Ellison's Media Empire, Ken Burns Joins, Cursor Mic Drop | Matthew Belloni, Gokul Rajaram, Nik Seetharaman, Raj Rajamani, James Everingham, Dr. Felix Ejeckam

You're watching TBN.

Today is Tuesday, March 3rd, 2026. We are live from the TBN OTRium, the Temple of Technology, the Fortress of Finance, the Capital of Capital. Let me tell you about Ramp.com. Time is money. Save both. Easy to use corporate cards, bill pay, accounting, and a whole lot more all in one place. And you know who's going to need a Ramp card? David Ellison, because he owns Warner Brothers now. So David, if you're listening, just, just do it. Just onboard to Ramp. We have a fantastic show for you today, folks. Let's pull up the linear lineup because we got It's Hollywood Day basically on TVPN. We got— that's right— Matthew Belloni from Puck. He's the founding partner coming on to help us understand David Ellison's mission with Warner Bros. Discovery, Paramount, HBO, CNN. He's got CNBC, CBS. He's got everything now. Then Ken Burns is coming on to talk about his whole career in filmmaking. And of course, we have a fantastic film.

Film.

Yes.

For anyone listening that might be like myself.

Quickly, Linear is the system for modern software development. 70% of enterprise workspaces on Linear are using agents. Film is the process of capturing images digitally or using celluloid. Do you know about celluloid?

No.

This is right, right?

Right?

The filmmakers in the back know. So you used to need a specific chemical process. You'd probably be very hands-off on this process, but you used to create a strip of film, the thin, clear, and you put some chemicals on it, and then when light hits it, it— what does it do? It makes— if a lot of light hits it, it makes it black, and then that's the negative, and they flip that around by shining light through it to then expose the final picture, and then you get a beautiful movie that you get to watch at the cinema. Interesting to know how Netflix will change their tune on theaters. So Ted Sarandos has been basically trash talking movie theaters for a long time, being like Netflix is the future, tech is the future. He hasn't been super rude or anything, but he's been like, it's not really key to our strategy. Then as soon as he was in the deal for Warner Brothers, he was like, yeah, the theater's amazing. It's not going anywhere. Like, we're excited to invest. And we will talk to Matt Bellini about how long that lasts until he starts talking trash again. But he's already—

Netflix is up 22%.

Oh, yeah. The stock— the shareholders hated the idea. They hated the idea.

Loves.

Yeah, market loves.

Getting your enemy to overpay for a handful— a basket of legacy assets.

Yeah. There is another thesis on why Netflix stock has mooned. And it's mostly around this concept that Ted Sarandos just stunned in a pair of jeans at the SAG Awards. So look at this. I mean, you see this. How can you not want to buy the stock? This is not financial advice, but you see a guy pull up in a fit like that, you're like, this is going to moon. I got to get in. I'm going turbo long. I'm top blasting. You're buying the dip.

Top blasting.

You're top blasting, and you're buying the dip when you see a man pull up in $2.8 billion jeans, because that's the amount of money that Netflix got wired because of this deal. That was the breakup fee. For some reason, I had it in my head that the breakup fee would be the other way.

Not too shabby.

In fact, Netflix got paid, and Matthew Bellamy got $1,000. 1,000 likes on this post. So my question is, there's a ton of debt going into this deal. Will I be able to watch The Sopranos on Netflix? Will I be able to watch The Dark Knight on Netflix? It's a complicated question because content licensing deals are not— they're not one size fits all. They're subject to windowing and certain markets. So if Netflix has really high penetration in Germany and HBO doesn't, it might make sense to license it there, but not elsewhere. But I just want to know broadly, what do we expect from Paramount, Skydance, Warner Brothers, Discovery, CBS, CNN? TNT. I think they have Shark Week now, right? That's in Discovery. They got the Food Network, HGTV. You can watch the Property Brothers, maybe on Netflix one day. Who knows? But this is like the most extreme scenario. And we talked about a post where someone was like, oh, like masterful 3D chess. Netflix got Paramount to overpay for this. And now they're going to be so indebted because the deal got so big that they're just going to have to come to us and license 100% of the catalog on day one. I don't know if that's true, so I wanted to dig into it. Let's talk about it. So just to recap and set the table, David Ellison, who is the chairman and the CEO of Paramount Skydance, agreed to raise his offer for Warner Bros. Discovery to $31 a share. A lot of people, myself somewhat included, thought that Netflix was at least going to counter a little bit, but they folded immediately. And so the co-CEOs, Ted Sarandos and Greg Peters, they didn't counter and they said, you got it, wire us the $2.8B. And they did. So the result is that David Ellison is handing over $111 billion in exchange for Warner Brothers, HBO Max, CNN, and the other cable networks. There was this whole political angle in DC about how interested would Trump be in the deal. And apparently as soon as he found out that Netflix was not interested in the cable TV assets, he became less interested in the deal. 'Cause I think he watches a lot of news, but he probably doesn't watch a lot of Batman reboots and doesn't really care what happens there. And so, it's just like a different thing. It's like, he's not turning on the latest DC Cinematic Universe film and be like, "They're taking shots at me." But if he turns on CNN and they're taking shots at him, he's like, "I care about that," right? So in terms of the financial situation, they're levering up. They took our advice when we started the show. We told everyone rule number 1 is lever up, and they did. So Paramount is already levered. $10 billion in net debt. Yeah, to the gills. $10 billion of net debt. It's like $13 billion, but they have $3 billion in cash. So $10 billion in net debt, $3 billion of adjusted EBITDA with which to service that debt. Reasonable, but still pretty high leverage, 3x. Paramount will be adding $60 billion of debt. So all combined, the company will have something like $70 billion of net debt, $79 billion of total debt with roughly $12 billion of adjusted EBITDA. And so when you're operating north of 6x leverage, that leads to different decision-making, higher discount rates, higher interest rates. It can shift the focus to near-term cash. So Netflix is a logical counterparty here. Because it can pay, it has the cash, it has a lot of cash flow to pay for global rights at scale and has a long history of paying top dollar to deploy capital to known franchises. And also Netflix is in a particularly interesting financial position. So they've been an incredible case study for operating leverage, a true overnight success, something like 20 years grinding up the subscribers, grinding up the revenue, investing more and more in content. And if we can pull up the chart, either scrolling down on the tbpn.com newsletter or going to the third post in the timeline, you will see the small bar, the red bars are the revenue, and the little dots that have the numbers on them, which you can't really see, show how much they're investing in content. And so for a long time, for something like almost from 2002 to 2018, so 16 years, Netflix was like, we made $100 million, let's make $100 million worth of content, let's buy $100 million worth of content. And then they were like, wait, we made $1 billion, you're never gonna guess what we're gonna spend on content, exactly $1 billion. And then they were like, okay, we made $10 billion, This is going to shock you, but we're spending $10 billion on content. And so they would just spend exactly what they made on content. But that changed. That changed in 2019, 2020. COVID, the amount of, the amount of subscribers spiked, the amount of revenue spiked. Then they launched the ad model. Took them a while to figure that out, but eventually worked. And the business kept growing. The top line kept growing, but they ran out of stuff to buy. They ran out of stuff to make. And even though they were paying, you know, top dollar for all these different assets. They sort of held their content budget flat. Their revenue went up. And that's, of course, operating leverage. So profits, profits to pay for Warner Bros. Discovery assets potentially. And so they have the money to spend, but also their direct competitors in streaming with Paramount Plus and HBO Max. And so there's a question about how friendly will they be, you know, if you license all the good HBO stuff, people will unsubscribe from HBO.

Yeah, or Paramount.

Or Paramount.

Yeah, exactly. It just makes the Paramount subscription offering less competitive.

Totally, totally. So it's this delicate balance. There's also potentially, and this was from the president of Fox, there's potentially this idea that regulators might attach a condition to the deal, or at least signal or put pressure that the combined entity should maintain its, its tradition of licensing and selling content.

Yeah. Lachlan yesterday said— this is Fox CEO Lachlan Murdoch— expects that regulators will impose a third-party content licensing condition on Paramount and Warner Bros. Discovery's $110 billion merger. He said, we wish them the best of luck. We've seen this regardless of whether it was Netflix acquiring Warner Bros. Discovery or Paramount acquiring Warner Bros.

Discovery.

There will be conditions put on this transaction. Which would require a producer of that size to continue to sell their content to third-party platforms. This was yesterday up in SF.

Yeah. So I ballparked it at like maybe 70% chance that in the next 2 years, there's some sort of meaningful content licensing deal between Paramount, Warner Bros. Discovery, and Netflix. And so, but I do think that there will be limited windows. Some selection in markets where content is available, and pretty much everything will be non-exclusive. So if, you know, some, you know, like the original Dark Knight trilogy goes on Netflix, it'll probably still be on HBO and on Paramount Plus. The real key brand assets like The Sopranos and the DC Universe properties, those are much less likely to be licensed, and there's basically no world where Netflix gets the first streaming rights to like tentpoles. You wanna maintain those, you wanna give the superfans a reason to stay subscribed, to begin a subscription on your now two streaming platforms. There is a lot of potential upside for the Ellison empire. I think of this like buying a house and then getting a roommate. Like you're levered up, you got a crazy mortgage.

That's such a good example.

Right?

You buy a house.

You're living large, but.

But you got a roommate.

Yeah.

And the roommate's Netflix and they're paying the rent to you. You pay the mortgage.

You kind of hate your roommate.

You kind of hate your roommate. But, but you're long the real estate market. And then as you get promotions, as you get more cash flow, as you get more adjusted EBITDA, you can pay your mortgage. And as you pay your mortgage down, you wind up with a really cool asset, which is this nice house that actually has the extra bedroom.

They're like, look, we had a good run. Yeah, I really love the way you would put dishes in the sink and not do them. I really love the way you would watch movies loud into the night and yell. But I think it's time for us to—

It's time for you to move out and I'm going to take the mortgage myself. I'm good for the full bill every month. I'm good. But there are a lot of similarities, right? Because what are you getting with Warner Brothers? You're getting a whole library of IP that's very valuable, potentially for like 100 years. Porky Pig is from 1935, and I still know this guy's name. And I know, like, I could draw a picture if I needed to. And so, if you believe that intellectual property will have a very, very long lifetime, lifespan, and it's incredibly valuable, but you have to finance it with debt, you have to work really hard to pay the bills in the short term, you could wind up with a really great asset even if you have that roommate for the short term. The other interesting question was, what does this reveal about the Ellison family's worldview for the future? Because on the one end of the spectrum, you have Larry Ellison, who seems incredibly AGI-pilled. He's going crazy into AI data centers. You disagree with this?

Well, I think, remember, I think we did like the chart of who's the most AGI-pilled.

He needs AGI versus AGI-pilled.

So he needs AGI, but he's not AGI-pilled.

Why is he not AGI-pilled? Oh, just because of his rhetoric. And he just doesn't have any podcasts. Get him on Dwarkesh. Yeah, yeah. We'll see what he says. Maybe his timeline's like ASI next month, you know?

Who knows?

Yeah, he's like, you need to sit down your family today.

He's obviously very AGI-pilled.

Does he have a Mac Mini? That's the real question. Funding data centers is generally aligned with a future where AI is an important technology.

Yeah, levering up massively, right?

Yes, levering up.

He believes that AI is going to be big, right?

Yes. Yeah, exactly. He believes AI is going to be big. But then he also believes, or, you know, in concert with David, his son believes that, that AI will not destroy legacy Hollywood assets, that like no matter how much, no matter how many dollars you have, no matter how much compute you have, no matter how many generations you can use the latest video model, you will not just be able to create a, you know, a superhero that kids will dress up as at Halloween. And I think that's true. And I believe both of these. You don't? You're going to be dressing up as Slotman next Halloween.

Slotman.

Look at Slotman. You're going to be like, whoa, you don't know this superhero that I prompted myself and no one knows but me? It's a highly— don't you know the internet's highly personalized now?

No, but you can dress up as like a superhero who has like these black ears and like looks very similar to Batman, but it's not.

Batman. Not the non-IP infringing Batman. I think the—

yeah, literally describing Slotman.

Yeah, Slotman. You're going to dress up as the knockoff Slotman. I don't think anyone's going to do that. I think Batman suit and pork and bacon.

I think broadly media becomes like— every piece of media, the audience gets smaller and smaller. Everything becomes more and more personalized. Ah, fair enough. Like maybe, yeah, I think IP still has a fairly big role in that. But you want to be on the cutting edge.

I just think it's like the beginning of shelling points and things that— Like, yes, I can have a video that's perfectly tailored to me, but I like talking about movies with Ken Burns.

They're calling you Timu Batman.

Yeah, Timu Batman over here. Get out of here, Timu Batman.

I don't know.

But, but so maybe that's not your worldview, Tyler, but it's clearly the worldview of the Ellison family because they believe in this barbell effect. They believe in the value of—

Yeah, I would, I would, I would. You know, another way to look at it is like you could get this hyper-personalization, but it could be around existing IP. So you get to watch a personalized version of Batman content that's targeted just for you individually, but the kids are still dressing up as Batman.

And we've already seen a glimpse of this with, with video games. There are video games that leverage existing intellectual property but allow you to customize your character in some way, put more skill points into intelligence and cast spells versus strength and use a sword, right? Like, you can have your own experience if you play— even like the Arkham Batman series, it's pretty linear, but you can play it more aggressively, you can play it more stealthily, and you can experience a Batman world that is more stealth-focused or more fighting game-focused. And that allows you to experience something that is unifying and we can discuss, OK, you played the same game as me. You understand the same character as me, but we had a different experience. And I think AI accelerates that and makes it more valuable, not less. I don't know. I'm somewhat pilled on this, but I respect your differing worldview that you will be enjoying Slotman.

Enjoy.

So by the loosest counts, This is the 9th or 10th time that Warner Brothers has changed hands since it became an independent movie studio in 1923. There's some of those acquisitions that have been like divestitures or reorganizations within other companies, but just like a fascinating thing. And I do think that David Ellison is gonna go the long distance with this. I think with a little careful financial management, and probably some few, a few content licensing deals. Like, this might be the last one. It might, it might, he might, he might be running this organization for his entire career. He's been working in, in Hollywood since he was a kid and has been very dedicated to this particular industry for a very long time. And so I don't think he's going anywhere. And so get ready for a whole bunch more content. Apparently, they're going to make like 30 movies a year. Like, which is a dramatic step up. I think that they're going to try and actually ramp up the production.

Are you sure that's not per day?

I hope so. With AI, it's totally possible. As long as Tyler over there is buying tickets to see Slotman number 17, Revenge of the Slotman, I think that there's money to be printed. Let me tell you about Revenge of the Trough. Lambda is the superintelligence cloud building AI supercomputer for training and inference that scale from 1 GPU to 100,000.

Slotman 2 Revenge of the Trough. Revenge of the Trough is going to hit the kind of brakes.

For sure. Let me also tell you about Vibe.co, where DTC brands, B2B startups, and AI companies advertise on streaming TV. Pick channels, target audiences, and measure sales just like on Meta.

And if you are inspired by the whole Netflix-Warner Brothers saga, you should head over to the Goodwill in downtown Brooklyn. Tom from Vanity Fair says for $14.99, you can Zaslav Max this spring. They're selling a nice Port Authority puffer vest from Warner Brothers Discovery. That is a good pickup.

Zaslav is a deals guy, Hall of Famer for sure. He was involved in launching CNBC. He also helped create MSNBC, became the CEO of Discovery in 2006, and just a whole bunch of other crazy deals. Love to have him on the show once all this once all this settles down, like his full career is fascinating. And he's been loved, he's been hated, he's been everything in between. Lots of people have opinions, but there's no way to be unhappy if you've been riding with him as a shareholder because he got every penny out of this thing.

He certainly did. According to the Washington Post, there was a researcher skeptical of Havana syndrome. So he tested a secret weapon on himself. In 2024, Norwegian researcher skeptical that pulsed energy weapons could do damage to human brains built a device and tested it on himself. It didn't go well.

So this is crazy.

Built, uh, built it, tested it.

Working in secrecy, working in strict secrecy, a government scientist in Norway built a machine capable of emitting powerful pulses of microwave energy. And in an effort to prove such devices are harmless to humans, in 2024, he tested on himself. He suffered neurological symptoms similar to those of Havana Syndrome. The unexplained malady that has struck hundreds of US spies and diplomats around the world.

Yeah, we don't know how to make scientists like this anymore.

This is, uh, glad they're still out there. Extremely bullish for the tinfoil hat industry because I believe that the, the, the actual underlying basis for the idea of a tinfoil hat— of course we use it, uh, somewhat sarcastically here, uh, when we're discussing hypothetical conspiracy theories, but the tinfoil hat is supposed to eliminate microwave radiation and various EMF pulses. And so, if you're worried about getting Havana Syndrome, maybe pick up an EMF-proof tinfoil hat or maybe just some sort of other— there's probably a better way than just tinfoil around a ball cap, but startup idea, startup idea, build a real tinfoil hat that actually protects you from Havana Syndrome.

That's the real American dynamism startup. Thomas Maxwell says, "We have to stop this. I'm not eating a, quote, 'Jacob.'" Eat a Jacob. I haven't seen this one before. It's 20 grams of grass-fed protein, no seed oils, nothing artificial, sweetened with organic honey. This definitely resonated. 24,000 people agree. But I think there's— Tyler, you should do a taste test of all the different protein bars named after just regular dudes. David Bar, the Jacob Bar. There's probably more. You got to get on it.

There are a lot of these. Yeah, the name. I mean, Lucy, named after a person. It was a whole trend for a while.

It was. Cursor would like a word.

Yeah, this was staggering.

This was great. They came out yesterday. They had heard the FUD on the timeline around Cursor. They came out yesterday, and Bloomberg got some of their data. Their annual revenue topped $2 billion in February, according to a source, a figure that underscores the fast growth of the coding assistant. Sasha summed it up well. He said, Cursor sees the timeline turning against them. Quietly give Bloomberg a $2 billion ARR press release. No post from the company or founders. Haters squashed. True comms masterclass. And Michael Truel actually did come out this morning and says, we believe Cursor discovered a novel solution to Problem 6 of the first Proof Challenge, a set of math research problems that approximate the work of Stanford, MIT, Berkeley, Cursor solution yields stronger results than the official human-written solution. Notably, we use the same harness that we built a browser from scratch a few weeks ago. It ran fully autonomously without nudging or hints for 4 days. This suggests that our technique for scaling agent coordination might generalize beyond coding. So very, very cool to see the progress from them. And clearly, doing something right in the enterprise, right? Just because everything that you see on the timeline is like so many enthusiasts.

Totally.

One of my buddies is Kyle Russell. I worked with him like a decade ago. He's at a company called Valen. And he posted— so Valen, they make SaaS for mortgage services. He says it's very boring, but we're basically getting them all to flip from legacy software and are rate-limited by biz ops people being able to onboard them. So we're gonna try instead to think about it in terms of what tokens do we need to extract from the org in order to deploy fast? So he's in this like AI deployment lead role, and he went, I think, pretty viral saying this morning, one person of his team said, hey, can you unsub me from Cursor? And somebody says, done. And then a bunch of people said, same, same, like I don't need it. I'm like, I'm happy with another program. And Kyle said, today we announced we're removing 90 Cursor seats because they haven't had any use in 2 weeks. And it is like, it's flipped from like the most cutting-edge thing to, like it's clearly very sticky because you see the ARR numbers that this like ultra-frontier coolest, hottest thing is getting some like FUD. And then Sasha broke this down where he was like, the timeline's turning against them. So you have to provide some evidence that like, you're not cooked because the timeline is very much like, oh, well, you're not the hottest thing anymore. You're not the coolest little thing for the people that are on the most frontier. And so no, just squashing the haters with a $2 billion ARR number, which is fantastic.

Up from $1 billion in Q4, which is just insane.

Yeah, and so, yeah, I mean, I think there's just something about the— it goes back to diffusion. I was debating this with Tyler last night about— how much of AI adoption is just actually getting the forward deployed engineer in, actually getting people to change their workflows, onboarding. Like there are certain people that will just bounce from the most frontier thing, oh, this model's better, I'm on Codex, I'm on Claude, I'm on Codex, I'm on Claude, back and forth. But for a lot of companies, they need a little bit more handholding. And so there's a whole massive chunk of the economy that can be transformed by developing great products and then actually getting them in the hands of businesses.

R for Rock is hearing Cursor maybe in the midst of a new raise at $50 billion. That's crazy. That would be a meaningful markup from $30 where they were sitting again in Q4.

Yeah.

What does DD say? Everyone in the Tech X bubble thinks they are wholesale ditching Cursor, but enterprise diffusion is glacial. Most of the world just got a hold of it. It's a good point. It's a very good point. Well, in some good news, Mark Zuckerberg has purchased a mansion in Miami for $170 million. This was such big news that we had to pull it forward from the mansion section on Friday to talk about today. But there is a little bit of a black pill here. We can go into it. First, let me tell you about Railway. Railway is the all-in-one intelligent cloud provider. Use your favorite agent to deploy web apps, servers, databases, and more while Railway automatically takes care of scaling, monitoring, and security. And let me also tell you about Cisco critical infrastructure for the AI era. Seamless— unlock seamless real-time experiences and new value with Cisco.

So of course Mark got this because of California's wealth tax, which it feels like if it goes through, he would still be subject to at least a one-time payment, you know, which would get litigated. Of course, what we need now is a picture of him in his backyard recreating the Ben Affleck smoking meme. Why? Because there's this new national wealth tax that Ro Khanna and Bernie Sanders are pushing. So he's like, he just, he just moved. He just got this new place across the country. Yeah, probably overpaid to a degree.

People are just going to be like, actually, I live in international waters full time. I don't live on— I don't live in America. I don't live in any country, actually. I live in space. I live I live on the International Space Station or something. Very, very tricky. $170 million sounds staggering. A 5% tax on Mark Zuckerberg's wealth, over $11 billion. That should have been his shopping budget. Because in dude math, if you save money, you just have a free license to spend it. So you're like, by moving to Florida, I'm effectively saving myself $11 billion. I can spend that, right? Isn't that how that works? Exactly. Meta Chief Executive Mark Zuckerberg and his wife Priscilla Chan have paid $170 million for an under-construction mansion on Miami's sought-after Indian Creek Island. The deal closed Monday. The purchase set a record in one of the country's most expensive to date. The current U.S. record is held by billionaire Ken Griffin. He spent $238 million for an apartment. That is a crazy amount of money for an apartment. But I guess it's in a good building.

There's a lot of apartments.

The Florida record was set last year when a waterfront compound in Naples traded for $225 million. I feel like this isn't counting Ken Griffin's compound that he's building because he's built—

Yeah, because it's multiple properties.

He's bought multiple properties.

Okay, here's where it gets crazy. The sellers are Dr. Aaron Rollins, a cosmetic surgeon to the star, and his wife, real estate agent Maureen Rollins. The Rollins paid more than $30 million for the roughly 2-acre site in 2020.

Wow.

Flipping.

Not real estate agent.

Flip.

Don't wanna go up against them. They know what they're doing.

Plans called for a 9-bedroom home measuring about 30,000 square feet with a dock and a swimming pool. Amenities were to include gym, hair salon, and massage room, as well as a 1,500-gallon aquarium and library with a secret passageway.

Zuck does need a hair salon in his house 'cause he's always changing his hairstyle. Sometimes he's got the Caesar going, sometimes he's got the fro going.

So this is the Gen Z.

This is elite for him. Who knows what'll go next? Maybe he'll have long flowing locks like Fabio. That would be the good next move for him, I think. How many gallons is the aquarium? 1,500. Is that a lot?

It doesn't seem like that much.

What can you put in there?

Can you put a shark in there? No way.

Some fish?

No way.

How big is Keith's aquarium? Keith's aquarium is massive in Miami. You can scuba dive in it to clean it up. Anyway, congratulations to the Zuckerbergs on their new property. It looks beautiful. And without further ado, we have our next guest in the Restream waiting room. Let me tell you about CrowdStrike. CrowdStrike, your business's AI, their business is securing it. CrowdStrike secures AI and stops breaches. And we have Matthew Bellini from Puck. How are you doing?

What's going on?

Doing great.

Thanks for having me.

Thanks for hopping on.

Great to meet you. I enjoyed your podcast with Dylan earlier today, and I wanted to dig further into the actual merger. I mean, it seems like we're sort of done now, but how has the process of covering this been for you? What were the key turning points? Like, what was most exciting for you?

Well, I'm learning a lot. Certainly didn't know what a ticking fee was until covering this deal. Learning a lot about the debt market. Learning a lot about the contours of raising money in the Middle East. All sorts of fun things. But yeah, it's— I mean, it's been a wild moment in Hollywood for the past 5 years, basically.

It's—

it's—

a time of tremendous change, tremendous disruption. A lot of people are really scared and don't know what's going to happen. So it's, I mean, it's been, it's been interesting for sure.

Is that, is that just a combination of like COVID plus streaming plus consolidation plus AI, or is there like one story that sticks out to you?

You can throw in a pretty crippling labor strike in the middle of that. And it's been one thing after another. About 5 years ago, the end of the peak TV era, the Netflix correction where the market decided that Netflix had to make money, streaming subscribers was not the end-all be-all. That created an absolute correction. And pretty much I would call it a content recession. Within Hollywood, and that's trickled down throughout the industry. Fewer jobs, fewer productions. You add that in with the flight of productions overseas, the everything from the strike that happened in 2023, which gave these companies an excuse to really reevaluate the level of production. Throw in the LA fires that disrupted so many different people. And it's, you know, you guys are in LA. It's a really interesting and, for many people, scary time.

Yeah. Explain the overseas thing, because I'm familiar with Atlanta and I believe Toronto, but are productions really moving to other continents now?

Oh, yeah. I mean, the UK has been extremely aggressive. They have, for many productions, a 40% incentive, and they're willing to give you incentives on what's called above-the-line costs. Those are the actor, director, writer. Those are the main talents in the film. That is a huge incentive because it essentially subsidizes you for Tom Cruise's salary, Ryan Gosling's salary. And even in some of the more aggressive U.S. states, you won't get that in California. So, you know, it just— it's a game changer when they are doing the line. I just reported on this movie today, Project Hail Mary, an Amazon movie that's coming out in 2 weeks. The gross budget for that movie is $248 million. The actual net budget is a little under $200.

Okay.

So you're saving 25% by shooting this movie overseas.

Interesting. And, and do those come as like tax credits upfront, or is it more like a tax deduction on profits where if the, if the movie costs $250 to make and it makes $500, then I'm saving 25%? Or do I just get the money basically upfront?

It's a rebate. I'm actually not sure. I think it depends on when you actually receive it.

Sure.

But it's a rebate for the qualified spend in the jurisdiction that is giving the incentive. Yeah. So if you are going to get an incentive from the UK, you actually have to have the work done in the UK, and that means hiring local crews. It means using a soundstage in the UK and having, you know, the production actually shoot there. And you can have films with multiple incentives from multiple jurisdictions for multiple aspects of the production, everything from the post-production to visual effects. Those can be subject to incentives for doing the work in a particular jurisdiction.

Really quickly, I think we have the wrong mic selected. Would you mind adjusting it? Because I think we're on your laptop mic, and I'd love to get that Shure SM7B working because I know it has higher audio quality if you can switch it. And while you look at that, I will tell the audience 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.

Any— How we doing there?

Is that better? No.

No.

Almost.

Almost better.

It's okay.

How about same as— Any luck? How about this? Is this better? Oh, that's much better.

There we go.

There we go.

I did a good job.

Wow, you have sound effects.

I'm just very jealous. Yes, yes, yes.

I wish I had an air horn.

Some of these are not sound effects on a soundboard. Some of these are real folks in the studio clapping with their hands. We have both.

I can't be trusted with an air horn.

Yeah, yeah, we had a lot of fun. So talk about the reaction. I know that we talked to Ashley Vance, who's a documentary filmmaker, and he was voicing a little bit of worry that if Netflix acquired Warner Bros. Discovery, that there would be one less buyer. And is this any better for the average filmmaker who wants to go out and get the best price for their product, for their— for the film that they've been working on? Are we still telling the same consolidation narrative, or is this somehow structurally a little bit different?

It depends who you talk to, but I would argue that this is actually worse for creatives because this is the actual smushing together of two legacy studios.

Yeah.

When you talked about Netflix buying Warner Brothers, Netflix is not in the theatrical movie distribution business currently. They do not do movie theaters.

Yeah.

So there is a certain class of movie that was going to remain a Warner Brothers movie. Now, Netflix makes a lot of movies on its own, but those are often smaller. They are not— they don't have as much ambition as the big theatrical movies. They do some of those too. But that Netflix was at least positioning it as we are going to leave Warner Brothers alone. They also were not in the production business in television for other buyers. Everything that Netflix made was for Netflix, whereas Warner Brothers makes shows for its own platform, HBO Max, but also for Apple and for Netflix and for many other different platforms. So they claim that that would be at least another opportunity for buyers here with Paramount buying Warner Brothers. This is going to be the combination of two television studios, the combination of two film studios, and ultimately down the line, that is going to mean fewer jobs. Now, if you talk to the Paramount people, They say they are committed to making 15 movies for theaters for each Warner Brothers and Paramount. That's 30 movies a year, which would be an escalation for both of those companies. So they're saying they're going to be more. But the history of the movie business has shown that when two studios merge, ultimately the output is reduced and that trickles down to creators.

What do you think about Netflix's future theatrical plans? Ted Sarandos was talking about the future of movie theaters, painting a more optimistic vision. Do you think any of that sticks?

You've got $2.8 billion burning a hole in the pocket.

Yeah, I mean, I don't think Ted is going to use that $2.8 billion to all of a sudden make movies for theaters. Yeah, I just don't think so. I think it's going to be a continuation of their current model, which in many ways is designed to put the theaters out of business. It's designed to give you first-run A-list filmmaker, A-list star movies at home. Even 10 years ago, a movie like The Rip with Matt Damon and Ben Affleck, that would have had a nice robust theatrical run before hitting home video. Not true with Netflix. Now, what Ted is saying is that they have had some constructive stunt experiences with theaters lately. They put K-pop Demon Hunters in theaters. They've done a couple of other things. They did Stranger Things in theaters. And I could see Netflix doing more of that, using using theaters essentially as a marketing device for Netflix, but I don't see Ted all of a sudden saying, oh, you know, through this process of looking at Warner Brothers, we have determined that the theatrical model is a good one and we are gonna all of a sudden give our movies a 45-day theatrical window.

Don't think that's happening. And do you think, how much of that do you think is Netflix just, well, if we have the, we have a crown jewel asset, we want people to go subscribe to Netflix and see it there first versus, A lot of times they don't know what's actually going to work. And the beauty of the Netflix model is they can stuff something like K-pop Demon Hunters on there. It can effectively go viral in their algorithm. And they don't need to do the billboard campaign, the theatrical rollout, and take all that risk. So focusing more on downstream, we got something that's already working. Let's make a musical sing-along version or a promotional version for theater.

Yeah.

That is absolutely the benefit. And Sarandos has said multiple times that that the best marketing platform for Netflix content is Netflix. Now, I recently had the head of Sony Pictures on the town, and he, I think correctly, said you do not make stars on Netflix. You do not make franchises that last for decades on Netflix. And the way you do that is in movie theaters with a big, robust marketing campaign that tells the world about the product so that even if you don't see the movie, you're at least aware of it so that when a sequel comes out in 4 years, maybe you see that one. But if a sequel to a Netflix movie comes out, nobody except the people on Netflix know about it.

Yeah, yeah, yeah. I saw Ryan Gosling on the Hail Mary billboard and I was like, oh, like, we should go see that. And somebody was like, well, what is it? And I was like, it's a Ryan Gosling movie.

Right. What else? And that's a perfect example. I mean, Amazon had the choice whether to make movies just for Prime Video or to do theaters. They are going big on theaters this year with a full slate of theatrical releases. And they and many others in town believe that the best way to create movie franchises and to create demand is to first put the movie in theater and then debut it on the service.

Yeah, it's interesting. It reminds me of like an apparel brand doing pop-ups specifically, not, not as like a real revenue driver, but to just deepen relationship with their—

Now, don't get me wrong, these traditional studios, they need these movies to make money in theaters. That's their whole business. And the reason you are spending $50 to $100 million on a marketing campaign globally is to generate profitability in theaters. And then you get the waterfall, we call it in Hollywood, the waterfall of revenue that first you go to premium video on demand, then you go to subscription video, then you go to TV. Then you go to free TV, then all of a sudden you're syndicating the content all around the world and it's in your library forever. The reason why Warner Brothers is going for $111 billion is not because they had a nice run in theaters last year or because HBO has had some hits lately. It's because they have a 100-year-old library of films that is incredibly valuable to exploit across all platforms and because HBO has an amazing brand.

Yeah. So talk about content licensing at the new entity, the combined entity. It feels like with $79 billion of debt, cash flow is going to be important in the short term. What's the probability that we're watching Sopranos on Netflix or watching something from the catalog on Netflix in the next couple of years?

It's funny because Sopranos has been the one title that Netflix has not gotten.

Yeah.

And HBO has kept that. Not the, not the one, but the one big one that everyone's waiting before.

Yeah, yeah. And, you know, the, the—

this is—

the debt issue is not new. The Warner Brothers Discovery transaction, putting those two companies together, that created $50-something billion in debt, and they needed to make money fast. And the past 3 years have largely been about servicing that debt and bringing it down. So you've seen tons of Warner Brothers shows and movies on Netflix, and they do it because they have to. I mean, at one point last year, I looked on the Netflix top 10, and 6, I believe, of the 10 movies in the top 10 were Warner Brothers titles.

Yeah.

And they do it because they need the money. I mean, we know Sony Pictures just did a re-up of their Pay One deal. We call it Pay One because it's the first pay platform for subscription viewing of these movies. They got $7 billion out of Netflix over a long period and globally. But that is for the first-run pay-one window for Sony Pictures movies to go to Netflix. Incredibly lucrative. That's what Netflix was buying when they wanted to buy Warner Brothers.

So Sony and Netflix already have this relationship. Is there any chance that Sarandos goes shopping with his $2.8 billion bonus and all the cash flow that he has? He's sort of, you know, kick the tires on one big deal. Is Sony Pictures in play at any point in time?

The Sony Corporation in Japan has said that they are not interested in selling Sony Pictures. It would seem to be the obvious choice because Sony is the one Hollywood studio, the traditional studio, that does not have a general interest streaming platform. They chose not to enter the streaming wars, and that would make it an attractive buy for Netflix because Netflix doesn't need a global streaming service. They already have one. They could pick off the Spider-Man franchise. They could pick off, you know, all the production and distribution capabilities that Sony has for theaters would seem to make sense for Netflix. But there has been no indication that Japan is willing to sell.

Yeah.

Is any part of the Ellisons' strategy— does AI kind of factor in at all? I've had this kind of like loose theory that this, this IP, Warner Brothers IP, could potentially be more valuable as AI gets better because you can create so many different variations on top of that IP.

But it's also unique. It's not replicable.

That's it. That's it. It's like you have a monopoly on the IP and if it becomes substantially less expensive to create new film and television, TV shows, movies, etc., maybe the IP is more valuable. But what's your take?

Yeah, I have always thought that one of the reasons why Netflix wanted Warner was because they saw what's going on with YouTube and the volume of content that YouTube is soon going to be able to host and eventually create themselves. And Netflix is saying to themselves, okay, So if anything is possible in streaming and with AI, you can create AI-generated shows, music, content, movie stars, whatever. What are we doing? What do we have that is defensible here? And they're looking at Warner's and they're saying, okay, well, we don't have a Batman, we don't have Harry Potter, we don't have, you know, this 100-year-old library of known quantities that cannot be replicated by AI, at least not yet. So we need that. And Netflix has had a hard time creating franchises. If you can count on one hand the Netflix franchises, the things that they own that are meaningful in the world— Stranger Things, they own. Wednesday, they don't own. Amazon actually owns The Addams Family because they bought MGM. There are not that many franchises. You guys probably don't even know what they are.

Yeah. Yeah, yeah, yeah, it's fascinating. Can you take us on a tour of who is actually involved financially in this deal? You mentioned that it was sort of like a whirlwind tour in the Middle East, a lot of different funds coming in. I imagine that it's some overlap with the tech investors that we typically talk about on this show, but—

Apollo is in this. I believe Blackstone is. I'm not 100%. There are 3 big PE firms that are providing a lot of the debt financing here. And then there is an interesting question of whether the Middle East money that was initially floated as being part of this deal is still in there. There was, there was a Saudi fund, there were Qataris, there was one other fund, and they were not in the press release. Okay. So I actually need to ask Gerry Cardinale of Redbird whether they are still in this deal. Sure. Because Redbird has been kind of the, the you know, big financial backer and consultant on raising this bid together. And obviously Larry is backstopping the whole thing. This— we would not be having this conversation about a company with a $15 billion market cap, which is Paramount Skydance, acquiring a company like Warner Discovery in $111 billion deal if not for the backstopping of Larry Ellison himself. That was the key factor. He did not want to personally backstop it. Then once he did, the levers started moving.

What is Larry Ellison's history with investments in Hollywood? I know it is not just David that has been involved. He has invested before. Obviously, David has almost multi-decade career at this point. What has been Larry's strategy? What can we learn from his past adventures in Hollywood?

It's a good question because people think this is out of nowhere and it's all of a sudden Larry Ellison deciding to dip his toe into Hollywood, although if you can call over $100 billion dipping your toe. Big toe.

He's dipping the big toe.

Right.

But Megan Ellison, David's sister, she had a very artistically successful production company called Annapurna, which she still has. But there was a period in the 2010s when she was the dominant producer of so-called arthouse and, you know, independently financed, auteur-driven movies. And she actually got 2 Oscar nominations in the same year for herself for Best Picture when she had, I believe it was Her and The Master in the same year. And she's done— I mean, she's done a ton of movies with big-time auteur-driven, filmmakers. And then the wheels came off. She just lost too much money. And Larry basically had an intervention with her and said, the money's not coming in anymore. You're done. I'm not doing this anymore. And he had his people step in. They pulled way back. They downscaled the company. And Annapurna still exists. They had a, they had a big Sundance movie this year directed by Olivia Wilde and starring Seth Rogen. But they're not doing a whole lot. So if we learn from this, it's that Larry is happy to indulge his children, but there is a limit on what he's willing to do. And she hit that limit. We, you know, we'll see what happens with David.

Yeah. Is— sorry, you can go. If you just— just— I'm curious if— if— I mean, it sounds like Annapurna had a very different just financial underwriting strategy in the sense that Warner Brothers is very much like this basket of assets that are so proven. Oh yes, it's high leverage. Yes, it's expensive, but it's this mortgage on land that if you pay the bills on time, you will get the value of because there's this catalog value that you can license. You can continue to merchandise these assets. So it feels like a different financial strategy.

Absolutely. I mean, there aren't enough— there are not a lot of legacy studios with 100-year-old libraries sitting around. It is a pretty unique asset, whereas Annapurna was creating its library from the ground up, which is a much more difficult business. The, the, you know, the corners of Hollywood are littered with the corpses of companies that try to take on the studios and do this on their own because you have to have have that 100-year-old asset to float you when you have bombs. Because this is a portfolio business in the sense that if you release 20 movies, 10 of them probably not going to make money, if you're lucky. Yeah.

And it's—

you got to make a lot of money on the ones that do make money, and those will make money forever.

Yeah. Can you bridge that to A24? Uh, we're more familiar with that because there's some tech investors involved. I personally like a lot of the movies, but what, what is the strategy of A24? How do they fit into the modern Hollywood landscape?

A24 has been incredibly successful on the low-budget model. They, they are one of the few film studios that have created a brand for themselves. I would argue that the A24 brand is more meaningful to moviegoers than the Warner Brothers brand. I mean, everybody knows what Warner Brothers is, but what does Warner Brothers stand for? What is that brand? You don't really know, and you never had to know because that was just the— it wasn't a direct-to-consumer business. They put their movies on other branded platforms, whether they are theaters or whether they were television or now streaming on HBO Max. But A24 has created a brand. They sell branded candles and hats and things that are meaningful to their fans. But they succeeded. They succeeded on betting small on, on specific auteur filmmakers, and they had a nice run. They also raised a bunch of money, as you mentioned, and they have recapitalized a couple of times now. And because they got what I believe it was, a $3.5 billion valuation a couple of years ago, they have essentially now been forced to go bigger, and they are releasing movies. Marty Supreme cost $70 million. Yeah, uh, The Smashing Machine with The Rock, that also cost a lot more than their typical A24 movie. And they're still buying the small movies at festivals and releasing smaller movies, but they now have more of a portfolio of bigger budgeted movies. And the jury's out on whether that model is going to be successful for them, because again Once there's many companies that have had small level success in Hollywood, and when they try to scale up and compete more directly with the traditional studios, they've faltered. And we'll see if A24 is able to pull it off. They got really smart people there, so I would hope so.

What is actually— what is Hollywood actually doing with AI today? Not like, what do they want to do? Not what are Silicon Valley companies pitching them on doing? But how is it actually being used today?

Well, there's two answers to that question because there's one in the general what is AI doing for all businesses, and that is, you know, streamlining processes and, you know, creating faster workflows. That's all the boring aspect. What you're probably asking about is how is Hollywood using AI to create content instead of in place of a camera? Right, exactly. And there it has been much more limited. The toe is in the water and each of these companies has an initiative to incorporate more AI specifically and more aggressively in the animation divisions. The goal is to bring down the cost of animated movies significantly. And you do that by using AI to automate a lot of different functions. And they're being coy about exactly how, but I have heard like the storyboarding process, and the physical production of the animated images. Once the artists have done their work to create the character design and the movements they want, AI can be used to essentially effectuate a shot. And you're seeing more and more of that. Where they're treading a little bit more cautiously is where the traditional labor unions are involved, where actors, and directors and writers are involved. You're not seeing studios greenlighting scripts that are written by AI. You're just not. Screenwriters are using AI as a tool—

That they know of. That they know of.

That they know of, sure. But they are— screenwriters are using AI as a tool, and they will freely admit that, much like they were very happy to use Google when it first came about, and they didn't have to have a researcher that would go to a library and tell them what the interior of the FBI looked like. They don't have to do that anymore. And with AI, they can create scenarios or create drafts or create things that they— that will help them. And the— all that the guilds want is for their members to get paid. So the studios are not allowed to buy a script or greenlight a script without paying a human writer. And that's a big deal. It's the same for actors. You can use AI to accentuate or to, you know, do a quick reshoot shot. But you have to pay the people involved. And we're headed into a new labor negotiation where the studios will try to chip away at that. And the guilds are going to try to be more aggressive and make sure that their members are getting paid for any use of AI here.

Where do you think— what kind of scenes do you think AI will end up coming for first? Is it the Animation. Is it the stunt? Well, animation, but then what about stuntmen? I feel like there's a—

Transformers-level CGI where it's so expensive.

Something like that, yeah.

Visual effects.

Visual effects. Do you think that the dynamic between the studios and the guilds could allow kind of like an entirely alternate kind of like industry to form that is sort of broken away from traditional Hollywood norms? Like, you know, you could imagine a group of people in, you know, some suburb of LA that aren't a part of the guilds, that have never engaged with the unions in any capacity, just decide, "We're gonna start making movies ourselves," and kind of just ignore the legacy process.

[Speaker:SETH GREEN] I think that's already happening, right? I mean, you see it. But the difference is, I mean, we all went nuts over that that Seed Dance video of Brad Pitt and Tom Cruise fighting. But that's because it was Brad Pitt and Tom Cruise. If it was two dudes fighting, nobody would care if it was two AI-generated dudes that did not look like Brad Pitt and Tom Cruise. And that's the power stock footage Hollywood people have. Yeah. Yeah. It's— I mean, there's going to be— I think what's going to happen is there is going to be a class system of AI-generated content that will live and have an audience and generate revenue at scale, but it will be perceived as lesser than the more professional and union-supported Hollywood-style work. And I could be wrong about that. It could just bowl over everything, but there's still value in Ryan Gosling's face. And for now, the unions and the Hollywood studios have been able to protect that from AI.

Yeah, you can imagine the luxury houses versus fast fashion, like a similar dynamic playing out where like, yeah, you can get a jacket that like, functions in the same way as something that Gucci made. Maybe it's the same even material, but it costs less and it has obviously much less prestige.

And I think that there will be a middle ground too, where where movies, the cost of movies is brought down because of advances in AI where things can be done without outsourcing something to 5 artists to work on at a visual effects company for 2 months. It could be done in-house for less money with an AI tool. I think that's already happening. And there are much lesser guild protections for that aspect of the business.

Post the OpenAI Disney deal, I was expecting to see another deal of the similar shape happen. It hasn't happened yet. Are you hearing rumors? Do you think other studios with big intellectual property and character libraries will take a similar approach?

I think that there are a lot of talks going on and they're trying to figure it out. Disney, Disney was in an interesting spot there because it was— that deal with OpenAI was as much about putting pressure on Google to come to the table and to not go rogue as Google has been doing. Sure. As it was to do OpenAI.

Got it.

Also, there was a preexisting relationship there. Bob Iger, the CEO, was— he worked for Thrive, which, you know, is a big investor in OpenAI. And Disney is, you know, if you're going to go after one, company, that's the one to go after because they have the best characters. And if you know about that deal, that deal does not include any voices or any likenesses of real people. Even though Luke Skywalker is the animated Luke Skywalker. Yep. So it's a lot less compelling for people to play around with if you're playing around with Lightning McQueen or Yoda and they don't sound like Lightning McQueen or Yoda. It's a lot less compelling than if they did.

Interesting. Yeah, that'll be interesting to see how they roll that out. Maybe that happens down the, down the road. Jordy, anything else? This was great. Thank you so much for taking the time.

Great to finally meet.

Great to finally meet. Anytime.

Fantastic reporting. Come in in person next time.

Yeah, we'd love to have you here. Anytime. We'll talk to you soon.

You can listen on The Town or you can read me at Puck.

Yes.

Perfect. Thank you so much for hopping on the show. We will talk to you soon. Goodbye. Let me tell you about the New York Stock Exchange. Exchange. Want to change the world? Raise capital at the New York Stock Exchange. And let me tell you about Plaid. Plaid Power is the app you use to spend, save, borrow, and invest securely, connecting bank accounts to move money, fight fraud, and improve lending now with AI. And Chris and Kyle has a post here. Credit where credit is due. China really nailed the Bond villain aesthetic with their Antarctic research base. Qinling.

Qinling.

Okay, these are just renders.

These are renders. This is what they're thinking about.

Because if it was real, I'd be telling Tyler to get down there, get a boat immediately. Tyler was in Arizona over the weekend and sent us a very cool picture of him out at TSMC.

Give us a review. Did you feel the power? Did it feel like visiting Arrakis?

I mean, yeah, it was like 91 degrees. Okay. Extremely hot. It is massive, massive facility. They're still building. Like, there's like, there were two kind of main sections. I couldn't actually tell like what parts they were. There were signs there. There were a bunch of fabs. I think I took a picture next to like Fab 21 or something.

Wow.

But there's like, there's so much building going on, and it's just like, it's basically north of Phoenix. Yeah. And so there's like 10-mile radius where it's just like basically just empty land. They're flattening the desert. Whoa. But yeah, very, very cool. Gift shop?

Fun to walk.

Any gift shop?

No, they didn't let us in. There's like a ton of security around at every gate. I wanted to like go in and like take a tour or something.

I want to get, I want to get a wafer that didn't make it through quality control and get some of the chip CEOs to sign those. I think a Lisa Su-signed TSMC wafer would be amazing, or Morris Chang-signed wafer. That would go pretty hard. I would love that. Love it. Moving on. What is it? The billionaire tax has gone national. It's gone national. Bernie Sanders is proposing bringing it to the United States broadly at the federal level. Bad news for everyone who migrated to Miami or Florida, because it is going to follow you wherever you go. Ro Khanna and Bernie Sanders are proposing a national wealth tax on billionaires going even further than California. They want 5% unrealized wealth tax to be annual.

Every year.

Every year. That would be seriously controversial.

And I've seen some people running the numbers on like, oh, if this tax had been in place since 1999, Jeff Bezos would still be worth $61 billion, and he could still afford his $500 million mega yacht, just not actually processing all the negative externalities of something like this. So yeah, I just—

Amazon turns out into a PE-backed shell pretty quickly because you lose control, right?

Yeah, that, or Amazon never even gets the level of investment that it got because of capital flight. I would be so much more sympathetic to to Ro Khanna and Bernie Sanders on this if they had like, here's 5 case studies where wealth taxes worked. And like, they can't come up with a single one. And so it just feels like—

50% over 10 years, that is a lot. But the implementation of this tax would ultimately create wealth flights, says Chris Withens. The wealthy are wealthy because they fight to preserve their wealth. So in year 1, the base is already decimated. Then every single year afterwards, the remaining billionaires leave or hide their assets. Meanwhile, the markets are all declining all this time because they see that America has been infiltrated by wealth-destroying communists. Chris is not a fan.

Dave Friedberg has a bunch of kind of good, good content around this, talking about how there's something like $8 trillion of billionaire wealth and $160 to $170 billion of middle-class wealth. So that is the real prize pool. And again, hopefully, hopefully our lawmakers and voters process how ridiculous this is.

Find another way. And reject it. Write the balance sheet of the US government. Let me tell you about Vanta, automate compliance and security. Vanta is the leading AI trust management platform. And let me also tell you about Graphite, code review for the age of AI. Graphite helps teams on GitHub ship higher quality Data, higher quality.

Sorry. Clue Show Investment says, "Her, the market is only down 0.08% year to date. Things must be really calm, right?" This is actually crazy.

If you asked me where the market was and I hadn't looked, what is this from? What movie is this? Is this AI or something? This is a crazy scene. I really like this. This is a good meme template. I haven't used this before. Just RPGs flying around. All-out war. It is all-out war. I mean, a lot of stocks are down like 50%, other stuff's way up. Things are all over the place. It has been absolute turmoil. A lot of red days in our favorite companies. Is it from Maze Runner? Oh, maybe Maze Runner. GoPro is one of the stocks that is particularly far down. We did a deep dive on GoPro very early in this show's history. They are now at a $150 million market cap. Josh Wolfe has a, has a quote here. GoPro's global market share of action cameras dropped to 18% as of September 2025, compared with 84% in 2022, according to Beijing-based consultancy Meritago Services. At the same time, DJI share rose to 66%, and Insta360 is up at 13%.

Yeah, so it doesn't even count Meta Ray-Bans. Just out-executed on the product side, but cell phones are probably even a bigger factor here, kind of capping the overall TAM. Cell phones at some point just got really durable and the video quality got really good. I remember back in the day skiing or snowboarding, I was a little bit more hesitant to be like, I'm gonna take my iPhone out and use it while skiing, and just because it wasn't waterproof, and then they got waterproof, They got really durable. And most people aren't good enough at what they do to actually warrant needing effectively bulletproof device.

The dust collection problem is very real with action cameras. You buy it thinking you're going to be shredding, going downhill, you're going to be surfing. And then you go for a hike once. The video's kind of underwhelming. And then it collects dust. Skiing this weekend, a friend of mine— I didn't see any GoPros among my friend group. One friend of mine was wearing the Meta Ray-Ban Vanguards and enjoyed taking videos and images with those. I've been seeing a lot of Meta, Meta Ray-Ban content on Reels. It feels like they're promoting that. They have a little tag and it feels like that's getting a little boost in the algorithm. So if you're a Reels creator, or maybe go pick up a pair and then post some stuff on the algorithm because you might get a little boost. But should Meta buy GoPro? $150 million? That's like, what, 1/10 of an AI researcher and you get a whole suite of products. And I just have no idea if they would get anything through approval. It just feels like this is a company that you can't do a zombie acquisition with because you need the assets, you need the IP, you need the supply chain, You need the products, the distribution, you need the whole company. And Meta can't acquire like a VR fitness company without getting this approved. Granted, it's a different FTC, so who knows? But if you were Mark Zuckerberg and you were trying to get into the action camera market broadly with Meta Ray-Bans, would you buy GoPro for $150? Absolutely not.

Absolutely not. You partner with Oakley.

Partner with Oakley.

It's a better— it does exactly— It's just the future. Yeah, it's just, they made a pair of Oakleys, the action sunglasses that function just as well as a GoPro for 99% of people, and it's integrated with your phone, and you can watch Instagram Reels while you're backflipping off of— in the terrain park. But the crazy thing, GoPro still did somewhere around $670 million of revenue last year, so trading at— at less than a quarter times revenue, but absolutely brutal out there. Some other people were suggesting like a foundation model lab should acquire the company for like— and that felt like a— Feels like a stretch. A silly idea. I don't know. Doomer says the past looks more like the future than than the present does. Let's pull up this. What is this? PM's Phoenix, an expandable van you can build for $2,000.

This is sick. What company is this? PM's? How old is this? Is this AI-generated or something? This looks like incredible. Most companies and most campers and vans these days that are roomy enough to live in are too large for comfortable driving. Conversely, the compact, fun, drivable rig are usually too small to live in. And no matter what its size, the seller will probably want all the money you have now, plus most of what you'll make in the next 5 years. This is a good copy.

So this was in Popular Mechanics in 1978.

This is a good daily.

Someone should pick— built on a Volkswagen minibus chassis.

I like that they restyled the front cab as well. Someone's, someone's mentioning that this— the Cybertruck camper back end and looks pretty similar. But it still feels odd to take an electric vehicle into the wilderness. I don't know why, because it's not like you're bringing a bunch of— it's not like there's a gas station out in the middle of nowhere. But this would give me range anxiety, even though I think that's deeply fake at this point.

New model alert.

New model alert.

What we got? Introducing— Logan's introducing Gemini 3.1 flashlight, a huge step forward on The boundary of intelligence beating 2.5 Flash on many tasks. Tyler, you should try to do a speed check. OpenAI also came out this morning with 5.3 Instant. Okay. Touting increased accuracy.

Is this in ChatGPT? Because I feel like I've had 5.3 in Codex, but not in, but not— oh, I got— yeah, I have 5.3 instant in ChatGPT.com now.

Interesting. I do not yet.

Very cool.

And then they're also teasing 5.4. They say 5.4 sooner than you think.

So let's see. 5.3 is designed for very, very low latency. It still feels slower than the Cerebris response that I got. From 5.3 Spark because I think that the Cerebras chips, although they are faster, they're probably in more limited supply, more relevant to the long-running coding agent models. And so they are still rolling that out. But I'm excited for Cerebras in ChatGPT as a product. And yeah, 5.4 sooner than you think. With a capital T. Very fun.

Interesting.

Let me tell you about MongoDB. What's the only thing faster than the AI market? Your business on MongoDB. Don't just build AI, own the data platform that powers it. And let me also tell you about Figma. No matter where your idea starts, Figma Make, Cloud Code, Codex, or a sketch, the Figma canvas is where ideas connect and products take shape.

Partner with Bezel. Oh yeah, to allow people to trade price movements of iconic watches. So we can pull up this video here. This basically allows you to trade on what the— on the prices of Rolexes for the month of March. Basically, is price going to move up or down? So Quaid over at Bezel has been telling us about this one.

So you could, you could in theory buy a watch on Bezel and then take out a short position on Kalshi and have a market-neutral Submariner or something, right? Because if the price goes up, you own a watch.

It's not just any watch. It's a market-neutral Sub.

It's market-neutral because if the market crashes, you make money on Kalshi, but you lost money on the watch and you're neutral. So you can hedge yourself now.

That's very bizarre. Yeah, I saw somebody pushing back. Oh, you can. This is silly. You could just buy the watches and actually trade them. But obviously it's quite a bit more complicated.

Well, you can't sell one that you don't have. Now you can short watches effectively. True. Let me tell you about AppLovin. Profitable advertising made easy with Axon.ai. Get over to get access to over 1 billion daily active users and grow your business today. And I believe we have Ken Burns Ken Burns in the Restream waiting room. So we will bring him into the TPN Ultra. What's going on? Ken, thank you so much for taking the time to join us today. How are you doing?

My pleasure.

I'm great.

Great to be with you.

I would love to start with the first project you worked on ever, because I'm fascinated. Your career has such clarity, but finding a life's work is often the hardest part of going on the type of journey that you've been on. And I'd love to understand who you were as a child, where you found a love for filmmaking, for documentary work, for history, and sort of go through from there.

Well, I was born in tragedy. My mom had cancer almost all of my life and died when I was 11. And a few months later, when I was 12, my dad, who had a pretty strict curfew, would allow me to stay up late at night watching TV shows, even on a school night. And he started to cry at an old movie called Odd Man Out. And right then and there, I'd never seen him cry before, not when my mom was sick or not, you know, when she was— when she died and not at the terribly sad funeral. And all of a sudden I said, I want to be a filmmaker. That meant Hollywood. But I ended up going to Hampshire College in Amherst, Massachusetts, a brand new and experimenting— still there, still experimenting—

school.

And most of the teachers were social documentary still photographers and reminded to me there is as much drama in what is and what was. And all of that sort of combined by age 22 to realize that I had a latent and untrained interest all of my life in American history. And 50 years later, I'm still doing the same thing.

Walk me through your process. Do you have a list? Like, has your filmography played out the way you would have predicted a decade ago? How many How many twists and turns have you run into in your career? Zero.

I mean, obviously, life throws a lot of stuff at you. I live in rural New Hampshire. That's what I made a decision to leave New York City because I thought that becoming a documentary filmmaker, strike one, in American history, strike two, on PBS, strike three, meant I'd taken a vow of anonymity and poverty. And I moved up here, finished the first film on the Brooklyn Bridge, got nominated for an Academy Award, and everybody said, "You're coming back." They'd warned me that if I went up there, I'd die. If I stayed, I would die. And I listened to neither. And it's made all the difference. So, I've been able to, working in public broadcasting where I couldn't have made any other films, any of the films anyplace but, one foot tentatively in the marketplace, the other very proudly out, that I've been able to have director's cut for the next 40 films, all of which have been broadcast cast first on PBS, streamed on PBS. So it's been good. So you might have a 10-year plan with lots of different projects. Sometimes you may, before you'd really gotten into something, switch one subject to another. But we've been hitting the marks for a pretty long time. The real struggles are within these, and you want to leave the kind of room to experiment, to be thrown a curve, and to figure out how to do it. The recent film on the American Revolution, which now has over over 4 billion minutes of watching. The first time PBS— first time PBS ever entered the top 10 of streaming programs at the end of November with what I thought was a ridiculous number, 565 million minutes, is now over 4 billion. So there's some pent-up curiosity about who we are, where we've been, particularly what our origin story is. Uh, so it's, it's been a long, interesting thing, but every single one of the films, a director cut, and that's only possible because I've stayed in public broadcasting.

And then in public broadcasting, I imagine, we were just talking about some of the changes that are happening in Hollywood and how there's often a pressure to go bigger, and that usually means bigger budgets. Okay, you made a $20 million film, your next one's gonna be $70 million. How have you thought about where the business changes, what more money means, what more success, what more confidence means? How has the business side of your work evolved throughout your career?

Well, we've tried not to let the technological tail wag the dog, to pick up one aspect of your excellent question. You know, we waited 10 more years when many of my colleagues had already switched to digital editing around 1990. We waited till 2001. We didn't stop film until 2009 and then switched to digital stuff because we didn't want that tail to wag the dog. At the same time, our model is, you know, an interesting one. It's all grant-funded. We don't have investors. So you write a budget. It pays for your salary. There's no profit margin. There's no contingency. And you empty that bucket, the film's either got to be done or you got to be raising more money. So they're very expensive. These are labor-intensive things, particularly when you take, you know, 10 years to do something like the Vietnam War or the American Revolution, so they're pushing $30 million. At the same time, I raise it all, so I retain that creative control and I relieve PBS of the burden of having to fund more of it than they already do. And so it's a kind of a good situation to be in, and it's really outside of the pressures of the marketplace. So I don't think about increasing budgets. I can go from a $30 million budget to $10 million, and I'm still— these budgets are are high for my contemporaries, but they don't get to spend. And look, I could have gone and saved myself all the trouble of raising, say, $30 million for the Vietnam series that came out in 2017. 10 and a half years. I spent 10 of those 10 and a half years raising money for it. I could have gone into a streaming service or a premium cable and gotten all that money in one pitch, but they wouldn't have given me 10 and a half years. And that's the point. I can't tell the story that I want to tell with the depth and complexity. I mean, that film is now— let's do the math— that is 9 years old, and it is still probably one-stop shopping, even among scholarly works, to aggregate all of the new scholarship into one place. And nothing certainly replaced it in film, and nothing that I know has replaced it in, in scholarship. So we're feeling pretty be good about that.

That's great.

Has fundraising gotten any easier? Is it just as hard every time? Did you ever kind of hit a moment where you decided to maybe pick up, return the calls from people that wanted to be more traditional financial backers in a particular round?

No, no, because the sacrifice in control is You know, so we— as the films got longer, you know, in the beginning, the first film on the Brooklyn Bridge cost $180,000. I probably made 2 cents an hour over the 5 years I did it. So there are more zeros on the other side. So I guess it's proportionally harder. I do have a good reputation, and one I wish to maintain. And so we're not looking for compromises or ways to sort of get that cash in hand. You sort Part of the duty of running the marathon is to run the marathon. And the reward is in not only crossing the finish line, but every single moment along the way, the million steps that you take where one part of you is saying no and another part, maybe weak voice, is saying yes. And so these are often really complex struggles in which the analogy of a million problems is not an exaggeration. You just can't see a problem as pejorative. You just see it as inevitable friction to be overcome, like picking up that step and making another step.

After the adoption of digital editing, what was the next technological advance that you found useful?

I don't know. I'm a Luddite. I don't know what I'm doing on your show, but I'm really glad you reduced I'm really glad you reduced it to initials, because, you know, what is— there's nothing— people say PBS, nothing could be less, you know, sexy than that, but Technology Business Programming Network? I mean, KFC has behind it Kentucky Fried Chicken. You know, I knew all my life, or most of my life, what it was. Now it's KFC, and I kind of— but this is a really good move. No, I went reluctantly to digital editing because During the '90s, a lot of the interns that would come and work for us, they could see working in analog as some sort of badge. But about halfway through the '90s, I realized these kids want to be race car drivers, and I'm telling them how to shoe a horse. It just doesn't make any sense. And so eventually, almost for them, we switched to digital editing. It's been incredibly interesting. There are still some things, you know, stuff I might tell my editor that I knew would take 2 weeks now take 2 days and things that took 2 days take 2 hours and things that took 2 hours now take 2 minutes. And there's, you think there's salvation in that way, running to daylight there, but there are as many traps, as many places of quicksand, so you just want to be sort of careful about how you approach it.

What is your opinion on digital restoration of archival footage, archival imagery, technology seems to be advancing there, and yet there's risks.

Yeah, there are a lot of risks. You know, a lot of filmmakers are going back to original film masters. They are taking a digitized thing, and one of the reasons why is that you may have a 1 and a 0, which is absolute in one way, but as you change the programs, you need to have the machine that did the original thing, and now they're not there. They're in a museum someplace, and you have to go get that thing and figure it out. So I know some filmmakers that are— I can't do it just economy-wise because I've got 250+ hours of programming, you know, and that's reels and reels. Every half hour is about a reel, you know, of stuff. You can do the math yourself. There's probably, you know, 500 reels in all of this. So it's almost a prohibitive expense. And so you're hoping that your next iteration of Masters will be able to read not only the contemporary films that you're working on, but the ones you made 20 years ago and update it. But the restoration part of it, to be able to painstakingly do that sort of stuff, has been really terrific.

Can you tell me the story of the Ken Burns effect?

Yeah, sure. It's pretty interesting. I got a call in November of 2002 from a guy claiming to be Steve Jobs. You know, and I went, okay, yeah, right. And then I realized very quickly it was Steve Jobs. I live in rural New Hampshire. I said, okay. He said, would you come and visit me? And I said, yeah. So a few weeks later, in early December of 2002, I visited him in Cupertino at Apple headquarters. And he ushered me into a room with two engineers. And he showed me this stuff. And as I said, if I'm a Luddite now,— you can imagine what I was like in 2002. I could barely do word processing. And he said, "So look, we figured out a way to download or upload whatever it is, your pictures that you take, and that we can do all this." And they showed it to me and I was kind of like, "Okay." A fairly superficial version of what I've tried to do to wake up the past, not just visually, to explore energetically the surface of the photograph you're doing. In a way, the feature filmmaker I used to want to be has a master shot, a long shot, a medium shot, a close, a tilt, a pan, a reveal, insertive details. But here was just a very simple sort of panning, zooming through stuff. You could add it, you press a button and add a music thing. And I said, you know, okay, cool, you know, whatever. And he said, well, you know, Next month, in January 2003, every Mac computer from now on will have this. I said, okay, cool. And then he said, and we want to keep the working title. And I said, what's that? And he goes, the Ken Burns effect. I said, I don't do commercial endorsements. And he went, what? And the two engineers kind of blanched. I think they were, you know, aware of this titanic temper, which I never saw. "Come with me." So we go back to his office, we talk for a while, and we end up becoming friends for the rest of his life. And when I was in the Valley, I'd stay at his house on the futon above the garage and the flats at Palo Alto and knew his wife and his daughter. We even gave an internship to his daughter, Lisa, the famous Lisa. And, you know, it was a terrific friendship. But what I did is I walked out an hour later with about $1 million of hardware and software, which I turned around and gave to nonprofits. Usually Final Cut Pro to schools that needed it. You know, computers that nonprofits needed. I think a couple fell off the truck because we needed a computer. But 99.9% of it went to a good cause. And I think he was intrigued that somebody in a world in which it was all a 1 or a 0 for him, or somebody would say, okay, I'm going to— you should charge. When the world tells me, oh, you should have charged him for this, I said, then you call it the pan and zoom effect. I wasn't interested in having my name on it insofar as it could help other stuff. It ended up giving me— and way too short in that regard— a really good friendship with not arguably one of the 5 most important people in the United States in the last 150 50 years.

It's remarkable. What is your process for truth-seeking and understanding history?

So, I have in my editing room, and I've had for years and years and years, a lowercase neon sign that says, "It's complicated," in cursive. And that can mean a lot of things. One is when a filmmaker has a scene that's working, you don't want to touch But finding out new and destabilizing facts obligates you to do that. It's not true of every place, but it does in our shop. And so, I think pursuing truth isn't the question. It's there. You can find out the facts. You just have to wade through a lot of shit because today, everything is just dump truck after dump truck of manure dumped on top of other manure. It's just hard. It's one of the reasons why— mostly it's to manage a complex narrative with lots of different balls in the air and hundreds of characters in the case of the American Revolution. Not just the boldface names who are, you know, just statues, you know, out in the park collecting pigeon shit and not real dimensional complex human beings. We try to do that, but introduce you to scores of other characters and set a narrative in motion that's at its heart a military one, but is also adorned with very complex political and social and geopolitical and economic sort of factors. And so that takes a while, but also making sure you're right. I mean, we lock the film, meaning we're not gonna do any more work on it, we promise. We promise the sound editors, and we unlock it, 'cause we found out that word 16 battleships, deaths, months. We've got a little footnote and it says 2 scholars of repute said that that that, you know, making this up, said that that was right. And then you find out a third scholar who says, we're not sure it's 16. So you scour— the narrator's already read and gone home and you're not going to get him back. So you scour all of the written stuff, you know, hours and hours of stuff, and you find a perhaps, and you copy it and you move in front of perhaps 16, and then you sleep better. Like, that's the kind of level of truth-seeking, as you put And I don't know any place else that kind of is that kind of concerned about that level of accuracy. But if you add that up a billion times, then all of a sudden you have something— get ready to blow your whistles— that has 4 billion-plus minutes of watching.

Amazing. Another sound effect.

Thank you.

Can you help me understand some of the throughlines, emotional elements, or takeaways, or the fingerprints that you try and leave on your work across projects, not just one single project.

That's a wonderful— thank you. That's a very thoughtful question. When I was working trying to raise money, and I looked like 12 years old, I was trying to get people to buy the Brooklyn Bridge. I remember writing a sentence that I was not interested in the mere archaeology of dry data. And facts and events. I was interested in an emotional archaeology. Now, I didn't mean sentimentality or nostalgia. That's the enemy of good anything. But I'm looking for that thing that animates art, that animates faith, that animates a lot of the most intimate stuff that we care about. And that's where you say the whole is greater than the sum of the parts, but you never ask yourself, what's that difference? Where 1 and 1 always in your work has to equal 2. 2. In my work, you want it to equal 3. You want to have that unknown factor. So I guess once when we were working on our jazz series, the trumpeter and composer Wynton Marsalis said, "Sometimes a thing and the opposite of a thing are true at the same time." I'll relate it back to the Revolution. George Washington is the indispensable man. Without him, we don't have a country, period, full stop. And in a world in which we're more interested interested in bottom-up than top-down. It's exhilarating to be able to resurrect somebody. But he's also deeply flawed. He owns hundreds of human beings in his lifetime. And as the writer Rick Atkinson says, you know, you can't square that circle. He is rash on the battlefield, risking his life and therefore the entire cause. He also makes some pretty bad tactical mistakes at the largest battle, Long Island, of the war. And then later the same tactic, leaving a flank exposed at Brandywine, which the British take full example of. But, and this is the biggest but I know, he's able to hire subordinate talent that's better than him, and he's not threatened by it. He's able to inspire men in the dead of night to fight and to remind them that they're not from New Hampshire or Georgia, but they're a new thing, Americans. He knows how to defer to Congress. He has incredible humility. And more than anything, he just waits it out. He understands he doesn't have to win, he just can't lose. And he does that successfully and defeats the greatest empire on earth. And even more than that, he gives up his power twice. The second New York is the British leave New York in November of 1783. That is like 8 and a half years after Lexington and Concord. He goes to Annapolis where the Congress is meeting and resigns his military commission. Commission. He's brought back out of retirement to run the Constitutional Convention, and then is unanimously made president. And after 2 terms, he walks away from it. And as George III, the King of England, said, if he did that, then he's truly the greatest character of the age. And so, you can have both. And a problem in a binary world in both computers and media, we want to have a black hat and a white hat. We want to say good or bad, and you We can't do that in our own personal lives, and we ought to extend that to our history.

Yeah, the dialectic. I love it. You have been very fortunate in that it feels like there's no project too big for you to undertake. And I'm wondering if there's ever a desire— how you balance going big with a massive story that everyone has at least some conception of versus going back to something like the Brooklyn Bridge and refocusing on something very narrow. Is that something that appeals to you? Do you feel like you've graduated, you've done that, you never want to go back, or is there something— Talk about the tension there.

Yeah, well, that's another excellent question. I mean, within these big, huge epic series are all the intimacies that were contained in the Brooklyn Bridge. And we do have a kind of exhalation. They're big series. We could just say Civil War, baseball, jazz, World War II, national parks, the Roosevelts, Vietnam, country music, and now American Revolution. But within them are— I think that's fewer than 10 films. There are at least 30+ films that are shorter than that. And so, they experiment more with what you're talking about, that earlier experience. But again, the more important point is that any given minute in a film, say a series, a lot of people would say, "Okay, you're the director, but you'd assign a different producer for each of the," in the case of The Revolution, 6 episodes. We don't do this. We all do work on all of them, and they all are shaped. Same with the editors. They cross episodes, and they do more than one. And so what happens is you have a kind of unity of presentation, but within each Each scene, you have the same kinds of intimacies, the same kinds of small moments that you would find in Brooklyn Bridge. And you're also elevating not just George Washington to a more dynamic and complex and understandable and real, truer version of who he is, but you're introducing people to scores of others that you've never heard of, 'cause I never heard of them. You know, a 10-year-old girl named Betsy Ambler, who's a refugee most of the war. We don't think of them Americans as being refugees unless you realize how bloody this was, what a civil war as well as a revolution it was, what a global war, and you're dealing— that involves more than 2 dozen nations, European as well as Native American. And when you say Native American, you don't say them. There are all these nations to our west that are as distinct and separate culturally, economically, diplomatically, militarily as France is from Prussia at that time. If you think you're going to just make it them and not say French, Prussian, Delaware, Lenape, Cherokee, you've made a huge mistake and left out an important dynamic. Or take half the population, women, and say, yeah, we don't really care about that. Or the 500,000, 20-plus percent of the population that is enslaved or free Black Americans. All of that makes an unbelievable dynamic. And if you add the violence to something that we say is this sort of bloodless gallant myth that attends the Revolution, just guys in Philadelphia thinking great thoughts, and they are great thoughts, as great of thoughts as ever been thought, but it ain't the whole story. And when you try to integrate that whole story, man, you start to have something that fires on all cylinders. And that's what wakes me up every single morning of every day.

I think about what has changed in America throughout my life. I go to Apple, I go to technology, I go to business, and your work has focused, you know, you've obviously touched on business, but you focused on politicians and generals and athletes and musicians and artists, and I'm wondering about your perception of the business community, the technologists, over the history of America, their interaction with your other subjects, how you want to integrate them into the stories that you tell?

Well, you know, I've done 4 wars. It seems disproportionate because they're all big long series, but wars are huge sponsors of technological change.

All you have to do—

I made a film on the Second World War called The War, and it ends because of one of the greatest technological changes that you could possibly imagine. But the Civil War happens because back in the 1700s, an industrialist, Eli Whitney, who will invent the kind of exchangeable parts in the beginning of an assembly line thing, goes down and sees a cousin who's lamenting how long it takes her enslaved people to card a pound of cotton, and he invents the cotton gin, which revitalizes Cotton and slavery and set in motion the wheels economically and technologically that will sponsor the Civil War that's upcoming. So it's all the way through that. And we're constantly thinking about and dealing with those changes. And one of the stuff in my head full of ideas, if I were given 1,000 years to live, which I won't, I won't run out of topics in American history is invention and technology. And I'd start I'd start with Eli Whitney because there's an underbelly to it. And I'd go to Steve Jobs because there's an underbelly to it. I got kids. I am— we're all addicted to this. And I mean, my new statement is, and it's such a Luddite statement, I apologize in advance for offending your sensibilities, but I say social media isn't and artificial intelligence is. They're not social. Ever been in a room of people or a subway car or a block where everybody's on their phone? That is not social. And artificial is artificial is artificial.

Yeah, no, that makes a lot of sense. I'm sure you get asked this question all the time, but walk me through some of the advice that you give for aspiring documentarians.

It's really so bad and platitudinous. That I'm embarrassed almost to repeat it because there's no laws. And everybody I know that is a working documentary filmmaker, meaning they don't have a side teaching gig, you know, to pay the bills, that are actually making, you know, living, supporting their families, have all come at it different ways. And so I realize that there's no set career path the way there might be for a feature filmmaker or a doctor or a lawyer. And so what I tell them is one is Socratic. You got to know yourself, right? Who am I? Film is intensely glamorous for teenagers and at a college level, the first, you know, the introductory classes are jammed with people and then all of a sudden next year, it's pretty winnowed down and there ought to be a mechanism within us. This is the Socratic dimension that says, you know what? I don't think I have something to say. I thought this was going to be easier than it actually is. I see how difficult it is.

Is.

And then the other one has to do with that difficulty that there's way more talented people than there are possibilities of making a film in this economy, even with the explosion of documentary films now, and some of it quite dreck. So obviously, some of those people who didn't deserve to get a budget got one.

Is that—

is that you have to persevere. I mean, I used to keep on my desks 2 3-ring binders filled with hundreds of rejections just for that first film on the Brooklyn Bridge, just to remind me, you know, that it doesn't come easy, you know. And that was sitting juxtaposed from a little thing on my door, a quote from a theater impresario in Minneapolis Globe Theater named Tyrone Guthrie. He said, "We're looking for ideas large enough to be afraid of again." Not a very good English sentence, but it just means just bite off more than you can chew and learn how to chew. And a lot of that is what I I try to impart when people ask me the questions and I warn them in advance that it's just going to sound like— I'm sure I've just described things that you guys know about what you do. And I'm sure everybody who's listening knows, "Yeah, those kind of apply to what I do." And so, I think there's— basically, in his essay on self-reliance, Emerson says, "Do whatever inly rejoices." You need to actually How do you actually figure out what this thing is that is, you say, "I" to, of which there are millions of "I"s, and which one is the authentic? And how do you proceed to do whatever inly rejoices, as he said?

Is it refreshing to change centuries? Do you think your life would be different? Do you think the projects, the quality might suffer if you had at some point when you realized that this was gonna be the career, you had started linearly and done American Revolution, Civil War, World War II, then Vietnam, and you just go year by year as opposed to bouncing around a little bit.

It's not even bouncing around because only in retrospect can you do that. Yeah, exactly. I just— my best friend was a book salesman and I was sick and he threw on my bed a paperback, a trade paperback version of David McCall is The Great Bridge, the epic story of the building of the bridge. I read it in one gulp. I said, we're gonna make a film about it. And then, you know, we're thrashing about to what to do, and my future wife and I are driving through western Massachusetts, and we pass a Shaker village outside of Pittsfield, Massachusetts, with this beautiful round stone barn. Slammed on the brakes and said, who are these people who made these buildings? That was the second film. Another guy was trying to convince me to do a film on, on Huey Long, the turbulent Southern demagogue. And someone else said, "Boy, you know, you did a great film on the Brooklyn Bridge. You know, the Statue of Liberty's 100th birthday is coming up. Why don't you do something on that?" I said, "Yeah, but I'm not gonna make your film. And so, if you give me the money, I'll do it, but I'm not—" You know, so Liberty Mutual Insurance, before they got the fucking emu, gave me money for the whole thing. And I made a very complex film, which also got nominated for an Academy Award. Award. But each one of those first 4 films come from different things. And so you begin to sort of make choices. So at any given moment, you know, and I didn't think until I was in the middle of Vietnam— not the middle, at the end of it, we were locking it— I was looking at a map we made of the Ia Drang Valley in the Central Highlands, and you could see the positions of the Viet Cong and the NVA and the Americans. And I thought, boy, this could be the British moving west on Long Island towards Brooklyn. And I looked up, I said, "We're doing the revolution." So, everything comes that way and it's much better. It's like a batter, you just got to read the pitch. And at that given moment, it's a fastball, so you're going to hit it this way or hit it to the opposite field, or you're going to try to choke up because he's throwing junk and you got to figure out how to respond. I mean, it's literally— I'm so glad there's not a plan. Or chronological stuff. And you can do is take those 40 films and then put them in chronological order.

And that wastes some of your time. I just have to pass something along. A lot of people in the chat are saying that they love your work, obviously, but particularly that they became much closer with their fathers through watching documentaries.

Yeah, this is true. And mothers too. Particularly with regard to baseball. I don't know how many people come up and said, I've seen baseball more than you have. And I go, what are you talking about? They said, well, every single January during the Hot Stove League, my dad and I would watch it. And my dad just passed away, but my son is old enough. I mean, I've had so many moving things about that or mothers taking kids to baseball games or going over battlefields. So there's a nice— we talk about how divided we are right now. All you have to do is look at my films and see we were way more divided at other times. And what brings us together are stories. The power of stories neutralizes that dialectic that you were mentioning, neutralizes this false notion that there is a simplistic binary that we can impose on everything and figure it all out. You can't. And the complexity of us in the end reminds us that there's only us. And know them. So I make films about the US, but I also make films about us.

The chat also wants us to ask about AI. I already have an idea of your reaction. I appreciate your approach of not letting technology guide the work, but are you at all excited at some point in the future of being able to recreate a scene from history using image or video generation tools? Not really.

I don't think so. I think every once in a while when you're trying to run down a portrait of Benedict Arnold, who's not a popular person in the United States, so there's no real visual evidence left of him, AI can be helpful in locating that. You know, I use it when I Google a question. I notice how quickly the AI comes up, I have to take it with a grain of salt right now. It's not that it's wrong, it's just you may not be offered all the nuances and possibilities. That's going to change and whatever, but not to write a thing or to generate an image. There's a sanctity. I mean, if you say a picture's worth a thousand words, then if you make it up, then what is it? I mean, when everything is permissible, what's true? And I just— I mean, that goes back to your question. You know, what is true? And I think it's really important to at least have— you know, it's like a kid. You remember this when you're young, 4 or 5 years old, and your mother or your dad sitting on the bench at the park and you're playing, but every once in a while, you got to go and touch home base. I want to be that person. Yeah.

Very cool. You mentioned McCullough, and I'm a fan of Robert Caro.

Yeah, we just interviewed him for a film we're doing on LBJ and the Great Society.

I can imagine. Are you excited about where that genre of historical writing is going right now?

Oh, God, yes. I mean, there are so many heavy hitters. I've been out on the road for the last year, a lot of it with Rick Atkinson, who's two-thirds of the way through his trilogy of the American Revolution, is the first talking head you see accidentally in our film. No accident. And the last talking head you see, there's Doris Kearns Goodwin. David McCullough has left us, Stephen Ambrose has left us, but I was just on a podcast with Ari Emanuel and Walter Isaacson yesterday talking about, you know, what the Mount Rushmore that wasn't a president would look like, and having great arguments with him and his friend Ben Persky, and the four of us, you know, arguing and fighting for how the fast fastest hour that I, that I've spent. So there's so many, there's so many good writers. Walter's influenced two of my films, you know. I, I, he's one of the biographers of Benjamin Franklin, and he also did our— he also wrote a biography of our first non-American topic, uh, Leonardo da Vinci. Yeah.

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

This is very informative.

This is really— yeah, John has done a great job keeping it together, but, uh, I I think he's definitely one of your biggest fans. And you've come up a lot. Oftentimes, when I can't get a hold of John—

He's probably watching. And yeah, we watched so many documentaries together. Together?

It was fantastic. What's her name? Marjorie. Marjorie, thank you for watching. Good job.

Thank you. Great stuff, Ken.

Thank you for joining. All right. Be well.

Cheers. Have a great rest of your day. Goodbye.

Thank you.

Let me tell you about Gusto, the unified platform for payroll benefits and HR built to evolve with modern small and medium-sized businesses. And let me also tell you about Gemini, Gemini 3.1 Pro. I'm gonna have to update this soon, uh, with a more cable-based baseline. It's great for visualizing super complex tasks, synthesizing data into a single view, or bringing creative projects to life. And without further ado, we have Gokul Gurram from Marathon Management in the studio.

How are you?

What's going on?

Good to see you.

Great to see you, John. Welcome to the show. Is software dead? First, introduce yourself. Sorry.

We'll get to that. Hey guys, I'm a huge fan, a long-time fan and listener to you guys, to TVPN. I'm Gokul. I've been an operator and investor for 25 years in the Valley.

And yeah, Excited to be here with all of you.

So yeah, it's about time.

Yeah, about time. Long overdue. Yeah. Walk us through how you have been processing the SaaS-pocalypse, the Citrini article, all the different narratives that have kind of shaken the tech industry, both to the upside in many cases and also to the downside in others throughout the start of this year.

Look, I think software is going through a change. I don't think it's dead. Software is still going to be around. In fact, every industry industry is going to use software at its core. I think what is going to be challenged is companies that are surface level, where they're going to be lightweight, which are easy to rip and replace. I think of, I think of software as two kinds. One of them are lightweight, seat-based, easy to rip and replace. A good example, I always say I don't want to pick on them, is Zendesk. Zendesk is customer support software. They sell on the basis of seats. And use it for delivering outcomes for customers. The problem with Zendesk is that if I'm a customer of Zendesk, I'm buying 50 seats. I could reduce it to 20 seats and use 30 seats for AI. In the same budget, 30 seats, I can use AI. So I can literally compare Zendesk's performance to AI agents in parallel, and Zendesk won't even know this is happening. So you might be running an A/B test, and you can literally run Zendesk in parallel. And that's a challenge for a company like Zendesk because it's not not deeply rooted can use it part-time for only part of their employees or part of your customer service requests. The other kind of software is stuff that is not outcome-based, that is truly running your company, that's really running a process. A good example is NetSuite, which is an ERP, which is literally at the core of your general ledger, is running your accounting. A lot of your business is running off it. You can't really run another version of NetSuite in parallel to NetSuite. You've got to And so NetSuite is in a much better position than Zendesk. Now, at the highest level, the reality is every piece of software has to reinvent itself over the next— but the time frame, it's much more urgent for Zendesk to do that in the next 6 to 12 months, while something like NetSuite probably has years, 3 to 5 years to think about it. So if I were NetSuite, what I would do, I would create AI agents, which I'm sure they're already doing. And I would, as they say, I would say commoditize my complement. I basically make it free to actually have a bunch of interesting workflows, which are offered through AI. And I would basically still charge for the data. Very, very different.

And how did you, there's a whole new class of AI-native ERPs that have come online and raised over the last 1 to 2 years. How did you process them? There's certainly like a long graveyard of startups as well, but part of the challenge is there's just so much software to build with a system like that, and maybe now is the right time given that we can create software a lot faster than we could historically.

Right, I think the— I know many of those companies, I think they're great entrepreneurs. I think the challenge for them is that there is no pressing burning reason for somebody using a Netflix— a NetSuite, I should say— to rip and replace NetSuite. So what you have to do is you have to go with a company that doesn't have an ERP today and get them to start using you as a soft— as a software company, as a vendor, and then grow with them. And so I think it is— so that's the strategy: build your customer base from net new companies which don't have an ERP today. I think going to somebody who's using a large ERP and say, well, throw out your ERP and replace it with me is hard. So you've got to either come up with a very, very, very narrow wedge where you can inject yourself in and then use it to over time take on NetSuite. But more interestingly, I think go after younger, fast-growing startups which will need a NetSuite or an ERP system in a few months, offer it to them for free. Some of them actually, some of these ERPs have been saying if you're funded you're funded by one of these VCs, you can use the first year of my product for free. Sure. And that's a smart idea because you're using it for free, you're not going to generate much revenue, but in exchange, they become sticky. And when you actually start generating revenue a year from now, they're there for you.

Yeah. What do you think is happening in the private markets for software companies where their public comps have declined 50% there. We sort of went through this, and I'd love to know how you process the end of ZIRP. I remember reading Logan Bartlett on this at Redpoint a lot. He was pulling public comps and seeing how multiple compression had happened in the public software SaaS market and the effect that that was happening in the private markets. It feels like the private markets haven't fully digested the SaaS-pocalypse in the same way because a lot of the startups in the private markets that are going after those comps were sort of AI native or earlier on the AI narrative. But at the same time, there's like a bigger question there. That's right.

The way the private markets work is what can the private markets do? They can't re-rate the company on a daily basis like in the public markets.

So what they can do is— That term sheet I signed is due Do you for a wire? At what valuation? But I'd like a 50% discount, please, because the comp is down 50%. Yeah, you can't exactly do that.

You know, that happened in 2008 where you did see when the 2008 Lehman thing happened, you did see people pull term sheets right after that happened that were outstanding. But those kinds of things that happened on the edge. But the biggest thing that's happening is that a certain class of business models that have maybe raised a seed or an A are now unable to raise a B or will be unable to raise financing. And those are business models that are primarily application layer. I think if you're just doing workflows, it better be a deep and complex workflow that is deeply, deeply, deeply embedded. Because now with OpenClaw emerging, you guys might have seen this, yeah, OpenClaw actually does the work. Yeah, it's not just kind of workflows, it actually goes and does the job. Yeah, so we're just putting software workflows that's enabling other people but not doing the job yourself, it becomes harder to justify that you will exist. Because remember, the VC time frame is not a year. It is 5 to 7 years. And if you think about compounding over 5 to 7 years, that's insane. Why? Because the capabilities of these models are doubling what, once every 6 months, say? Yeah. And if once every 6 months, it's quadrupling every year, that means in 5 to 7 years it's 2 to the power of 12, which is 1,000-fold better than what it is today. And my human brain is unable to comprehend what it means means for these models to be 500 to 1,000 times better than they are today, what it means for software. So I think everybody's trying to figure out over a venture timeframe, what does this mean if you extrapolate the problem of these models and make them 1,000 times better? What survives and what doesn't? What survives? Things that are deeply rooted, that have atoms. Yeah, we think as software investors that things that have atoms— yes, there are humanoid robots, but I personally feel general-purpose humanoid robots that machinists or HVAC repair people, whatever the case is, that's going to be slightly further out. So things that have atoms are generally protected over the 5 to 7 year timeframe. By law, things that have regulatory moats are protected. Things that have money flowing through them are protected. So there's a few classes of things that are protected. Yeah.

If somebody came to you and wanted to vibe code DoorDash, what, what advice would you give them about the challenges that they might face on building a competitor?

You know, I'll give you— I'll give you a story. I'll give you two stories, actually. Assume it's June 2028, right? Which is what I think Citrini had for 30% drop in stock. Assume you're hungry. You open your phone, you ask Claude, say, or whatever your new startup is, the food ordering startup, to order your sandwich with gluten-free bread. It routes you directly to the restaurant because it's figured out how to connect to all the restaurants and it saves you $2, but it takes an hour and a half to show up. Up, it was cold, and they forgot to make the bread gluten-free, so you can't eat it anymore. Who owes you a refund? Does Anthropic have a customer service number? No. So basically, you open DoorDash, the sandwich is the first recommendation, you tap on it, 90 minutes later you're holding your sandwich still fresh. And guess what? They remembered the gluten-free bread. So that's one story where if something goes wrong in the physical world, who does— who takes care of it? Yeah, DoorDash is a managed marketplace. I'll back to that. The better, more pertinent story is around 2016 or 2017. Google launched a product called Google Food Ordering. You might have seen this. Google Food Ordering, you can actually order from a restaurant, and it literally— it was driving at that point multiples of traffic to any restaurant compared to what DoorDash was driving, because Google, still massive back then, was massive. You could search for a restaurant, you could order from directly from Google. However, the retention of people who ordered Google Food Ordering was basically zero. It was, it was negligible. Why? Why? Because after you place the order, there are tons of things that go wrong in the physical world and you couldn't really call Google. The restaurant didn't take responsibility there. So that product doesn't exist today. And I think that's the fundamental reason. There are things that happen in the digital world that are not enough to replicate something like DoorDash. DoorDash manages not just aggregation and supply of restaurants. Manages everything afterwards. It manages figuring out how to get you out to the right, right dasher. It manages how to basically refund if there's something that goes wrong. It manages customer support in case you want to use it. There's a collection of systems that work together. Discovery is something that I think agentics will do well, and hopefully DoorDash will incorporate many of those. But discovery is not enough. If you think about Google and what did Google disintermediate, what do we use Google for today? It did really well with Expedia and flights and hotels and also things like mortgages. So the types of marketplaces that Google has been able to disrupt or search are things that are more lead gen or just, I would call, light marketplaces where payment happens to Google. You can book hotels or flights to Google. Yeah, you don't have to take care of what happens afterwards. But a managed marketplace in the physical world is the most durable thing. It is not a piece of software. And this is the thing I think people get wrong about DoorDash. DoorDash, the soft— DoorDash, the first piece of software they built after the website. If you remember, DoorDash had this Palo Alto delivery, it had this website with a PDF. The first piece of software they built was not a consumer app, it was a Dasher app. It was the app for Dashers because that was what was needed to get the food properly, to make sure the Dasher got the order, accepted the order, picked up the food, delivered it before the consumer.

Can you talk a little bit about what it takes to build liquidity on a marketplace? It feels like that's also underrated. If you're— if you need to spend billions of dollars to actually create the scale of a marketplace, that that's something that can't necessarily be vibe-coded. Exactly.

So it's the biggest moat here. The one of the moats is density. Density leads to liquidity, which means the more orders you have in a given geography at a given time, the better you can batch deliveries. Which increases driver utilization and reduces cost per order. And that's what you have to do. If you were to actually pay the actual cost of delivering, it would be $10, $12, $15. The reason you're only paying $2 or $3 is because these orders are batched and you're basically not having a single person. You're basically batching it to the best Dasher driver. That's why you might see in, in busy, busy times that there are drivers waiting outside certain restaurants because they literally know that that order, they can pick it up and there is massive density of orders. So density and liquidity, that's what drives marketplace economics. And so whoever wants to replicate DoorDash doesn't just need to improve aggregation at the consumer level, but build a massively dense geography by geography marketplace in every city and every suburb that DoorDash operates in. And that's a massive endeavor. That's what DoorDash spent a few billion dollars building.

Yeah, just a few. What does it take to be a great public company board member?

It's interesting. I think, you know, it has changed over the last several years. I think a few years ago, a decade ago, say, the profile of a board member was quite uniform. It was people from the finance or corporate worlds. And that's still important. I think you need a finance person who runs your audit committee. You need a CEO. I think every public board member, if you would go to any public CEO and say, what is the archetype you want on your board? Most of them want CEOs. They want to see you on the board because you want to see who you can confide in and you can ask for help from. But there are new archetypes that have emerged over the last few years. One, product engineering leader, someone who's led product and engineering at a company. Why? Because every company, like I said, has become a software company, a technology company. So every public company wants someone with product engineering jobs. Second, you want someone who represents your customer. At DoorDash, for example— oh, sorry, at Square, we had the CEO of Shake Shack. And why? Because Shake Shack was a prototypical customer for Square. Uh, no, Danny's— it was Randy Garutti. Oh, Danny was a chairman. Yeah, so Randy was— Randy was a CEO. I mean, yeah, Danny was a founder and Randy was a CEO. And so he was excellent. He always— he would really assume. And Typically what happens is, especially if it's a company that came through venture capital, typically one or two of the VCs that have been early in the journey of the company stay with the company and the CEO trusts them. In Square's case, it was Roloff Botha at Sequoia. In DoorDash's case, it's Alfred Lin, ironically again from Sequoia. But in Pinterest's case, it was Jeff Jordan from Andreessen. So each company has one VC probably who— and then in Coinbase's case, it's Fred Wilson and Marc Andreessen from Union Square Ventures and Andreessen Horowitz. So we see product engineering chops, voice of the customer, CEO, finance person. Those are the 5 archetypes. One of the best practices today that I've seen companies using is a notion of a board buddy. What it means is every exec at a company should actually be paired up with a board member who's an expert in the same area. So for example, I typically will work with the head of product and the head of engineering at the companies I'm on the board of, and I meet with them once a month outside of the board meeting. Those meetings are actually really good. I am essentially listening to them in a judgment-free zone and trying to be helpful and getting feedback on them. How can we be helpful? And the same is happening with the finance person, the finance-oriented board member. They meet with the CFO, the customer board member, they move to the CRO, etc., etc. So I think, uh, that board buddy system I've seen helps dramatically because that's what your board member should be helpful, not just the CEO but the exec team. It also helps get exposure for us to the exec team. Who they are. And as we think about succession planning and stuff like that, we can evaluate who these people are and who should succeed the CEO, who are the people.

Yeah. What are your forecasts around M&A in 2026? Last year was the year of the licensing deal. The zombie acquisition with the ghost ship. Yeah, the ghost ship acquisition. But it still feels like we could see a lot of activity. This year.

A lot of cheap companies out there. GoPros at $150.

There's a lot of— well, yeah, in the public markets, but also, but also public companies that want to bolt on fast-growing teams.

Yeah, I think there's going to be both. I mean, you saw, for example, MyFitnessPal. Yeah, bought this company Cal AI, and that thing is 18 months old, right? $50 million. And so I think there's going to be a tremendous amount. Why? Because like you said, there's a few things happening. One, software companies are valued and the valuations has gone down. And so I myself, in a portfolio, have seen if the CEO has courage, they can actually— you're— the stock market is already telling you that the terminal value of your business is zero. So there's really no downside to burning the boats and going all in on AI. Yeah. And I have a company, Podium. I mean, you should actually get the guy, he's amazing. He's similar to Intercom, very similar to Intercom. Sure. He had a business that grew to $200 million was kind of growing very slowly. And he said, you know what, I'm going to go build an AI agent. I'm going to— it's basically a product for SMBs. And he's basically gone from $0 to $100 million in 24 months and is accelerating by just saying, you know what, I'm not going to worry about the legacy business. I'm going to start a new business. So I think— but the assets are there. These companies have great assets. They need that. And so it might be PE buying these companies for cheaper, but the PE can no longer have the playbook of we'll operate operationalize. No, it's not about operationalizing or improving costs. It's about building a new product, a net new product, an AI-native product. And then, like you said, I think there's a tremendous number of complete— for example, inference clouds are a new thing. Fireworks, Modal, Base10. There's multiple $5 to $10 billion companies. I have to assume the hyperscalers are not sitting around saying inference in AI is happening on these completely different cloud platforms. Yeah, different I have? What the hell am I? I'm going to be a commodity cloud platform. So almost certainly I think you're going to see one of the hyperscalers buy one of these neo clouds. I think CoreWeave was one kind of cloud, but there's Together AI, there's like 6 or 7 different next-generation cloud companies. So cloud software, you saw chips with NVIDIA buying Groq, there's a bunch of next-gen chips. I think you're going to see— and of course the whole Warner Brothers thing at the application level, the consumer side, there's a tremendous amount now that I think we are in an administration which is much more open to this, for lack of a better word, you're going to see— I think there's going to be amazing activity next year.

Do you think that the Square— you were at Square for almost 6 and a half years. Do you think the Square riff from last week will be an anomaly, or are you expecting 40% to be the new 10% riff? The latter.

I'm almost certain there's going to be multiple 20% to 30% cuts. Maybe even 40. The reality is nobody's hiring. I think almost every public company has basically— net new hiring is almost zero. Why? Because you can do much more. Every function can do more with AI. So it's equal to a 30% boost in productivity. And that's the first— next year, if you get another 30% boost, you need 30% fewer people. So over the next 18 months, almost certainly every public company is going to have a 30% If they don't, I question the leadership.

So yeah, so a lot of people's reactions to the Square riff was, okay, AI is just a convenient excuse. It kind of is a narrative that the market will be able to process. You think there's more? You think they're actually seeing the efficiency to be able to make a decision like that and it's not just marketing?

100%. I think they realize that they could basically do the same job and probably more efficiently with fewer coordination. The fewer people you have, easier it is to coordinate, as you know. I mean, the two of you run the show. Assuming you had 4 hosts, how hard it would be? Like some of the ESPN shows, right? It's just the two of you. Every time you add a person, it's M times N. It's like, like, it's, it's increases exponentially the coordination problem. Yeah. So I think that 12,000 person company or 6,000 person company, night and day, I bet they're going to move faster, be more innovative, and drive more productivity than they were doing before.

What's, what's the update on Marathon? Like, where are you guys focused in Q1? What's most exciting for you?

We are an early-stage software and fintech-focused firm. We focus on some of the best companies who do are at the intersection of the two. But in software, we look at infrastructure, security, and application. I think the themes we're looking at are some of the themes that we talked about. One is what are durable companies at the application layer. There are lots of interesting and, you know, things on the security side we're looking at. For example, phishing is something which is going— I mean, I have been phished now, I mean, I feel almost on a daily basis I'll get an email that looks like it's coming from a human being, but it's actually AI. And then the third area is around infrastructure. Like I said, there's a lot of new infrastructure that has been built because every piece of infrastructure that existed till 2 years ago was built for humans. Every piece of infrastructure has to be rebuilt for AI agents. On fintech, you know, stablecoins is just— I think I'm in love with stablecoins. I think USDC— I have so many people I know outside the US now who basically want all their money to be not in their local currency but in USDC, a stablecoin, because it's pegged to the US dollar. It gives the opportunity for everyone in the world to not suffer a 10% daily inflationary, basically reduction in their purchasing power, but to hold money in a way that prevents respect to the US dollar. So I think stablecoin is one of the greatest. So far, I think it's the best real world, outside of tokenization, is the best real world application of the crypto industry in some ways. And I'm extremely bullish on it. Amazing.

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

Yeah, great to finally have you on the show. It should have happened a long time ago.

Yeah, always happy to join you, man. Huge fan.

Yep, always welcome. It's been amazing. It's been 2 years since you started.

It's been incredible. No, it's been 1 year. Isn't that crazy? Yeah.

That is incredible. Congratulations, guys.

Thank you, guys. Thank you so much. We'll talk to you soon. Cheers. Have a good rest of your day. Let me tell you about Sentry. Sentry shows developers what's broken and helps them fix it fast. That's why 150,000 50,000 organizations use it to keep their apps working.

Board member Coinbase, board member Pinterest, board member The Trade Desk, spent 4.5 years at DoorDash, 6.5 years at Square.

Overnight success.

3 years at Meta.

Let me tell you, 4 years at Google. Fin.AI. Wow, Google, that's insane. Uh, the number one AI agent for customer service. If you want AI to handle your customer support, go to Fin.AI. There There is a talent shakeup going on at Quen, at Alibaba's AI lab. I want Tyler Cosgrove to take us through it, but there are a number of posts. Junyang Lin says, "Me stepping down, bye my beloved Quen." And another individual highlighted by Zephyr says, "Bruh, Alibaba Quen disintegrating in real time." Why?

Yeah, I mean, it's kind of unclear, right? The Chinese labs are like very not open about what they're doing. Also, it's not like an independent lab, right? It's under Alibaba. Yeah. But yeah, it's very interesting. I mean, Qwen has like recently has been like kind of the main— they're the staple in open source and like in Chinese open source, but in open source broadly, right? Even the American open source labs are still like pretty far behind Qwen. So yeah, it's very interesting. People are saying that there was like new leadership. They apparently— Alibaba brought over people from DeepMind, and then maybe they're just like completely transferring the leadership there. And then some people are saying, oh, these guys are all going to come to the US, or they're going to go to XAI or something like this. Yeah.

But yeah, it's definitely very interesting. Do you know where these folks are located physically? I believe they are in mainland China, correct? I think, yeah, I think that's reasonable to assume because there is a world where there's a whole bunch of AI talent that's missing or looking for the next gig. You sign up for some golden visas, even at $5 million a pop, that's completely affordable for an AI lab these days. It'll be interesting to see where these folks land. Yuchen Jin says, the end of an era, Quan lost his job its tech lead when we launched the Qwen3 Next endpoints on Hyperbolic with Junyang and his team. They were still online at 6 AM Beijing time. They were grinding. Thank you for pushing open source AI forward. Do you think they were caught distilling other American models and then they were fired?

I mean, busted. Yeah, I don't think that's like a fireable offense, right?

That's the best way to do it. That was probably the top-down mandate.

Please spend as little money as possible.

Well, maybe they were fired for distilling fast enough. Maybe, maybe. Nathan Lambert has an extra take here. He says the gaping hole that Quen imploding will leave in the open source research ecosystem will be hard to fill. The small models are irreplaceable. I'll do my best to keep carrying that torch. Not that I've reached the level of impact of Quen by any means, but every bit matters. It will be interesting seeing how Deepseek responds to this, whether these folks go to Deepseek, whether Deepseek adjusts their strategy when it comes to small models. That feels like something something that Highflyer could potentially fulfill. But also, who knows? We don't know how large the Quen team, how sizable this is, what they have planned. Sometimes there's large departures because there's a whole bunch of new people coming in. And so they might have just hired a ton of people that just haven't announced. It also will be interesting if this is one of those situations where it's clear that there is a raid going on, but people aren't announcing where they're landing. For another couple weeks. And so they just announced their departure. We see this a lot at the American AI Labs where people will say, it's my last day. And then 2 weeks later they'll be like, I joined the competitor. And then 2 weeks later they'll be like, I'm back at the original company. And they'll go back and forth all day long. Anyway, 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 every agent, secure any agent with Okta. I wanna talk about fast AI.

Fast food CEOs. Yeah, let's pull up this video of McDonald's CEO.

We never watched this all the way through. I want to, I want to see for myself. This is McDonald's CEO Chris Kempczynski. Uh, he went viral after seeming reluctant to eat his own burgers. He takes a tiny bite, looks uncomfortable, and calls the food product. That's the size of it. With— you've heard about it— the Big Arch. There it is. The Big Arch. This is something that we have tested already on humans, Germany, Canada, animals.

I love this product.

It is so good. I'm going to do a tasting right now, but I'm going to eat this for my lunch, just so you know. So here we go. First, I can survive a single meal. God, that is good enough for you. We've got a very unique kind of sesame poppy bun on it. Public presentation. Got two 2 quarter-pound patties, a delicious Big Arch sauce. It's probably his first TikTok, but he's doing it. Oh, there's so much going on with this. First of all, let's try to get this thing— I don't even know how to attack it. Got so much to it. There's also some crispy onions on here as well. I see those kind of coming out. He's stoked. All right. I met the CEO of McDonald's, the previous CEO, at South by Southwest in, like, 2013. No good. That's a big bite for a big artist.

That's a huge burger. Distinctively McDonald's.

Undeniably not that big of a bite. Yeah, not that big of a bite. He just kind of— but eating on camera is so hard. If you just start housing that thing, there's going to be a whole different set of backlash.

It definitely looked like his first ever bite of a burger.

Yeah, it's not a strong performance. Lulu said layup opportunity for the number 2 ranked fast food chain CEO to film a brutal—

and I think somebody did. We can pull up another video here.

Someone, someone did it.

Uh, there's someone saying Japanese mega rider trying to write Americans. The CEO would perhaps compete over who would take the manliest bite of a burger. No, that's ridiculous.

Surely America cannot be like that. Okay, watch this. Oh, that's a real bite.

That's a real bite. Yeah.

I'm a Burger King guy now. He won me over very, very well. It's— this guy seems like he can flip burgers.

He can take bites of burgers. I saw another video of Chris, the McDonald's CEO, coming out on his personal Instagram saying he eats— he wanted to clarify that he does eat McDonald's 3 or 4 times a week, he said. Okay. So he is defending his honor. McDonald's is great. But yeah, saying Gabby Goldberg called it out. I love this product. I love this product. That's what every chef says after they bring out a meal. They say, I love this product. I think you will, too.

He's a business guy. I don't have a problem with that. Refer to it as product. Go off. I stand by it. I think it's great.

I think it is a product. Even though it technically is a product, I think people want their food to be food. Food.

Yeah, probably should just use food next time. Well, good luck on the comms front. What is happening with Claude? The traffic is spiking due to all of the news and people are having fun with it. Has it actually been going down? I haven't— I've been seeing mixed things. I checked Claude to see if it was down. It was fine for me.

It went down last night for enough time for people to pay attention, but—

To the tune of 4,000 likes. I do wonder if there's like the compute debate over how many servers you need, how many chips you need to marshal, how much investment do you need from various hyperscalers? Does that equation change if there's a new sudden influx of consumer adoption? And it's unclear that consumers are driving, you know, insane inference loads compared to, you know, your OpenClaw folks who might be firing off 12 hours of agentic programming. That feels a lot more inference heavy than people that are like, yeah, I'm replacing my daily driver LLM app. And then they like fire off the usual like one query a day. That's maybe not, but at the same time, lots of people coming in, scaling challenges comes for us all. We need a new whale for AI. You remember the Twitter whale, the fail whale? Twitter had such a funny whimsical culture. There was this thing when Twitter was scaling, they had this monolithic Ruby on Rails app, I believe, and they couldn't scale the app fast enough because, Things would go viral, big moments would happen in the news, in the media, basketball NBA Finals would happen, everyone would get on, be talking about it, and the site would just go down for everyone. And so their 404 page, their page that you would land on when the service was not responding or populating tweets or showing you any content at all, had a fail whale, a whale that was being lifted up by some birds if I remember that correctly. And it became like loved by the Twitter community. And the rest is history. And eventually they did scale. Our next guest is here. First, let me tell you about Labelbox. Reinforcement learning environments, voice robotics, evals, and expert human data. Labelbox is the data factory behind the world's leading AI teams. And without further ado, we have Nick from Wraithwatch. He's the founder and CEO.

How are you doing, Nick?

What's going on?

Good to meet you. What's going on, fellas?

Good to be here. Welcome to the show.

So you have a big gong, you've got a big deal. Yeah, yeah. Kick us off with that.

No, introduce yourself. Introduce yourself first. Tell us what the company does. Yeah.

Thanks, guys. Well, I'm Nick. I'm Wraithwatch's CEO, one of the founders, former CIO, Chief Information Officer at Anduril. Joined Anduril when it was about 100 people, scaled it through when it was about 4,500. I think they've doubled since then, since I left a couple of years ago. SpaceX and Palantir before that, leading cyber defense teams at both companies. And then prior to entering the private sector, was responsible for pre-assault electronic warfare and cyber warfare in support of United States Special Operations Command.

So you worked at Palantir, SpaceX, and Anderal? Yeah, that's insane. Wow.

Well, thank you for your service. So it was really hard to raise—

it was really hard to Thank you for your service to the venture-backed defense tech industry and the LPs. Thank you on behalf of the LPs.

Yeah, it was tough. Sprinkle some AI in the deck and it was all good.

The rest is history. So yeah, where is the product today? What are the key, like, landing zones that you're going after?

Because cybersecurity is pretty broad these days, right? Yeah. So fundamentally what Wraithwatch does is We build autonomous cyber defense systems for the United States government and its allies. That includes private sector and Fortune 500 as well. So we're deployed today to Fortune 500 commercial nuclear, aerospace defense, manufacturing. But the byline on the screen is our recent announce with the federal government where we just landed a $30 million deal to deploy WraithWatch Wraithwatch in parallel.

Clean hit.

To deploy Wraithwatch to almost a dozen federal agencies in parallel. So pretty, pretty exciting time for us and ironic time given that we're in the middle of a shooting war with Iran, which is known for its offensive cyber warfare capabilities. Yeah.

Can you explain to me cyber defense? I mean, like, you can go and look for holes in systems and scan code bases. You can also be rerouting, dealing with firewalls if you're getting DDoSed. What's the nature of the threats that you're seeing? What's in the wheelhouse? What do you leave to other cybersecurity firms at this point?

Yeah, so just to set the stage here a bit, so historically, cyber, is kind of the last mover when it comes to any kind of, um, advance in software engineering or tech in general. Like, we're always the last to adopt stuff for whatever reason. Maybe it's because we're just, at the end of the day, a bunch of skeptical neckbeards.

I don't know why. Well, is that because, like, basically people want to develop a new capability and that's really exciting, and then, and then people realize, okay, now we need to defend against against said new capability.

Now we need to secure it. And so it definitely— the Gartner hype curve applies to the cyber adoption curve, I would say. And it's precisely because of what you just said. Like, OpenClaw is a great example, right? There's all this excitement around it, and then 2 days later, it's like, well, shit, this entire ecosystem is open to attack. There's API keys floating around in the wild. There's people injecting prompts. There's all of stuff going on. And so what you'll find is just a default sense of skepticism across the industry when it comes to anything new. And AI is no stranger to this. Makes sense, right? But what we're seeing is that the attackers on the attack side of the equation have no such compunctions. They are generally willing to try and take on any new technology that comes on board, and they certainly have have adopted AI in order to industrialize and weaponize their exploits and their attacks and campaigns in general. And they're moving at warp speed. When we started the company a few years ago, we were met with skepticism like, do you guys really think this is a thing? No one really knew what an agentic workflow was with AI. To everyone, it was still a chatbot. Now everyone knows what an agent is. Everyone knows you can string them together. And the adversaries have been kind of one step ahead of everybody, and they've been wiring these things up to offensive frameworks, offensive tools. And now you're seeing this kind of asymptotic attack pressure that's targeting organizations, and that's just coming up with novel new exploits, attack techniques, and so on every single day. Now, the problem is there's no corresponding defensive counter-pressure. So you have this mounting attack pressure on the outside. And then organizations and their cyber defenders are still moving at human speed on the inside. It's like, what do we do? Every 6 months, a red team comes around, tells us, you know, how screwed we are, and then we scramble around and then deploy a bunch of controls. Okay, what knobs do I have to turn over there? And as this is happening out here in the attack landscape, moving at human speed in here just isn't sufficient. So what we intend to do is essentially drive a similar asymptotic curve for defenders internally. Yeah, that's ultimately, you know, what the strategic landscape is and the way that that needs to happen. And I don't care if it's us, I don't care if it's another company. Essentially, the same problem that Palantir set out to solve— and we can get into the reasons why cyber resisted it for so long— is that the data lives in silos. It doesn't talk to each other. The tools don't like to talk to each other. Your CrowdStrikes don't like to talk to your Oktas, don't like I'd like to talk to your Splunks. Splunk in general is like an old model of let me just send data there and then worry about it later. And so Splunk to me is where data goes to die. And so what we do is tackle multiple things in parallel. Number 1, break down all these silos, force these tools to talk to each other. Number 2, for any dead data sources that are sitting around your organization, whether it's Splunk, whether it's data in S3, all right, let's light those things up. We can dispatch swarms of agentic, whatevers to get in there and start asking questions about, okay, what kind of data exists in here? What kind of aggregations can we do? What kind of— let's just build swarms of these little mini data scientists and throw them at these dead data structures, light them up, and then bring them into the greater cyber defense ecosystem. So you can almost think of the thing that's necessary as we move forward into this future is like the cyber version of JADC2. 2, essentially, for anyone that's familiar with the defense tech ecosystem. JADC2 is Joint All Domain Command and Control System. And it was an initiative by the US military— still is— to tie together sensors, fuse that sensor telemetry into a common cohesive command and control layer, and then allow the warfighter to make decisions at speed against it. And cyber defense just has no analog to that capability. And that's, that's the capability that we're bringing to the table for our USG partners.

Very cool. Walk me through a little bit more of like why I would want to investigate dead data. Is that like, if I'm a business, I have a bunch of data in S3 and it happens to have customer credit cards, if that got hacked, that would be bad. And so maybe I wanna lock that down, delete it if I'm not using it, or is there some other shape of like the threat of this like dead data concept?

Yeah, so dead data exists in one of two dimensions. So the data exists in the sort of pre-attack dimension of, okay, there's data that could tell you about latent risks in your organization. So I'll rewind you guys back to 2019. And there was a massive malware attack called Triton against oil and gas facilities in the Middle East. And the reason that attack was able to go down is because no one put the pieces of the puzzle together that, hey, we've got a senior engineering dude with a system that has weak controls applied to it that also serves as a jump-off point into the OT or critical infrastructure sector of the network. Now, to put those 4 pieces of data points together, you need a bored security engineer, you need a red team to come in and find it for you, and they might not ever find it even though they might spend 3 months trying to look for it. So, you know, that could be 1 of 100 attack paths that exist in your environment. Environment. So that's kind of the pre-attack latent risk piece. And then on the post-attack side, you've got data that's just being emitted from everyone's devices inside these environments. And what that data can tell you is, all right, all this latent risk back here, is anyone actually taking advantage of it or exploiting it? This is where the second half of the equation is, where you discover that. But what we've done over the past 20 years, because I don't know, Big Four consulting firms of the world have told us to do it is we send data to these data stores like S3 or Splunk or what have you. And then we just don't have the bandwidth to come around and then do something useful with it. You need some kind of sophisticated data analytics people that think along the lines of how to make use of this data appropriately. You need to think along the lines of joins and tables and objects and ontologies, and that kind of skill set seems to just not be there when it comes to cyber defense. Palantir, I think, was the first one to kind of break through and show people that this could be done for cyber defense as well. And there's a multitude of reasons why it didn't work. Foundry was early at the time when we were trying to break into cyber. And so what we're trying to do is close the gap and just have a turnkey solution and say, okay, you've got dead data here, you've got dead data here. So you don't even know where the adversary is going to come from. Like, we can't predict the next breach because no one's looking at it. We can't tell you whether a breach is actually underway because no one's looking at it. So let's break all of this down, fuse it into this cohesive layer, and then we can tell you where the next breach is going to come from. And we can tell you if it's actually going down right now as we speak. So that's, that's where we view the kind of like lighting up these dead data structures because otherwise, you needed a person to kind of sit there and come up with, okay, what queries are we going to ask and what automations are we going to put into place? And it's like, dude, when the dimensionality of some of these attacks is so large, you can't predict what queries you want, right? You don't know how that attack is going to go down. You don't know how it's going to manifest itself until it does. And so the only way to solve it is by generating and evaluating those queries and their results at machine speed.

Can you give us a brief history of cyber warfare with Iran?

Yeah, the Iranians have a long history of executing advanced cyber warfare attacks against the United States and our allied partners as well. So, you know, the cyber companies love to give the Iranian cyber warfare groups all these cute names like HydroKitten and all this shit. But at the end of the day, they are a well-known, you know, advanced nation-state cyber warfare outfit, like, like on the same level as kind of Russia, GRU, or PLA cyber warfare units, DPRK cyber warfare units. They were really not happy when Stuxnet was a thing. So back in 2010, for anyone who doesn't know, Stuxnet was kind of the first, hey, let's have a cyber thing that can be implanted in a physical system and then have kinetic effects in the physical world. So that thing went into the nuclear enrichment facilities, or the one facility that it was targeting, and then made the centrifuges rotate to the point where the hardware just broke down. So that was the first kind of example of that. The Iranians took that as an inflection point, kind of ramped up their offensive cyber capabilities. And for the past, I'd say 15 years, they've been been on a campaign to kind of bring those capabilities to bear targeting US government and private sector entities pretty much across the board. So all of that happens, and it not only happens with Iran, it happens with Russia, it happens with North Korea, it happens with China. It kind of all happens as this cosmic background radiation against which the regular commercial industry and private sector operates. And there's very few people that actually realize that there's all these nefarious characters in the background that are continually targeting our organization, whether through external cyberattacks, whether through trying to flip insiders, corporate espionage, or what have you. And so there is a cyber cold war going on. Iran has been a key player in it. And despite decades of trying, it took, let's call it for what it is, an administration that Enough is enough, and we're going to put some bombs on target. So hilariously, the Cloudflare founder tweeted out a few days ago that basically the Iranian cyber warfare outfits' activities have ceased completely since the strikes have gone down.

So we'll see what happens. Yeah. Well, good luck out there, and congrats on all the progress in the business.

Business.

Thank you so much for taking the time to come on the show. Cheers.

See you soon.

All right, let me tell you about ElevenLabs. Build intelligent real-time conversational agents. Reimagine human technology interaction with ElevenLabs. And let me also tell you about Cognition. They're the makers of Devin the AI Software Engineer. Crush your backlog with your personal AI engineering team. And without further ado, we will begin the Lambda lightning round. Let's see Cloud in all of its glory.

There we go. That is the moment that we've been building towards all year.

We're working on this. We're doing live TV and we have our first guest of the Lightning Round.

What's happening? How are you doing?

Raj, great to meet you.

Raj, great to meet you. Good to meet you, John and Jordy.

Thanks for having me on your show. Thanks so much for hopping on the show. Since this is your first appearance, please give us an introduction on yourself and the company. Sure.

My name is Raj Rajamani. I'm the co-founder and CEO of Jetstream Security. We just came out of stealth this morning, so super exciting times and we are building an AI governance platform to help companies scale their enterprise AI adoption. Okay.

How much did you raise?

$34 million.

In a seed round? Yes. Seed round led by Redpoint Ventures and the CrowdStrike Falcon Fund.

Yeah. So tell me about how this fits in with with CrowdStrike, Okta, Wizier, you have a lot of experience at those firms, some investors in common. How do you see yourself fitting into the market and going to market?

Yeah, we think we are building something very unique and differentiated. And this is something I've discussed with both George and Mike, who are the CEO and president at CrowdStrike. So before starting Jetstream, I was leading the products initiatives at CrowdStrike., for the last 3 years or so. Cool. And we had a lot of discussions about AI adoption and adoption challenges. And at Jetstream, we believe companies have an AI trust problem more than a technology problem. And by that, what we mean is with AI being probabilistic and being able to create compilers, their own social networks, and even deleting the odd email or two without being prompted, we have serious trust issues. And we think there are certain tools that you need in order to know what type of AI products you are using, understand them, and even control them. And once you have all those in place, you are able to trust it and move it into scale production.

That is what we are trying to help customers do. So yeah, what's an example of this? Because if you build a tool in-house using AI, you probably know exactly what model you're using. You might not have the weights, but you have, you know, you can read all about the various model, the capabilities, the context window size, everything else will be, you know, made available to you in the API. Is this more for SaaS products that are vending in AI features that you might have less insight into or more interpretability of the models themselves?

I would say the number one casualty in AI adoption is institutional wisdom. By that, what I mean is documentation is never anyone's priority. It's seldom accurate or up to date, which means that we have subject matter experts within companies who are responsible for understanding how these systems operate. Frequently what happens, especially with AI adoption, is that everyone's role and responsibility is going to be much larger. And what that means is that the institutional wisdom gets diluted. And the way we counter that is by delivering what are called AI blueprints. These blueprints are the operational contract of an AI workflow. They define what AI is supposed to do and flag any deviations from the intended purpose. They also track all the runtime activities, the permissions they are using, the data it touches, the services and tools it calls, and so on and so forth. So you have everything that you need in one single place. And just like blueprints, like the ones that we use for building homes and other buildings, it helps you see things that are not obvious to the naked eye. It helps you understand the specifications and tolerations that you as a company are willing to agree or use within your enterprise. Will you be learning to drive a Le Mans prototype so you can hang out with George Kurtzman I'm afraid I'm more of a swimmer than a driver. And the last time I was on the track with him, he chided me for driving a Ferrari the way I drove my Prius.

I've been driving Prius for the last 15 years. That's amazing.

You're a swimmer, though.

You're a swimmer. Yes.

Interesting.

I try to get in at least 2 or 3 times a week. So I had a swim yesterday and hopefully one later this evening.

Open ocean or you— 24 hours.

Of the Pacific Ocean.

24 hours of Alcatraz, just swimming around.

It's more like 24-hour fitness lap pool. Yeah, yeah, yeah. You have to start somewhere. That's fantastic. How big is the team? How quickly do you want to go to market? How are you seeing enterprise versus the mid-market, small startup adoption?

Where do you see all this going?

We are about 40 employees now. And the interesting thing is almost every last person has worked with everyone else for an extended period of time. So this is a team and crew that knows each other really, really well. And we are open for business. The product is in POVs already. We were doing demos and I was getting texts from some of my investors saying, hey, I want to introduce you to these accounts. That's great. We are obviously targeting the large Fortune 500 accounts initially, but we think this is a universal category in the sense that regardless of geo region, vertical, or even the size of the company, everyone will need an AI governance platform, much like how everyone uses an EDR product today. And if we are able to solve it in a graceful and elegant manner, I think there is a lot of room in this market for both large enterprises and SMBs.

Yeah, that makes sense. Well, thank you so much for taking the time to come chat with us. Congratulations. On the fundraising round and say hello to George. Sure, you'll be back on soon. We'll talk to you soon. Have a good one. Let me tell you about Restream. One livestream, 30+ destinations. If you want to multistream, go to restream.com. Jay Butard on Twitter or on X says there is a 0% chance regime change is accomplished in a short time frame, referring to Iran, but there is a 100% chance Trump Trump quits this war ASAP if Intel keeps gapping down 6% a day and the Dow Jones stays losing 1,000 points per day. We have priorities in this country. And so watch the stock market if you want to know where the war goes, I suppose, is Jay Buttar's point. Trump said that you'll be finding out very soon who's in charge of Iran. And of course, Ani Ayyengar said, Ahmad Sheer. The brief CEO of OpenAI, which would be, of course, the funniest outcome. So never write it out. There is another big story in the technology and venture world. Thrive Capital and Andreessen Horowitz are leading a $4 billion round into Anduril. Very exciting. And congrats to everyone over at Anduril. What is— Still in the works. It's still in the works, but They're— Bloomberg is reporting Josh Kushner's Thrive Capital and venture firm Andreessen Horowitz are co-leading a funding round for Anduril that could nearly double the defense startup's valuation. There had been rumors about this, this round. It's $50 or $60 million, correct?

So the fundraising is finally going to be worth more than Figure.

That's right. And they're getting up into the defense tech prime territory. It was a big moment for Palantir when they passed the legacy primes on market cap.

Well, still got a ways to go with Lockheed Martin sitting at $150.

$150, but within the sights. I don't know. Josh, Andrewel seeking to raise about $4 billion. The fundraising expected to value the startup at $60 billion. The company is raising capital as it pursues projects, including a facility in Ohio to mass-produce aerial and maritime drones and other products. There was a piece in the New York Times about Palmer Luckey that was going back and forth, kind of digging into honestly just a lot of history on the company, but also sort of, you know, trying to understand how deployed the different, the different products are, where things are going well. I don't think it was the most friendly piece, but certainly painted an interesting picture of Palmer and a nice graphic in the header. Well, without further ado, let's bring in our next guest. We have James from Guild.

How you doing, James? What's happening? Doing great.

Thanks for having me on. Welcome to the show. Since it's your first time, please introduce yourself and the company.

Sure, sure. I'm James Everingham, the CEO of Guild AI, building an agent control platform Fine.

Okay.

This is important because my agents are out of control. Yeah, yeah.

They're going to get worse. We're here for you. Yes, I was very interested in orchestration, what's happening with Gastown, some of the stuff we've been seeing with OpenClaw. How long have you been working on this? And then where do you think— like, what do you think the correct user interface, the correct user experience paradigm is? Yeah, absolutely.

So I think you just wrote— yeah, I was reading your post about 2026 being the year of agents and orchestration. And first off, completely agree with that. And what we're seeing though is once you start getting these agents in your infrastructure and they're starting to come in pretty, pretty heavily into a lot of companies, the next problem becomes controlling these things. You know, it's like, it's like Gremlins, you know, the first one's fine until they start multiplying and taking over and pulling levers in your infrastructure. You know, they can cause a little chaos. So what we have built is an infrastructure layer that provides governance and control around these. You know, you can centralize them, you can run them efficiently, you can give them and control what they have access to and what they don't. You need observability in order to figure out what they did, what they touched, and, you know, who can get access to configuring them. So that's where we're coming in. We ran into this problem, a lot of us came from Meta. We were working on the developer infrastructure teams, And we were working on getting the engineers to use agents to embed into the infrastructure. And well, we got that to work. And that basically highlighted the next problem is like, okay, now you have lots of these. We think there'll be more than 10. We think you'll have thousands of these in your infrastructure at some point. Then suddenly, security pops up. And if you're in a regulated business, compliance pops up and you need a layer layer to control this. Agents are non-deterministic, and you need a deterministic layer if you're going to have a stable infra layer.

How much is just cost-benefit important these days? I've seen crazy bills from people that are orchestrating a ton of agents, and, you know, it starts to creep up on you. Like, wait, you spent how much on tokens for that to-do app? Like, I don't know if that was a good trade. And I imagine that in the enterprise, if you're talking about deploying fleets of agents to hundreds or thousands of employees, you could be looking at a really big inference bill and monitoring that seems important too. But is that something that is just too abstract for clients to demand at this point?

Not at all. That's actually one of the top requests we're seeing. Like, you know, some of the customers that we're engaging with early, like literally one of them had this problem is they had a monthly budget and and it's built for single player so that engineers are off running these on a separated server or on their laptop. One engineer blew through their entire budget in 12 hours and they didn't know. So, building a centralized place where you can put circuit breakers and say, hey, turn this thing off if it gets too hungry, or be able to report and understand where your spend is, but not just like where the spend is, What am I getting for that spend? And like, you need to be able to measure the impact of these as well.

I feel like you're at the tip of the spear. What type of companies are being most aggressive about adopting fleets of AI agents these days?

Yeah, you know, it's, it's the timing is really interesting because we're seeing like many of the customers are now just leaning into it. It's like, you know, last year it was like experimentation. We've got one agent. It's a really interesting demonstration or example, but now they're all starting to scale. So we've been talking with small companies and large companies, and interestingly, they all seem to be at the same phase where we've now figured out that the models are powerful enough to operate tools and to be able to be part of our infrastructure. So we're seeing it across the board. It seems to be hitting all these companies at once.

Okay, so evenly distributed across across scale and stage. What about industry? Are there particular industries that are more rapidly adopting fleets of AI agents?

Yeah, I think that there is some little bit of differentiation there. Of course, you know, some of the more, you know, the companies that are a little more tech-forward, you know, they map up to what we traditionally would look at as the early adopters and the laggards. It's the same. So the highly regulated businesses are slower by design because they, they, you know, they can't take risk. And one way to think about it is the companies that we see really leaning into this is, look, you know, there's a lot of pressure for competition and being able to move faster right now. So I think that the calculus between risk and the consequences of failure has changed. People are willing to take on a little more risk and try these things. But, you know, our aim is to allow them to move fast while still reducing that risk in their infrastructure.

What kind of tension is there between— like, I can imagine a world where Guild provides all of the enterprise infrastructure so somebody can just bring— somebody can create an agent and effectively give it access to Guild, and that gives it access to the enterprise. But a lot of companies will want to actually own the full stack. How do you think about about that tension as well as opportunity?

Yeah, I think, well, one, I think it's an opportunity. Right now, what we're seeing is that a lot of companies are off building sort of like sort of small versions of this internally, and it's not their core competence. They don't want to build it. They don't want to maintain it over time. It's not going to get better. They don't have the specialized expertise to do it well. So what we think is that companies are really open to using this as a service. And one thing that's important is that, you know, being vendor neutral and, you know, model neutral is also important because people don't want to be locked into one vendor because the competition right now is things are, you know, they're leapfrogging each other and they're hopping from model to model. So, you know, vendor neutrality is also important here. Totally.

What about the types of activities, the types of use cases for fleets of agents that you're seeing? I'm interested to know about this. There's a lot of narrative around like, oh, you're just going to go vibe code your CRM and rip out that legacy software. But it just feels like with agents, you're so much more likely to just go pull your roadmap forward and go deliver more value value for your customers. Maybe you'll go ask for a contract renegotiation, but what are you actually seeing? Yeah, sure.

So the interesting thing is, like, people— we're seeing the most innovative agents come from our customers and not ones that we thought of. I like some of the favorite ones that I've seen so far is this onboarding agent. There was an agent that made the code base sentient, and so instead of saying, hey, I want to check out these files, you can say say, check out this feature, and it would say, oh, that's these files. Would you like a system diagram? Sure. And we paint you a system diagram. Do you want to know the history? And you could have a conversation with it. Another one, or we're seeing these agents, and by the way, agents can get executed not just by a person, but by a tool and by an event. So we had a risk analysis agent that when it gets pulled, like when the CI/CD system pulls from source, it would analyze the risk to have how likely is this diff going to take the system down? And it can let low-risk diffs go through and high-risk ones stop. So things like this can help eliminate holiday code freeze of the codebase. So those are a few, and they can expand way outside of just development. They can go into all of these different workflows, anywhere where there's an integration prompts and evals, you can make use of this. We have one in ours that like doesn't allow duplicate bugs to happen. So, books into Git, fires an event, the agent goes and says, this is a duplicate bug already, it'll just mark it. With the Guild system though, we were able to just go in and natural language interface, just say, hey, don't make it, make it so it's not possible for duplicate bugs to exist in our company. And it just went and did it behind the scenes, installed the agents, set it up right, And boom, like that wasn't possible for any engineers in the company.

Talk to me about where Guild sits in a world where I build a product that is effectively a fleet of agents. Maybe you call me a rapper, whatever, but my customers love the product, which is heavily agentic. I'm using Guild, your product, to orchestrate. And then I wanna vend that into a company. Am I like white labeling you? Am I using you internally and then my customer doesn't know? Are they going to be wanting to use Guild on their site? Like, what's the—

where do you fit in that stack? Yeah, so one of the things that we found is that like a way for these agents to be— for engineers and for companies to learn how to use these is by like looking at them and seeing it. So like we have something we call Agent Hub. Okay. And Agent Hub will be a public surface. So companies can have internal agents that are private, but they can also push them externally and they can be a series of agents that work together. We look a lot like GitHub in that way where you can go, you can find agents right on the surface that might work for you. These could be distributed by third parties or people that are building agents as products. And within minutes they can just authorize, set up a workspace, click and boom, have these things working in their workspace. So we're big open source believers. We've been in that space our entire career. So, you know, there is an aspect of that that we're bringing to this agent marketplace.

Very cool. Well, we want to ring the gong for you.

How much did you raise? Oh, thank you. Yeah, we— it's kind of tough following that $4 billion one that you just announced. But, you know, we will go. We raised $44 million. It's our second round. Apologies, it was $30 million plus $14 million in our seed.

We raised both of these rounds within 4 months. That's incredible.

Who participated?

Yeah, any like companies I might have heard of involved? Yeah, sure. Like our lead investor was Google Ventures. There we go. So GV, they read— they— we have NFX involved. We had a Khosla Ventures came in this round. Scribble Ventures, Accru Capital, Web Investment Network. I think I got them all.

You will have missed one, and that one person will hold a grudge against you forever. Oh, of course.

That's a great line. The rule of lists. You always forget one.

It'll probably be the biggest one too, right? But thank you so much for taking the time to chat with us.

We'll talk to you soon.

Thanks for having me. I appreciate it. Goodbye. Let me tell you about Public.com, investing for those that take it seriously. Stocks, options, bonds, crypto, treasuries, and more with great customer service. And let me also tell you about Phantom Cash. Cha-ching! Fund your wallet without exchanges or middlemen and spend with the Phantom card. And without further ado, our final guest of the Lambda Lightning Round, we have Felix from Akash Systems. How are you doing, Felix? I'm good.

I'm good.

How are you? We're good. Good to meet you. Since it's your first time on the show, please introduce yourself and the company.

Sure. My name is Felix Adracham. I'm founder and CEO of Akash Systems in the Bay Area.

Amazing.

What do you guys do? We are a venture-backed company that makes diamond-cooled servers for AI and cloud service providers.

Okay. Diamond-cooled servers. Very interesting.

Yes.

I love it. Very exciting.

I love it. I'm familiar with air cooling. I'm familiar with water cooling.

What is diamond cooling? So diamond is interesting because it is not only a very beautiful material, but it is also the world's world's most thermally conductive material. It is extremely thermally conductive. If I held a stick of diamond in my hands, it would take the temperature of my body and move it very quickly through the entire material and propagate it to whatever is on the other end of that material. What we do with Akash is we take that diamond that's very pretty but also very thermally conductive, and we place it right on the GPU. So a GPU, for those of you who don't know, is the hottest component of a computer or server. So the things that answers your questions on ChatGPT, on Claude, the very essence of the thing that does those questions is the GPU. It gets very hot. You've probably heard people talk about that. And cooling it down can be very expensive if you use liquids. But what we do is we take synthetic diamond grown in the lab with thermal connectivity. We put it on the GPU, it reduces temperatures by 10, 15 degrees Celsius. That reduction causes an acceleration in the inference of the GPU. It also reduces the cost of cooling the entire data center.

Yeah. So once the heat's transferred from the GPU die to the diamond, it's got to get shuffled away at some point. Yeah.

Where does that go?

Is that air or water?

How is it getting moved away? That's air. Yeah, so we're compatible with liquid cooling. If you want to have, if you already have an infrastructure of liquid cooling, we're totally game for that. But if you don't have the capital to invest in a liquid cooling infrastructure, then you don't need it because you've got diamonds that we provide on the GPU. And so it goes from the diamond to a heat sink, and then that goes to air. And so—

People that might've built a gaming PC, you put something directly on the CPU. Is that copper typically?

What, what, what? That's typically copper. And by the way, with a DiamondCool server, you don't need to have your air conditioning on. You can literally turn off the AC in the room and the room can get as hot as 120 degrees Fahrenheit.

And the servers literally like— Because it's shuffling it away so fast. Correct.

Got it. Okay. So fast into the ambient and you'd be fine. We just released today a server made with an AMD chip chip that is cooled by diamond, and that can operate at 120°F, and you would not need any liquids, no air conditioning, and it would just be—

the computations would still be fine. Okay. What—

how much, how much actual diamond is like—

what is the cost? Is it like one very thin diamond, or are we talking about something that could fit on a ring?

I love your questions. Yes, it is a very thin layer of diamond that placed on the material, on the GPU. And to your question about cost, so the effect of this diamond cooling is that it generates about $1 million by our conservative calculations, about $1 million of revenue and some cost savings to the operator of the GPU. And that's over typically a lifetime use of around 3 to 4 years. So we're not just giving you compute, we're giving you cash. So roughly about a million. And so we take a slice of that to add on to the cost of a traditional server. So our servers are— the cost of our servers don't compare to the million dollars that you would get back in returns.

Yeah.

So I imagine you're literally— you're icing out the GPUs.

I love it. So I imagine you— I imagine that a Diamond cooling system has got to be more expensive than the legacy legacy system, but you're saving money on the, on the, on the energy cooling, the AC, the water flow, all of that.

So less energy intense. That's correct. And we're talking, I'm going to give you some numbers here. So a typical server, maybe like $250,000, $300,000 hours would be on the order of, let's say $400,000. Okay. So, but on the other hand, over the 3, 4 years of use life, you're going to get $1 million.

And then also I imagine that you're on, sort of a Moore's Law-esque scaling curve of your own costs because lab-grown diamonds have gotten dramatically cheaper over the last decade and you expect that to continue?

Okay, explain that. That's exactly right. The cost of making diamond has literally collapsed through the bottom of the floor.

It's so low and that's just improved our economics. Okay, so for actually making the diamonds, you're working with a contract manufacturer who also makes diamonds for saws or jewelry?

We source diamonds from all over the world. We buy diamonds from wherever we can get them very cheaply. We have ways of growing them ourselves, but the interest of scaling rapidly, we buy diamonds from everybody. As long as you can meet our both quality and cost requirements, we're happy to buy from you.

Yeah, I remember seeing the first lab-grown diamond, I don't know, maybe 15 years ago, and it was basically like a massive dome that was putting an immense amount of pressure and it would come out with a diamond-shaped diamond. It was something that would belong on a ring. Now, I imagine as wafers and GPU dies have gotten bigger, you need to make a bigger lab-grown diamond and then slice it. Can you walk me through like, how are you dealing with like the size of the GPU die that I would imagine is bigger than like a 1-carat stone? Out. Yeah.

So you're, you're very perceptive. The— what you described is the old-fashioned way of making diamond where you try to replicate the conditions of the core of the Earth. Oh yeah. Very high pressure, high temperature. Okay. What we do is something called chemical vapor deposition, a more advanced way of making diamond where you don't need to get a massive thing, uh, to recreate the core of the Earth. Uh, these reactors that we use are very small, typically And the amount of diamond that we need is relatively small.

Yep.

So it needs to be cost effective and it needs to fit into the traditional size of a server.

Folks who have purchased a diamond, maybe for a loved one, might be familiar with the 4 Cs. Do any of those matter to you? Can you use a lower quality diamond or is there a benefit to heat dissipation if you have a particularly clear diamond? Great question.

The answer is yes. We don't quite use gem quality, beautiful, spectacular, twinkling little star diamonds. Diamonds that we use are actually what we call industrial diamonds. So they're not pretty, but they do have an extremely high thermal conductivity. And that's the thing we care about.

And so we use that for our work. Are diamonds forever? Do you need to replace these at some point?

Diamonds are forever. We reclaim the diamonds. So at the end of the life of the server, we're happy to take the server back and pull the diamond out and reuse it for our needs.

Yes. Talk about flexibility of integration. Congratulations on the AMD deal, by the way, but it feels like it's not rocket science to get this working on a different chip.

Walk me through that. It's not. So it is, but it isn't. It's taken us a long time to get here. Actually, this technology that we've deployed with the AMD chip actually started off in space. So we actually started off doing diamond cooling in satellites. In fact, last year we launched about half a dozen satellites with diamond cooling. But then when we saw what was going on in AI and that the power densities in the GPU are actually a lot less than what we deal with in space, and the fact that in space the conditions are much more harsh. We're talking about solar radiation, space radiation that's a lot more difficult and challenging to work with than what we have on a terrestrial. We said, wait, if we can address a problem in space, we can definitely address a problem in a GPU. And so that gave us the confidence to replicate what we now have on, on ground. So, and getting to space took us a long time, and we're leveraging all of that to do this. So So we are— the technology that we've deployed here is covered by dozens of patents. There's a lot of material science and device physics, but in the end, we're able to do it cost-effectively and scale rapidly for our customers.

Amazing. I heard that there was a secret shopper who stopped by the diamond store, placed an order for a huge amount of your product. How much is the deal? Deal. I know you can't tell us who it is with, but I'd love to know the size.

I want to ring the bell. It is a $300 million deal.

You said the biggest deal. Congratulations. That is amazing. This is amazing. What a remarkably sci-fi and amazing product. Yeah, incredibly sci-fi. I love this, and thank you so much for taking the time to come talk with me. Us. Do you have any other questions?

This is great. Yeah.

Are there any other interesting applications for diamonds that you can think of? I've heard of the diamond drill, the diamond saw. This was not on my bingo card, but should we expect big things from diamonds in the future if they get really, really, really, really, really cheap?

Definitely. I mean, we get calls all the time. People want to put our stuff— I mean, because we have the space heritage as well as now servers and GPUs, people want to now do data centers in space. Of course, that's super sexy, amazing. Uh, to be honest with you, just trying to execute what we have in front of us, sure, uh, is already quite a bit of a full plate. But yes, there are applications in, in space, as we discussed, uh, applications in, in other areas, industrial, uh, areas, uh, such as with Bitcoin mining. Uh, that's something requires a lot of compute. And so we do get pings, but right now we're very focused on just getting compute to data centers and cloud service providers and AI companies.

Well, Felix, thank you so much for taking the time to come chat with us. This was very fascinating and good luck to you. We'll talk to you soon. Cheers. Have a good one. Let me tell you about Console. Console builds AI agents that automates 70% of IT, HR, and finance support, giving employees instant resolution for access requests and password resets.

Uh, the Germinator says it's official, Apple launches the M5 Pro and M5 Max MacBook Pro, M5 MacBook Air, new Studio Display, and new Studio Display XDR.

This is what I always say, it's not official until Mark Gurman writes about it. I, I don't actually go to Apple. It's not official when they post about it. I like that he's like, he's got the scoop, uh, but I believe him and this is exciting. Um, uh, the M5 Pro, M5 Mac, M5 Max, Macintosh Pro.

Uh, yeah, so apparently there's still redesign rumors for the MacBook Pro. I might, I might, I'm gonna upgrade this. I might upgrade.

Uh, 6:15 AM Pacific tomorrow morning. Tomorrow morning you can order. Yeah. Yeah, the Studio Display XDR is gonna be awesome. The Pro Display XDR has been dominant for like a decade. And you will go on YouTube and look up, should I buy an Apple Pro Display XDR in, and it will just autofill 2023, 2024, 2025, 2026. And the answer is always yes. Like it's still a great monitor. And so it's always, the only problem is just it's been really expensive. And it hasn't been refreshed in so long that a lot of Apple folks that want a beautiful display, high refresh rate, they— there's really like no trade-offs with it. Whatever they have planned there, I think will be very, very exciting. So buckle up, Apple fans, there's gonna be a lot of good news tomorrow to react to. Let me tell you about TurboPuffer, serverless vector and full-text search built from first principles on object storage, fast, 10x cheaper, and extremely scalable. Definitely scalable. Uh, the last thing I think might be worth mentioning is that Ted Mabry over at Palantir is issuing a call to former Palantirians, former hobbits, as he says. If you've ever considered returning to Palantir, this is the week to do it. The world is demanding every last unit of creative energy we can muster. If you return, you will be on a plane one day and committing code that matters within hours of getting your laptop, uh, from the foxhole to the factory floor. To the surface area to support something incredibly meaningful immediately is immense. I promise you nothing other than a sense of satisfaction that comes from the purpose and intensity of the most intense deployments you've ever worked on. If you have been chasing that spark, come find it again. The Shire is calling, saying come back and work at Palantir. Very fun.

Anyway, well, time to plant the bomb. I think so. Okay. Actually, let's not plant the bomb. Okay.

Yeah.

But thank you for being with us. We will be back tomorrow, 11 AM. I can't— I cannot wait. I cannot wait.

Uh, have a wonderful evening. Leave us 5 stars on Apple Podcasts and Spotify.

Subscribe to our newsletter.

Call your parents. Tell them you love them. Yeah, tell them you want to watch Ken Burns with them. See ya. Do it. Goodbye.

Bye.