Intel Rips, Cursor's Plan, Thrive's Giant Bet, GPT 5.5 | George Kurtz, Professor Sendy, Gary Vaynerchuk, Yoland Yan, Ben Horwitz

You're watching TVPN.

Today is Friday, April 24th, 2026.

We are live from the TVPN UltraDome, the temple of technology, the fortress of finance.

We have a great show for you today, folks. We have George Kurtz coming on from CrowdStrike, Professor Sendy from Instagram. You might have seen he's the expert of Wombos. Gary Vaynerchuk's coming on. We have a lightning round going on with a couple founders. It'll be a fun time. But the big news—

wrapping up with Ben Horowitz.

Ben Horowitz.

Horowitz.

Horowitz. Horowitz. The big news, of course, is Intel. Intel is absolutely ripping. Intel jumped 20% after hours on the back of $13.6 billion in Q1 revenue, only 11% above analyst estimates. But there's 5 key factors that are coming together to create a new narrative around Intel that are driving the stock much higher. And it is almost—

the main factor is Bubble Boy on X.

A lot of people have called this and it is long overdue. It's the revenue is only up 7% year over year, but next quarter is already guiding to be better somewhere north of $14 billion probably. But there's still big losses under the hood. This quarter they lost $3.7 billion. Not good, but that was mostly driven by one-off charges related to Mobileye and derivative payments tied to the US government's 10% stake. So if you strip those out, Intel actually earned $1.5 billion, which is much better than what people were expecting, which was basically breakeven. So the cruise ship of Intel is starting to turn somewhat, but the narrative has already completely shifted. So Intel is working again is the idea. The AI trade has mostly been Nvidia, Nvidia's memory suppliers, TSMC, power equipment, cloud CapEx, and a few software names that can prove real adoption. If you're accelerating your top line, you are an AI winner. That's sort of the rule of thumb, of course, Of course, things might play out differently, but that's what's been happening in the market. Intel was the most embarrassing missing piece. Why isn't Intel booming when we're in a computing boom? It made no sense, but there were good reasons and we'll go through them. The company that invented the modern CPU era, they missed mobile, they fell behind TSMC, they failed to produce a competitive AI GPU. They did fab a GPU at one point. They tried to get into gaming at one point, but they never really found traction, especially in the data center and the servers for AI GPUs. Training and so they have fallen behind. And then they spent the last few years trying to convince the world that Intel could still matter. But now, oddly enough, the rise of AI agents, it's giving Intel a second shot. Do you have thoughts on Intel yet? You want me to keep going?

It's funny because at the moment that the US took position in Intel, it was like, it felt like a bailout, right?

It felt like that, yeah.

And it felt like, You would expect, okay, over time, this is gonna be very, very good for Intel to have that vote of confidence. But I don't think anyone was really predicting that it would go up quite this much in the span of just half a year.

I think Ben Thompson wrote a pretty strong bull case for the Intel-US deal. Basically, you had to grapple with this idea of like, This doesn't feel like free market capitalism.

Yeah, I guess the only thing is like there wasn't this narrative around CPUs at the time that the deal was happening, or at least it wasn't a very public narrative.

Yes, yes. The bull case at that time that I heard most loudly was not the CPU boom. It was just overall chip. It was the 14, like the leading edge fab that basically by having the US government as a shareholder, you see those dinners with all the AI lab leads and all the hyperscaler CEOs and, there's a world where there's another one of those meetings and the government administration says, hey, we are backing Intel. I need all of you to commit to buy, you know, a ton of supply if Intel actually goes and builds the leading edge fab, which is gonna cost them billions of dollars. But if they have the demand guarantees, then they will actually be able to go do it. Tyler, what was your take on that?

Yeah, I think Leopold, or sorry, Intel has long been kind of the Leopold take, which is that like, You know, this is the clear— like, this is the company you want to own in the kind of nationalization world of like, okay, Taiwan risk.

Oh, sure.

What are we going to do if we can't make chips, you know, in Taiwan? Intel's like the very clear strategy. Yeah, maybe it's like not working right now, but like at some point this is not just like economics. This is like, okay, this is, you know, national security. Like, it's extremely important that we have this capacity in the US. And Intel is kind of the only one that's even like remotely in the conversation.

Yeah, yeah, no, that makes sense. So there were some predictions from AI 2027 and other folks in the AI forecasting world around the rise of agents that agents would arrive at this time. I didn't see too many folks who were predicting strong, valuable, effective AI agents really predicting the CPU crunch, but that's exactly what's happened. So AI agents need CPUs to do things. Training frontier models, that's still a GPU story, but running Running agentic workflows across data centers, orchestrating tasks, routing jobs, managing memory, handling inference workloads, coordinating servers increases demand for the boring old central processor. Intel's data center segment produced $5.1 billion in quarterly revenue, beating the $4.5 billion that analysts expected. And Intel CEO Lip-Bu Tan said the next wave of AI is moving from foundational models to inference to agentic AI, and that shift increases the need for Intel CPUs, wafers, and advanced packaging. On the earnings call, he said that the CPU to GPU ratio is closer to 1 to 1 CPU for every 4 GPUs versus 1 for every 8 in prior years. And there was an interesting forecast from, uh, Liputan as well. But VK Macro says, uh, CPU to GPU ratio flipping from 1 to 8 to 8 to 1 is absolutely wild. That's just a completely new world to what we've had so far. This is from, uh, Evercore. ISI's Mark Lapisas has upgraded Intel directly from neutral to outperform and mentions that as AI workloads shift further toward inference and agents, towards inference and agents, the weight of CPU demand will rise sharply and the CPU to GPU ratio could flip from 1:8 to 8:1, which is a massive, massive switch.

Yeah, I mean, I don't know, like, That's kind of a crazy ratio.

It is.

Because like the lesson from tier models is that like the models are going to keep getting bigger. You're going to need way more like chips to inference them. Like Dylan Patel was on Patrick O'Shaughnessy. I think it came out yesterday or the day before. And he's just saying like, you know, the models are going to get really expensive. Tokens are going to get super, super expensive.

People are going to price out.

So I don't know if I fully believe the narrative that like you need all the CPUs.

Yeah.

You're going to need way more CPUs than—

Whoa.

I mean the— And that's all folks. It really lines up perfectly. His camera's the main one that fits perfectly. The glasses land on it.

Oh, I love the Friday show.

Yes, that's a good point. But there is this, I think at least in the midterm, like the— Yeah, there's like capabilities overhang. And just like one really smart AI system that's being inference on GPU can spin off a ton of CPU workloads. And do a lot of things that require CPUs. Yeah, I think this takes a lot— Even in the SaaSpocalypse, like you vibe-coded Slack or whatever, it's like, well, that vibe-coded system is running on CPU, and if it took you, I don't know, 1,000 GPU hours, but then that system runs on CPUs that are running constantly for every user, you still wind up with creating more CPU demand out of the GPU.

I think this take makes a lot of sense if you are kind of, you know, long open source. We're gonna have a lot of these open source models.

Yeah.

They're quite small. You can run them on like, you know, small amount of chips, but you just have a lot of them. You have a bunch of agents doing a bunch of stuff at the same time.

Yeah.

Rather than the single model, you know, inference on a massive amount of chips.

Why is the horse here? Anyway, whether it's going to 8 CPUs for 1 GPU or staying at 1 CPU to 4 GPUs, it's still, twice as good as it used to be, 'cause it used to be 1 to 8. And so that is bullish for Intel, and Lip-Butan is making that clear to the market and to the investors on his earnings call. Then there's the wild card, the TeraFab Elon Musk project. It's very exciting, but it's clearly further off. Elon Musk wants to build a massively vertical integrated chip manufacturing operation with Tesla, SpaceX, and possibly other Musk companies needing huge volumes of chips for self-driving cars, humanoid robots, and even space-based AI data centers. Intel is supposed to help design, manufacture, and package chips for the project. Now, the Wall Street Journal has a more cautious view today. Elon is aiming for TerraFab to reach 100,000 wafers a month and then eventually 1 million wafers a month, which is just insane scale. So let's put it in context. 1 million wafers a month is about 70% of TSMC's total monthly output across all fabs. TSMC's largest fabs put out roughly 100,000 wafers a month into production. So you're talking about 10, you know, leading class TSMC fabs at Intel just on the TerraFab project plus whatever else Intel is doing. So a pretty massive scale up. And, and TSMC's CEO, CC Wei, has said that it's just not that easy to build a fab overnight and get it up and running. So fabs take 2 to 3 years to build and then another 1 to 2 years to actually ramp. And we've seen this with TSMC Arizona, which we've been very excited about, but it just takes time. And there is a bottleneck and the bottleneck has been discussed at length. Ben Thompson wrote about TSMC risk in Stratechery, arguing that TSMC needs to spend more on CapEx. And I think that should be clearer now. We'll see what happens at the next TSMC earnings call because maybe they will be waking up. But overall, Intel now has a collection of plausible demand stories, even though the demand itself is not going vertical. The stock is going vertical on the back of these 5 key demand stories that are all pointing in the same direction, directly up. 1, AI agents need more CPUs. 2, AI systems need more advanced packaging, a higher ratio of CPUs to GPUs. Intel can help with that. 3, the US government wants to rack domestic leading— a domestic leading-edge foundry. This is just a national mandate. 4, Musk wants an impossible amount of silicon. And 5, Also, the hyperscalers want more supply. And so all of that is, is good news for Intel. There are more companies that are getting into the CPU design space. We talked about ARM recently and ARM, of course, it feels like they will be going with TSMC in the short term, but who knows what happens in the long term. So as CPUs continue to be important in the AI story, all good news for Intel. So suddenly investors are more willing to entertain a messy, expensive strategic chip story than they were 5 years ago. Intel famously missed mobile, which meant TSMC ran away with enormous manufacturing volume and left Intel with a demand problem. Volume was destiny, as they say. As fab costs grew by orders of magnitude, meaningful demand was a hard prerequisite for continued growth. You can't grow if you can't build the fab that's on the leading edge, and you can't build the fab unless you actually have the customers. And so if you, missed mobile, you just have this gap and you have to jump over it with the help of the government, a bunch of other people in the AI narrative and all these different 5 key pieces. So the pieces of the puzzle are coming together now, and that's good for Intel. It's also good for America's chip manufacturing prospects. So good news overall. And congrats to all the Intel shareholders that were believers early on and rode the wave.

Everyone, every US citizen.

Every US citizen.

Congrats to all Americans.

Every US citizen and every, every US debt holder. So, even the international folks that own Treasuries are in a better financial position today.

[Speaker:TREY_LOCKERBIE] You got to think about what this does to Trump's confidence level. He's like, I enjoy a bailout. He's like, I'll indulge Spirit Airlines, American Spirit. I think he's excited about the potential there. But why stop at just bailouts? Why not start taking a position in hyperscalers? He's like, look, I think I can move your market cap. Cap by 3x. Yeah, I've done it before.

Yeah.

Intel was— Intel was falling apart.

Yeah.

Imagine if I— imagine if he applied all his wisdom to a company already winning.

Well, certainly possible. Jim Cramer is excited about it. He said in 13 months, Lip-Bouton took Intel from a possible and unthinkable bailout candidate to one of the wealthiest companies in the chip industry. There is a Big Three of CPU: AMD, Intel, and ARM. And the agents need far more CPUs than these 3 can produce, so that means prices are going up. And they got the God candle.

Yeah, look at this candle from Facebook.

We love a green candle. Should be white suits today. Well, before we move on, there's some news. Justin Bieber brought Bieberchella back from the dead. Later in the show, we'll tell you what this says about late-stage venture reacceleration. We're going to tie those two together. For more than a year, Silicon Valley has buzzed about Cursors growth and whispered about its margins. Now on the cusp of a $60 billion bailout, Laura over at The Information—

Wait, they called it a $60 billion bailout?

Buyout, buyout. I'm sorry, buyout.

I already anticipated it.

We reveal the hot vibe, totally aggressive financials.

$60 billion acquisition, calling it a bailout.

It's not a bailout.

Their margins were rough.

We were just talking about— it got wormed into my head. Cursor had -23% gross margins earlier this year. That, Amir, says that is low for a company generating as much revenue as it is. This is the question of how sticky will Cursor be as an entry point to AI, as an entry point to inference demand? Can they reroute? It feels like with Composer, they have been able to retain a lot of customers. The company is still growing and we know Cursor users who have stayed with the, with the product while changing inference based on what their plan gives them. So they're subscriber and they will use whatever the plan, whatever gets the job done for them within the budget of the plan. And so obviously like you see -23% gross margins and you're like, whoa, And then you look at the SpaceX deal and the XAI deal and all the compute that's sitting at Colossus 2, and you think, oh, well, like, what will the gross margins be once the inference is happening on a Cursor-trained model or XAI-trained model? Something that merges their lessons from Kimi and Composer with Groq and all the different pieces of the puzzle come together. Are there gonna be higher gross margins? It's pretty easy to imagine that the gross margins would increase, but, uh, we'll have to, We'll keep tracking it. Did you want to play this piece from Patrick O'Shaughnessy?

Let's do it.

Let's play this clip of what Dylan Patel said on Invest Like the Best presented by Ramp.

Everyone's like, OK, so Anthropic just won, right? And Mythos in February, they never even released it because they didn't feel the need to. They're already sold out. Their revenue is already adding $10 billion a month. And then you've got Opus 4.7 today, all before OpenAI's alleged spud release, which media such as The Information and others have posted about. So clearly Anthropic is in the lead, right? And OpenAI is cooked. What's interesting is because Anthropic has such bounds on compute and they can only grow it so fast and sort of to the point of Dario used to gloat about how OpenAI was being too aggressive on compute and Anthropic was more sensitive. Irresponsible in their scaling, and now Anthropic is like, fuck, I wish we had a lot more compute. OpenAI is able to pay the bills perfectly fine. In fact, they've raised a ton of money to get incremental compute in addition to the irresponsible levels of compute that they were buying from Oracle and CoreWeave and SoftBank and all these people and Microsoft, such as Trainium. Now they're getting Trainium as well from Amazon. But what's interesting is if you were to say Opus 4.6, let's ignore models getting better over time, Let's just take diffusion of this technology. You and I may jump on the model immediately day one, but other businesses take time, and it takes time for people to learn. And so by the end of the year, let's say a 4-6 Opus tier model the economy would spend $100 billion on. I don't think that's unreasonable. It's spending $40 billion right now. Anthropic won't have enough compute to do that. And so, and presumably OpenAI and Google will hit that tier soon enough. Whoever hits that tier next, Sure, Anthropic may get to charge 70-plus percent gross margins, but if OpenAI hits it next, they charge 50% gross margins, they still get all of this incremental demand, and probably they also won't have enough compute to serve all the users. Sure, maybe Mythos is a model where if the world had enough compute, it'd be $500 billion of revenue or something crazy. There is such demand for these tokens and such limitations on compute, you know, and we see this with H100 prices skyrocketing and the useful life of these GPUs continue to extend, it's pretty clear even the tier 2 lab is going to be sold out of tokens, let alone the tier 1 lab. The tier 1 lab will have better margins, but the tier 2 lab will be sold out, and probably the tier 3 lab will also be close to sold out. Economic value that the best model can deliver is growing faster than our ability to actually serve those tokens to people via the infrastructure. And so this gap will continue to grow, and the model labs will continue to have expanding margins until people in the hardware supply chain, infrastructure supply chain, are like, wait, no, why don't I just jack up my Oh yeah, I love that.

Jam is such a good song.

I love that song. Bill Gurley, let's see what Bill Gurley, let's check in with BG.

What's he saying?

He says, I find this conversation foreign along with the argument that we are data center constrained or energy constrained. Historically in markets, price is a leveler of supply and demand. If you have a constraint, you price higher, you don't have surplus demand. But in this market, VC dollars act as subsidies as they did in consumer internet. Everyone believes if they have high growth, they get unlimited VC money. The biggest fear becomes losing market share. So growth at all costs becomes a game on the field. With that reality, you all— you are always going to have some constraint because you knowingly— you are knowingly choosing pricing that is out of whack with balancing supply and demand. It will continue until the major player feels they are forced to reconcile unit economics and profitability, as eventually happened in ridesharing when Lyft went public. Until then, you by definition have constraints. We can't disentangle true demand from subsidized demand yet. Some of the incremental demand is being engineered By excessive VC money forced into the system, uh, and the competitive dynamic between two companies that are losing massive amounts of money. No one can argue they aren't losing tons of money. Amazon and Uber maxed out around $2 billion a year. These companies could lose $10 billion or more in 2026.

Yeah, it's a good point that they are— that the big labs are losing money, but it doesn't feel like, uh, that the revenue of OpenAI and Anthropic are VC subsidized like it feels like the—

Yeah, it's funny, Semi Analysis is a good example. They spend millions and millions of dollars on tokens. They're not VC-backed.

They're not VC-backed.

Yeah, that's what I'm saying.

And their customers are not startups necessarily. It's like hedge funds and banks and stuff.

Hedge funds and banks that are printing by betting on semis right now.

Yeah, which is a different— which like, I'm totally on board with like, well, the Mag Seven are drawing down on cash flow and issuing debt and doing layoffs, and they're funding the— like, they're subsidizing demand, right? But the idea that VC dollars are the biggest drop in the bucket of subsidized demand feels like— I don't know, it just doesn't quite math out to me, because— the actual revenues and demand from Fortune 500 companies is so high, and the really big dollars in all of these rounds. I mean, Anthropic just raised something like $40 million from Google today, and that's not a VC subsidy. I mean, it is a subsidy.

Yeah, I'm surprised he's not talking about like big tech subsidy.

Yeah, yeah, maybe he is using the term VC dollars like broadly, which is like fair.

Including— Yeah.

Yeah.

And like, wasn't SoftBank in— was SoftBank in Uber? I think maybe. But there is an idea that you get to a certain scale and it becomes like sovereigns.

Guess how much SoftBank is up in the past month?

How much?

Guess.

200%.

Come on, John.

73%.

That's not bad.

That's a big number.

Yeah.

I mean, the ARM holdings alone, probably huge. So if he's using the term VC subsidies, VC dollars broadly, I think that does sort of— it is a reasonable point. But at the end of the day, you talk to most companies and most paid chat subscribers, and they're just like, I like the thinking model, so I pay $20 a month. I like the pro model, and I use it to create value. Or whatever subscription I'm on, I'm getting that much value in like software for my business that probably doesn't even exist. Like all the vibecoded software that we use here, like it's just not available off the shelf. We're building like completely net new products. And I think that that's like generally what's happening in the AI economy. There's been this like discussion for a long time. Martin Scully was sort of debunking it around like, oh, is this like, are these all like circular deals? And he was like, no, like the economy company is so broad that you have, you know, come, yes, there are, yes, like Google might take a position in Anthropic, Microsoft might have a position in OpenAI, but you also have completely main street customers and just average Joes who are paying, seeing ads, buying things. There's companies that are profiting off of running ads and other inference providers. Like there's like everyone is—

It's one big circle.

It's, yes, it's one giant circle of life that is, is actually self-perpetuating because it is a true economy that hits 25 different categories. Well, sorry, you might think it's over, but Elad Gil does not. He says, my view is the AI boom will only accelerate and is a once in a lifetime transformation. This is orthogonal to whether many AI companies should exit in the next 12 to 18 months. As some may lack durability versus labs, new entrants, or weird market shifts, but he's extremely bullish. He also posted a funny thing about his new life plan. He said his new life plan is to move to Brooklyn, get a neck tat, ride a bike everywhere, cold brew his own coffee, also start drinking coffee. That's an odd thing to jump straight to the coffee process.

Was this, was he like trying to, is this like a cipher of some sort? He wants to adopt a dog who loves grapefruit. Is there a secret message embedded in this post?

Maybe.

I don't know.

What's the Straussian reading? He wants to vape mango pods, but ironically, and wear a scarf everywhere during summer.

We have to pull this up.

Sarah Glow says, "No thank you." Pull this up.

We got to break it down.

Sourdough. I don't know. I have spent like a few weekends in Brooklyn. It's a nice town. I am not a fan of tattoos, so. No. I would stay away from the neck tat. Although, who knows? I've seen some cutting-edge neck tattoo technology that allows you to remove it very easily with certain light frequencies. So it can be sort of semi-permanent, easily removable tattoos. So if he wants to do this and then roll it back, I'm sure he can do this. It sounds like he might just want to spend the summer in Brooklyn, ride a bike everywhere, brew some cold brew, and then maybe get back to capital allocation.

Yeah, it's interesting. I mean, it sounds— I mean, it sounds like he's advising his portfolio companies.

Yes.

I don't care how hard you're ripping.

Yeah.

A lot of you should probably exit.

Yeah.

Just given how much uncertainty there is.

Well, later in the show, we have a fun story. A bear wandered into a backyard and took a dip in the family pool.

Cool.

We'll take you through it coming up after, later in the show. Uh, what this tells us about edge computing— get ready for it. Joe Weisenthal says defense stocks have been trading not like— not just like the Iran war is over, but like all war ever is over, says Barclays Capital's Alexander Altman. I would not expect this. Yeah, Lockheed is down significantly. Even pre-Iran War buildup, it fully round-tripped. Well, maybe that is good news. The goal of these companies is to bring about peace and defense. And so perhaps it's over. I don't know. It's hard to track.

Has Jane Street achieved AGI internally?

I think so. They definitely have—

They did $40 billion in revenue last year, more than all the big Wall Street banks, and with only 3,500 employees. Employees.

Yes, let's give it up for Jane Street. Fantastic. Capping a stunning ascent to the peak of the industry, the firm flew past global investment banks after reaping $15.5 billion in the year's final quarter, according to people with knowledge of the results who asked not to be named. They're leaking. With only 3,500 employees, it beat the nearest rival, JPMorgan Chase, by 11% during the year. Tether did $15 billion at 200 employees. So maybe, maybe they aren't the hottest shop in town, but that is very, very impressive.

That's like one of those questions like, would you rather do $40 billion of revenue with 3,500 employees or $15 billion with 200 employees?

Well, we have some exciting news out of Thrive Capital. Josh Kushner announced a new fund, Thrive Eternal, a permanent capital holding company that will be concentrated in a small number of assets that we can own and steward over many decades. Across Thrive Capital and Thrive Holdings, we are building and investing through a moment of exponential change, backing emerging technologies, the infrastructure that powers them, and the businesses they can transform. Increasingly, he sees a fourth category. These are assets with qualities that cannot be replicated by technology— iconic franchises and cultural institutions rooted in tradition. And he says his first partnership is expected to be with the San Francisco Giants, an institution built on more than a century of shared identity and community, and among the most iconic sports franchises in, in America. I looked it up before the show started. The San Francisco Giants, if you're not familiar, it's an American professional baseball team based in San Francisco. Baseball is a bat and ball sport played between two teams of 9 players. Each taking turns batting and fielding. The game occurs over the course of several plays, with each play beginning when a player on the fielding team called the pitcher throws a ball at a player on the batting team called the batter, and then the batter tries to hit it with a bat. And so they play this game, they sell tickets to the game, and that's how the San Francisco Giants make money. And Josh Kushner, someone could make a CBN-style show.

Yes. For whatever you just described. I feel like it would be very cool. That would be cool if you had, that would be great, you know, on-screen graphics and you could provide kind of live coverage maybe while that's happening.

Extremely educational. Well, we have our first guest of the show. George from CrowdStrike is back in the TDPN UltraDome.

He's in the waiting room.

George, how are you doing?

I'm doing well. How are you guys?

We're doing fantastically. Welcome Welcome to the show. I would love to kick it off with just an introduction on Project Quilt Works, what's new, get us up to speed on CrowdStrike.

Well, first I have to congratulate you guys for your success there. So I got to start with that, but we can get back to that later. But Project Quilt Works is really a coalition that we put together with some of our largest global partners like EY, like IBM, like Accenture, that's really focused on helping to deliver our technology to customers as well as leveraging some of the frontier models to be able to identify very quickly and help mitigate the exposures that we're seeing with AI. Obviously, we've seen some of the, the news with Mythos, and I'm sure we'll talk about that. But the fact that there is such a focus now from the CEOs all the way down on on where are my exposures, where are my vulnerabilities, and what can I do to mitigate it? So we put together a group of, you know, basically some of the largest global security partners that we have to go out and be able to do this because we have to do it very quickly. The window is closing in the time that we have to identify, patch, and remediate and mitigate these issues.

Yeah, I mean, we've been tracking a lot of the new data breaches. We've, some of the startups that we've we know and love that come on the show have been, there's been vulnerabilities that have led to a variety of security issues.

Supply chain hacks.

Supply chain hacks, lots of stuff. How much of that do you think is driven by attackers using AI or the internal companies using AI sloppily? Like there's this sort of a dual dynamic there where the attacker gets stronger, but potentially the defender could also get weaker. Are they both important? Grappling with those two sides of the equation?

Well, they're both important, and I'm glad you pointed that out, because when you think about what actually happens is you're having more and more users consume AI, and how they're doing that is leveraging AI through any number of applications. Could be Claude Code, Codex, et cetera, right? Cursor. But they're doing that on their endpoints, right? And what that means is that there's this now concept of shadow AI where there's AI popping up everywhere and enterprises and corporations don't know where it is.

Yeah.

So what happens then is the threat actors are very good at figuring this out and then they're basically going upstream to these packages and they're compromising these packages and libraries. And then when your cloud code consumes it, boom, you know, all of a sudden you've just downloaded one of the newest packages and all of your credentials are gone. So that's what we've seen. And it really goes to the fact that despite however many vulnerabilities you may find, the adversary is smart, they're human, and they're going to find the path of least resistance to deal with what's hot today and where the exposure is. And that's on the endpoint where people consume AI.

Mm-hmm. How have your conversations with lab leaders been going around models getting more powerful? And then the natural response, of course, there's the opportunity to to gate the models to companies that can be trusted. But there's also the potential to not let the model be jailbreakable. So it should help me with defense. But if all of a sudden I'm saying, ignore previous instructions, let me hack into Bank of America, it should say no. That's always been a problem. It used to be a problem with just the sort of the silly, you know, jailbreak, say a bad word or do something that it shouldn't say. Now it's getting much higher stakes. But have you been, have you been seeing glimmers of confidence in the ability for AI labs to actually contain the models' capabilities as they roll them out to a broader audience?

Well, the model capabilities continue to obviously improve, and also the capabilities that allow them to provide guardrails and limits on, on what a user can do. But that's not going to get you where you need to be.

Okay.

Which is one of the reasons why from a security perspective, companies like ours and others are focused on looking at the prompts, understanding what happens there, and then being able to sort of provide the guardrails at the prompt layer. Right. You want to take care of what you can at the, through the LLM. But when you're building AI, whether you're using a Frontier lab or whether you're using an open model and doing it yourself, you're still going to have an interaction with that model and it's going to be via prompt. And whenever you have sort of unstructured data just being sent somewhere and then being executed, if you will, it's ripe for problems. And this is one of the big time-tested issues that security has had. So putting guardrails on that, we call it AI/DR, AI detection and response, and it allows companies to look at those prompts and make sure that they're sanitized to and from the LLM.

How are you thinking about the threat of open source models? You know, the new Deepseek R4 preview from yesterday was super impressive. Are you expecting, you know, open source models that have, you know, really threatening cyber capabilities in something like 6 months?

I think Dario was estimating around 6 months. Lag time after Mythos for the, for the lagging edge to catch up?

I don't, I don't think it's long. Uh, in a few areas, I mean, if you look at, uh, Kiwi 2.6 that just came out, right? I mean, these are impressive models that are out there. And what you have to realize is even the frontier models themselves, the ones that are public, are still very, very capable. Like, it is extremely capable. So you can find a lot of vulnerabilities. You can basically chain these together and create sort of an automated attack. And that's what we're seeing. Obviously, some of the private models have the ability to actually chain these sort of vulnerabilities together more like a human. They've got more reasoning capabilities, which is one of the, you know, really powers of their models. But the other models are, A, are not far off. And the open source or the open weight models really are going to be focused on catching up very quickly. So the window, as I said when I started the program with you, is very small. And that's why we started this coalition with Quiltworks to be able to find these at, you know, again, with the, the board mandates all the way down, find our exposures, fix them, and make sure bad things don't happen. And there's not a lot of time to do it, unfortunately.

How are you thinking about AI-powered but potentially non-automated cybersecurity risks. I'm just, we were talking to a founder who's working with, you know, engineer recruiting and he was having luck using AI agents to send outbound emails. And you can imagine that, you know, phishing attacks will get more sophisticated, the ability to, hop on a Zoom call with someone that looks like your boss, but is actually an AI avatar. All of these new risk factors are popping up. And I'm wondering if this is something that we potentially need to be more aware of and not get distracted by, not get overly distracted and forget about the human element of security.

Well, the human element is really the weakest link. I call it a layer 8 problem, right? The problem is between the, the keyboard in this chair. Yeah. And, uh, that has been the case for a long time. So let's take your example. Yeah. One of the most prolific groups, threat actor groups, that we actually track is called Silent Chollima, which is North Korea. So North Korea has been very active in getting hired into a company, right? And then basically the laptop that you send them gets sent somewhere in the U.S. to a mule, and that mule takes it to a laptop farm and then the North Koreans control that. So they already just bypass all of your security because you just handed them a laptop.

Wow.

Okay, so we're seeing lots and lots of those. And it's funny because we buy— we really— I started identifying this before anyone even knew about it, about 2 years ago. And we first started notifying customers like, hey, we think you have a North Korean. It's not really Bob in Texas, it's a North Korean. Which obviously is always a little dicey when you're You're talking about an employee that may not be an employee. So we told everybody and then finally they tracked it down. They said, yeah, you're right. And they went to the hiring manager. They told the hiring manager like, hey, we need to get rid of this person, it's a North Korean. And his response was, well, they do really good work, can we keep them? So you have to think, I mean, I can't make these stories up.

But yeah, wouldn't there be an incentive at some point if they did get somebody on the inside to just actually staff a bunch of super talented engineers on that one individual employee. So you have 4 remote engineers doing the job of 1 real employee, right? So it looks like, wow, this person is insanely high output. They're working around the clock. Why would we get rid of them? They're incredible. There's no way they're just— you know. Crazy farming us out.

That's exactly, that's exactly what they do. So we're seeing that, we're seeing agents go rogue, which is problematic. Um, one of the, one of the big challenges where AI is, uh, is goal-seeking. So you create an agent, you know, one story is, uh, is a customer that created like 100 agents. One agent found some issues in code, wanted to fix it, but it didn't have access. So it went back to the Slack channel where the other 99 agents were hanging out and said, hey, I'd like to fix this issue. I don't have access. Who has access? One agent put its hand up and said, I can fix that for you, and happily fixed it and basically worked around all the security boundaries. So goal seeking— I don't think we talk enough about goal seeking, because when you put an agent on it, depending on what model you're using and how it's set up in the harness, they just go nuts until they actually get to the goal. Yeah, they'll actually steal credentials out of your keychain, you know, if you didn't give it to them. So they can keep going.

Wow.

Yeah.

That's what we're seeing.

Yeah, because in one way that, you know, if the goal is achieve the task, it succeeded, but at what cost? And it needs to know that there are rules for a reason. What are some of the other cybersecurity tasks that are being successfully augmented with AI or AI agents these days? I'm thinking of obviously vulnerability testing. That's something that, that great cybersecurity researchers, great programmers have been able to do for a long time. Now it is amplified a millionfold with artificial intelligence, but are there other maybe more bespoke or soft skill tasks that cybersecurity researchers are able to amplify their efforts with by working alongside AI?

Well, AI is gonna change the way the SOC works, the security operations center. And one of the things that I've talked about in our last user conference was really helping to try to pioneer getting to security AGI, which is a fully autonomous SOC, what I would call a Level 5 SOC, right? If we think about car autonomy, you've got 5 levels. You map that into security, 5 levels. The 5th level is it just does everything for you. Now, as an industry, we're a ways away. We still have human in the loop. But what happens is then you're really taking a lot of this voluminous information and you're letting AI agents just grind through what you're really good at, and it cuts down time and effort. And then you're elevating these Tier 1 SOC analysts and you're turning them into Tier 3. And that has paid dividends. But the soft skill is it would take one of our customers, give you an example, would take them 4 days in writing a report to, you know, what they call a sitrep report, the situational kind of response, what happened. And now 4 days turns into like an hour because it can all be automated using something like Charlotte, which is our agent capability. So all of those things that were problematic, you can just get through through and you let AI do what it's really good at, grind through lots of data, look at lots of sort of information, write reports for you, and then your AI agents can begin to take autonomous action. Some again by themselves, some with human in the loop, but it's really gonna make the SOC a lot more efficient.

Jordi, anything else?

More fun question than, less scary question. How is, How is AI changing Formula 1?

Well, there you go. This is a great question. And it's interesting because I think there's different levels of adoption of AI between all the different teams.

Yeah, I was gonna say, yeah, one team might be, they might be spending a ton on inference. Another team is like, ah, we use radios and we're doing it the old-fashioned way.

Well, what's interesting is, there is some talk, and I'm not sure exactly where it stands, where AI will be part of the cost cap. So if you think about wind tunnel time and all the expenses, so, well, you have an incredible technology that can, you know, dramatically automate things. All of a sudden now it's part of the cost cap, and then they're going to have to rein it in. So we'll see where it all lands, but that was some of the talk, uh, that has been happening. And I hope it doesn't happen because I think AI can really help things, But we'll see where it all goes. And obviously, I'm trying to help out the Mercedes team in those areas as well.

That'd be great.

Yeah, do you think there will be, 5, 10 years from now, do you think there will be, like you'll be able to look back and be like, okay, there's a pre-AI era and a post-AI era in Formula 1, or will it be more subtle because a lot of these processes are just already so micromanaged and optimized?

I think you'll see like a pre and a post. The challenge is you're going to have to run through a lot of the different engineers that have been in Formula One for many years, because just like in the corporate world, you know, there's resistance to, well, what is AI going to do for me? And there's resistance in Formula One in some cases from an engineering perspective because, you know, they're doing it themselves. They're the ones that are coming up with the calculations and that's just the way they did it. And it doesn't mean they're it's right or wrong. It's just the way they operated for so many years. So I think just as we see kids coming outta school, they're AI natives, right? They just gravitate towards that. And when you start to see AI-native engineers, you're gonna see more and more adoption of that in Formula 1.

That makes a ton of sense.

Makes sense.

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

Yeah, you guys, it has been a wild, wild quarter, but come back on soon.

It's great progress.

We will.

Great to see you guys. All right, congrats again to you.

Goodbye.

Thank you.

Uh, after our next guest, we will be diving into the Summer House reunion, which has gone off the rails. Uh, we're gonna explain why Bravo is the cleanest lens for understanding US-China decoupling. But first, we have Professor Sendy. He's a word combo guru. He's a wombo. Guru, welcome to the show. How are you doing?

Hey, well, thank— good to see you guys.

Good to see you.

Thanks for having me on.

I appreciate it.

We're thrilled. We're huge fans.

We're huge fans. Honestly, I discovered— I, I think I was pretty early. I will say, when I discovered your account on Instagram, I, I did not have a single other person that I knew following you.

Okay, so you're day one. Oh, so you were one of the OGs on the channel.

I think I, I think I'm an OG, and I, and and I kept sending them to John, and I'm like, this is the funniest thing in the world, right?

It's so good.

And so we've been trying to—

That's awesome to hear. I love the original fans. You guys are amazing. You guys, like, really, you're the ones that fueled the fire for me to keep going, so.

I'm glad, I'm glad.

Okay, so let's get right into it. I, like, you're the professor, like, interview. So we, this is a technology-focused show. We've been using Wombos in the show, but a lot of people in our world in the technology industry are not using Wombos very aggressively. Aggressively yet. We think that's going to change. You know, there's this explosion of AI adoption. We're expecting an explosion of wombo adoption.

Let's hope so. That would be amazing to see. You know, I've seen these wombos be spread amongst not just young people, but, you know, the older generation is starting to use them a little bit too. And I'm thinking about expanding it to sort of include some of the older generation slang as well. And I've I had some people request that, so I'm trying to expand this and try to make it more universal.

Yeah.

What, so what was the first wom— what, like, what, what really drew you to Wombas? What was the first—

you know, honestly, I saw a video and I basically just— it was, they had put the low kirkentologic, a low state, like, word together, and I was like, oh, I can just make this longer. And I just went ahead and added a word to it, and then it kind of blew up, it took off, and I was like, okay, I guess I'm pretty good at this. So I just kept going with it, and then before you knew it, I had basically the world's longest word ever said in one breath. So now, you know, I have this, you know, the— this crazy long word now that, you know, people are trying to learn. And, and it's just grown from there. And now I'm just creating original Wombos. And I'm sure you guys have seen, and everybody's seen, that, you know, people are using these Wombos all over the place in schools and, you know, just out in daily life. And so it's really been a, a movement that I was not expecting.

I love it.

Do you—

can you, can you one-shot your longest word? Like, is it on, is it on command, or do you, do you need to iterate through it?

No, I can do it on command.

Do you want to, do you want to do it?

Sure, I'll try it. Okay, here we go. Lulu, lemon, Kirkalistic, low, hieroglyphic, go express, Sopotamian, Kirkalistic, workalistic, with the logic, wow, there we go. I wasn't expecting to do that, so I can't believe you pulled that off.

Uh, are, are you actually a professor? Like, what, what is your background? How did you get into this? Was this just a side hustle that just went mega viral? Like, what is your story?

Yeah, okay, so basically I don't, I'm not a professor in real life, but I have a master's in education, but I don't work in education.

Okay, got it.

So I don't have a background in words either. I just, I used to be a rapper a long time ago. People ask me, how are you so good at doing this? How do you put these words together? And I just tell them, you know, it's a natural skill I have. And then I used to do it back in the '90s. I was like one of the original white rappers from back then before Eminem was even around.

That's amazing. Amazing, wow.

No way.

Yeah, so I've been doing it for a long time, so.

Yeah, we were reflecting earlier on—

Born too late to explore the Earth.

Yes.

Born too early to explore the stars.

Yes.

Born just in time to be the leading expert on wombos.

Yes, for sure.

That's perfect, yes.

Yeah, we were reflecting on the fact that, I'm not sure how familiar you are with this, but like Silicon Valley in particular went through like a series, —call them portmanteaus back then, but wombos—were very popular. You have Pinterest, Instacart, Instagram. And there were like 1,000 other companies where that was the default pattern for naming a company, because the dot-com would be available. It would be very hard to go get, you know, pin— you know, interest.com, or, you know, order shopping— like, you know, grocerystore.com. But Instacart was available, and so people would sort of mash these up. And I'm wondering if you've, if you've gone back and looked at any historical portmanteaus or wombos that have broken through and maybe need to be like revisited and re— reignited, or is it all just forward into the crazier and crazier phrases? You know, it's—

I think it's mostly forward-looking. I've seen a couple that were old school, so I know there was one that was like a calculator It was like an old ad from the 1979 or something like that. Yeah.

This is a calculator. This is a calculator that's a lighter too. Like if you want to smoke while you're doing math. Boxy calculators and it's got a little lighter thing coming out. We need one of those.

That's amazing. Yeah. And so another word. So I haven't really visited a lot of older words, honestly. So there was a word that came out in the dictionary the other day, choppelganger. So that word is actually not mine. Okay.

I'm not, I know, but you, Merriam-Webster needs to hire you. Yeah, like immediately and start getting—

I think they do, you know, I want to get, I want to get quiche peenimus in there. Yeah, that's one that I really want to get.

Explain that wombo for people.

Explain that one. Yeah, so, so basically quiche peenimus is like one of the big ones that people are using, and so that's like quirky, niche, and then peenimus is peak cinema analysis.

Sorry, I just, I was like, I'm like the overachieving student, you know, I studied, so So we, we, we integrate these words. Yeah, we, I personally, I use, I use, uh, I use Quiche Lorraine more.

Yes, quirky niche lore explained. Yeah, that's how you want you, Professor. Yeah, so I love Quiche Lorraine.

Um, I actually have a new version of that. I don't know if I can actually say it on your show. I, I don't know the, the limits, but, um, so it's Quiche Lorraine-visions. Oh yes, so it's basically like you take Quiche Lorraine and then you mix it with anal vision. And I apologize if I can't say that here.

We're trying to keep it a little bit clean.

We're a family show, but it's analysis vision, to be clear.

Is that right? Yeah, it's analysis revision. Revision. So it's basically, if you do one of those, you're basically just doing an analysis revision of it.

Okay. That's correct. Yeah. How much time are you spending trying to come up with, like when I think of quiche and Lorraine, I think of, you know, short usable phrases that actually could work their way into the American American vernacular, whereas with the low-content ideological, I don't think anyone's actually going to pick that up. It's more of like a stunt. It's impressive, but it's a different style of content.

How much— I use Lorraine all the time. I'll say, John, Lorraine Intel's earnings.

Yes, yes, yes. But how much time are you splitting between these? Do you think that there's something innate about being terse and short that will actually allow a wombo to break through and ideally make it into the, uh, Merriam-Webster's Dictionary.

Yeah, I actually have a theory behind like how to make effective wombos like that. So I think like you're saying, short wombos are good. Yeah, something that sounds good. Yeah, it has to sound good. Yeah, so it's like the quiche, it sounds good. I, I don't know, it just, it sounds, yeah, nice when it sounds like a real word.

Just don't work. Yeah, it's almost like, uh, where, you know, the, the pharmaceutical companies, they come up with these names for the drugs and you're like, that does sound like a thing, but I never would have guessed what that is. And it doesn't sound just like you pounded on the keyboard. You got like ozempic, which like sounds like a drug for some reason. It's just like, but it's from nothing. And those things tend to break through. Also, I feel like choppelganger really jumped straight into Merriam-Webster's because it feels like with a very little amount of work, you can reverse engineer it. And that also seems to be key to getting some wombos to break through is, can someone take a guess and get 80% of the way there and then sort of pique their interest and then go a level deeper and understand the whole lore.

Talk about the origins of nonchalash out. That's one that we've been using a lot.

That's a really good wombo.

Once you know that wombo, which is nonchalant crash out, you just see it everywhere. You do. It's constant. It's everywhere. Every time I open X or Instagram or YouTube, someone's having a nonchalant out. Yeah. Yep. Pretty much.

So yeah, nonchalant out is basically when you look nonchalant on the outside, but then on the inside you're crashing out. So I mean, I think a lot of, uh, people can, can do this. Uh, you know, for example, that, like, the CEO at McDonald's, you know, when he was eating that burger, I'm pretty sure he was having a nonchalant out when that was Yeah, he looks pretty calm on the outside, but on the inside, I'm pretty sure he was freaking out.

He was fighting. He was in a dark place.

So, so with something like nonchalant or any of these other examples, how much of it is— there is a phenomenon that— because a nonchalant is a thing that we didn't have a word for in English before, uh, but it has existed probably for, you know, entire human history. This has happened, uh, but How much of it is you see something and you say, that deserves a wombo, versus you're just sort of grappling with different phrases and then you piece it together and you find examples.

Another example from the professor, cap-solutely. Cap-solutely. I would use this personally if I know someone's just basically BSing, kind of like lying to me, I'd say cap-solutely. Yes.

You want to agree. Well, here's the secret, OK? So I'll let you go. I'll let you in on this. So when you, when you make a word, if you find an opposite word and you kind of make the word a paradox— so like a nonchalashout doesn't really make sense, right? Because it's like, how can you be nonchalant and crash out at the same time? But when you put them together, you really think about it, all human behaviors are kind of paradoxical. They all, they all have like, you know, this aspect and this aspect. So like another example would be liguratively. So that's figuratively and literally. How can something be figurative and literal at the same time? Time. Yeah, but it's kind of possible if you really think about it. You can like literally mean something figuratively, or you can figuratively mean something. Yeah, you can literally mean it halfway, you know, the Hegelian dialectic, the thesis and the antithesis create the synthesis. Uh, jam them together and you can really come up with some beautiful wombos that way. That's, that's really kind of the secret to a lot of it.

Yeah, for sure. Uh, where do you want to go with your creator journey? I, because I, I was thinking Instagram Reels, notoriously hard to run ads or monetize. I don't know that there's an obvious like Patreon paid podcast play here. Maybe I would be interested to listen to you talk about etymology and language for hours, but I was thinking like brands should be paying you to create Wombos about their products or their brands. And that would be, be sort of synergistic with like what the audience expects, what would actually go viral and what you could actually get paid for. But are you early in the creator journey? How are you thinking about this as a business?

Yeah, so right now I'm working on a dictionary. So I'm hoping to use that as a, you know, as a stepstone to launching my products. That's great. I've got it.

Yeah, so I'm watching YouTube. I think coffee table, like a coffee table book. Yeah. Like the great book of Wombos. Yes. Something you can proudly place on. A coffee table.

Yeah, it's a conversation. I want to have a book like where I can put it on the table and people can just like flip through it. I want it to look real nice and I want to be like a really great item to have.

But also, you know, like a coffee dictionary, a coffee table dictionary, a coffeectionary. We're working on these. Sorry.

There you go.

You're coming up with stuff. I'm working on it. But yeah, what else?

You gotta keep on going. Yeah, for sure. But, but you know, I'm early on in my journey, so I'm looking to, you know, work with some people, and I'm looking to make some books and— cool. I'm just looking to keep on building what I'm doing right now. And I've only been growing for like 4 months now almost, so I mean, I really haven't— you know, I was, I was barely there at— in January, and so it's just exploded. So it's, it's still kind of the beginning right now. So yeah, um, yeah, there's so many opportunities, so it's really looking up.

Yeah, we had someone on the show whose, uh, whose whole job is coming up with names for, uh, tech companies, products. Didn't he come up with the name Sonos and a few others? And I feel like that's a— your grasp on the English language is sorely needed among some companies that have very complex product names. And so I see a bright future in a bunch of different ways, but eagerly await about the coffee table book. Anything else to No, this was super fun.

We're having Gary Vee on in just a couple minutes and we'll ask him ideas on how to monetize Professor Sandeep in the Wombo Universe because I feel like he could, I think he'll have some ideas.

Yeah, that would be amazing. And yeah, if you're, yeah, mention me to Gary Vee, that would be very cool. That's awesome.

Yeah, I appreciate that. Do you experience anything different across the platforms? We follow you on Instagram. I know you're on TikTok. Uh, is there anything, uh, that you've learned from like the different platforms?

Definitely. So TikTok I think is more for discovery. Okay. So if, you know, if you do something that's completely novel, that video can really blow up. Yeah. I think Instagram is much more generous with its views. Mm-hmm. Um, but I think you can really build a solid base more on Instagram. I just feel like TikTok's more for discovery. Instagram's more for growth and personalization.

Of your brand. Have you put these on YouTube Shorts yet? What's that? Have you put them on YouTube Shorts yet? Just because it's the same platform.

I haven't. Um, I used to do YouTube and I just, I don't know if it's going to fit on, on that platform. I just, I don't know. I'm, I'm hesitant to go there and I've been hesitant to go on Facebook Meta.

Oh yeah.

Um, just because I don't know if it'll work and I just, I, I used to do that and so I'm just, I'm not sure about it yet. So we'll see. We'll see.

Okay, well thank you so much for taking the time to join the show. Fantastic to talk to you.

And thank you for all your analysis and teachings. Yes. I, uh, it has brought, brought both of us so much joy. Yeah, like not just watching the videos but using them.

We got our lawyer to use, uh, quiche Lorraine once, and we were very, very happy about that.

That is, that's a high compliment. I appreciate that. Thanks for telling me that.

It was, it was remarkable.

Yeah, Jordi and Jon, I appreciate you guys having me on. I really do.

Yeah, yeah, you're the man. We'll talk to you soon. Great hanging. Have a good day.

All right, thank you.

Bye-bye. Uh, Greg Brockman's a part the apartment is for sale where they started OpenAI. Uh, Shiel Monod says, I do not think there is an opportunity to build an OpenAI-themed Airbnb, by the way. Uh, but, uh, who knows, maybe some filmmakers will pick it up. Uh, it is a, uh, 3-bedroom, 2-bath in San Francisco, $1.545 million, uh, 1,800 square feet. Uh, it is a— it is an incredible piece of, uh, of lore that OpenAI started in this particular apartment.

Yeah, this feels like it will go quickly. Yeah. Like there are so many people in San Francisco that might need a place and to just snap up the place that OpenAI was effectively formed.

Yeah, I remember some of the early photos. Good times. We should go to this video from Zain Shah. Imagine every pixel on your screen streamed live directly from a model. So this looks like a website. And when I saw this, I thought that it was a website. I saw a URL at the top and I was like, oh, this is a cool demo for some sort of, you know, tour of Notre Dame website, probably vibe-coded, a testament to what the modern coding models can do. But it's not code. This is— what you are seeing is actually not code. It is— there's no HTML, no layout engine, no code. It is a diffusion model that is rendering this as the input is given. And so, you can think about this almost like those world models, those generative video games where you walk around, but this is sort of a very, very early glimpse into sort of what the future of generative UI might look like. Jordi, why are you laughing?

Because the chat is doing a bunch of wombos.

Oh, okay. We got to work on wombos. We will be integrating those more in the near future. But we have our next guest in the waiting room, Gary Vaynerchuk from VaynerMedia. Gary, how are you doing? Welcome back to the show.

Do you know about wombos? I do not, my friend. You got to learn about— You got to learn about wombos. They are the next meta. They are the next meta. If you're not doing wombos in 2026, You're getting left behind.

So, wombos are word combos. So the example would be a queesh is quirky and niche. You put it together and that's queesh. Or Lorraine, lore plus explain.

Queesh is a, that was a risky start for the first word.

Yeah, that's a risky one.

Lorraine is a better one. Give me, if I say Lorraine. Boys, first and foremost, congratulations. Thank you. Thank you. I really mean that. Thank you. Thank you. I remember our very first call. I think it was in Q1 of 2025. So we were just maybe like a little, just a few months into the show. And you, even at that point, we were very, very, very small. I think we had maybe just gone live a few times and you told us, you guys, you said go harder. We were already going pretty hard, but you said go harder. And, uh, we certainly did.

So, uh, and go multi-platform. That was really huge too. You, you, you were, you were very early in like, why don't you have a newsletter right now? Why aren't you on YouTube in multiple ways? Why aren't you on Instagram yet? And so we did the uncomfortable thing of posting pretty subpar content for a while, but it started the compounding very, very early. And now it's really much better.

Everything starts subpar, right? Like when you start working out, or when you start singing, or draw— like, you know, like, I— for everybody who's watching right now, every business, every personal brand, every B2B, every B2C, the— to not take advantage of the attention media landscape that's in place right now, um, it's really incomprehensible to me. There's never been a time in the history of humanity or business where the cost of distribution is zero for the distribution. It's— there's costs in the content and the way you do it, but the upside is so extraordinary against the investment. And when you frame it up properly, multi-channel, multi-format, the, uh, the business outcomes can be extraordinary. And, um, I'm happy to see you guys get one.

Yeah. Well, thank you. Are you—

is it, is it, is it funny that, uh, is it funny to you that that it feels like in some ways we're like a decade into live streaming, and yet people are—

Two decades by some measures.

Well, yeah, two decades by some measures, but like Twitch has been big for a long time. There's been so much attention there. I mean, I wrote this book in 2008.

It came out in '09, Crush It. In the back of the book, I talk about Ustream and live streaming. Yeah, I mean, it ebbs and flows. And to your point, like, I mean, I can't even comprehend the economic impact live social shopping is going to do over the next 10 years for humans and consumer businesses. Yeah, it's all the same stuff, brother. It's always been around forever and it's always the beginning, right? Like, it's just kind of the way consumer behavior and human behavior works. Yeah, it does. It does surprise me how many more people have not replicated the very basic Strategic framework that you guys executed in this genre?

Well, the interesting thing is there's easily been 100 shows that have like taken some level of inspiration from what we're doing. It's still very, very hard. It's still very, very hard to break through, right? It's not just the overlay and the format, but I have been surprised. I've been very confident. I've said this on podcasts for, you know, we've talked about this going back probably 6 to 8 months ago. You should take this format of a livestream and you just add an extra level of production beyond you just have a laptop there and you're hanging out. And you apply that, the example we'd always talk about is like cooking, where all you need is like a cool kitchen. That's your set. You have a couple of cameras, so you have different angles. And then you just say, hey, every day at 3:00 PM, I'm gonna cook dinner. This is what I'm gonna cook. Here's the schedule. You can make dinner with me. And I think that that's like an entire niche that you could build around. You could integrate guests into something like that too. So you have different conversations of content creators coming through every day. And then I think you can apply that to a bunch of other things. Sports has honestly been the most advanced in terms of live streaming. So credit, credit to them. Yeah.

Yeah. I mean, it's, it's, It's there for the taking.

A lot of things are there for the taking. Yeah. How are you thinking about how branding and marketing will evolve? I had this interesting moment this week with the GPT Images 2 launch where suddenly any brand, like basically fully democratized high-quality product photography. Now there's still categories where like, I don't want you to AI generate the product photography. Like even apparel's actually interesting where like, fit matters a lot. And so if you generate a bunch of AI images of your shirt and then I buy it and it doesn't fit well, like I'm not gonna be happy about it. But for a bunch of categories, it's cool. And I always kind of have this like kind of strange moment where it used to be you could identify an entrepreneur's ability by their product photography in some way. 'Cause even if somebody's like bootstrapped and scrappy, like they would find that friend and say like, "Hey, do me a favor, help me get some great images." and you could kind of categorize like a company that you're seeing online, like, okay, could this person figure out how to get great product photography or not? But now the bar is just so much lower. You're like a couple prompts away from it. And so it feels like it's gonna be harder than ever to stand out online. Well, I mean, there's a lot there.

I mean, it's, oh, we just got done talking about how hard it is to stand out online, right? Like the cost of entry is zero. But, you know, this is a competition. Like, you know, everyone's trying. And so everyone will make content, video, picture, audio, at a level that is incomprehensible to all of us that were born prior to 5 minutes ago. And this will be another transition. I think, you know, there's, there's always going to be a timing game to this, right? So right now, you know, like this open claw, you know, allows me to do things that a lot of people aren't thinking about right now. And me knowing that and me playing with that. And then even more interestingly, how creative or strategic am I with the Agentic agents and what are they doing for me and how do I understand and blah, blah, blah, blah, blah. So I think it's kind of always gonna be the same thing. Whether it was electricity, and some people put electricity in their home and other people were scared to because there was demons in it, or, you know, the car, or the, the, you know, the typewriter is one that I've like been fascinated by, the competitive advantages of the companies that actually brought typewriters in versus, you know, putting penmanship on a pedestal, you know, like, and computers and the internet and mobile devices and open source versus closed source and social media. Now, AI, you know, look, this AI thing is no joke. We all know that. It's big stakes. There's a lot to it. But, you know, I still think whatever that human was to being scrappy to find their friend for the photography, you know, that scrappiness is gonna be deployed into something else. Totally, totally. And so I also think, right, and I also think that we're about to see the explosion of analog. Right? Like, I think this barbell that I keep thinking—

Totally. I, I've been thinking like, okay, you want to start an apparel brand? You want to go on Instagram or TikTok and duke it out with like 10,000 other brands? Or why don't you like find somebody that has a retail store that they can't rent out and like do a deal with them and just try to get big in your hometown again? You know?

And, and my argument would be, would be, and yeah, right? Yeah. Like, to me, this part, like, I just see Extreme AI, I think, is creating extreme analog. I really do think it's a barbell. I think in the next 10 years, obviously it's 2036, but I feel like it's going to feel like 2050, which actually is bringing the rise of 1950. I couldn't be more impressed with what Ari Emanuel and others are doing that are investing in all these analog businesses. I could not be more interested in physical retail. In event-driven businesses, in concerts and venues. And I think the rise of analog— I have a restaurant business, part of a restaurant group. I keep pushing my partners who are really operating, let's open a restaurant that makes people check in their phone as soon as they walk in. Let's put people in group tables. I like that. You see what's happening with flip phones. Yeah. Gen Alpha buying them. We see vinyl sales. I've been very at the forefront of collectibles. I felt collectibles was something tangible and was a gateway drug to community. If you've never been to San Diego Comic-Con or the Sports Card National or Fanatics Fest. So I think there's a lot of interesting non-digital realities that are coming as a countermove to the insanity of AI advancements. We're literally within a half decade not believing a single video, not a single fucking video that's on the internet. In 5 years, if we're having this video, this interview right now, most of the audience is trying to figure out if we're real or not. That is very real and has real substantial counter opportunities. The photographer who's sad when they hear that, I'm like, no, no, no, you might actually crush in a different way. I'm curious to see what the counter-scaled moves are going to be of the next decade. I think for any real entrepreneur, They're not crying about AI killing them. They're curious about what AI at scale is going to create opportunity for them.

Okay, totally. All right, let's go deeper on analog. I want your reaction to this headline that for the first time this century, vinyl music sales eclipsed $1 billion in a calendar year.

Sales— yeah, they're finally—

records are up up something like 10% in 2025.

Yeah, they're effectively like 10% of like global streaming revenue is just vinyl still. It's bigger than CDs.

It's making a bigger comeback than other formats. What is your reaction? Is it people looking for wall decorations? Are they actually listening to the vinyl? Do you have any idea of like what we should read into that idea of this nostalgic format coming back?

I mean, do you see my shelves here? Only thing I've been thinking about for the last 7, 8 years— or not only, but like very hot on this. Yeah, I mean, the answer is yes to both. Some of it's happening with decorative, a lot of it's happening with collectibles. Even more is happening with subconscious counter push to extreme digitalization. Sure. Yeah, there are people— this is, by the way, one of the reasons we're not going to see unlimited television commercials and social media content that's just pure AI from the biggest companies in the world is every time they try to go there, you probably have touched on your show, McDonald's, others, they get such backlash because the whole world is so scared that AI is going to take their job that the consumer is pushing against it. And then that's one part. And then the second part is, yeah, people are starting to do counter behavior consciously and unconsciously. And, and then there's just swings of swings of trends. Like, you know, like music sounds great in a great record player, and like people go, this is kind of cool too. And like, it becomes behavioral. It, you know, it starts in the same stuff. You know this. It's like the Brooklyn extremists are like, let's go, you know, hippie cool culture. And they get their little thing in a little sub pocket, it bleeds out a little bit, then the Manhattanites feel like they're not cool anymore. So they do it to keep up with the Brooklynites. And then we have that version everywhere in California, Texas, and Florida, and everywhere in the world. And so it's just— it's normal consumer behavior. But it is clearly a counterpoint to extreme in-feed, in-digital consumption. And I think that's great. It speaks to the thing I most believe in, which is humans correct themselves at a level that we are unbelievably underestimating. The sheer adaptability of the human race is extraordinary. Extraordinary. Let's give it up for us.

Uh, sports. Well said. I want to talk about sports. Yeah, I mean, sports are a good example. We were talking earlier, a friend of the show, Josh Kushner, is looking to acquire stake in, in the Giants and out of his new Eternal Fund, which is, uh, basically a— will seemingly be a collection of like, you know, evergreen bets that aren't— that can't be disrupted by AI. Because I don't want to watch —simulation of the Giants. The Sphere is another good example. It's crazy.

Everyone was thinking that that was not going to go well. They were worried about the debt, and the stock is up like 3x. It's done very well. How are you thinking about both opportunities in new physical locations like the Sphere, and then also in the more legacy brands like the San Francisco Giants, the New York Mets, the different sports franchises? Out there.

Yes. I, you know, it's really interesting to tie a couple things together. This distribution channel that you and I are on right now, right? The fact that people watching. Yeah. In a way that 40 years ago, somebody would have to say, you guys are good faces and you look good and you're charismatic and you might get a shot on this distribution this decentralization of distribution, along with this analog and not disrupted by AI thing, is why 6 or 7 years ago, I started investing very heavily in alternative sports. So, I made a big bet on pickleball that I think is going to work out quite well. Unrivaled, the 3-on-3 basketball league, AJ and I, my brother, early investors in that. The wiffleball league, the sailing league, slam ball. I've been very aggressive on alternative sport investing. Padel. Yeah. So I'm very bullish on it. I think Josh obviously plays at a very heavy economic level with his funds of that nature. And everyone like him and others, especially him, I have so much pride in his building because he's done such a great job. And New York City-based when we were always like, and he is the best. So great to see him Rideshine. I think all anyone who does not realize how substantial the analog opportunities are is really missing the plot of what's happening. And so it gets exciting. It's exciting to think that you can invest in both something that could get as big as, you know, Anthropic in an extreme digital world. But also forget about the obvious ones like the Giants or the Sphere. I'm talking about people rolling up 900 local restaurants. Of building up a huge— like, like it compounds 2x what anyone in PE would have thought because people want to go physically out and eat instead of just do seamless. You know, somebody buying up shopping malls not to make them data centers, but because they think they can make a half-experiential, half-old-mall shopping environment that people are going to be yearning for. Here's a left-field one, like the drive-in movie theater, right? The stuff that dominated the '50s and '60s. When I say that out loud, of like, do you think a modern drive-in movie theater, which has done really well and executed super well, do you think that a lot of people would be about that life?

I think we can all agree that yes, 100%. Yeah.

And then what happens to the person that does 39 of those and has a real business and flips it? So I think it's a really cool time for a lot of us that are watching this show. We're like, oh, Crap, we can play on either side of extreme digitalization or analog.

Be creative. Well, yeah, the beauty, you know, give, uh, the example of somebody making like a, a modern chain of drive-in movie theaters is like you get all the benefits of AI still. You get to, you get to use it for to instantly respond to any customer message and help solve problems.

Now we're nerding out. What if you have 40 of them and now this 6 to 7 years from now and you start showing your own films that you make in AI and build your own IP through your own analog distribution.

Yeah. Oh, that's interesting.

Yeah, I could see doing that for kids. Yeah, totally. You know, like maybe slightly lower bar, but you create your own kind of kids' content.

That's again, a lot of things I do get more obvious later. I got made fun of writing that book, "Anyone Can Build a Personal Brand on Twitter." Now it doesn't look so funny. No. I mean, I'm saying, Who's laughing now?

Who's laughing now?

Yeah, but when I started building VeeFriends and like building this Pokémon Sesame Street thing, yeah, yes, there was a lot of blockchain in it, but it was a lot of my smartest friends in Silicon Valley 5 years ago like, this AI thing is percolating, you know, it was a little more slang for machine learning, and then it obviously got accelerated real quick. But like, yeah, I believe in a lot of this stuff.

Well, speaking of people that are watching the show, there is an army of people named Ryan in the chat and they want to be acknowledged by us. So I'll ask you a question about a Ryan from your life. Can you tell me something you learned from Ryan Serhant or Ryan Holiday or maybe Ryan Harwood? Do any of these names conjure any interesting stories or life lessons?

Yeah, the Ryan Mafia is good. I mean, Ryan Harwood, I've learned nothing from. Zero. Absolutely nothing. So let's just put that one to death. Uh, you know, Ryan Holiday was definitely the individual that I was like, finally put a word to kind of how I was living my life, right? I know my mother's my hero. She taught me just how to be a really happy human. You know, I thought it was immigrants, and I thought it was simplicity, and I thought it was, you know, love and health over everything. And truly, truly, not just cliché bullshit, but live your life that way, even if you're an entrepreneur and a capitalist and want to compete. Me. But when Ryan Holiday started talking about being a Stoic and Stoicism, I was like, oh, like, you know, like, I was like, oh, shit, that sounds like— oh, that, that, that's interesting. Yeah. Um, and so I think that, uh, that's one for him. Uh, Sirhan, you know, is a fun one for me because a lot, you know, he's obviously playing at such a big personal brand level now. And he had that TV show on Bravo, and when it was off the air, his curiosity and humility was amazing. He was always around our ecosystem, and he really learned to play the playbook. I've lived, you've lived, he's lived so much of— now he's back on Netflix, but he had a really great era in between his television moments of really winning on social and going all in. And I'm really proud of him. And Hartwood is just a dear actual friend. He's family, not even friend. And so he's taught me that.

I'm glad you circled back, because I was like, damn, he's taking shots.

You know, he's a Long Island boy. I'm a Jersey boy. I think anyone from our part of the world knows, like, taking shots to your brothers is like the ultimate form of I love you more than anything. I think literally from 1985 to 2000, 98% of the words out of my best friend's mouth was highly disrespectful in my direction.

Of course, that's the best. That's the only way to do it. Ryan Holiday wrote "Trust Me I'm Lying: Confessions of a Media Manipulator at 25." Should more young people write books? I talk to some brilliant young people every day on the show. A lot of them fantasize about writing a book, but it always feels like it's something that needs to come with wisdom, that needs to come with being 40, 50, 60.

Yeah, do you get a lot of wisdom? But, but part of it's like you're generating wisdom through writing and thinking deeply. Do you feel like—

I think of it this way, man. I think it is— that's weirdly back to my barbell. I think people that have something to say will probably be best at the barbell, you know, meaning it's kind of like someone's first album. You had your whole life to write it, right? Yeah, yeah. And And so I don't know, for me, speaking for myself, again, back to the books, when I look back at that, I'm like, how— and I'm pretty like— I feel good about myself. And me and me have a good relationship with each other. But even I, I look at it and I'm like, how the hell did you write that in 2008? And then to your point, I think my next best piece of work will be when I'm wrapping it up and I can synthesize all the good stuff. So I would wildly encourage, especially the kind of characters you guys hang out with, they have a lot to say because much like I had in my youth, I saw things others couldn't see because I wasn't playing by yesterday's rules. I always say fresh eyes are dangerous eyes. You know nothing, it makes you dangerous. You can really innovate. And then I think those people that, you know, have sustained careers and can have years of of wisdom and chapters, have a lot to say that can really wrap up and bring value. So I think most people, if they're going to be in that game, are probably gonna write their best book first and last.

Mm-hmm. You partnered up with MasterClass. Yeah. Tell us about it. Tell us about it.

You know, 15 years ago, anything that looked like MasterClass, like selling stuff, was really scammy. Like I grew up in the Web 1.0 you're at. And like, I stay away from a lot of subscription— Patreon, OnlyFans, Masterclass— like, selling information was really dirty in '97, '98. $150 ebook that was straight garbage. I want to give Masterclass some flowers. Like, when they first kind of hit the scene around that era, we were just getting into the earliest stage of where we are now, where premium content, you know, Substack, Beehive— like, now it, now it Respect it, which is amazing, and I think is great for a lot of individual writers and contributors. Um, you know, it was really good company. Like, I like who they had for this class. It's addressing something I believe is real, which is like getting an MBA in real life and AI at, you know, pennies versus taking on extreme debt and getting an MBA from a high-class university that you think that logo is going to get you financial opportunity in a way that I believe has passed us by. It just feels like I want to be part of things that are historically correct or are just a little more practical and common sense. And so, yeah, I mean, they've come to me a whole bunch of times. This was the first time I was like, you know what, I can join this crew. I have something to say about AI marketing and production. And the strategies of that. And I hope that me or Cuban or any of the other people they threw at it might get someone to stop and not take on $300,000 in debt that's not gonna be ROI positive.

Yeah, well said. What are you hearing from Fortune 500 CMOs, larger company marketers around if anything they should change in the world of AI on their marketing mix?

Um, because we're gonna wrap up here, and I'm actually gonna take your question but try to give an extraordinary amount of value to everyone that I know listens to this epic show. It's not what they're saying, it's how they're acting, which leads me to the following sentence. Everybody who's watching, highly invested in big companies that spend a lot of money on marketing, your marketing department is wasting 93 cents of every dollar Okay, we are living in such a radical transformation of the mid-funnel dominating, not the upper funnel and the lower funnel. Sure, us three right now, we're on mid-funnel.

Yeah, we are producing content on in the mid-funnel.

I mean, look, I mean, this is what's going on in my room, right? This, this is a production day. Yeah, this will be posted produce into creative that will go in the mid-funnel, organic social. And the things that do well will go up for brand and down for performance. Every brand on earth should be spending 20% of their entire marketing budget just on social media organic production because the mid-funnel TikTokification of social has eaten up the world. And we should spend no media dollars against creative that we're guessing for the big campaign. And then when I tell you every Fortune 5000 company is still doing guessing on the upper funnel, sponsorships, ridiculous shit. And then the lower funnel, 2016 A/B testing, beta frameworks, when the mid funnel ate up the whole world and you guys are winning, Ruins is winning, and any personal brand and many other businesses. And so that doesn't even get into the AI of it all. That to me is tooling and infrastructure to be great at what I just said. But Jesus, there's a lot of money being wasted by the companies representing the eyeballs that are watching the show right now, and that makes me sad.

Yeah, makes a lot of sense. I, I can't stand wasted marketing spend. Yeah, it makes my blood boil. Well, thank you so much for taking time to come chat with us. Congratulations.

We barely got to talk about live commerce.

Yeah, there's a lot of work to do.

Next time. Yeah, we—

you need everybody here. Anybody selling to the consumer needs to understand that China's going to do a trillion in GMV, a trillion in GMV this year in stuff sold. And TikTok Shop and whatnot are doing real numbers in the US. And Meta's clearly going to be popping off any second with what they're going to do, which will lead YouTube to do what they're going to do.

Sorry, I know you got to leave, but give me your 30-second outlook on TikTok post kind of acquisition, change in ownership. How's the platform changing and what's your outlook?

There's— I felt nothing. I haven't done any real homework on any like corporate structure or what is going to be the vibe.

I know, but will the new owner— will the new owners be willing to like subsidize TikTok Shop? Is kind of what I'm getting at. They don't need to.

It actually works on merit.

Okay. You only need to subsidize shitty shit.

Mm-hmm. Ah, always a good time.

Thank you so much for coming on the show. Great to see you. We'll talk to you soon, Gary. Goodbye. Later in the show, we'll be telling you about a local man who grew a 900-pound pumpkin in his backyard. And we'll tell you what it means for AI scaling laws. But up next, we have Humble Robotics. The founder and CEO is in the waiting room. Let's bring in Eyal Cohen to the TBP and Ultradome.

How are you doing? What's going on?

Great, how are you doing, guys?

We're doing fantastic. We're doing great.

We're having a great day. Having a show, having a show.

What's going on in the background? Yeah, take us on a little tour.

Yeah, what's going on here?

You can see the truck that we built. Ooh. The vehicle that we built is a— Class A electric autonomous vehicle for moving freight. But actually I wanna hear about this 900-pound pumpkin. When do I bring that up? That's actually—

It applies to your business.

Yeah, yeah, yeah.

You started, your demo was probably the equivalent of a 45-pound pumpkin. Maybe that truck in the back is a 300-pound pumpkin, but this is all in service of getting to the 900-pound pumpkin, the elusive 900-pound pumpkin. And you raised some money to help you build that vision. How much did you raise? Who's it from? Tell me about the round.

Yes, okay, so we raised $24 million and—

Amazing. Blocky clearing order here.

Continue. It's so satisfying, really. We raised $24 million, the lead investor was— investor was Eclipse. Fantastic. We also had, we also had EIP in there as well as some other, other friends. So that's great. Yeah, it was a great round for us.

Okay. $24 million. It sounds like a really solid seed.

Eclipse is one of the highest signal. Yeah. Probably the highest signal if you're building in robotics.

Yeah. At the same time, you're building something big in robotics. $24 million. That's great for a seed, but how, how much does it cost to build one of these? What are the comps like? What, what, what will the next couple of years like be like of actually getting these on the road?

Yeah, so $24 million is a start, right? Yeah. It's a, we are, we're doing hardware, we're doing robotics, we're doing autonomy software. That does cost money, but it costs way less than it used to. I think, you know, you look at some of the players in the space and, you know, Waymo took a very long road, spent a lot of capital, but we're standing on the shoulders of giants now, right? A lot of work has been done in this space. I've been working in this space for, 10, 11 years now autonomous freight. And we get the benefit. I love it. We get the benefit of a lot of work that's been done in many and over the years. And so it's just far cheaper than it used to, to sort of think about a great new idea and scale it and scale into production.

Okay, talk about the key trade-offs. No cab, no driver. Are you pro-LIDAR? Are you pro-teleoperation? And like, are you gonna do a crawl, walk, run thing? Is it a straight shot to full self-driving? What's the thesis?

Yeah, great. So we, you know, we're not dogmatic about technology here at all. It's just what's the best technology that we can use at any moment. Most of my career has been spent working on LiDAR technology, but actually where we see vision-based tech on cameras is incredible right now. I think especially in the last year or two. I always joke with our head of autonomy, I have 7 startups under my belt now. This is the second one I founded. And so we've been kind of doing a lot over many years. This is the first time the company has really been vision camera focused in autonomy for me. And what I'm seeing on that side is incredible. It's honestly kind of magic. But we still take the same approach that you see in autonomy, right? It is a, what'd you say, crawl, walk, run. We take that approach too, right? First, you test in private roads and tracks, you validate the hardware and the software together, and then from there you start pilots, you supervise those pilots. But we will take any technology that is beneficial for the project, for the program, and we'll apply it. That's the idea.

Average 18-wheeler costs something like $120,000, $200,000. Maybe $275K at the high end. Do you need to be at that price point? Can you go higher? Do you have flexibility there? And then the actual autonomy package, I feel like you could literally strap an iPhone to every possible vantage point and you don't need to do that 'cause you can just buy the actual lenses and cameras. So the cost of the vision stack has to be pretty low, but are there GPUs on board? Onboard? Are you doing on-device inference? Like, what's the economic balance? How is this— how should a customer even think about this in a different mindset?

Yeah, so what do you get from removing the cab, right? What's the benefit of having a cabless vehicle other than it's cool? I think when we were conceiving the project, the question was, how do we move freight? Given where technology is, how do we move freight at the lowest possible cost, right? And the lowest possible cost to me was, okay, we think about making it autonomous, right? We could save about $1 per mile there. We make it electric, right? So that lowers the maintenance costs. And now, especially with like some of the volatile fuel issues, you know, we get benefit there, right? And, you know, the cab itself is cost to the equipment that is necessary in autonomous world. And so by removing the cab, you make the whole vehicle lighter and you make the vehicle less expensive. And so for me, the target is we want to have the equipment be comparable in price to what you see a tractor and a trailer today because this is effectively combining a tractor and trailer together, right? It's both the semi-truck part and the trailer part in combination, right? And this for the 40-foot version of it that we're looking at. So we want the equipment cost to be comparable and then the cost per mile is just dramatically lower. That's the idea. Yeah.

How do you think about supporting infrastructure? I mean, there was a massive build-out of charging infrastructure as Tesla ramped Is that beneficial? Is that enough? Is there a lot more that needs to be done? I imagine even if I just have a warehouse, I might need to check if my warehouse is compatible with a cabless delivery and doesn't have the wrong configuration to be able to puzzle piece in the container.

Yeah, that's right. And actually part of the reason that we launched, I mean, the company's very young. We started about 8 months ago, right? We built this vehicle very quickly. So yeah, we have a— we have an absolutely killer team there. They're the most fantastic people I've ever worked with. And we try to move very quickly and we launched, you know, pretty early in the journey. We were public about what we were doing. And the reason is that you want to build with your customers. It's really important for 8 to operate within the constraints of the logistics space and understand what people need and how they operate. So we were being public about what we're doing because it's our message to our customers that we work with them on that front. And that goes from everything from how does this vehicle load, go back up into a docking area? It's one of the advantages of our vehicle, by the way. It's like we could put cameras behind it, we can back up into docks, right? And so how do we organize that kind of effort with them? How do we think about charging? How do we think about plugging into their software, right? Which, you know, sometimes every shipper has a different software stack that we have to work with. So we were starting those conversations really early, and as a company, we just move very fast.

What's it gonna take to get you on X?

On X? I actually signed up for an X account right before the launch. I'm on it. I just need to figure out how to use it, but I'm there, I'm there.

Okay, good. Win, we'll tell Nikita. One more user. One more user. One more user for Nikita. Yeah, where's the company based? Where, what does actually scaling the company look like? Are you gonna make Detroit great again? Can we do something over there? Motor City? What are we thinking?

We're in San Francisco. We're in downtown San Francisco. And actually when you walk into this warehouse, you're greeted by this 40-foot vehicle, which is like, before we were public, you know, people would walk into our warehouse and their eyes would just get really big. They're like, what is happening in this space? Space right now. But, you know, we're in San Francisco. That's where I live. That's where I'm based. That's where I've been working on tech for the last 20 years.

And it makes sense as a R&D and prototyping hub. You have a warehouse, but you don't have a gigafactory yet, and that might be somewhere else. Correct.

Got it. Yeah, and that can happen anywhere in the US, but we're designing in the US, we're building in the US, we're part of the reindustrialization movement of the US. So that's the plan.

What have you learned about the reindustrialization efforts that are talked about a lot specific to the supply chain of autonomy components, electric vehicle components? I'm sure you're grappling with all of this, but what has surprised you to the upside? What has surprised you to the downside? Give me some white pills and black pills.

So I started, I started my career working in electrification, electric vehicles, and actually a lot of the technology was here in the Bay, a lot of it developing battery packs, motors. And we watched over the last 20 years as that shifted out of not only just the Bay but out of the US entirely, right? Yeah. And I— and now you see the sort of the rush to get it back. And I think, I think there is— it takes time, but there are a lot of smart people that have been working in the space for a long time. And there's a lot of excitement and capital pouring into this. You know, our backers at Eclipse are are just the most phenomenal people you could be working with, and they get this and they understand it. And you'll just see a lot more investments in this space in general as we grow. And yeah, I think the supply chain's coming back here. It just, it takes time, like everything. It's just gonna come back in automated factories this time around.

Yeah. We are at time, but you need to read this deep dive in The New Yorker about these insurance scams that are happening in New Orleans around 18-wheelers where people are crashing it— have you read this? Are you familiar with this story?

I'm a little bit familiar with the story together, the money for the first round.

Oh no, no, no, that's not what happened here! But I mean, it is something that would be at least easier to fight back against in terms of fraud if there were cameras all the way around the vehicle. So a little bit of an odd knock-on effect to the positive. Part of the reason we designed it the way that we did.

Thank you.

And thank you so much for coming on.

I am, just based on this conversation, I'm extremely confident that this company will be wildly successful. And I'm glad you're doing this. Yeah.

Terrific.

Thank you so much for coming on the show.

Hope to have you back on soon.

We'll talk to you soon. We'll see you soon. Bye. After— missed it completely. There we go. There it is. There we go. After our next guest, we will be taking you into a story about Coachella. It ran past curfew and it got fined. You will learn why music festivals may be the best way to understand data center permitting risk after our next guest. That's a narrative violation. We have Yolande Yan from Comfy in the waiting room with an iconic entrance. Let's take a look. Let's go. What are you holding here? Please introduce yourself and the company.

What are we looking at?

Let's fucking go. Yeah.

Please.

Welcome to the show.

I'm surprised no one's ever pulled off an entrance like this. No, no.

This is exquisite. Please introduce yourself. Tell us a little bit about what's going on.

My name's Jolin. I'm the co-founder and CEO of ComfyUI. I apologize. I've been instructed that I can be anything but boring for this show. So here I am.

That's true. You studied. You studied. Yeah, I did. Tell us your entire life story from inception. No, what— I think there are too many points for you to start.

Art over here. Let me, let me remove the distraction for you for now.

What is the distraction? Give us the backstory, like why that particular creature.

The distraction is a character created by my co-founder. It's actually the origin of this tool. Yeah, yeah. If it weren't for this character, we wouldn't be raising and becoming a half billion dollar company today. So congratulations, shout out to the Fennec character.

Yeah. Uh, so for those who don't know, ComfyUI, node-based workflow for visual design, video production, you know, DaVinci Resolve level performance, but for the modern era, is that roughly correct? How are you thinking about introducing the company?

Yeah, yeah. Comfy is a node-based system that aims to give the highest leverage of control and quality for content creation. I think if you guys think of, you know, let's say Midjourney or text prompt-based content creation using AI. It's the polar opposite. It gives you every bit of control and possibility to adjust the model and control the camera angle, control the posture. And that's exactly why Comfy ended up being adopted by these top studios around the world. The recent Coca-Cola ad or Super Bowl ads were all made using Comfy.

Yeah, and then also inject a level of deterministic and programmatic variation into the generative AI workflows so that if you want to create multiple assets and you want to swap out a bear for a cat for a dog, you can do that in a way that doesn't just require you sitting there prompting, waiting 5 minutes, prompt again, 5 minutes. And getting, you know, actually parallelized output. Yes, that makes sense. So yeah, if you're working with big companies, like are they just, forking the repo, downloading the open source project? Are they actually calling you? Are you doing implementation? And how are you making money from all of this so far? Yeah, yeah, great question.

It started first as an open source project, actually. It's a personal project to begin with. And then originally, a lot of the companies actually just ended up using it running on their own hardware, most of the time locally. Yep. We introduced Comfy Cloud as a product about 5 months ago. Oh, cool. And the main differentiation in that world is that you can actually tap into a large cluster of compute that we provide in the backend. It's incredibly difficult to stitch everything together to provide this infrastructure level of support, especially for like, you know, professional creative companies or studios. So yeah, that's kind of like, you know, how we monetize off of it.

How big of a deal is Seedance 2.0? The videos that I saw were incredible. It felt a little bit inaccessible. Is it diffusing rapidly? Yeah, yeah.

The entire visual AI ecosystem is evolving very rapidly. These foundation models are making amazing progress and you can access all of these models on our platform as well. We are generally agnostic in terms of whether it's open or closed source model. But just to comment on the progress over here, we're super excited to see where the trend goes towards to when these CDANCE models get launched onto Comfy platform. You see people kind of building on top of creating even better workflow using these models. And I can't wait for one day we'll just have avatar being created and someone in his bedroom and backyard.

Okay, when? Over/under 2028. Give us your Hollywood predictions. When can I actually watch a 2-hour movie?

A 90-minute movie? Because I've seen some incredible short films, some incredible vibe reels. Not even short films. I've seen, yeah, maybe 2 minutes, 3 minutes.

Yeah, everybody's now seen like—

There's probably longer stuff out there. I haven't gotten sucked into anything. I did watch all of Harry Potter Balenciaga. That was great. You know, but like I'm waiting for something that's Harry Potter, Balenciaga level entertainment, like hold my attention to the end, but 10 minutes, then 20 minutes and then 90 minutes. And when the 90-minute thing happens, I think a lot of people are going to be like, wow.

Yeah, this is actually a part of an insight that most people don't really understand right now. It's one on the quality side of the area. People don't want to create long-form content because by the time when you finish a feature film length of content, the model has shifted so much that you're running into this James Cameron problem of like, oh shit, I have to go back and then rerun these using different models. So a lot of times people focus on short clips in terms of the content. And then the other portion is actually it's already being adopted in a lot of these creative creative studios and agencies using Comfy. What it doesn't do is it doesn't become production. It just enters the previs portion.

Oh yeah.

It just enables people to rapidly iterate and then create this content right away that you might have someone acting in a pure green screen and then the output of it in 5 minutes, you see the entire 3D CG rendering using AI on the system in Comfy. And in that world, it's already influencing the direction on how they act. It just doesn't necessarily get to the quality where it completely could become a film production level of content.

Is there more to be learned from the Cognition model or the Cursor model in terms of training your own models versus being at the application layer? Both companies have been wildly successful, huge. Yeah, but very different paths.

Yeah, I believe the visual ecosystem is quite different. The main differentiation is there's a huge amount of taste. Even if someone can prompt Seedance 30.0 to generate a feature film tomorrow and it gets to that level of quality, for the real creatives or for the people who want the taste and control to be in their own hand, it's still going to be a very iterative process. It's just going to make their life a lot easier. And that's what Comfy enables you to do. I think those are the kind of differentiation where in the coding world you might not care in terms of, you know, is this piece of code actually good enough? But in the visual world, you absolutely care. The camera angle, the facial feature. The control layer. And that's, in my opinion, what differentiates the, you know, AI slobs from the Comfy ecosystem of creation.

What's the easiest AI feature to pitch to a Hollywood AI skeptic?

The easiest way I would say from our perspective, this might not be for everyone is the fact that we have this creative in the loop, a human in the loop centric belief and mechanism. And it's actually true. In a lot of these case scenarios, when people are using Comfy as a tool, it is actually empowering the creatives to do their work, to skip all of these boring stuff of colorization, of, you know, kind of upscaling, downscaling. Downscaling manually in the traditional process, it enables you. So those are the world where we're not actually taking jobs, we're creating jobs. If anything, you go search on, you know, any job site that today there's actually a huge amount of people hiring for ComfyUI artists and specialists.

Yeah. Yeah. I think like the Mount Everest is like, oh, well, one shot the movie, replace the actors, the easier stuff is, what if your green screen was a little bit cleaner? You know, what if your color grade was on your dailies as well? There's a whole bunch of interesting places where you can plug AI to work.

Yeah, and I've said this before, and at least some people were trying, were calling me an idiot online. But I've had this idea that I think AI could actually help Hollywood. We're in Hollywood. There's a bunch of empty studios around. There's a bunch of people here that that have been working in entertainment that haven't had great job prospects over the last few years, especially if they weren't willing to travel overseas. And it just feels like we're gonna enter into an era where 10 really smart people could come into one room and if they work really hard and leverage AI properly, they'll be able to make an incredible short film, film, make beautiful advertisements, whatever they wanna do. So I think we will enter into a golden, age of human-powered, AI-powered humans doing incredible creative work.

You mentioned that you're currently valued at half a billion dollars. How much did you raise? Tell us about the funding news.

Yeah, we raised $30 million from— oh man, I've been looking forward to this.

Yeah, we raised $30 million. The round was led by a Craft Venture, a huge fan of folks over there. They've been tracking us for a while. Good job. Yeah, been amazing working with them so far. Good job. Super excited. We're here to destroy our enemies and make open source creative tooling win. Yeah. I love it, dude.

Imagine being your enemy right now logging on to a desktop app.

Oh, I think we talked to his enemy earlier this week. I'm not going to say who it was, but I think we know who he's coming for. and I'm rooting for them.

Well, imagine logging on to xd.com and seeing Mr. Yan and just being like, oh, it's over for me.

But real, real problem.

We have no idea because in our mind we don't really have enemy. The only enemy is ourselves.

Oh no, I'm not talking about startup competitors. Oh, I see. Gotcha. I'm talking about like— Adobe.

Never heard of that word.

Yeah. But congratulations. Thank you for everything. Everything you do. Incredibly cool tool. If you haven't already checked it out, go download ComfyUI, make some cool stuff. It's a lot of fun. And have a great rest of your day. Goodbye. Yes, great hanging. See ya. Later in the show, did you hear about the cow that got loose on the highway and backed up traffic for miles? I did. You did? Well, we're gonna tell everyone what the AI industry can learn from that story about supply chain bottlenecks. But up next, we have Ben Horowitz from Dorm Room Fund. He's an investment partner and is the creator of Sincerely, misspelled. You'll have to figure out how to spell it yourself if you want to go check it out. But we'll let Ben introduce himself now. How are you doing? Hey, guys.

Thanks for having me on. Love the show. Thanks for coming on.

Also an MBA at HBS. Oh, yes.

People were taking a victory lap on that. Soon-to-be master of business.

They thought it was impossible. For an HBS grad to start a business, but you're proving everybody wrong.

Thank you. Yeah, Twitter flamed me a little bit for it, but I'm still happy I'm here.

Wait, wait, so why were people upset?

Oh, it's like a non-technical thing, probably, right?

Yeah, I think Gary talked about it too a little bit.

Yeah, yeah, just the idea that like, I mean, HBS grads have started so many amazing companies, but it's—

Oh, I thought you were talking about the product.

But the HBS thing is typically not like the, like, go and build a software product in a weekend or in a week, like very quickly. Like that's been more like, oh, you're a CS grad and you're a dropout and then you go do YC. Like that's been the more established path, but you're bucking the trend. Tell us what you built.

Yeah, so I built Sincere.ly. It's spelled with two Es. So S-I-N-C-E-E-R-L-Y. And it's really like the anti-Grammerly. So I am naturally a terrible typist. I don't know if I can cuss, but I mess up my emails all the time. And I found that people really reply to them. I don't think it's just because of the typos, but it's definitely a factor. Yeah. And I was so sick of AI slop in my own inbox. Oh yeah. People would email me. It all looks the exact same. And that coupled with the news about how other people have been emailing across social strata inspired me to basically create the anti-grammarly, something that really messes up emails. Are you—

you're talking about like CEOs that have terrible grammar because they're just typing quickly and it doesn't really—

well, there's the famous story about Ron Conway types in all caps. There's a whole bunch of these examples.

Typing in all caps—

all caps is elite because it just sounds like someone's yelling at you.

It's kind of the anti-Gen Z in a lot of ways, the all caps.

Yeah, we don't have a Ron Conway mode yet though. Yeah, but I had, I, I had this, I had I had this experience maybe a year ago where someone sent me a very lengthy, you know, multi-paragraph deep dive on some topic. And I could tell that it was AI-generated. And I said, you know, you should just send me the prompt because I can also do the instantiation. Like I can also expand bullet points into paragraphs in my head. And so just send me the bullet points. And there was that also that meme of like, Oh, turn this, turn these bullet points into paragraphs. And then the other person on the other end saying, turn the paragraphs into bullet points. Are people using this specifically to write emails or is there an email summarization feature coming? Where do you see this going?

I guess I think what's kind of crazy is people are copy-pasting. First they're prompting, then they're copy-pasting their prompt into email. Yep. And then they're using Sincerely to make their email not sound like AI.

And that's a crazy loop, right? Like it's human.

Humans using AI to make AI more human.

And that's why Intel is up 20% for this specific use case. No, there are more useful, there are more important use cases. But we were talking about this history.

Was this a fun stunt or?

Yeah, somebody called it a drop and I think it's a business. What do you think?

I think it started as social commentary. This is like, it's satire and playful. And I didn't want to pay for everyone's token. But like, I do— like, I love making products that make people think twice about technology. Sure. And in this case, it's also a little bit like the pushback on how we're using technology today. So, you know, there's a $4.99 a month model that if you want to not bring your own key, you can pay me. I think it's probably double-digit people who are paying and not using their own key, but it could be a business.

Yes, for sure. We were talking to George from CrowdStrike earlier. Are you ramping up security? What are you doing to make sure this is enterprise-grade ASAP? Because when I download a vibe-coded viral stunt, I want it rock solid.

I learned a lot in this process, as you might have imagined. I'm a huge fan of CrowdStrike too. We have pushed a few security fixes and just made sure that people know what they're getting themselves into when they download this. Yeah, but ultimately, like, this is a— it's satire meets something real. Um, so I hope that, that people get it.

Um, yeah, yeah, that's the nature of like doing fun things on the bleeding edge of the internet. We all went through this with OpenClaw and understanding, okay, there's some real trade-offs here. You probably got to be at least a prosumer if you're going to dive in, but it sounds like you're handling it well.

What is sentiment like in the current— in your current HBS class Oh yeah. How are people thinking? Are, like, with the acceleration in AI, are there people thinking about dropping out? Like, or are they like, I'm actually glad I'm in school right now 'cause it seems kind of gnarly out there? Okay. Yes.

I have friends who have dropped out for YC and friends who have stayed for the exact same reason. Yeah. It's kind of, you know, swim or sink. I think that people are building— they're not building funny things, I would say. They're really trying to build things in the HBS sort of business-first way. And my girlfriend's at GSB, so maybe there's a little bit of a different comparison that you could ask her about.

Uniting the great houses.

Do you think there will be a sway? I mean, you go back to HBS in like the '80s, and you imagine that it was churning out like oil and gas executives probably. And I'm wondering if you think there'll be a swing back. Like we saw Josh Kushner announce Thrive Eternal. He's taken a stake in the Giants. And there's a whole different world where you take the HBS learnings and playbooks and you go apply them to things that are extremely AI resistant, but maybe like they weren't— they haven't been trendy over the past decade, but maybe they're going to come back in style in a big way if you're, you know, yeah, yeah, I went to HBS. And now I'm like, you know, VP at GE Vernova. That's probably a pretty good gig.

Yeah. I mean, the friends who I'm thinking about who actually dropped out to go do YC— you're gonna laugh— are focused on HVACs and putting AI in the back of HVACs.

It's not a roll-up. It's not a roll-up. They're like, true HBS.

Why has no one identified the opportunity in HVAC? HVAC is important.

Have you ever not had air conditioning? No. One hot day in this studio and I demand HVAC. It is universal.

It's sort of— that's another loop that's tricky though, right? Like, are we making the world warmer? Are we cooling it off? Are we putting AI in the back of HVAC?

It's all an AI play at the end of the day. Interesting. Interesting. What were you doing before HBS?

I started a company with my friend Reid that was a sort of like a philanthropic robo-advisor, like Wealthfront or Betterment. Yeah, yeah. And we sold that to Charity Navigator. I worked there. Wow. And yeah, hopefully I'm a free agent in like 3 weeks. So heading to SF and fulfilling my manifest destiny.

So in 3 weeks, is that HBS grad? You're saying you're finishing your earnout or HBS?

Yeah, finishing HBS. The proverbial debt earnout.

I don't know.

Yeah, yeah, yeah. Awesome. That makes sense. All right, well, it was really great to meet you. Yeah, great to meet you. And please let us know when you figure out what you're working on.

And let us know when the next drop's happening. I feel like you got more of these in the can. I can tell.

We have something really funny coming up.

Yeah, I can tell. By we, I mean Rawlings.

Amazing. Thanks, guys. Yeah, let's talk about it.

Yeah. Have a good one. We'll see you soon. Thanks, Scott. Goodbye. There is a ton more news. What should we go through? 4 says, what's the charge? Front-running an overseas coup d'état, succulent Venezuelan coup d'état. But this was a crazy story. Brandon Gurel on our team was, uh, obsessed with it. A U.S. Special Forces soldier involved in the capture of Venezuelan President Nicolás Maduro was arrested for allegedly betting on that operation, netting him $400,000 in profits. This is when the betting on yourself meme goes wrong, right? You should always be betting on yourself, but not literally if it's against the law. If you are involved in a government capacity, you should probably not be gambling on the outcome of your government work. Shame on you, but still an interesting story.

Even Trump was like, he was betting on himself, and then he like referenced some baseball player that isn't in the Hall of Fame because he was betting on himself, but Trump was seemingly implying that that wasn't so bad.

But anyways, people are really, really having fun with the new ImageGen model. Oh, this is so rude. But this guy's effectively playing Minecraft through ChatGPT because he has it generate him an image of Minecraft. He says, of course, he's doing the island in Little St. James, the very controversial one.

And then he says, okay, walk closer. And then just generates a new scene. I'm sure there's some kids out there that have figured out how to play Minecraft in ChatGPT, but this is very funny. I found this post when it had 40 likes this morning. Sniped. Sniped. Invested early. Cohere is merging with a German AI —lab called Aleph Alpha.

I'm the biggest fan of Cohere. I love Aiden Gomez. Transformer paper alum, Death Grips enjoyer. What's not to like about the legend Aiden Gomez? We got to get him on the show. But this is a congratulations moment for Cohere and Aleph Alpha.

I think I'm saying that correctly. GPT-5.5. Yes. Kits is saying it might be AGI because If you ask it, bro, the car wash is 5 minutes away, should I walk to it or go by car? And it says, car, bro, it's a car wash. Let the plot make sense. Wait, what? Like it just immediately clocks it.

Oh, I like that it's, I've seen this exact test done in normal language, like with a proper prompt, not the casual bro slang. And it actually is more, it feels more like feels more like AGI if it is able to pick up on the lingo and mirror that back. Car bro, it's a car wash. That's very funny. That feels very remarkable. It also got the Rs in strawberry correct, right? And how many strawberries are in the letter R? Zero. The letter R is not known for its berry storage. These are good answers. The being a helpful assistant but simultaneously rejecting nonsense has particularly been difficult for LLMs. So good progress here. We'll figure out more about what went into this and what the next generation of AI impacts are from this.

Tyler Hodge says, found a post from 5.5 years ago. Elon says the rate of improvement from the original GPT to GPT-3 is impressive. If this rate of improvement Improvement continues. GPT-5 or 6 could be indistinguishable from the smartest humans. Just my opinion, not an endorsement. I left OpenAI 2 to 3 years ago. I am a neutral outsider at this point, Greg says. Thank you. A lot, a lot has changed, but, uh, he was accurate in his prediction.

Funniest outcome is the most likely.

And then the last, the last thing I wanted—

Best Bros and do a buddy cop film together.

The last thing, there is a new app from X. It's called XChat. XChat. And I'm excited to try this out because the current— yeah, the DMs on X have been pretty brutal. Yeah. Right?

I haven't had that much of a problem. I do have this weird bug where when I click the chat button on desktop, it sort of loads and then it resorts. Shorts after it loads. And sometimes if I haven't been in it in a long time, it needs to reshuffle several, several times, and so that can be a little jarring. But overall, it seems like they made the migration to encryption. I haven't— I don't know. I seem to get messages. It seems to work fine. It's cool that there's a new app. I probably need to be better about answering DMs. I have a lot of them that are unread, so having a dedicated app for that, that sounds It was cool. People were taking shots because the Everything app's supposed to be everything in one place instead of 3 apps to get everything in the Everything app. Sort of silly. Who cares? More apps the better. I like apps. So go download it. Go check it out. And go— go like Nikita's post because he's been on a generational run. Also, GPT-5.5 is available in the API now. That's breaking news, I guess, from Greg. So anyway, we will see you on Monday. Have a great weekend. Great weekend. Enjoy the weekend. Leave us 5 stars on Apple Podcasts and Spotify. Sign up for our newsletter at tppn.com, and we'll see you tomorrow. Flashbang!

Have an incredible weekend.