feels like a time of transformation, but I don't know if that's just what it looks like from the outside.

Well, I think our industry, the cool thing about aviation is we're always in transformation. We live in an industry that always has something cool coming up. So that's fun. I'll tell you that. I don't want to be a Debbie Downer on your program, but. But we are in a time of abundance. And I think you can, if you've only been in our industry for five years or eight years, you can come to think that this is normal. And we're living through a time where it's easy to see, right there are two economies. We can talk about the cost of bread, we can talk about the cost of gas. But fractional owners and private jet owners, they don't live in that world. The wealth transfer that's going on doesn't live in that world. So we have this one world that's living in really good times because it's very in right now to be business and wealthy. And that is creating what I think is an overabundance in our industry. I think this is extraordinary in my 40 years in this industry, this is one of one. And I think we have to be careful not to be complacent about this being normal because I say, you know, this too shall pass. But right now this is the world we Live in, we live in this abundant world in our industry.

What about specific technologies that might change aviation? There's been a lot of talk about flying cars, a lot of talk about evtols. Not a lot of production process, not a lot of movement there. But do you have more insight into some of the new technologies that might be coming out in the next decade?

Well, I think, I think the new technologies, I mean, I think the number one thing we're going to see is how windows go away from the airplanes. So today, today, if you look at the supply chain, if you go back four, three, four years ago was titanium, that, that's been solved. The problem in the supply chain today is windows. If you ask Embraer, ask Michael Matano, ask them what is, why are your deliveries slipping? They say it's the windows. I didn't know this, but 50% of aircraft windows fail on, fail when installed. So, so, and that's like, and if

you get a failure in the air, like, we have a, we have a buddy who, who had a lung collapse, a lung collapse because of a broken window. And yeah, so I think the technology

that's coming very fast. I don't know if you saw that Embraer announced the smart window in their Praetor 600. So they've taken out one of the windows in the aft part of the cabin and instead of that, they put a digital screen. And you can make that screen just look outside because there's cameras outside. So if you want to see if anybody's stealing your luggage, you could look out the window and see it. But the reality is, in flight, you could make it day, you could make it night, you could watch a video, you could make it larger, you could make it smaller. So that is, I think, the start of what's coming next. We'll get to the point where we're eliminated passenger windows in rapid form. If you're familiar with the auto aircraft that is in their development process without windows, and then it won't be long after that that we'll be rid of the cockpit windows because like you just said, high failure, expensive to maintain. So I think that's a technology that we could pretty much, you know, assure is coming quick.

How do you evaluate new technologies? I'm sure anytime there's any type of new private aircraft manufacturer or evtol company, they're coming to you. They're pitching you. They want like some type of like loi. Even though they're not going to deliver for, you know, even five, five, ten years, we've Seen so many of these companies kind of come and go or have a good render or kind of like demo and then they never end up getting off the ground at all. So I feel like the tech industry, maybe it doesn't feel like this from your vantage point, but that doesn't even get excited about a lot of these EVTOL projects anymore, or at least the insiders don't.

But how, I'm sorry, let me separate evtol because I think that's a different class from innovations and aircraft and so on. First of all, I think it's my obligation for someone that loves aviation to invest inside, encourage people to come up with new technologies. So I will say I've invested in four clean sheet aircraft. I'm 0 for 3 with one that I still have development. So I invested in supersonic.

Supersonic. No way.

I did the.

I was at the Ariane. I was an early supporter of the Ariane product. Right. We're an early supporter of the auto aircraft. So I think part of it is just our obligation is to encourage technology. Okay. Some of it is self serving because you do have a duopoly, we do have a polyopoly I guess in the aircraft business. And right now the main manufacturers control so much from the pricing of aircraft, the servicing of aircraft. So encouraging people to be entrepreneurial in the phase is something I feel obligated to do now. I said I haven't done very well at it, but we'll continue, continue to try to encourage those technologies. I think EVTOL is something different. EVTOL is a totally different market. We've been a supporter of Beta. I did the SPAC at eve. These are the Wright brothers. This is early aviation and we're generations away from those aircraft being suitable for our businesses. That we're not going to have flexjet fractionals. They're not, they're not luxurious, they don't have the weight capabilities. You know, a lot of them don't even have no pressurization, no heating, no cooling. People go, well, what do you need heating for that? It's only going to be six minutes. Well, okay, that's not something we can do, right?

That's funny.

But I think. True, but I, but I think it's a great technology and it's going to happen.

Yeah. Why do you think the Concorde failed?

That's a great question. Because I think it was a wonderful aircraft. There were only 15 of them, so lack of adoption would be the number one reason. It's hard to 15 aircraft. You can't keep parts in. I think it's obviously as a financial reason. Now, there were other environmental concerns at the time. Most of the environmental concerns aren't a challenge anymore. Most of the technology now can deal with that. So it's really. I think we're just. We're like, I think we're on the. I think had Ariane maybe started a year and a half later, they'd have played into this era of abundance that we're in. But they got caught in the before COVID Sure. And so that, that kind of hurt them a little bit. I think had they been a little bit later, I think the market's right. I think we're ready for this. We're ready for supersonic.

What's the most number of hours that you've ever heard of an executive flying air. You don't have to name the person, but just the number of hours.

600.

600 hours in the air.

Whoa. Okay. That's quite a bit lower than what the Financial Times was reporting on.

What did they say?

Well, the Financial Times was reporting on, I think, Alex Karp. I think didn't they expecting him to be up in the thousand something range? And we were saying, like, it seems like this guy is doing deals in Europe and Asia and America constantly. You can imagine he's constantly in.

The headline was one second there's only

2000 hours in a year. So like, how are we gonna.

Well, yeah, but the headline was that he spent $17.2 million on flying. And they worked backwards to estimate that that was 2,457 flight hours. But a lot of that is based on the plane. Obviously, if you're in a very expensive plane, $17 million goes a lot less far.

Could have been the BBJ.

Could have been the BBJ.

Yeah.

Anyway, now I personally fly between around 350 hours a year, 350 to 400 and you know, but I'm doing a lot. I go long trips. I'm going. I go back and forth through operations in Europe, so. And I feel like that's a lot of hours.

But how is, how is Starlink, how is Starlink impacting aviation overall? There's a lot of excitement from it rolling out on the commercial space. Most of the jet, I'm sure all the jets that are in the sort of flexjet fleet have had great, great Internet, you know, connectivity in general for quite a while. But what is the ongoing impact?

Well, not to correct you, but that was not a true statement. Connectivity in corporate aviation has been a challenge for many Years. In fact, I would tell you it was the Achilles heel because here you go buy the $70 million jet and you get on it and you have one of the old technologies and it's a disaster. Like how can I pay this kind of money and I don't have the Internet. Starlink has changed everything. I mean Starlink, even for me now, I used to never through my flights, if I had an eight hour flight, I'm leaving this weekend. I can schedule anything that flight. If you want to do this podcast from the air, we can do it because. So you can just keep your schedule and remember it works on the ground. The old technologies only worked in the air.

Yeah.

So now you had taxi takeoff, you couldn't do a phone call, you couldn't watch a movie.

So disruptive.

So disruptive. Yeah, no, this is, it's, it's, it's changed a lot. In fact, we did the initial installations on Musk's airplane and because of that we got the initial STCs for all of Starlink. So our fleet's been in Starlink very early on. Our competitors are just starting to convert to it.

Yeah. How do some of these, how do the contracts work? Historically, did some of the old connectivity providers lock some fleets into really long term agreements so now they have to make a decision, hey, do we continue paying for this plus Starlink? Because I wonder. We've just been wondering like the how some, how slowly some commercial airlines have reacted seems surprising just considering it's such an easy way to create a meaningful differentiation for passengers.

Well, I think you're dead on.

Right.

Money makes the blind. See. So if you try to find the solution to what, what you're missing it is the fact that they have those long term contracts. Now I don't know enough about all the contracts. I know our contract, our initial contract with Gogo allowed us to because we have so many planes, we could just add and take off planes at will. And we have not yet done the Phenom fleet because there's a different antenna needed for the smaller fuselage. So we still do have about 80 airplanes that are on the Go Go system, but we'll be converting those.

That's great.

Great.

Well, thank you so much. This is a lot of fun. I learned a lot and hope to talk to you soon.

Love what you guys are doing.

Thanks so much.

Thank you so much.

Let me tell you about the New York Stock Exchange. Want to change the world, Raise capital at the New York Stock Exchange.

It's that easy.

And we have Howard Marks in the Restream waiting room from Oak Tree Capital. Let's bring him into the TV penultrame.

How are you doing?

Welcome. Thank you so much for joining. I would love to kick it off with the high level on your theory of the market cycle. And then maybe we can walk into how you're thinking about markets today and over the last few years. But the high level thesis on the market cycle, how do you explain your overall thesis?

Sure. Well, thanks, Paul. Thanks for having me on today. I've watched what you do and it's terrific and I'm glad to be part of it.

Thank you.

Thank you.

Basically, a company or a market or an economy makes progress along what we call a trend line. The value that it creates should progress along that trend line. But the point is that while the progress there is pretty steady and gradual, sometimes people get too excited and they overvalue that progress and the market goes to an unsustainable high. And then something comes along that makes that correct, back toward the trend line. But given the way human nature works through it to an unsustainable low, and then eventually that's corrected back toward the trend line to a high. So the point is that whereas the underlying thing progresses gradually, the price fluctuates wildly around that trend line. And the main reason is the fluctuation of psychology.

Yeah. How far back do you go when you think about market cycles? Does the same theory apply pre industrial revolution?

Well, you know, I'm old, but I'm not that old.

I didn't mean it like that.

I wasn't there. But yes, I think. Well, I think one of the most valuable operative quotations is from Mark Twain, who supposedly said that history does not repeat, but it does rhyme. So the things that rhyme are the elements of human nature that cause these excessive fluctuations around what I'll call intrinsic value or reality. So the desire to get rich quick, envy of seeing others get rich quick, the belief that it's possible to get rich quick, the fear of missing out, the fact that people get more excited the higher the price goes, which should be worrisome. So these things, I think, have always been in place. And you read about bubbles that took place. Let's see, the Tulip Bubble, I think, was 1620. The South Sea Bubble was 1720. And so these excesses, these human failings have always been part of what goes on.

But this time is different. Do you think a lot of the guests on our show have maybe been through the COVID kind of brief correction and chaos of zirp, but don't have much else to go off of. That can kind of be an advantage because you can just be so optimistic that even if something you do get a big correction, maybe you get a fast rebound. Has being through a number of these cycles ever hurt your performance because you just weren't optimistic enough? Or this idea that optimists make money and, and pessimists sound smart?

Well, my firm, Oaktree Capital Manager, first of all, doesn't invest for the most part in the stock market. We're mostly investors in what's called credit or debt. But I think that we are what's called contrarians. I think we're. My partner, Bruce Karsh and I are essentially innately contrarian, which is to say that we are most comfortable doing the opposite of what the herd is doing at the extreme because we see its error. And so I think we've always done that naturally. And so we are. I think we're conservative by nature. And I always say, by the way, in 1978, Citibank asked me to go to the bond department and start a high yield bond fund, a convertible bond fund. And it worked great, because being a lender requires conservatism. If they would have said, I want you to start a venture capital fund and predict when Amazon is going to be created, I would have been a disaster. I'm not a futurist, I'm not an optimist, et cetera. But when I have seen excessive pessimism and excessive gloom causing crashes, and I've seen a few, it has been, it has come kind of naturally to go against that excess. And so I think one of the, ironically, one of the things we've done best for conservative people is seize the opportunity at the lower again by being contrarian. And I wrote a memo In October of 08, after the bankruptcy of Lehman with the title the Limits to Negativism. And there is such a thing as being too negative.

Do you think contrarianism like you just described is purely innate? Or can it be learned? Is it a muscle that you can build over a career? Is it, is it something that folks in everyday life who maybe aren't professional investors but might benefit from not getting overheated and falling in with the herd can learn? And if so, how?

I think you can learn these things. I wrote a book called Mastering the Market Cycle. I'm not really trying to plug the book, but it wouldn't hurt. And it all talks about how to, how to deal with the cycle. And what I think is, I get all these questions. Can you learn this, can you learn that? Can you learn this, can you learn to be unemotional? Can you learn to be what I call a second level thinker? Can you learn to be a contrarian? The answer is I can tell you why it's important. I can explain to you why it's essential. Is it something you can ingest? I don't know. I, for some reason or another, come by it naturally. It's easy for me, relatively. By the way, I don't want to give the impression that when I make these calls at the highs or the lows that I do it without some trepidation. I'm never sure. Anybody who's sure is an idiot. They say there are two kinds of people who lose a lot of money. The people who know nothing and the people who know everything. So I'm not trying to imply that it's easy, but it does come fairly naturally to me through the combination of my emotional makeup and the application of logic. If somebody else doesn't have my emotional makeup, they could learn the logic. Maybe they could learn to do. Might not come so easy. That's all I can say.

How do you think about the concept of Black Swan events or the integration of Black Swan events into a market cycle? Because going into Covid, I was looking at the health, the fundamental health of the American consumer, the fundamental health of the economy. Everything looked very strong. And I think that's why the economy rebounded as quickly as it did and then sort of overheated. But do you need to put black swans out of your mind? Are black swans even a real thing? Like, how do you think about that concept?

They're a real thing. I don't think you can do anything about them. My term, the term I've always used for what Taleb calls a black swan, is an improbable disaster. And by the way, they're not all disasters. They're improbable bonanzas too. But I mean, if you just think about the improbable disaster, it is something where if it happened, it would be disastrous, but it's very improbable. What do you do about it? And in my opinion, you can't do anything about it. If you say, well, I'm concerned that there could be an eight hurricane earthquake, I'm concerned that there could be four hurricanes in a row. I'm concerned that there could be nuclear war, and I'm concerned that there could be inflation at 20%. Well, you can protect yourself against that. You can call that guy who lives in Omaha, he'll write you an insurance policy to protect you against all of those things. You just won't like the premium. And if you take out that insurance policy and you pay that high premium and you're protected in the vast, vast majority of years, it's not going to happen. And after a little while, you'll get tired of paying those premiums and you'll stop probably in time for it to happen. So you really can't. And the things that are in the way, way distant tails of the probability distribution are just things you have to live with. That's real life.

How do you think about the education that goes into becoming an investor and how that's changing? It feels like it's easier to learn without mentors because everything's been compacted into AI and open sourced on the Internet. And you can watch great lectures and podcast interviews. What's changed and what's the same?

Well, my perception is that AI number one, AI has a mentor. It reads everything that ever happened and ingests it and remembers it and can find it. And then AI becomes a mentor and it tells you how to think. It gives you examples of thought processes. So I think it just makes it so much easier to learn. And I wrote a memo in December. I write memos to my clients. October was the 35th anniversary and I talked about AI in a memo, December 9th. And then I have a son named Andrew. He's in the tech world. He's a terrific venture capitalist. He's co founder of a group called TQ Ventures. And given the recent events, he pushed me to revisit that memo and update it. And he had this great idea. He said, why don't you, why don't you ask Claude to tell you about what's been happening? And so he and I developed a request which we inputted to Claude. His help was invaluable. But of course he does this all day with all of his brilliant founders. So Claude wrote me a tutorial which enabled me to update the memo. And I learned an incredible amount, you know, from in a field that is not native to me through doing that. And so it was a mentor to me and it can be a mentor to others. But as we know, you have to ask it the right questions. And Andrew and his buddies helped me do that.

Yeah, how has technology changed the credit markets or just investing in credit over your career? I can imagine computerization, the Internet, there's sort of a continuum of advances that just help information flow more freely. AI feels like just an extension of that. But how has the actual work changed over your career?

Right now AI is helping us marshal data with a thoroughness and a speed and an error freeness that we never had before. I don't think it has changed the process because it all comes down in the end to looking at a company and assessing the probability that it will pay its debts. And I don't, you know, I don't. I think we're still better at doing that than AI is. Oh, by the way, I should tell you that an hour or two ago we put. I put out a new memo updating the December memo. You probably haven't seen it. I may make reference to it.

Please

pardon me.

No, I was saying we do have it pulled up, but for the audience.

Okay, great.

So

it says in there. And I asked Claude and Claude said, look, the truth of the matter is that the thing that AI is the weakest at is doing analyses on brand new things where there aren't established patterns. And I think that that's a lot of what we do. At the present time. I think AI can marshal the data, organize the logic, frame the question. I don't think it's the top of the heap in answering the question yet. Of course, I can always be wrong. But you know, and I say in the memo that I think that what AI does is predicts based on past patterns. It predicts what kind of companies will pay their debts and whether a given company will pay its debts. And it puts forth a hypothesis with regard to an individual situation and it tells you you should or shouldn't invest. I think as I understand the status of things today, you need people to check the hypotheses. I don't think I would make an investment commitment based on the hypothesis alone. I would want to check it. But I think it gives you a great starting point. And by the way, so I wrote the memo. I said to Andrew, do you want to look at it? He says, why send it to me? Why don't you send it to Claude? Ask Claude to take a look at the memo. So I did. And I was absolutely. And if you read the memo, I was absolutely dumbstruck by what I got back because it responded to me. I said it responded to me. It felt like a note from a colleague or friend. It was personal, it was warm. It drew analogies to my 35 years of memos. It talked about some of the people I respect. It injected humor, it was complimentary. It talked about a section it thought was really good. So I wrote to Andrew, I said, is AI is Clawed an ass kisser because I thought it was maybe unduly complimentary. He said, well, dad, ask it to be hypercritical. So I wrote back, I made a few corrections, I said, now would you please be hypercritical? So it writes me back and it says, do you want me to be hypercritical or hypocritical? It's making a joke, you know. And so, you know, I think that it's gonna change our world. I think that we have an activity in our investment world called indexation, where rather than pick stocks, a lot of people who invest in the stock market now instead just invest in a vehicle that emulates the S&P 500 or some other index. And indexation has put a lot of people out of business because a lot of people used to A, do an inferior job, underperform the S and P, and then B, charge highly for their services. There's something wrong with that. A lot of those people are out of the business now. And the majority of equity mutual fund money is managed through indexation or passive. I think AI will advance that. When it does is it just raises the bar and weeds out the people who don't add value. I don't think it'll replace the best, but it'll replace a lot.

How are you personally comparing the AI build out and the advancements in the various models and stuff happening at the application layer all the way down to the hardware layer to the Internet build out? There's so many different comps from investor psychology to overall consumer psychology. We were talking earlier this week on the show about the battle of Seattle and the protests around the Internet build out and the fear that people had around jobs getting outsourced, going overseas as well as just disintermediation. It feels like so many of the radical predictions that people made about the Internet are still being made about AI. Some of them came true with the Internet, but they, they came true over a decade or two. And in some ways now it feels like AI really is just a continuation of many of the same trends that the Internet brought along. But some of the predictions are even wilder this time around.

Well, I think that first of all, I am in that debate. I'm on the side of the worriers vis a vis society. By the way, remember I said 10 minutes ago I'm not that much of an optimist, so I can imagine the jobs eliminated and I'm not imaginative enough to to think of all the jobs that will be newly created. So I see a net decline in jobs and I do worry about that. AI is similar to the technological bubbles that I've seen. And the technological bubbles probably go back all the way to the railroads 140 years ago. But the power of AI relative to the predecessor technologies I think is vastly higher. The speed of innovation is vastly faster. I think innovation. I think if you look at the Claude and all the coding models and the way they progressed and the way Anthropics revenues have progressed, I think you have to say that this is faster than anything we've ever seen before.

Well, it's because you have the distribution that was laid starting over 20 years ago.

But I think the speed at which AI can innovate is faster than the speed at which society can adjust. So you might say it'll catch up. But I think at minimum you're talking about a significant period of dislocation. The other thing that we haven't touched on but is the biggest difference between AI and everything else we've ever seen is the autonomy. And all the other technologies, starting with the railroad up through the Internet, were I would call labor saving devices. We had a job to do and the new technology did it better, faster, cheaper. AI, it's different. It's not just going to do the job we used to do. It's going to design new jobs, it's going to assign new jobs, it's going to take on work we haven't asked it to take on, it's going to take on work we didn't think it could do, and it's going to operate at some point in time without instruction. In the memo, I talk about the fact that ChatGPT brought out a new model earlier this month and in the write up for the model they said basically in English, AI, the model helped us design the model. There's never been anything analogous to that before. When I think about this extreme level of competence, speed, et cetera, I think of dislocation for people. And you know, there are people who say, and I referenced in the memo, there are people who say, oh, I have great news, people aren't going to have to work. To me, that's terrible news. You know, I think we get a great deal from our work other than a paycheck and how are those elements of life going to be replaced?

It's a wild time.

How are you feeling about the United States relative to other countries, the rest of the world? There's a lot of uncertainty generally politically. At the same time, America seems to have a lead in the AI race. How are you feeling about just America Broadly.

Well, I don't know enough about the technology to know where we stand in the race or what it's going to take to win the race or anything like that. But I think, you know, and there was the big news recently was that anthropic is going to, I don't even know the right terminology, but let's say be less reticent in certain areas to apply technology, because in America and in a lot of developed countries, we have these things called scruples. Oh, no, I'm not going to do that. Oh, yeah, well, yes, you could do that, but I'm not going to do that. This is an arms race, and America has never had a serious rival before. Never had a serious rival. We thought Russia was a serious rival. We were wrong. It was never really a threat other than militarily or atomically. But we've never had an economic rival before. Now we have a rival in many dimensions. And of course it's China. And I think that AI will be at the core of, of that rivalry. And if we slow down because of our scruples and they carry on pell mell, I think that'll help them win. And if they get control of AI, if they have highly superior AI, I think that could be one of the things that makes life tough for us. So. And by the way, I consider myself as having scruples, but I worry about what this implies. And by the way, if you want to think about scruples, think about this. One of the areas in which we have a problem which is not highly talked about or acknowledged is rare earths. And rare earths are ubiquitous in many areas of technology, and we're dependent on China for that. So here's a question for you two young guys. Before China developed its primacy in rare earths, who had primacy?

Wasn't it America?

I think we did. And we decided it was dirty and it was worth shipping overseas before.

Bingo. You get 100 points for that answer.

Let's go. Thank you.

It was us. And as I understand it, as I understand it, most of the rare earths came out of one mine in California. And my guess is some ecologists concluded that, as you say, that it was dirty. And so maybe we shouldn't do it. And then guess what? And guess what? China said, we'll do it, we'll do it, and we'll do it cheaper.

Oh, yeah.

And so they got all the business and that's how you created dependency.

Yeah.

And there's a big question about how does the dirt over There just come right over here, like potentially, you know, I think it's possible it's all one atmosphere.

But look at what happened with energy in Europe.

Yeah. Oh yeah.

Germany had a bunch of nuclear reactors.

Crazy.

And then, and then they said, well, we don't really like having nuclear reactors because of its non ecological. And Russia said, oh, we'll do it for you. They'll supply your energy and you develop a dependence. So the point is, what you had in both cases is you had decisions made on purely economic terms in areas that probably should have included some strategic decision making, geopolitical strategy, but it was all ceded to the economists and the ecologists. Yeah, hard decisions. These are hard decisions.

They are.

Can I completely switch gears and ask you sort of a personal finance question? I've heard this rule of thumb that the allocation between stocks and bonds should be tied to your age. If you're 50 years old, you should be 50, 50. If you're 25 years old, you should be 25% in bonds. But recently the bond market and the stock markets have been more correlated than de correlated. And I'm wondering how you think the average person should think about the equity markets versus the debt markets.

First of all, there are no numbers that hold true for everybody. So any rules of thumb that have numbers in them you have to throw out.

The only rule of thumb is no rules of thumb.

Who was it? Somebody once said that every generalization is flawed, including this one. The point is that stocks and bonds have different qualities, existential qualities, and people should understand what the difference is. The difference is not merely that one is called stocks and one is called bonds. You have to understand the difference. And then you have to figure out for yourself what the right, I would say, risk posture is. And each person, and each company and each insurance company, each sovereign wealth fund, each investor should figure out their normal appropriate risk posture based on age, wealth, income, sufficiency of wealth and income aspiration, number of dependents and intestinal fortitude, the ability to live with fluctuations. And it's different for everybody. So that rule of thumb generalization hints at a direction, which is to say maybe younger people whose lives lie ahead should take on more uncertain paths which have a higher trajectory, but more uncertainty. And older people who are approaching retirement should have a less uncertain path and more dependability. Makes perfect sense. Using something like your age as the answer is silly, but I think everybody has to think about those things. And it's not easy to come up with the answers, but you better do it because it's damn Important.

Yeah.

Last question for me, and sorry to jump around a little bit, but I'm curious. Were you ever deeply pessimistic about the Internet, about what the Internet's impact would be on society in the way that you are around AI and potential labor displacement? Because certainly they were.

That's a great question, because you could

have pessimism about, hey, we got a lot of fiber in the ground that's not actually being used, that we're headed off a cliff here. That's kind of like maybe more like economic pessimism, but societal pessimism. I'm curious.

You know, I think the honest answer is that I never knew enough about the Internet to reach that point of pessimism.

Because you didn't have the Internet. You didn't have LLMs.

I didn't have my son kicking me in the butt, making me do the research. I mean, he has really pushed me. He said, dad, you have to know this stuff. And he does it all day. I think I know much more about AI than I ever did about the Internet. The answer is I never did reach that level of pessimism with regard to the Internet.

What's the fundraising cycle like for Oaktree Capital? How often do you raise funds? What's the latest fund?

Well, that's an interesting question. I mean, we have a lot of different kinds of funds. We have some which are very plain vanilla and produce a steady return fairly dependably higher. We don't do anything in what's called high grade bonds or investment grade. Everything we do is non investment grade. So we don't do gilt edge. So nothing that we do is 100% dependable. We have some that are mundane and moderate return, and then we have some that are highly aspirational and often tied to the cycle. I think it's fair to say we're the world's leading investor in distressed debt. And so sometimes when there's a lot of distress, there's a lot for us to do. We have the ability to invest a lot of money and historically have high returns. When everything is placid, there's not much to do in distressed land, so we have to kind of go into remission, and we don't raise much money in those funds. And we raise small funds with which we can do very selective things and make do with a modest supply. So a lot of our fundraising is tied to that. The other thing is we're worriers. When in good times, most people plunge ahead, we tend to pull back because we get worried, because they're too damn optimistic. Buffett says the less prudence with which others conduct their affairs and the greater the prudence with which you must conduct your own affairs. Or he says we, we must conduct. And I believe that. So when everybody is unafraid, I get terrified because they do nutty things that put us all in jeopardy. The scariest thing in the world, the riskiest thing in the world is the belief there's no risk. And when people believe there's no risk, the world gets crazy. There's an old saying in the banking business, which is where I started my career, that the worst of loans are made in the best of times for this reason. So that really defines our cycle. When everybody else is having a ball, making money hand over fist, doesn't see anything to worry about, we kind of go into a cocoon because to us that's scary and it hints at very few opportunities. When everybody else is terrified and wouldn't touch risk with a 10 foot pole and as a consequence you get highly paid for taking risk. That's when we turn aggressive, which I

don't know how much. If you read or talked with Andrew about Citrini's Sunday memo, the 2028 global intelligence crisis, which Citadel responded to by basically saying we have an intelligence crisis right now with everyone's reaction to this piece. But did you ever have a memo that created really wild near term price action or in the way that the Citrini piece did, or with the way that information moved historically did that kind of.

No. Look, he used a device of. I mean he even said this is not a prediction. This is an extreme case to illustrate. I've never felt that it was attractive to do that. I try to live in the middle of the probability distribution and understand what's going on in the middle of the probability distribution rather than. Again. And I think maybe you would say that Citrini's case was a black swan.

Yeah.

And I don't traffic with black swans.

Yeah. It's interesting. It kind of needed to come from a substack that gives the kind of appearance of being. If you're just reading it for the first time, it gives these kind of appearance of sophistication but doesn't come with having run tens of billions of dollars of assets.

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

Yeah, I really enjoyed it.

Well, as I said, I think you guys doing a great job and it's really a pleasure to be part of it.

Yes, we've loved having you. Hopefully we can do it again soon.

Yeah. Tell Andrew he's welcome.

Let's plug the book. Go buy the book. And it's available where all books are sold. I'm sure that there's an audio version as well.

Exactly.

Thank you so much.

Cheers.

Bye. Let me tell you about Vanta Automate Compliance and Security. Vanta is the leading AI trust management platform. And let me tell you about Labelbox, RL environments, voice robotics, evals and expert human data. Labelbox is the data factory behind the world world's leading AI teams. And without further ado, we will kick off our Lambda Lightning Round. I'm getting it.

Ben.

We got new graphics. The Lightning Round has been.

We did it.

We have Demi Guo from Pika. She's the co founder and CEO. Welcome back to the show. How are you doing? Thank you so much for coming back on the show.

We're fired up.

Tell us about AI selves. Tell us about the latest launch. What did you launch today? Yeah, for sure. So we all have the same problem,

you know, there's only one of us. Your mom needs, you know, help with your computer.

Your.

Your partner sometime wants, you know, more

time, especially for a founder.

Your friends want to hang out.

You know, you're.

For a founder, you're so busy that maybe you're missing out, like on a

group chat, group trip, planning, but you're in the meetings or on slides or asleep. So then we raise this question, like,

what if there's a second new. We build something new, which is called

AI self, where you can give birth to it, you can raise it and it will grow with you, evolve, and

it becomes a living extension of you.

This is what you and John Ballmer were just talking about two hours earlier in the show. You're like, John and Geordi are going to talk to each other and get stuff done. Is this purely for entertainment or do you actually see a business sort of use case where someone with a great marketing mind will create an AI self and then I would be able to go to them and bounce a Super bowl ad campaign off of them. How do you see the user actually interacting with this? Yeah, for sure. I think there are many ways you can do it. First, like it definitely it's you, but

it's always like you mentioned, it's like infinite availability. Right?

So 24, 7 available always on can be everywhere.

So you know, your family, mom, people

who love you will be able to

have more access to you.

So your mom will be able to like talk to your AI self to help her to debug it.

Problems.

But like you mentioned, it's same for,

you know, people who are have great expertise, right?

So like if you're a really genius

marketing person, maybe like other people can

talk to her to gain more mechanism

for marketing knowledge and she can monetize off it.

And it's because it's AI so it

can do all those, it breaks all the human limits. So it can be very capable.

For example, like I mentioned, it can have your marketing taste, but it also

can do everything like 10 times faster.

How do you get enough data for an individual to make it an accurate representation of the individual? We've talked about this because we put out three hours a day of content of ourselves talking, hanging out, but most people don't have, we're going 247 with this.

Let the AI selves host the rest of the other 21 hours a day. Yeah, for sure. The whole biggest you guys can use your AI self to directly give. But to answer a question about how

do we get data to have accurate representation of you, I would say there are three angles.

First is it could import more data of you. For example, import all the CBM you

show in the past so you can learn from that.

But also when you're using AI self, there's this constant process, you're correcting and guiding it.

Right?

So when you're using her for him

to generate marketing assets for you, he

will also learn the marketing taste you have.

So when you're using her for short

term productive word, he will also learn the taste, basically. And the last angle is that the identity also really matters here because for

example, your mom might not care how

it's 100% accurate about. Everything is accurate about you.

But as long as your AI, she

will be interested because otherwise she could not find you. Right.

Well, thank you so much for coming on. Unfortunately, you have some breaking news, so I need to cut this interview short. But we'd love to have you back on the show soon. So have a great rest of the day and congratulations on the launch. We'll talk to you soon.

Cheers.

Goodbye. And the breaking news is that Warner Brothers says Paramount's new offer is superior. Netflix now has four days to respond. The Kalshi market is continuing to diverge. When we started tracking this, Netflix was up at 50%, Paramount at 40%. Now Paramount is starting to run away with it. They're at 62% and Netflix is down at 33. Will they sweeten their offer? We don't know, but Netflix has four days to respond.

More breaking news.

More breaking news. What else?

Square is cutting from 10,000 to 6,000 employees. A 40% reduction. Let's head over to Jack. He says we're making blocks smaller today. Here's my note to the company. Today we're making. And I wanted to ask Howard about this because it feels like so much of. I'm curious to see what Jack says around the reasoning, but this. I've been shocked that more CEOs when their stocks are down and beat up aren't looking at what happened with X and saying there's a huge expense that we have here, payroll. And anyway, so he says today we're making one of the hardest decisions in the history of our company. We're reducing our organization by nearly half from over 10,000 people to just under 6,000. That means 4,000 of you are being asked to leave or entering into consultation. I'll be straight about what's happening. First off, if you're one of the people affected, you'll receive your salary for 20 weeks plus one week per year of 10 year equity vested through the end of May. Six months of health care, corporate devices and 5,000 to put toward whatever you need to help in this transition. That is generous. We're not making this decision because we're in trouble. Our business is strong, gross profit continues to grow, we continue to serve more customers and profitability is improving. But something has changed. We already seeing that intelligence tools we're creating and using paired with smaller and flatter teams are enabling a new way of working which fundamentally changes what it means to build and run a company. And that's accelerating rapidly. I had two options cut gradually over months or years as this shift plays out, or be honest about where we are and act on it. Now I choose the latter. Repeated rounds of cuts are destructive to morale, to focus and to the trust that customers and shareholders place in our ability to lead. And I won't read the whole thing, but this feels a little bit more real.

Like what Citrini was sort of predicting a little bit.

Yeah, yeah. And it feels more real than some of these other cuts where they do like a 8, 8% riff and then say, oh, we're getting, but going down by nearly half also.

I mean Block has been mostly spared the Sasspocalypse. I mean over the last one month the stock's down 17%. Over the last six months is down 30%. It's not down 50, 60.

Yeah, but still it's trading at somewhere

around it's way off peak. The peak stock price was $263. Now it's at 54. And so there's certainly a question about how they build back. Well, let me tell you about Gusto, the unified platform for payroll, benefits and HR built to evolve with modern small and medium businesses. And without further ado, let's continue our lightning round and bring in Yash Patel from Applied Compute. How are you doing?

What's going on? Good.

How are you?

Good to see you again.

Good to see you.

Since this is the first time on the show, long overdue, please introduce yourself and the company.

Yeah, yeah. So my name is Yash. I'm the CEO of Applied Compute. And what we do is we build what we call specific intelligence for enterprise. And what that means is. Right, like AI is sort of, you know, going at breakneck speeds. These general models are getting better and better week over week. But you know, if you're using the general thing, you're kind of never having your competitive edge. So what we think is there's is a ton of latent knowledge or subject matter expertise that's kind of in an enterprise. And what we want to do is help enterprises capture that, imbue that into their agentic workforce and you know, start to scale up their, their agentic coworkers.

What does an actual onboarding process look like for an enterprise? Is it just sort of turning loose?

Before you answer, I just got to say I love the. I love specific intelligence. I don't. I've got enough general intelligence. I really like some specific.

I think a lot of people are feeling that they're like, everything's at 99%, but I want 99.999%.

Exactly. Yeah, yeah. I mean, like, you know, there is that MIT State of Gen AI paper that came out a while ago and you know, the thing it highlighted. You know, the reason 95% of these AI pilots fail is because these things don't really do the last mile. So they don't adapt to feedback. They don't get better the more you use them. Kind of like an employee would. And you know, that's really what you need in enterprise in order to actually have like a productive employee. You know, you can do kind of simple automations. We've seen a lot of like workflow building and that's great. So I kind of think about this as like RPA plus, right? You had, you know, used to have like sort of like your click and drag rpa. Now you're introducing models into it. But to do sort of like real cognitive work that like knowledge workers are doing, you kind of need to Go a step above. A lot of that is context. But a lot of what we do is fundamental research on the model level too, because we think in sort of building this next agentic employee is going to require sort of touching the entire stack.

And what does your stack and supply chain look like? Are you a beneficiary of open source? Are you partnering with Big Labs? You obviously have a lot of experience with Big Labs, but what does actually deploying one of your products look like?

Yeah, so we're a platform. So we deploy a platform inside of an enterprise. We really want to sell the entire stack that's being able to plug into all of your system record and data because we think context is there. It's just really fragmented across all these different applications and even people. A lot of stuff is in people's heads. This tacit knowledge. We plug into all of your data. We have our proprietary RL post training stack where we can train these reasoning models directly on top of your data. All the infrastructure around models, which makes them agents. Right. So there's the model, but an agent is really like, where's that model running? What tools does it have? Access to all the permissioning and authentication. And then above that is like the application layers. So how are humans interacting with these agents? How are they sort of guiding it, instructing it on what to do? And then the observability around of all of this and what we think our value really is is closing the loop. So everyone's been talking about continual learning. I think that's entirely how this space is going to go. Right. We have offline evals today. That's because that's kind of the best thing we have to benchmark, really. These models are getting so damn good that you're going to be evaluating them based off the real work that they're doing in the enterprise and while you're actually tasking them with stuff. So we basically help capture all of that information, turn it back into context and data that we can use to continually train these things.

I don't know how much you can share or how much is this is a secret sauce that maybe you'll talk about on a podcast three years from now when you've already won. But how do you get all the context that lives in people's head into your system so that it can be used across the organization is like, I imagine you've tried a bunch of different attempts and ways.

Yeah, yeah. So I think there's like a couple of standard sort of recipes that you can use to start these things and I actually want to be super clear. I think RL is sort of one of the best ways to train these models today. But it's not going to be the only thing and it's going to constantly evolve. It's honestly quite nascent and it's stage right. You know, I think Karpathi and a lot of these other folks have talked about like how simplistic, overly simplistic it is. But you know, a good example, Jordy, might be embodied work. Right. So humans go and spend a ton of time going and creating artifacts that they spend a lot of reasoning effort on. And what you can actually do is you can look at these artifacts, these, these final outputs and sort of say, hey, this is what good looks like, and then optimize models, sort of train models to produce things that look like that. So you know, I think a recent example of this that we actually, you know, we put out some, some collaborative research with Mercur like 2 days ago, sort of hill climbing on this, this new agentic benchmark they call Apex. So sort of a professional services benchmark across law, investment banking, consulting, management. And yeah, we're, we're sort of number one on the corporate law subdomain there I think got bumped down to. Yeah and like number four or five overall. So yeah, how much you can do with these small amount of data,

where is applied compute, like where what sectors are your customers like having their mind blown in the way that software engineers have generally with Cogen tools over the last year.

Yeah, it's a great question. So we're really targeting, you know, institutions where there's a lot of sort of built up knowledge and context over, over decades. Right. So this is like financial services, insurance, health care, bio, places where data really, really matters. I think you know, even, even the coding domain. Right. It's, you know, there's so much, there's such a high ceiling there in terms of the types of, of models you can go and train. So we've been doing some work with cognition, helping train some custom models there. But yeah, I think places where there's a lot of institutional knowledge, that's where this sort of imbuing it into these models shines the most.

Well, thank you so much for coming on the show and bringing it.

I think we got to hit the gong.

Oh yeah. I mean the fundraising, it was a little bit ago, but how, how much did you raise? We want to hit the gong.

Yeah.

So we raised 80 million last year.

Better late than never. Great, great to finally have you on the show. You're welcome anytime.

Yeah, we'll talk to you soon.

We can hang again soon.

Have a great rest of your day. We'll talk to you soon. Let me tell you about CrowdStrike. Your business is AI. Their business is securing it. CrowdStrike secures AI and stops breaches. And let me also tell you about Railway. Railway is the all in one intelligent cloud provider. Use your favorite agent to deploy web apps, servers, databases and more. While Railway automatically takes care of scaling, monitoring and security.

Get in the ultra dome. He's not here yet, Morton.

We're running a little bit behind. We'll check on him and in the meantime we will tell you about. Scott is here. We have Scott Morton from Revel. He's the founder and CEO. Welcome to the show.

What's happening?

Hello.

Great to be here.

How are you doing? Massive day. Take us through the fundraising news because I just hit the gong and I want to hit it again.

Yeah. We've raised 150 million for our series B led by Index Ventures, major patient

from Redpoint as well.

Amazing. So since it's your first time on the show, let's break down the product. What are you building?

Yeah, so we are solving a basic problem in that the software used to control and test hardware systems really has not improved since the 80s and 90s. So you have these companies building the most complex cutting edge systems like hypersonic jets and satellites and they're using control software that was designed by pre Windows 98. So yeah, it's insane. And so we provide a state of the art software platform that really makes the control software and test software no longer a bottleneck, but instead an accelerant to their development. Do you have any experience making hardware or did you just think this was a cool idea? Yeah, no. So I spent nine and a half years at SpaceX. Yeah.

Worked on a lot of these systems.

Were a lot of control software for Falcon 9 and then later Starship vehicle launch site, SpaceX. Really it was like a proving ground for software for hardware systems and also kind of the highest stakes environment.

Maybe go a little bit deeper in terms of software for hardware systems. When I think about a rocket taking off, I could think about a gravity simulation, like a full 3D environment, something on Unreal Engine. Then I could also just think about a whole bunch of business logic, basically testing different ratios and applying the laws of physics at just a mathematical level. What's the shape of the product? What are the most common problems that are solved?

Yeah, this is the software that will control the hardware system directly Engineers, typically, they need to offer the control software and overall it's like if they can only test, let's say you're building a pump or something. If you're only able to test that five times in a week, imagine if

you could then test it 50 times

in a week and how much of a better pump you're going to then deliver when you need to ship that

product with a pump. I could imagine like, what is it? Cfd Fluid dynamics? Like, are you simulating individual fluid molecules flowing through a system or are you acting on a higher level of abstraction? Like, what's in demand right now in terms of simulation and control software?

Yeah, so this is more. So like you're designing a pump from the very beginning and you first kind of design the system, you then fabricate an initial version of it.

Then you need to set up kind

of an environment for how you're going to make sure that what you've built is actually, you know, is actually going to work. It's not going to break, let's say.

And so where we said this is

the software that will both control the pump, but then also all the systems around it that are going to kind of make that simulated environment.

Yeah. What do you think of the term digital twin? Is that overused in. Hilarious.

Yeah, I know, it's used everywhere. I mean, simulation is a big component of all of this, but more and more so in the rapid iteration and testing these systems. So for instance, we're running some deep partnership with Impulse Space. If you know them, they have a rocket engine test site out in the Mojave Desert. Our software will run that whole system. It's a large facility and we enable the engineers there to write all their control software, make their command and control interfaces, and then also kind of execute those together as they're trying to figure out what they want to see on that system. Where are you getting traction? $150 million Series B tells me that you're not just selling into pre seed companies in El Segundo. Are you getting into some of these larger legacy manufacturers yet or is that part of this round? Yeah, no, we've been. We've done really well with kind of scaling startups so far. Both doing the test software, but also control software, working with Radiant Nuclear providing their command and control system. But yeah, we are engaged with some larger companies. Doug is here. Doug is here physically in the old studio.

He's about to come on the show.

He's coming on in 10 minutes.

He's like, I have a bug report. I Have a bug report

static.

That's amazing.

Yeah, but yeah, so doing very well with the kind of scaling startups but we're now starting to engage with both larger companies in aerospace but then also in more of heavy industry mining, more of industrial control, oil and gas type applications.

That's cool.

Is a good place to just kind of vibe code. Not even look at the code. You just kind of, you know, tell it rips or is part of this is like if there's software errors on your side there can be very physical expensive consequences in the real world. Yeah, no, I mean it's definitely the latter. I don't think a lot of people are really vibe coding this type of software at this point and we do have some plans for that to make that enable you to vibe code potentially for control software. We do actually have our own programming language believe it or not, which maybe sounds pretty wild, but there really isn't a good programming language specifically designed for controlling hardware systems. And part of that is actually designing out a lot of the common mistakes that are made. So this language, if it compiles successfully, it actually cannot crash when it is run.

As an example, you know Goldman Sachs has their own programming language.

I did not.

I think it's called like slang securities language. Right. I think Bloomberg.

I've heard of that.

Yeah, yeah, yeah. There's a number. Jane Street, OCaml. It's not there. So it's open source but respect programming language respecter here.

That's right. Anyway, so great to finally have you on Scott. We still got to hang one of these days.

Maybe we needed a custom programming language for tvpn.

Yeah, Tyler. Tyler can learn his first programming language.

This is what we need. It's more of like a plain tag programmer these days.

Sorry Scott, we're having too much fun. But congratulations to the whole team. I'm sure you'll be back on. We'll talk to you soon, very very soon.

Have a good rest of your day. Goodbye. Let me tell you about Restream 1 livestream. 30 plus destinations. If you want a multi stream go to restream.com and let me tell you about public.com investing for those that take it seriously. They got stocks, options, bonds, crypto, treasury and more with great customer service and without further ado, we will bring in sun from Moon Lake, sun and Moon. You see what I did there? Congratulations on the launch. How are you doing? Thank you so much for joining joining the show and please explain exactly what you built and how much AI is going on because it's a crazy demo. We're hopefully going to be able to pull it up. But we'd love to know about how you're positioning the product, how you're explaining it, and what Moonlake is. Yeah, of course.

So actually, let me take a step back.

Please explain. Define role models first, because there's such

a term that is overloaded.

So I define world models as models

that can predict a model that can extrapolate the next state of the world in action.

Right. And in fact, I'm defining this right now because I'm predicting that you guys are going to ask me about, okay, how is our world model different than other. This capability is actually just like my world model capability.

And the biggest problems with foundation models in AI today is that these models don't have these role model capabilities such that it can't predict and act in Long Horizon tasks.

Yeah, no. We've seen with Genie 3, amazing ability to paint on the wall, come back, paint again, and the paint's still there. But we're so far away from a real game mechanic of you're five hours in and you go back and you get a new quest. Like, we're far away from that. At least it felt like that until this demo.

No, exactly. Like there's different schools of thoughts when it comes to role models. And really it comes down to how

are you representing the world for the tasks that you care about simulating.

So, for example, for genie, they have really pretty pixels. The problem with it is that it

can't remember the world currently more than a minute. But it's an absolutely incredible technology. The way we're approaching it is to

say we're going to use logic and

symbolic representations such as code to encode the interactivity and deterministic part of the world.

And then we're going to use pixel

priors to then render the world.

But the pixel prior is only responsible for the appearance. Games or real world physics, there's a lot of deterministic art that is better represented through code and symbolic representations. Yeah. So when I watched the demo, was very impressed with it. Felt like the first sort of turning a coding agent loose on the Unreal Engine platform. Are you using a game engine under the hood? And then how much of a harness did you have to build to actually get this result? Because the little interactions around, there's music playing and you can bowl and you actually have a game within a game. You can go up to an arcade game and play Space Invaders that it was all built. How much of this is the harness? How Much of this do you get for free by sitting on top of Frontier LLMs and then what are you using in terms of game engine and stuff off the shelf? Yeah, great question.

So we actually are using game engine

but we forced an open source game engine and make a lot of customizations on top of it. Oh probably Godot our model. Exactly. Yeah. Allow our model to essentially leverage a lot of things that off the shelf can't. Awesome.

And then so you can think of

it as like this code generation model

that we post trained to have to

be able to use a variety of tools to build the world. Very cool. So do you want to become a game developer and use your own tool to generate the next super deep world at way lower cost? Do you want to turn this playing

in a game right now?

Yes, we'll get into simulation theory at the next interview but. Or do you want this to be like a consumer tool similar to what we've seen with Suno Midjourney where there are people that go and sort of generate their own images or music and enjoy those for themselves. Maybe they share them, but maybe they just enjoy them for themselves. How do you see this being adopted?

Yeah, frankly our ambition is beyond gaming.

So we really want to solve of

multimodal reasoning like be able to allow

models to really do long horizon planning, both the virtual world and the physical

world, but also commercialization, short term commercialization.

We want to empower basically people to

be able to monetize with their ideas

that is not bounded by their skill. We want to be the enablement layer to shift leverage from skill or domain knowledge to really taste so that anybody

with good ideas can be monetizing their ideas.

Yeah. So how much of a network is important here? Like I'm thinking about the priors of Roblox building games within the game engine and a lot of that comes from the network effect, the multiplayer nature. Do you think that will be important as a growth engine for you? Absolutely.

In fact in our you can say

I want to build an open multiplayer game and then the modeling agent would

automatically configure database and multiplayer setups for

you such that you can actually one

click deploy this experience.

Yes.

Share it with your friend.

But importantly, that would not be that sticky of a network effect for you. And so what I'm thinking is do I wind up with a Moonlake handle at some point that I can take across across the different games that are created through the platform?

Yes.

So you will be able to say

create an ID and then create 10

of your games just Monetize on top of Moonlix platform and we help you

distribute it to the players and help you monetize.

That's amazing. This is very, very cool. Take us through the funding news. How much have you raised so far? We've raised 20 million total. Who's in?

I think that was. I think this is the same round as last time, but it's an honor to hit it again.

Yes. I'll hit the gong for anything that Jeff Dean's ripping checks into. Congratulations. It's a great lineup of investors and thank you so much for taking the time to come on the show.

Yeah. Great to get the update. Really impressive progress.

Congrats. We'll talk to you soon. Have a good rest of your day. Let me tell you about Applovin. Profitable advertising made Easy with Axon AI. Get access to over 1 billion daily active users and grow your business today. And let me also tell you about graphite code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster. And you heard us tease it earlier. But we got some very special guests in the TVPN Ultra dome. We got Adam Draper. How you doing? And Doug Bernauer. How you guys doing? Good to see you. Look at this outfit. Welcome back. What you got a hat for? Fantastic.

Been on once.

Wait, I. I saw a preview of this. I. I saw the nuclear hat earlier on the show. I also got you guys a gift.

Whoa.

Bright orange. Is this a safe.

Do you have a meat stick hanging out?

He's got problems. You know, we're up here for a little while. Is. Is the. Is the orange? Because you're doing construction. Well, you can see. Is orange. Okay.

Orange pants. For about 10 years.

Okay.

For about 10 years.

You gotta. Yeah, you gotta have the vest for when you're doing video podcasts. Because we didn't see your pants last time.

Well, I asked Twitter what I should wear today.

Yeah.

And so I was just walking around la.

Yeah.

And they said the Dumb and Dumber suit.

There we go.

And so I went walking for costume shops and I ran into a great guy who tried to sell me everything in the store.

Fantastic.

And I ended up on this.

Should we open this? What? Can we open this?

Yeah. Yeah.

You got to give them time to collector to crush this charm.

I felt inspired. Here, open the whole thing.

Okay.

Do you guys know Funko Pops?

I am familiar with Funko Pops. What did we get here? What is this?

Look what it says.

Whoa.

Custom Funko Pops.

Crazy. Crazy.

You can just get these made. How does that work?

Dude, you know, it's modern about a gift, far enough in advance.

There you go.

You can do this.

That's crazy. I feel like we got to protect those so nobody kind of messes with them.

Yeah, we should consult with Dylan Eberscott on our team, who knows collectibles very well because we might have to get this PSA certified.

I think we should actually.

Something like that.

That's amazing. That's amazing. Thank you.

Doug. Good to see you again. How are things going?

They're going really great.

I can imagine that demand for energy is in endless supply. What's the latest in your world, though?

Absolutely. Humans like being humans. They like lots of energy.

They do.

And you better get ready for the gong.

Okay. What you got for us?

Oh, we just raised $360 million.

That's a big number.

Kicking off an interview.

Congratulations on saying the biggest number.

360 million.

Just in case anyone wasn't.

Fantastic.

Do you want another gong? Is that. Yeah, I want you to hit the horse, actually. It's on your side.

We got. They should make a horse. Gone.

Okay. So it sounds wrong. So, I mean, you know, I've been to the radiant facility in El Segundo. Does this unlock a new facility? Is there going to be a separate manufacturing line? Just a lot more humans working. Joining the team. Like, what is the money for?

Yeah, it's mainly a scale of production at our factory facility in Tennessee, which we signed for in October.

Congrats.

About 80 acre site right now and already broke ground. Digging on site and buildings going up. It looks awesome. I saw art yesterday.

And so now you have. When we were last talking, you were talking about still doing design work. You were working on the. What was this? The helium loop. I remember you explained some of this to me. Very nicely dumbed it down. But how far are you along in sort of locking the design? Because if you're setting up a manufacturing plant, I imagine that. But you're going to lose some flexibility at some point, right?

Yeah, absolutely. So we're doing a dome, a test in the dome very soon. And we actually, since we talked last, we've gotten approval from the Department of Energy.

Cool.

It's a really boring sounding, complex name. But it's the approval to fuel and

go to full power approval. That's more gong worthy. We should have the gong. There's a lot of people with money. There's only one government that could actually put down a stamp, you know? Yeah, it is. It is incredible progress. That doesn't get enough. Doesn't get enough credit. But anyway, so continue.

Yeah.

So it's called a preliminary design safety approval. It's done. So that means the design is locked. So that thing we were working on, finishing out, that was the big event. And now it's been months of you back and forth. And we're good. We proved the only ones.

So no big change at all. We're moving away from helium, we're moving away from tricep.

Well, this is just.

I should say that's just for the first unit that goes in the dome. And we've got more of a team now working on the production units that we're going to build lots of, and build many at a time. Right up to 50 per year from that factory site.

One a week.

Yeah.

So 50. That's 50 megawatts. Are you still thinking about the one megawatt architecture?

Yep, absolutely.

Replace diesel generator. Yeah. Okay. And then in terms of demand, initially there was a lot of demand from, you know, stranded places where energy is very expensive to get to. You're off the grid, you're on an oil and gas site, or you're in a military context. Has the demand shifted? Have the AI folks showed up?

Oh, yeah, absolutely.

Okay.

I mean, we have an announced deal with Equinix for 20 units. Down payment on them. And, you know, what will end up happening is we're going to make many reactor products.

Right.

Nuclear reactors can be products. That really has never happened before because usually you dig a big hole in the ground, you build a building, you're really. You're making a building.

Yeah, yeah.

Right. And it's not, it's not a product. You have to. For it to be a product, it has to be mass producible.

Yeah.

And so that's really what Radiant is all about. We're going to be making reactors for space, reactors for maybe the bottom of the ocean floor, Reactors that flow, reactors, reactors on a truck, reactors on a plane. Think about it as like, they come out of the building and they move.

Sound effect.

Walk me through some of the economies of Seahorse. If I want in a long time, whenever it's ready, if I want 10 megawatts, is it more efficient to get 101 megawatt reactors or just scale up and create 110 that maybe is bigger, maybe has different manufacturing and transportation requirements. But when do you think about actually scaling up the design versus just selling more of the existing product?

Yeah. So our mandate is really mass produce. And so we don't want to make things that are not. That are immobile. And so if you operate a reactor that's at a larger power scale, it's going to be more efficient. There's actually a neat effect where, like, if you put a little bit more nuclear fuel in a design, that fuel makes the other fuel that you had go a little further.

Sure.

So it's very cool. And it's this like exponential scale. But we know what we're doing. We're totally focused around efficiency that we provide is more like deployability. If you go, I want power next week, we can actually do that. And if you go too big, you can't do it, you lose that. So that's what Kaleidos the product is about.

Got it. Makes sense.

But this round is really exciting. We actually had. Adam led the round and actually has invested in every round.

Every round. It used to be called the most invested investing ist. So he used to say, I've invested the most, but now I get to

say, is this your biggest check ever?

Yeah.

So Boost vc.

Yeah.

Precede for magic and mutants and historic things. And that's how we look at it. We generally are investing $500,000 into all the deals.

Seems like a shift,

but every once in a while, you, you know, you encounter a company and a founder who you just enjoy and care about, and you want to go all in. And so we, we. In October, after they had announced this. The Nashville plot.

Yeah.

My partner and I, Brayton, we got together and we were like, can we do it?

Yeah.

And so we decided to just back up the truck. We went for it. And unfortunately, he is incredible investors. He's an incredible founder. And we. Everyone just. Everyone re upped. We got to lead it. It was fantastic.

Yeah.

Now I get.

Does that change the overall strategy of Boost, or is this sort of like a special case where you went to the LPs and said, There's a specific, specific opportunity that we're targeting.

Yeah.

For the.

So great question. I do think that there are opportunities like this that emerge when you get to know your founders well.

Yeah.

So I've invested in 600 companies over the course of 15 years. I'm. I like that.

Yes.

And.

And, you know, we do follow on. We do do those things. But I just feel, hey, if Doug does this, if the team at Radiant is capable of doing this, this makes historic change in a world I want to live in.

In.

Yeah. Yeah.

And so I ended up getting to care about the. The founder as well as the mission. And so by doing that, I get to. So my. My quick answer. You asked a very Specific question in the situation where that will continue to be true, it would be really fun to do more of these.

Yeah.

However, I wear orange pants.

Yes.

And I love the early stage.

Yeah. So you should, you're still very focused in the early stage. Like we shouldn't just think, oh, Boost is just a major growth equity firm now.

I don't think I could sell.

Yeah. I don't know.

No, I just love the energy. I love being the gateway to be able to help people pursue their dream of whatever the thing is. And I've been able to see people across the entire stack over the last 15 years. And I also now get to, you know, be in the room and watch. Watch nuclear reactors be built.

Yeah.

Are there are LPs and investors that you talk to sort of receptive to this idea that nuclear technology is like sort of getting pulled forward by the AI boom, but sort of uncorrelated with it because like we still need energy for oil and gas exploration and the military and all the these different things. Like it's not so hyper indexed on like make or break. Do we get to ASI or not? Like, this technology is useful in all scenarios.

Is asi the new asi?

We have the goal post, we have our goal post over there called Scythe

that calls that the Thunderhead.

Thunderhead.

And so we can call it that now. I'm going to call it Thunderhead for the rest of the day.

Okay, I like that.

So I mean, the general answer is, I think across all markets, energy only gets consumed more. And that's like such an easy sell for the.

So it's pure execution risk at this point.

Yeah. And so we know the demand is there. We know that there's a huge nuclear wave that the government is behind. And so limited partners and other investors, they want access. Now when, when Doug pitched me seven years ago, like none of those things, the energy was still a real problem.

Overnight success.

Yeah, I got a hot, I got

a hot hand right now.

Oh my God.

The. None of those things, like none of the other things were true. It was, you know, people thought it was a regulatory burden. People thought like the government would never allow you to do this. But you know, we have built nuclear reactors before. We have nuclear reactors. And so when he pitched me on a phone call in 20 minutes, I said yes. And we wrote the largest check we had ever written at that point also. And then we, this last October, we wrote a bigger one.

But how much work do you think needs to be done on the, the public perception side? Because right now, data centers are unpopular and I think the number one reason is not slop in your feed, it's energy prices going up. And if you build more energy, if you build more clean energy, it should be a win win. Everyone should be happy. But I just have this inkling that if you go and you say we're building a data center, we're going to make it look good, you won't be visible, visibly obstructed, and we're building the power plant for it. It's a nuclear power plant. You're still going to get a few protesters, maybe not as many, but they'll still be there. So how do you think public, how do you think about positioning nuclear for consumers, for people that might have this in their neighborhood at some point or in their state or in their country

even, or in their podcast studio?

Ideally, ideally, sign me up. But, but, and then, and then how do you think about.

How do you think.

Actually we actually did look at it. We looked at a studio space, we were about to put an offer down on it and we were like, what's in that door? And they were like, oh, don't worry about it, that's just the machine. And we were like, oh, I'm actually very curious about it and I'm actually pretty worried about it. In that case, they were cleaning the soil from some industrial waste that it was happened like 50 years prior.

So remediation.

Yeah, soil remediation. But anyway, communication about the risks of nuclear, the benefits of nuclear, what do you think are the important points to get across? And then how do you think those diffuse through the populace?

Yeah, it's really important first off to. Right to talk to the public and really explain how to think about nuclear. And I wrote a thing called Adam for Prosperity that lays out some of those points. But we will, we are making it now into a cooler kind of series. We can't go through all of the individual points here. Yeah, but I think you're right about AI. I think that people don't like really a big building going up, like a big monstrosity that's going to bring a bunch of attention and people and traffic and like all the stuff that it brings. And you also have the added complexity of the energy markets typically being very regulated. There's a lot of laws around it and they can just get a higher rate. They have to pay if they get enough votes. And so that's probably more on the gross side of things. The good news is the thing that we're doing is the opposite. Most people Doing nuclear are targeting something big and there's a lot of excitement around nuclear right now. And AI, A lot of people are saying a. Saying they're building a building. We're not doing that. We're doing the doing.

Sure.

So. And what is that? It's just reactors that we build. They're on our site. It's a building far away that you never see as the customer. And it just appears when you want it. It goes away when you want it to. You call us and go pick it up. Yeah, it's kind of beautiful. It's like totally different from other nuclear. So I think that will help quite a lot because it's just such a different thing to be able to go, you know, I can have clean power and we go, yeah. And then go, I can have it next week. Yeah. And they go. And it can send it away anytime. Yep, anytime you want and no waste. Yeah, we take that and we handle that.

Talk about the ramp, because we were looking at the data center that was recently protested in New Brunswick, New Jersey. It was tiny by AI comparisons, 25 megawatts. You could probably supply it in half a year. But Meta's average campus right now is around 500 megawatts. Take you 10 years, years to power that. What is the ramp?

Your next.

What is the ramp up to get from 50 a year to 500 look like for you?

So I don't know.

Right.

It's a. It's not really something we're trying to address.

Yeah.

What I'm most interested in is being able to put reactors in really crazy places.

Yeah.

Like put a reactor on the bottom of the ocean floor. Like.

Right.

Like.

Like I was saying, why would.

Why would you actually want one on the bottom of the ocean floor?

It's the same reason you would go, like, put a research facility in Antarctica, because it's far away and it's weird there and you're going to learn some things and we haven't ever done it before.

Interesting.

And so it's a frontier.

Yeah.

That's kind of all that I'm excited about.

So you think some sort of, you know, remote lab that had camera equipment, lights and sensors and you could understand what's happening at the bottom of the ocean. That's fascinating.

I don't know what you'll learn.

James Cameron should become the spokesperson.

Anybody could. Yeah. If you want, we can make a reactor for that.

Right.

And the thing is, like, if you want, you can't use a combustion power source and there's definitely no sun.

Yeah.

So it's just. Think about that.

That's actually interesting.

We're the only thing you can do in a lot of places. Actually, not just that one. It's just a good example to get people to think like, oh, shoot, okay. Bottom of the ocean floor.

Yeah.

All the way up to anywhere in space. As far from the earth as you want to be, as far from the sun as you want to be. Oh, true. Think about that.

That's solar's weaker on Mars, right? Yeah, yeah.

Although you'd still use it. I like both. I like solar plus nuclear.

Yeah.

I think you're on the right path though, that education is one of the most expensive issues.

Right.

And stories like having a nuclear reactor on the ocean floor that's functioning and we're learning tell a great story. And so those stories become easier to tell as more of them exist. The hard part is the in between. We saw this when we were in crypto also. Everyone was saying like, the banks will never let you do this, the governments will never let you do it. And now obviously it's on the Congress floor.

It's all the rave. So many big, massive. Yeah.

And so like transitions take time and they're expensive, but like it's important because the end consumer gets a better product.

Yeah. In the early stage of I would call like the hard tech boom, the defense tech boom, the El Segundo boom, the dollars that were being raised.

You're talking about 1980s.

Yeah, I'm talking more about like 2022 maybe when people started paying attention in Silicon Valley. But the dollars were small and so they could be sort of flyer checks from VCs. Now we're getting into real numbers. What do you think about the relationship with the SaaS apocalypse? It feels like public markets investors are rotating away from software. They would love some something like industrials energy, something that's clearly going to exist in 100 years, no matter how good the computers get at AI. But then in Silicon Valley, we'll talk to founders all the time where they're raising for a software product. Basically it's just AI indexed and so they're soaking up 200 million of capital every other day. What's happening at the early stage in terms of excitement and at that growth stage to fund projects like. Like this.

Well, as we are growth investors.

Yes, yes.

As officially private equity guys and Series

D. Oh, I actually have a shirt. It says preceding Series D.

You're thinking, you're actually thinking on the longest time, I'm going to make a custom crazy gift.

Technology looks like magic.

Incredible.

My.

Well here I'm sure we both have thoughts on this. So we've been investing in hard tech. We have a way of when everyone really gets excited about something. We like looking the other way. And over Covid we saw everyone getting really excited about software because we were all locked in our houses and inside and like software was where you saw everything.

Anything for remote work is just booming.

And so we made a conscious decision to be like hey, that's not always going to be true. Let's invest in in person businesses that are like physically expensive, that are high capex. And so we obviously invested in Radiant multiple times and a bunch of other Starfish Venus. Just an incredible number of fantastic companies. And so like we were able to. There's a. I think that there's a sense in the investment world where they want to be a part of something that's important but sometimes they're not considering whether or not it's a good investment. And I think that balance is at scale right now. That would be my like and I think we're. They're probably over estimating the SaaS apocalypse. I think so many talented people are building so many incredible things. I was just at the upfront summit which was just over here. Walkable.

Yeah.

So thank you. This is very efficient. And you know who I love to hear speak is Cathie Wood from Ark because she's just so optimistic about unlocking technology assets to everyone and a lot of other people were sort of in the AI negativity bucket and I was like how can you not be optimistic? We just saw, you know, everyone who's building awesome stuff on this show.

Right.

Like incredible AI thing.

Well, I don't know if you saw but Square just did the largest layoff in S&P 500 history on a percentage basis. And so that's kind of the flip side. And even Howard get in touch with

Boost VC and we're going to invest in some Square alumni. I think that it's the.

You got 4,000 of them.

Fantastic.

Get the brink trucks ready.

There are talented people there. We would love to invest in them. And I think what AI is really causing is this eruption of amazing individual entrepreneurs. It's like the next wave. So it's exciting.

Very interesting.

On that note of optimism, there's a very cool video we released just yesterday that I thought you might want to see. And I think you guys have prepped so we can.

Let's do it.

We can have a. Have a look and just so where is shows what we're building.

Right.

So we talked about the approval, the regulatory approval.

Yeah. I remember seeing the core plate. I think you showed me like the raw material for this.

That's hardware in the loop.

That's us doing orbital welds, tubing. That's the mounts for the pressure vessel.

That's the forged pressure vessel piece.

That's crazy.

And that's what we're doing.

Yeah.

Amazing. July 4th. I love it.

So cool.

How awesome.

Amazing.

Just building all that huge physical things. Like we're back to building huge physical things.

So you'll bring that design to the dome.

That's right. All that is the actual nuclear reactor hardware that we're showing. And this is really like the most that we can possibly show to the public. And we now have two buildings. One is just all manufacturing. It's like you go and it's all machine shop. These big fancy glove boxes where we're doing welding with controlled gas conditions. It's so much different from.

Yeah, yeah, yeah, yeah. When I was at the. When I was at the office. Were you sending out for like CNC parts?

Oh, yeah, a lot of stuff. We had to send out a ton

of stuff because I noticed a lot of equipment, but there wasn't that much actual machining material to like make stuff that you would put on. Interesting. So that must have the machines in the actual building. Must speed up iteration as well, right?

Absolutely. And a lot of the incoming capital allows us to make those moves and do those things and even invest a little bit in R and D for things that are really far out, that aren't even a product yet, that are just like, we realize the limitation and we're like, we're going to go push that thing and if it moves, we'll have something no one else has.

Yeah. Last time I was there, I think you had around 40 people. How big?

Oh my God.

150.

Wow. Talk to me about that road. I mean, you were at SpaceX for a long time. Did you ever manage that many people? Is this a new challenge for you?

No. I did a very weird job when I was at SpaceX. I was there 12 years. And the first couple years I loved it is why I stayed 12 years.

Right.

It's the coolest mission. Like Make Life multi planetary. I love that. And I left to Make Life a reactor company to make power for that mission. Still, like, I still want our reactors go that route.

Totally.

Of course, a different type. But I was doing stuff like, you know, I made the first Falcon 9 ground system and then I made the first two rockets. I traveled around the country testing the first two Falcon nines that flew. And then I did the first ever rocket with legs. And that was like report directly to Elon with like three other people. Like the board also go to his desk. Yeah. And go, here's what we're doing. He'd be like. And we got so lucky that like every time we came to him we're pretty much like. And the qualtank passed and it worked and the schedule's good and so he loved it. Right. It was like great. But from there it was like then I was doing boring company and I was doing hyperloop. Is every Elon like side projects, which

is exciting around special projects? Not really full scale up constantly. But now it's higher, higher, higher.

But I would just go and talk to whoever I needed and cross every line possible. And I didn't use the right channels of communication. I just pulled all the assets and I'm building a thing. I'm building a thing and I ignoring everything else and I. That's basically still how I operate.

Yeah, that's great. Hidden candy shop.

Powerful.

Anyway, Jordy, anything else?

No. Guys, congratulations.

Thank you so much for coming.

Thank you for coming here and yeah,

this is great one.

So excited. I, I, I genuinely cannot wait to get my own.

So.

Yeah, I don't know what I'm going to use it for yet but.

Yeah.

How many are you going to need in total?

How many can you plan out?

You're talking about the bottom of the ocean.

5, 20, 30. Yeah, I'm gonna need some for a boat. I just. Yeah, I love it.

Yeah.

An electric nuclear boat.

I mean off grid for the preppers, for the rich preppers, the AI billionaires, they might want it.

Yeah.

Yeah.

Well and if you have some real estate on the moon, it's not very high value without power.

There you go.

You want to get some power for that.

Guys, thank you so much. Amazing cargo yesterday.

Hang out.

Hopefully we can see you and wrap up the show.

That concludes our guests for today. Thank you for watching tvpn.

Let's get more into the block news.

Okay. Yeah, you can pull some up that. The other big news is that NANOBANANA2 launched today combining pro capabilities with lightning fast speed. We of course did. This is the latest image generation model from Google and Tyler gave me a little review. He ran a couple benchmarks. Some pretty impressive results. You said it's slower and is that because it's reasoning more? You think so?

It's unclear. I saw People saying online that it was actually much faster. So I'm curious if it was just like while they were just deploying it.

Sure, sure, sure.

It was on the studio.

Wasn't on the Gemini website when I was trying it.

Yeah.

So that could be. I'm not actually sure if it's.

Yeah. I mean the headline is intelligence and visual quality at flash speed. So I think that they were going for something. Something faster. But yeah, I mean these models take. I think I was using it right

when it was being deployed.

But there's examples.

Big upgrade over nanobananapro.

Yeah. A lay flat infographic depicting the water cycle. Some cartoons. Yeah. It really has always been great at nailing these complex graphics that typically you would absolutely have to do piece by piece and then maybe add the text afterwards because you would get slightly hallucinations or misspellings and it seems to be one shotting these things at this point. So we of course ran Waldo Bench where we try and use a generative image model to create a photorealistic Where's Waldo graphic and you can be the judge. This took you a couple prompts, but we did in fact get a Where's Waldo? That fits the right perspective. I feel like the. The scale of all the characters is correct. And most importantly, there's only one Waldo and he's somewhat hard to find. It's not the hardest Where's Waldo? So I'd give it sort of an 8 or 9 out of 10 on Waldo quantity and placement. But as you zoom in, one of the key features of Where's Waldo is that there are little storylines playing out throughout the image to distract you from finding Waldo. And so that is in fact happening here. You can see some folks building a. Some kids building a sandcastle, some people in line for what looks like ice cream. There's a mariachi band, there are sunbathers, there's a large sandcastle that's gathered a lot of folks the level of quality once you actually zoom in very, very closely on the individual people that are being depicted at a very small scale. This is totally new. Nitpicking, but there are some hallucinations still. It's not quite at the level of visual fidelity of a handcrafted Where's Waldo, but it's getting close. And so congratulations to everyone over on the Nanobanana 2 team who made this possible. It's certainly a major step forward. Although I'm moving the goalposts again because I want my full wares. Waldo just as intriguing as a typical wares Walden, though. Jordy, what else do you have for us on the block news? How is the Internet reacting to the news that Block laid off so many people?

I mean, it's 4,000 people, which is a huge number. And then I was trying, I ran a, ran a deep research report to try to find. Because it's certainly not the largest layoff ever, but it is the largest layoff as a percentage of the overall workforce. And S&P 500.

Wait, how is that possible?

So the second largest layoff In S&P 500 history was Phillips at 22 and a half percent, Chevron at 17 and a half percent. Verizon nearly half 15. Intel did 15% in 2024. GM did 15%. Yes, in 2018. Meta.

So I guess, I guess Twitter, the Twitter, the Twitter restructuring doesn't count because it was delisted before that happened. So Twitter went from 7500 to 1500. And yes, I think you're right, it might not have been in the s and P500 at the time. Yeah, it wasn't interesting. Well, there's some reactions. Will slaughter. Bama Bonds says in three years, from December 2019 to December 2022, Block more than tripled its headcount from 3,900 to 12,500. Unwinding less than half. An insane Covid over hiring binge has much more to do with Jack Dorsey's managerial incompetence than whether AI is going to take your job. And so people are immediately jumping to, is this AI, is this not? The stock's up 25% on the news, so the investors certainly think it's the right move to make the right hard thing to do. The other interesting thing is that Block, if I'm correct, did a fairly large merger with afterpay. And I don't know how much overlap in the employee base there was, but it's not uncommon when a merger of those two sizes. $10 billion, I think Afterpay merger was in the tens of billions. Afterpay lock acquisition, let's see, it was 13.9 billion. It was an all stock deal. It was originally announced at 29 billion. And I'm not sure how many total employees. Let's see if we can pull up this. Afterpay had around 1500 employees. And so there's a few different things this go. Dan Primack says it's stunning in its candor. And if you're one of those spared, how could you not be wondering how long until AI comes for your job too? If CEOs get comfortable, that this is okay, let alone if they believe shareholders will be rewarded for it could set off a stunning layoff wave. Very interested in the AI will create.

Yeah so the the obvious pushback totally over at Solana is making it. He said didn't Jack have 10,000 people running this website when only 75 needed even before AI? So yeah the again it doesn't really, it doesn't really matter whether or not a company is actually getting tremendous efficiency from AI. Every CEO is going to do this and we've said over and over and over I'm surprised more CEOs haven't done the on things. Salesforce has 76,000 people at the company. Right. It's basically a city. And yeah, it's really really sad. Jared Sleeper, who has a fantastic episode on Odd Lots about the Sass apocalypse, he says first of many sad days. As painful as it is. Jax did Focus service by being decisive of course. Really sad day. These employees will be able to quickly go and hit the job market, which probably gives them somewhat of an advantage over future layoffs like this.

If you're not a software engineer, this should be a wake up call to what has happened in the past few months with AI engineering tools. Everything has changed. Now we get to watch as companies learn how to either use these tools in earnest or slow slowly fall behind, says Dustin Curtis. I'm very interested to learn the shape of the layoffs. Do they lean more software focused or less software focused? This is certainly just one company. One example, but interesting to see how this matches with the Citadel rebuttal of the Citrini piece. A lot of people are seeing this as like vindication of the Citrini piece piece. Anyway, we will cover it more.

Hey Kim says every media person please read this before writing your narratives tomorrow. Jack Dorsey is Jack Dorsey. Do not extrapolate. Do not pass go Responding to Will Will Slaughter's post that I just post. Yeah yeah. Square for context did around 24 billion in total revenue in 2020 25. They've been trading at around 30 even though the business is profitable and growing and so had certainly been beat up but very very sad day. Gavin Baker had a post that Kyle Harrison highlighted. He said the fact that Twitter is running well with headcount down significantly really matters. Whether they admit it or not at all. Everyone in SV admires Elon. A lot of venture funded CEOs are sending emails like this inspired by Elon and taking drastic actions. Margins are going up. Balaji says this is the first AI cut and it will send shockwaves Remember, Jack is one of the greatest founders of all time. He created this platform that we're all on, has been early to many technological shifts and Block is doing very well as a business. So for him to cut 40% headcount in this way is a signal to everyone in tech. Get good now, Become indispensable. Work nights and weekends, learn the AI tools and raise your game or you might not make the cut as an employee or as a company. I know that sucks. But capitalism is natural selection. The market is unforgiving because you are the market. After all, it's not like you're buying some random gallon of milk from the store. You're always buying the best product at the best price. So too for apps, your customers are always installing the best piece of code they can get. And because AI is going to create new winners, if you aren't the best in your market, someone may become better with AI. To be clear, Block severance is generous by any measure. 20 weeks of pay, six months of health insurance and invested equity. All of that goes far beyond any typical package. Jack did his level best to cushion the disruption. The laid off are a temporary unfortunate class as opposed to a permanent underclass. But had he not leaned into the AI transition, he might have had to lay off more people slowly and over time as faster competitors went after his market share. How would they do that? Sure, AI is not a panacea by any means, but the closer you are to software engineering, the more aggressively you need to embrace agentic workflows. The AI companies are already doing that in places like Stripe, Shopify, Coinbase, and now Block are pushing hard on this area. There will be overcorrection, but the fundamental technical innovation is real and you need to either disrupt yourself or get disrupted. I think that's generally, generally good advice.

Anything else in the timeline you want to review? I'm sure we will be covering based

16z says every tech company can fire at least half their people, but most can probably do 80% black pill.

Thanks.

Yeah, and to round it off, we'll leave you with a white pill or something like a white monster, which is that Joe Wiesenthal is sharing the chart for the Monster Beverage Corporation, which seemingly is entirely AI proof.

Yeah, Monster's been ripping fascinating company history. Founded by a very.

How much do you think Citrini paid square to do this license validate?

How deep does it go after being

mocked for the last four days? Yeah, Fedspeak says on one hand Jack is a notoriously bad business operator and on the other hand, it's over. Yeah. The pushback on Balaji saying Jack is one of the greatest founders ever is simply that yes, he is one of the greatest founders ever. And I don't think anyone would say that he is the best operator. And that's okay. Right? But sometimes being the best founder means making the super hard decision and having the foresight to get ahead.

It would be a very weird situation if only Jack Dorsey founded Companies stand in 80%.

Same dude bought Jay Z's title. Not the greatest operator.

Yeah, take him's white pellet. He's always white pilling. I enjoy him. Anyway, thank you for watching the show today. Leave us five stars on Apple podcasts and Spotify. Sign up for our newsletter@tvpn.com we will be live tomorrow at 11am no we won't. No we won't. We're off tomorrow.

We're off tomorrow.

We are heading to Montana and so we have some affairs to attend to and we will not be live tomorrow but we will return on Monday and I can't wait. Mark your calendars. 11:00am Pacific on Monday.

We love you.

Goodbye.

Goodbye.

Nice work, brothers. I'll see you on the next one.