TBPN - Delian Asparouhov, Ridge Wallet Interview, Nvidia Earnings, The Sheikh of AI

Episode Date: February 27, 2025

TBPN.com is made possible by:Ramp - https://ramp.comEight Sleep - https://eightsleep.com/tbpnWander - https://wander.com/tbpnPublic - https://public.comAdQuick - https://adquick.comBezel - https://getbezel.comPolymarket - https://polymarket.comFollow TBPN: https://TBPN.comhttps://x.com/tbpnhttps://podcasts.apple.com/us/podcast/technology-brothers/id1772360235https://youtube.com/@technologybrotherspod?si=lpk53xTE9WBEcIjV(01:35) - Nvidia Earnings (21:17) - "Spy Sheikh" (43:51) - Stripe (54:19) - AI Growth (59:51) - Ridge Wallet Interview (01:19:05) - Delian Asparouhov Interview (01:30:13) - Timeline

You're watching TBPN live from the Temple of Technology, the Fortress of Finance, the Capital of Capital. That's right. Today is Thursday, February 27th, and this show starts now. We got a great show for you today. We're going over Nvidia earnings. We're talking about the spy chic, a man out in the Middle East who's making waves and AI investing. And we're also breaking down Stripes Letter from the founders. Lots of interesting thin tech data in there.

We're going to have two call-ins today. We got Sean Frank from the Ridge Wallet and Deli and Asperuha, from Varda and Friwa. founders fund. What are they going to be talking about, John? Sean is actually going to come on and talk about App Lovin. A lot of turmoil in the stock. I think they're down a bunch recently, right? I actually haven't been following that, but he's the expert. That's why we're having him on. Yeah, I'll pull up the chart right now, but Sean has been a big customer of App Lovin. There's been a lot of attacks against them recently. He's come out in defense of them saying that, you know, you can say that it's overvalued. That's fair, but it is a real platform. We've driven real results on it.

And so it'll be interesting to get his perspective there. I think he spent a few million dollars there last quarter. And so we'll get an update there. And then a very exciting day for Varda. Yeah. Second reentry, successful. And so Delling will come on and break it down and hopefully give us some color on what's going on in the space economy. And, you know, hopefully get political, explain what's going on with DOD procurement. and hopefully get political? Oh, yeah. We don't get political.

We let our guests get political. I don't know. He'll talk about whatever he wants to talk about. Anyway, let's kick out off with a report in the Wall Street Journal about Nvidia's earnings. They beat earnings. They're down today. If you want to pull that up on public, I'd love some extra context, but I'll start reading from this. The chip giant has stayed relevant even as Deep Seek and other reasoning AI models rise.

Nvidia faced a growing threat earlier last year. the artificial intelligence world was shifting in a way that invited competition as millions of people started using AI tools operating the underlying models to respond to their many queries became more important than the computing intensive work of training them. This is the training to inference shift that we saw happen, which was expected. You train the model and then you want to actually serve it, and so you've got to do inference. This had propelled Invidia to the top of the AI boom. Many expected that shift could give competitors, including advanced microdevices, AMD, an opening to pry away market share. But Nvidia was already preparing to adapt and stay at the

forefront of the AI race despite the shift away from creating models and toward operating them. A process known in the industry's inference. And I love this article because all the killers are here. They got Dylan Fetell quoted. They got a bunch of other people that will get into shortly. Its latest AI chips, Blackwell, are larger in size, have more computer memory and use less precise numbers and AI computations. They can also be linked together in large numbers with super fast networking, which Dylan Patel, the founder of industry research firm, a semi-analysis said led to breakthrough gains in inference.

So, full case for NVIDIA laid out by Dylan Patel over at Semi-analysis. Nvidia's performance gains for Blackwell are much larger in inference than they are in training, he said. NVIDIA's latest quarterly earnings report on Wednesday partly reflected its success in adapting to the industry shift. It included sales and profit that exceeded analysts' expectations coupled with an optimistic forecast for the company's current quarter. Inference has become a growing focus as AI evolves towards so-called reasoning models where a digital brain thinks through answers to users' queries step by step. That process can require a hundred times more computing power, Chief Executive Jensen

Wong said on a call with analysts Wednesday. The vast majority of our compute today is actually inference, and Blackwell takes all of that to a new level. We designed Blackwell with the idea of reasoning models in mind. So yeah, obviously when you're reasoning, so many more tokens are being generated hundreds of hundreds times more. And so inference is becoming much more important. And Blackwell has positioned an Nvidia well to take advantage of that. Got anything for me, Jordy? Oh, I mean, it's just, it's only a company like Nvidia could have an earnings beat and a very optimistic forecast in the stock trade down 6% the next day, which is where it's at, almost six and a half percent. So it wasn't just price to perfection. Price for it.

Price beyond perfection. And the stock price is paying for it today. But let's get into the rest of it. And then I'm sure we'll have other thoughts. So Colette Kress, NVIDIA's chief financial officer, added that many early developments of the company's blackwell chips were earmarked for inference work. That pattern was a first for a new generation of the company's chips. Among the companies pursuing reasoning models are OpenAI, Google, and the upstart Chinese AI company Deepseek, which we've covered on the show. The emergence in January of Deepseek, which said it had built sophisticated AI models that required fewer of invidious chips, touched off the significant scare for Nvidia since the AI boom began. So of course, it's sold off, then it gained back,

and now it's back down. Speaking of Deepseek, I don't know if this is in the stack for later, but somebody shared yesterday that they have off hours pricing. So you can basically, it's already extremely cheap, but then off hours, which are their sort of non-busy hours, it's like off another 50% and somebody was saying, great, we have TEMU AI now. Well, there's a ton of load balancing that does happen in these AI inference models. So mid-journey, when you're in America and you're using mid-journey during your workday, you are likely running that inference in Asia on an Asian server because of the time difference. They're not using it in the middle of the night and then vice versa.

And so that type of load balancing happens all the time. But it is funny how sometimes it seems like the models, if they're under stress, they still work, but they become like lower quality or there's a bunch of knobs that people are tuning. And it's basically like, yeah, your AI worker is getting lazy. Just won't try as hard. It doesn't try as hard, exactly. Get some protein. Yeah. So Wong brushed off that threat on Wednesday describing deep seeks advances as an excellent innovation that AI developers everywhere we're taking inspiration from.

I think that's true. A lot of those gains will be baked in, but people will still want to run them on really big, really big server farms and using a lot of Blackwell chips. Why not stack all the gains together if you can? In the past, Wong has suggested that inference and training will eventually converge as AI more closely aligns with how humans operate. People don't absorb new information and reference it separately.

He said at Stanford University last year, you're learning and inferencing all the time. Great point. And some of these models actually, it's what's called online learning, where you can continue to train the model even after it's been fully trained. Yeah. So, I mean, obviously there's the pre-training, post-training, but a lot of these models are now what's called online so they can be essentially updated.

And I think that the people that look at AI as, you know, we're kind of replicating the human brain, think about it that way, where there's like fast thinking, slow thinking, all these different models. And you need to replicate all of that. And humans are fixated on, you know, especially in our corner of the internet. at improving your thinking day over day, not in sort of quarterly installments. And yeah, it's interesting because right now people seem very comfortable with the idea of it would become a lot to process if, you know, Claude 3.5, you know, there's some people

today. I saw the cursor founder sharing. He still prefers Claude 3.5 over 3.7. And people are sort of oriented now around these sort of releases and having like fixed models. But there's a world in the future where they're just constantly morphing and evolving. And I think that that's what this point's getting at here. Yeah, I've seen some diagrams where people have said like, oh, there's too many algorithms now and too many like hacks on top of the core LLM to really call it AGI. Like it doesn't count. It needs to be like one simple algorithm. The algorithm for AGI will be just some simple transformer.

And then it'll just be massively scaled up and that will give us everything we need as opposed to what we have now, which is like, Yes, like it has the ability to use a web browser, but that's adapted to it and has like ability to look up data and index things. But another way to think about it is like in your mind, you have the access to go to the library and look up books and you have the ability to use a calculator. And so maybe we don't need to create an LLM that can do every single calculation from its head. It can just use a... That feels more like super intelligence, right? Yeah. Closer to the definition of something that doesn't need any... Anything else. External references. Exactly. And it's almost like an academic exercise of like,

can you just create some superintelligence with a single simple algorithm? Well, it doesn't matter because really what people want are just agents that do their jobs. And so it doesn't matter if it's function, you know, just going and writing some Python and then coming back with the correct answer. I just want the correct answer. Yep. So, Nvidia still faces strong competition in inference industry insiders say, well, Nvidia's advances in hardware and investments in its AI software have kept customers around, a variety of new chips from startups and more established chip makers mean it won't be easy for Nvidia to maintain its position at the top. Robert Wachin, the co-founder of AI chip startup etched. I actually spent three hours interviewing him last year. Great team, fascinating team,

two Teal fellows on that team, and they are trying to build an inference-based chip that's just a transformer baked down into silicon. So those would run much faster. and they've done some cool demos where they've been able to generate Minecraft essentially on the fly. And the only thing that it takes in is the input from the controller. And then it just generates the next frame that you would expect to see. And so you can just play this full game and it's all being generated. There's no game engine.

Yeah, exactly. Level the guy on X, I think levels I.O. He's doing something with flight like an alternate, like a flight simulator, fully generated version. We'll talk about that later, but I'm pretty sure what that. that is, is he had AI write game engine code. Interesting. It's different than generating the video. Yeah, exactly.

So yeah, but similar things. I mean, the pace of all of stuff is really interesting. Yeah, with the pace of development, some of these models, you can imagine a future where games are just perpetually generated. Yeah. In every, you know, in every possible direction, right? It'll be a totally new experience to play games versus historically there's a map and there's sort of boundaries to the game.

And then what happens when there's, there's no more map. Like the map is whatever you can imagine or whatever the game can imagine. There's no finite levels. Like you can go to chat GPT right now and just say, tell me a story about a night fighting for a castle.

And then just immediately be like, continue the story. Where does it go next? Where does it go next? And you can do that for millions of pages. Pretend, you know, you just finished the movie, you know, this movie.

You know, continue the movie. Yeah. And right now it might get boring. but eventually you can say, hey, hey, make another level for me. Make another level for me, essentially, and do it all in real time. And so he said Nvidia's chips were fundamentally limited by their origins as graphics processing units adapted for AI instead of custom made for the moment. Sharpening the Swiss Army knife only gets you so far.

Wachan said you have to build specialized hardware. If you want to get maximal performance, you're hitting a wall here. And we certainly saw that in Bitcoin mining. Bitcoin was originally mined on the CPU, then eventually GPS. use, then what are called FPGA's field programmable gate arrays, which are really kind of customizable silicon pieces. And then, and then after FPGA's, it went to ASX, which are, the algorithm is baked into silicon. The chip can only run one algorithm. And the algorithm that the Bitcoin A6 run is the Bitcoin algorithm that just mines, mine, mines. And hugely risky because

your chips yearn for the mines, much like the children yearn for Minecraft. True, true. Everything I think converges to mining in this world. Mining and tokens, crypto tokens, AI, LLM tokens. You know, there's these weird crossovers all over the place. But for the Bitcoin, folks, it was very risky because you have to spend all this money to make these chips. And then if the price of Bitcoin doesn't go up, well, you're not going to reap your benefit. Now, a lot of those...

Or the price of energy goes by that. Or energy, exactly. And so it's the same game that we're seeing playing out right now. And Etch is betting that Transformer, architecture, is here to stay and that AI inference is going to be here to stay. And I think that's a very safe bat. I'm very excited for what these guys are building. Of course, it's a huge task to actually go and build a transformer-based chip and get it manufactured. But the co-fathers are some hilarious stories. It's so interesting to think about the comparison between humans thinking. Just the process of thinking

requires a bunch of calories, which is one form of energy. Yeah. Infference, on the other hand, the same level of human thinking is going to require some amount of inference to accomplish sort of the same generalized sort of reasoning process but it also requires energy so it'll be interesting to think like I actually want to see comparisons of for a human to reason something out what is the energy cost what is the the comparable energy cost for a machine to do it I would think that right now it's very possible that it takes more sort of calories in almost energy in for the machine to do it but eventually that'll that could flip but it might be closer than

you'd actually think, but it's an interesting comparison. How many... How many David protein bars do you need to have a good take, right? For you, it's maybe like a quarter of one. Yep, yep, yep. We'll see you every time. Yeah, it really, it's, yeah, the Matrix is basically that story. So he goes on to say a number of startups have begun making inroads among large AI customers.

Sarah Bruss, which is a company we've talked about before, they design. the largest chips ever produced, wafer scale chips. By the way, I love this because everybody's like, we need to make the best, smallest, most performant chip. They're like,

no, we're just going to make the biggest one. We're going to try to make the biggest and best chip. Very American out of a French firm. I don't know. Oh, no, they're working with Mistral now. They're working with Mestral.

And it was interesting because Cerebrus, I think they might have gone public or they became a pretty big company. And they had one deal that wasn't with a company that people had actually used the product of. It was kind of more of like a hypothetical. Cerebrus is in Sunnyvale. So not French, but working with Mestral.

And so they're working with Mestral now to build the world's fastest AI chatbot. So because it's all on one chip, you can inference it very quickly. There's no memory bandwidth to deal with. And so there's advantages to that. Of course, there's a bunch of drawbacks. We talked about this when you're manufacturing at Waifer Scale.

One tiny defect throws out the entire chip, as opposed to if you have a wafer and you're getting 50 chips off of that, if you, you know, your yield is, you know, 49 out of 50. And so there have been a lot of risks to that approach, but I love to see that they're trying a different approach and trying to create just a new product that optimizes around a different vector of progress. And so Saudi Arabia's oil giant, Aramco, is working closely with AI-chip startup grok, which is with a Q, it's a different grok.

That's a tough one. It's very tough. And Sanbanova systems to set up large computing facilities for inference. And now the GROQ, they went viral with a demo because they ran Lama on their chip, which I'm not exactly sure. They did something where they optimized for like super fast memory bandwidth, but low total memory allocation or something like that. They made some trade off. And the result was that you could inference Lama and generate tokens, like extremely. extremely quickly.

And it would just, you know how like you're kind of waiting for chat GPD to load with the GROQ chip? It would just like spit out a full page like immediately. And that was like kind of magical. When you're so used to LLM's taking a little bit to think. Yep. And when it happens instantly, it's really shocking. It's magical.

And it also. Yeah, exactly. And I'm sure there's that thing about like Amazon, you know, every five milliseconds of optimization really matters. And so you could see that really driving adoption. it's unclear if that's if GROQ is like scaling up and successful in all the new models because some of the new models they might need oh this one needs more memory and it just doesn't work on their chip so there's always interesting when you're when you're looking at a reasoning model

sort of think something out right now it's paced out in a way that that it sort of you can see the sort of chain the logic chains and it and it's more or less as fast as as you know you might think through it if you were knowledgeable on the subject yet when it happens instantly you start feeling a little paperclipy. And it's certainly nerve-wracking. It is so funny that the first AI thing was clipy and then everyone's worried about getting paperclips. And it's like it all comes back to the paper clips.

I wonder if that's where the paper clips, the paper-clipper like AI doom scenario came from. Is it all tied to Microsoft's Clippy? I don't know. Well, it's like the etymology of like the paper clipping thing. I understand the like the thought experiment that like, hey, you know, you tell it to, it's a good example. You tell it to go make a paper clip and then it just turns everything.

into paper clips. So the, I think the original, the origin of this was that philosophers have speculated that an AI task with a task such as creating paper clips might cause an apocalypse by learning to divert ever-increasing resources to the task and then learning how to resist our attempts to turn it off. So I think that that's sort of the. But why did they choose paperclips? Is it because of Microsoft's clippy?

That's what I mean. Oh, yeah, yeah, yeah. Because they could have said staples. They could have said staples or they could have said hats. Yeah. So it's a 2014 thought experiment by this guy, Nick Bostrum. Okay. Yeah. And so Clippy would have predated that. So it's very possible. I don't know. We got to ask Bostrom. He does shows everyone so on. We should get him on. Ask him where he came up with paperclip specifically. I think that's interesting. I have to imagine that NVIDIA is going to drop some kind of inference powerhouse sooner rather than later because I think they will get eclipsed in that market.

He said it may take many years, but I think that's the direction things are heading. And this was Jim Piazza, an executive at IT management company and Sono, who formerly worked on computing infrastructure at Meta. He said that Nvidia might need to take further steps to directly address the competition in inference by developing chips specifically for it. So if you can think about this, like, you know, what is Apple doing in inference? Well, they have, you know, the M1, M2, M3, M4 chips. those are Apple Silicon.

They aren't really wired together in a way where you can do a large language model training run on it. I'm sure you could get it to work if you worked hard enough and you got enough of them. And we've seen some previews of this where like maybe they just, you know, networked a bunch of Mac studios together to do some sort of small training run. But it's not really designed to be like Blackwell competitive. Yeah. Or A100 H100 competitive.

But, you know, Google has their TPU. and I think Amazon has a custom silicon chip as well. And so there's a few different hyper scalers that have chips that are specifically, they're like, hey, we're going to do all our inference ourselves. We don't want to pay Nvidia their margins. And so Nvidia needs to keep up with that, cut it off. Yeah, and the competition for Nvidia is the grocs of the world, the GROQs, right? These sort of upstart chip design design companies.

But then it's also the hyperscalers. And that is honestly in many ways more risky because it's very difficult for, or startups can kind of pick a certain edge. But the issue with the hyperscalers is they're the big buyers of these chips as well. So if they start developing their own, that's a huge risk factor. The other thing is in the news this morning, well, we are the news, but in some other news this morning, Andy Jassy was on, I think it was Bloomberg talking about Amazon's new quantum computing chip.

And so every hyperscaler cares a lot about this category is innovating themselves here and, you know, in hopes to one day be, you know, running a lot of their data centers on their own chips, right, would be the goal. Cut out the middleman. Well, regardless of what you think about, NVIDIA, the best place to invest in NVIDIA is probably public.com. I would say so. Investing for those who take it seriously, they have multi-asset investing, industry-leading

yields, and they're trusted by millions. Go to public.com for stocks, bonds, treasuries, options, and more. And treasuries are. great option in this market when everything seems to be in freefall. They rolled that out when interest rates first popped in 2022 and have just continued to ship improvements on it. So yeah, good time to have exposure to historically much more stable assets. Yeah, so go do some research and move some assets over to public. Let's move on to the spy chic taking the AI world by storm. This is a fascinating profile on Abu Dhabi's Tanun bin Zayad al-Nahan.

He has more than $1.5 trillion to spend, and he's in a hurry. He's grappling with Mark Zuckerberg. The Abu Dhabi Royal is Sun-Ky. Your size is not size. You remember that line? I think this was when Luna was crashing. People were sort of poking fun at the founders. He had this, you know, his company did collapse, but he had a good line in there, which

was your size is not size. So good. People were saying they were shorting. But nobody's size is size when you compare this royal thrown around $1.4 trillion. Yeah, move over Mossa. There's a new. There's a new big dog. Big dog. Masa when he needs, I mean, it's funny to think it's funny when Mossa is talking about raising $500 billion. You know, this is a guy that could do it, but he's certainly not going to put 30% of his entire portfolio into one, you know, sort of potentially high risk project. So let's kick it off. The Abu Dhabi royal sometimes called the spy sheik is accustomed to using his vast wealth to master his many obsessions, which have ranged over the years from martial arts to chess to video games. Now, Sheikh Tanun, the United Arab Emirates National Security Advisor and brother of the president is deploying a similar playbook at a much grander scale. He wants to muscle his tiny emirate to the front of the race to develop and control AI systems that could transform the global economy. The world is taking notice.

The CEOs of Apple, Microsoft, and BlackRock have all paid homage to his elaborate Persian Gulf Royal Compound in recent months. You know we love Royal Compounds. Love compounds. I've been, I told you before we joined the show, I've been on an absolutely iconic Persian Gulf Royal Compound in the UAE. I don't think it was this one.

There are many. There's plenty of space in the desert to put these compounds, but they really are an oasis. the one that I went to had over 100 Ferraris stored on the property. So many. There was actually, it was actually to the point where he would have like one F40 in each color. Wow.

And so, and he'd have like, he'd be like, oh yeah, I like vintage rangerover. And then he would have the entire rainbow of just one, one car, same spec,

just changing the pink color. Yeah, yeah, so. I mean, it's like the Sultan and Brunei. At one point, the Sultan of Brunei was buying more than three cars per day.

for years and he wound up with a collection in the tens of thousands of my my he was a client of my father-in-law yeah insane he was making custom cars yeah he would go and get dominator the precursor to the bentega never sold to the public i mean a lot of i just asked just asked bentley i want an SUV and they said no there's so many we'll make 10 so there's a bunch of incredible cars that originated because somebody in the gulf said make me this car yep and we can thank them for that. The G-wagon was actually, so the G-class was developed as a military vehicle from a suggestion from the Shah of Iran back in 1972. That's crazy. I'd know it did. It started as an Iranian and is still loved by Persians specifically here in L.A. Yeah, whatever Iran deal happens

next, we got to make sure there's some cool cars involved. Yeah. Let's de-escalate and shift from, you know, the geopolitical jockeying to just who can design a cooler car. Yeah. That's the world peace that I want to see in the world. Exactly. We're working towards that. But this guy seems like an absolute boss. He,

on a fall trip to the United States, he grappled with fellow Brazilian jujitsu devotee Mark Zuckerberg and donned a sweatshirt and jeans for a visit with Elon Musk at Tesla's Texas factory. All of this is possible in part because of the 56-year-old

Tannoon controls more money than almost anyone on the planet as the chairman of two Abu Dhabi wealth funds with an estimated 1.4 trillion in assets and the steward of an enormous personal fortune. A new artificial intelligence fund he leads MGX is set to be infused with more than 50 billion from his wealth and other Abu Dhabi sources. And that's like a drop in the bucket compared to the $1.4 trillion. It's crazy. These numbers are so big now. According to people familiar with the fund, billions of dollars more are set to be spent on AI by Group 42, a company he controls. This is all

part of more than 70 billion, the UAE has pledged to invest in France and Italy earlier this month. Last month, MGX was one of the few names mentioned backing Stargate, $100 billion data center project. How are they calling this $100 billion data center project when the headline number just a month ago was 500? Are they dialing it back? I have no idea. I mean, it's possible that they started the road, they basically started the road show and said,

you know, shoot for 500. Even if you miss and landed 100, you're still in a good spot, right? you're starting to be more competitive with some of the hyperscalers, but definitely dialing it back a full 80%. I'm assuming that the Wall Street Journal sort of like fact-checked this and sort of confirm because for them to just call this, if it was still a $500 billion project and then they're going and calling it a $100 billion project, it doesn't look good on the journal.

Yeah, I mean the numbers just go up and down every other day. Yeah. like add 100 billion, take 400 billion off, who knows, it doesn't matter. And then is it over what period of time? And oh, then Zoc's doing 200 billion over here. And Big numbers. And Sotya, Microsoft is good for his 80 billion, but then maybe pulling that separately. And no one really knows how you're counting this stuff up.

But a lot of money's going out. Yep. Data centers are being built. And the products are getting better. So there, you know, the hype is somewhat. here, Saudi Ramco is working on their own project with GROC. Wow. Yeah.

Okay. So it is also written big checks to OpenAI and Musk's AI, XAI, and Amazon-backedentthropic. So he's in all three foundation companies. Good. Why not? And at that point, I think people are just like, I don't like that you're in my competitor, but I, who else am I going to go to for $100 billion? Yeah.

You're the only person. So, yeah, I'll take, you can have a slice of everything. You can create an index. I mean, it's super smart. Yeah. Even in a world awash with AI in funds, even in an AI world, a wash in funds, Tanoon stands out. While many are plunging tens of billions of dollars into concentrated areas, SoftBank is making a huge bet on Open AI and the tech giants are heaving money at

data centers. Tanoon is planning to spend more money broadly around the fledgling sector than almost anyone else. It's a risky time to be splashing so much cash on AI. Citation needed. Who knows? It might be the best time ever. Well, so the next, so they back this up by saying the low cost AI model from China's Deep Seek has abruptly shaken assumptions on the need for heavy investment. Got to put them in the true zone here. It's already come out that they spent way more than the headline number that they had. And so at this point, if they came out with a super low number, single digit millions, and then sort of revised it. And then since then they've probably revised it again, but the whole sort of

of world moved on, right? So we still don't know exactly how much it costs, but it certainly wasn't anywhere near the original number that they threw out. I'll keep reading here. Valuations across tech have skyrocketed and even optimistic venture capitalists often acknowledge the market is in a searing frenzy. Leading AI companies expect to lose billions of dollars for years to come, while pledges of future profitability rest on hazy assumptions of a tidal wave of demand. Many see parallels to the late stage.com bubble when cash splashers of the era that's a great one that's great gun slinger cash splasher cash splasher cash flasher is great you got to get cash flasher we got to write that down cash what are you splashing what are you doing so when the cash splashers of the era plunged tens of

billions of dollars into public into building out fiber optic cable networks based on projections that internet growth would be even more rapid than it was a glut of fiber and enormous losses followed. We've pushed back on this before. It seems obvious that we're in an AI bubble, but at the same time, our leading AI companies have hundreds of millions or billions of dollars in revenue, which the dot-com era companies did not have. And the actual adoption of AI seems like it's taking place much faster than the internet. Totally. And the reason for that is because the internet is the greatest distribution channel for software ever in history, right? You can launch a new model and get, you know, hundreds of thousands of users in the same day.

Yeah. I want to give some more color on who this guy is because there's a bunch of fascinating anecdotes in here. He says from his ornate palace in Abu Dhabi, Tanun personally tracks the progress of global large language models on a dashboard. His researchers built for him on his phone and computer. He urges subordinates at the giant network of companies he owns to rapidly incorporate AI. So he's probably like pulling up like Chapot Arena and like MMLU like every day basically

and seeing how it's great it's so much better when you're invested in every single company because chapbot arena is just your personal sort of dashboard right yeah i wonder i wonder what else he's he's tracking on the progress because it could be like da u's right if you can get that data i have this i have a situation right now where i have these two i have these two uh companies in the exact same category that i'm a shareholder in they're both growing very quickly yep they both raised a lot of money uh and and i just happen to be friends with with both CEOs That's how I ended up there. But it's sort of funny because every other week they sort of alternate between sharing, you know, one person's like, oh, I did this.

The other person's like, I did this. I'm like, I'm happy for all of you. It's fantastic. It's great. I like this one. So Tanoon was meeting with Satcha Nadella at Microsoft. Sotia mentions a book by George Washington University professor Jeffrey Ding that suggests AI could be the base of a coming global economic boom. And this is clearly what informed this book clearly informed Sotia's.

idea about, hey, let's get to 10% global GDP growth, 5% growth in the developed world. This is what we should be thinking about AI, not reading like the AI Dumer books, right? But the sheikh scribbled it down on his notepad, and soon after Ding was eating lunch in his Washington, D.C. office when a representative from the UAE embassy knocked on his door and said, I can't find your book in any bookstore. Can I get two copies? Amazing. Just like send my guy to just like go get the book directly from the same. source. We've talked about this before. Billionaires have guys. Yes.

He just used Amazon. Yeah. Yeah. Well, everyone has apps. Yeah. Yeah. But, but when you get to a certain level of like the super billionaire, the trillionaire class. Yeah, I mean, he controls a trillion dollars. He, I don't think he made it to the super billionaire list we looked at yesterday. I think that guys in his category pay did not end up on those lists. Yeah. Or it's just like they have control over capital, but they don't personally own it. And so it's all it's much. When you look at the list, all the people are people that you can basically clock their net worth by publicly available data. Like notably absent. You can see, we know how much of Tesla Elon owns. Exactly. So when the stock fluctuates by $20 billion, you just look at his share pro rata and you calculate, you know, what his net worth did that. I think they said there were 26 super billionaires ranging from 450 billion for Elon roughly to down to 50 billion for some other folks.

Vladimir Putin, Siegian Ping, not on the list. It's like, how do you not include those guys when they clearly control 50 billion worth of, you know,

economic power, no doubt. And they're not just, oh, they're around for a couple years doing the job of president. Like, this is their life.

Anyway, that is the nature of these things. But I love this guy because- I guess we're having a Zoom difficulty. So I'm going to just keep reading through this. And you can help Connor get the stream back up. Yeah,

read this story about the Age of Empires thing. Yeah, I'm catching up here. Sorry, folks. So I'm going to backtrack here for a second. But Tanun grew up in the shadow of his brother Sheikh Mohammed bin Zaid al-Nayan, the longtime crown prince who was named president of the UAE in 2022, while his brother rubbed shoulders with foreign leaders and crafted plans to remake Abu Dhabi's economy.

Tanun honed extra-curricular activities. He became so hooked on competing with his brothers in Age of Empires, a video game where players develop hunter-gatherers into powerful civilizations, that he hired an employee from the company that makes the game to coach him on tactics, former employees said. He turned a palace within his royal compound into a gym. He funds one of the world's top cycling teams, the UAE team Emirates, and he rides on an island near the presidential palace.

He ringed with cycling pass. He likes Superlight Kalnago bikes so much that he instructed one of his companies to buy the Italian firm that makes him incredible. Tannoon has poured money into building up Brazilian Jiu-Jitsu and other forms of grappling around the globe. A black belt, he started a well-funded championship tournament that has become the Olympics of our sports, says Bobby Razak, a filmmaker who is making a documentary on Tanoon and the event. He's definitely one of the founding fathers of mixed martial arts, he said, a sport that, quote, would not be where it was today without Tanoon.

The interest in AI has roots, his interest in AI has roots in the early 2000s when he was so intrigued by computerated chess. that he hired a sizable team of engineers to make him a program called Hydra that could compete with the world's best players. Erdogan's, a computer chess specialist for opening moves, said he often talked by phone for two or more hours a day with the Sheik as they monitored Hydra's games. That's incredible. So in late 2017, Tenoon's interest increased after he learned that Google's AI program Alpha Zero had beaten the world's best computer chess program after needing just four hours to learn the rules. It was far more powerful than Hydra ever was,

opening his eyes to the power of AI, he told associates. So the following year, he tapped Peng Shao, the former chief technology officer at software-turned Bitcoin firm MicroStrategy to start G-42, a company devoted to applications in AI. So it's very cool. He's been very obsessed with this for a very long time. And so at the time, Abidavi's wealth and economic strategy, was largely under the control of Tenoon's brother, Muhammad.

As de facto leader of the country, he pushed to siphon off part of the Emirates oil wealth and put it into investments in economic development, building the wealth fund bounty to Houn would later inherit. So the thing that I think people don't fully understand about a country like the UAE is it's almost more of a kingdom that is forced to define itself in certain settings as a country, right? So when you think about the wealth of the country,

of the UAE, you should be thinking about it as the wealth of a kingdom, which is controlled by the royal family. Right? And these families are massive, right? And so you can visualize the sort of family tree. And the closer that you get to the top, the more of the wealth that you control. If you're farther down, you might control some small aspect of the kingdom's operations. You're going to have like a team that works under you, but you're the ultimate decision maker.

And so what's interesting is if you want to get. big deals done in these in these sort of kingdoms you actually have to you know if it's a big enough deal you have to get the sign off of you know more or less the king right and so you can see these sort of decision trees that trace up and that's why you know the venture world has had this idea that you can just sort of show up in abo dobi and expect to raise you know billions of dollars like it's some easy task but if you talk to any large allocator they'll tell you that these relationships are developed over years, sometimes decades, and you have to, you're going to have to genuinely be, you know, invest time into those relationships and ultimately get the sign off of, you know,

the most powerful members of the family. Yeah. I think, isn't like every Emirati gets just like a payment? Yeah, there's, there's effectively a UBI. So education is funded. There's enough, you know, sort of a monthly sort of payment to cover a lot of living expenses. This is part of the reason why cars are actually so popular there. One is because it's an oil producing nation. So gas has always been super cheap. There's like an incredible car culture there. But two, if your education and your, you know, house or your apartment are covered and you're working, what are you, and you don't. And you have a perpetual income by being a citizen of the kingdom. You're going to spend it on cars. Exactly. And it's actually interesting. There's a dynamic, you know,

Speaking of cars, cars are actually, you would think they would be more expensive in the UAE. They actually end up being cheaper because they, they, they, one, these cars can't sell back in the states. But two, there are people that use cars to export wealth from sanctioned countries. Let's say you're an oligarch in Russia and you want to get wealth out of the country. It can oftentimes be easier to move a LaFerrari than it is to move $5 million out of the country. So there's a bunch of cars there that sell below retail. and below their sort of market value in the U.S. because people just want to get them into the UAE, move them, and then they have, you know, real dollars that are sort of protected within the kingdom.

Yeah. I think they were also struggling to set up like top tier universities. And so that's where they came up with this with the idea of like, we'll just pay if you go to school in the U.S. If you get into Harvard, we'll pay full tuition for you. Yeah. Because like, yeah, then it's just. Yeah, there's a lot of that. It's as good of an investment as trying to build the next Harvard in the UAE.

Yep. interesting. Well, did you get to the section where they had to rip out the Huawei routers? Do you see that? No, we didn't get that far. The UAE has long seen itself as something of a Switzerland, a trusted point for trade for east and west, and choosing sides on AI would risk damaging relations with China. But the allure of the U.S. lead on AI was too much to ignore.

Tanoon picked the United States. Let's go. And he says, Huawei routers were ripped out of G42's offices. and Western equipment put in its place. No more spying. Do not look at our internet traffic, China. While Tanoon and his aides continued negotiating with the administration on details

to enable the deal of the administration paired Tanoon's company with Microsoft G-42 agreed to use Microsoft's cloud for AI work and also to comply with U.S. export restrictions as part of the deal. And Microsoft invested $1.5 billion in G-42, joining its board of directors. Still, China Hawks and Congress are marine concerned. that the UAE's overall ties to China remain strong, fearing the country could allow powerful

U.S. tech tools to make their way to China. Those lawmakers could be an obstacle if Tannoon comes back for similar deals. Yeah, I mean, the challenge of being Switzerland of the Middle East in many ways is that, you know, the UAE does have a lot of economic ties with, from what I know, Iran, right? There's a lot of Iranian money there, Russian, et cetera. But the cool. thing about being in the UAE is if you go there you'll sit down at dinner and you're out of table with with it's the most diverse group of people I've ever been a part of I got a chance to go a couple times in what was it late early and then late 2023 and you'll sit down at a table with somebody from Iran somebody from Israel somebody from Russia somebody from the UK

an American and everybody's getting along because everybody's just generally there to, you know, in that setting, enjoy F1 and, you know, everybody's, you know, generally pro-business there. So it's a great environment. Yeah. I mean, the capitalism and business has been like the one thing kind of holding the Middle East together. Universal language. With, you know, UAE and Saudi Arabia kind of playing nice with both sides and kind of keeping everything together and a lot of pressure to not let anything fall into just absolute chaos, which is obviously risky. So speaking on a podcast in December,

they're talking about Josh Kushner, or Jared Kushner, recalled a 2018 meeting when he was visiting the region. Tanun told him, Jared, I'm not meeting with you unless you spend two hours after with me to talk about AI. Wow. And so he's trying to build his...

He's going podcast mode. He's like, you've got to talk to me for at least two hours, the shortest Joe Rogan episode. Yeah. And we'll get into the details. That's great. And so, yeah, he's been investing in a lot of stuff. And I'm sure he'll be someone.

We've got to have him on the show. Whose presence only grows over the next few years as he continues to make big investments. Anyway, if you're looking to take a trip and stay in a nice place, maybe impress a chic while you're on the fundraising tour, we recommend going to wander. That's right. Find your happy place, book of wander with inspiring views, hotel grade amenities, dreamy beds, top tier cleaning in 24-7. It's a vacation home, but better. And we have some big news from Kyle over at Wander. They just crossed 600 total locations.

And you were in the chat asking, weren't you at 500 a month ago? It said yes. Yep. It's pretty great growth. It's amazing. It is incredible. It's amazing. And what's cool is Wander is a platform where every incremental property makes a platform

more valuable, right? You know, I think that, you know, we dealt with this before. I was trying to stay to wander in when the fires were happening. They were all booked. I'm sure, you know, they've added multiple properties within Southern California and the pace that they're operating at is just incredible. So go check it out. They actually are running a spring sale for the last day today that you can get $400 off plus

a wander hoodie with code spring 400. So go check it out. And you won't regret it. It is unbeatable. And we are going to be staying in at least one wander in early March. and very excited to record a podcast. We'll have to do the show from The Wander because their Wi-Fi is so fast. For sure.

It makes it all possible. Live stream it. That's great. Well, let's move on to an analysis by Shiel Monat, breaking down Adyen and Stripes' 24 annual reports. They both are out, and so now you can benchmark these two companies against each other. Ad-Yen processed $1.34 trillion and was up 33% year-over-year. Stripe's growing faster at 38% year over year and is now at $1.4 trillion. Ad yen's valued at

$56 billion, public company. Stripe is valued at $91.5 billion. And in terms of employee count, ad yen has $4.3,000 employees. Stripe has $8.2,000. They're both profitable. Ad yen, EBITDA, $1 billion. Note, Stripe has higher margins. And so this comes to us from this a wonderful Stripe Community Letter that we will read some of this stuff, but it highlights a lot of interesting applications for the company. And we'll go through some of the highlights here. Yeah, the big, I mean, the big thing that stood out to me from Shields analysis, everything, you know, the two, the two numbers that jump out. One is Stripe is almost double, valued more than double. Yep. Yep. And Stripe has almost double the number of employees, right? And, and,

So anyways, there's the big story here is how, you know, how is Stripe worth twice Adyan, despite having less EBTA and twice the employees and all these things. But yeah, the letter here is fantastic and just shows the level of and the leadership that Stripe has with the Collison brothers. So their, Stripe's mission, of course, is to grow the GDP of the internet, right? Yeah, fantastic. And payment volume represents 1.3% of global GDP now. That's insane. Absolutely insane. It's one of those things like they could have said, hey, over time, we think the

internet's going to grow a lot. We think internet payments are going to be big. We think we can actually get 1% of the global GDP running through. And it's the most ridiculous number because every founder has thought about a world, oh, if we just get 1% of this market, you know, we're a multi-billion dollar company. Stripe actually did it. And, you know, an example of one of the most hyped companies of the last 20 years living up to the hype, right?

With the way that they're growing, there's a world where there is a scenario where they have double-digit percentage of global GDP running through their rails, which is just obscene. Yep. A straight billing alone is on a $500 million revenue run rate. We use straight billing. over 300,000 companies are on Strip billing. And Stripe was profitable in 2024 and is expected to remain so in 2025 and beyond.

And of course, they were one of the companies to kind of move quickly during the last down turn. And kind of, I don't know if they had to lay a lot of people off. I think they actually mostly just froze hiring and kind of grew into their valuation. And it set them up for profitability and then kind of scaling from there. So they kind of know what the base, you know, how the business operates. how it runs. Obviously, they're very big, and they can kind of grow into their valuation. Nearly 200 million active subscriptions are managed via Stripe billing. Stripe is trusted by half of the Fortune 100, 80% of the Forbes Cloud 100, and 78% of

Forbes AI 50 companies. One in six new Delaware corporations incorporates via Stripe Atlas. That's crazy, because I always thought of that. I mean, there was Clerky years ago, and those products, they're very easy to get. get, you know, popping in like small Silicon Valley cruise, but you don't get to one out of six without being like a mainstream product. The dynamic with Stripe Atlas is, is fascinating because it doesn't seem like they do a lot of independent marketing. It seems like a product that Stripe built truly for the community. It was a pain, you know, it makes sense. It's a perfect wedge product for payments, right? What do you need to do before you can accept payments online?

You need to have an entity. They built what I think is the best way to create an entity. online. You know, we used it for TVPN. It's a fantastic product. It's still super simple, despite it being a relatively complicated process to create an entity. The product is fantastically simple. And it's turned out, you know, it's turned out to be an amazing wedge product for them, just given the scale. Yeah. It's got to be a hard product to advertise, right? Because like, you only incorporate once. And not everyone's incorporating it all the time. And it's not like there's, So, yeah, I'm sure they're all over Google ads and stuff for when people are thinking about incorporating. But outside of that, it's hard to do a lot of awareness for that product, but they've done a very good job getting penetration there.

Hertz, and they give some case studies here, Hertz increased 4% of online payment authorization rates after switching to stripe. Turo saw a 4.7 boost and recaptured revenue, translating to an extra 114 million per year. I had no idea Turo was that big. Wow. That's huge. Intercom improved conversion rates by 2.1%. Forbes saw a 23% revenue uplift within six months after switching to Stripe.

There's a ton of examples of this. Major AI companies including OpenAI, Anthropic Mid-Journey 11 Labs are Stripe partners, Stripe's agent toolkit and virtual card tools enable programmatic spending and control. More than 14,000 vertical SaaS platforms utilize Stripes payment services. I mean, it's just like the default in Silicon Valley, at least.

Stripe Capital provides fast funding with funds arriving in under 48 hours. They've really like just built every possible fintech product on top of. Yeah. And if you look at the actual three lines of code. Yeah. I have to try to pull this up. But if you look at their product timeline, there is a pretty amazing.

You can actually see, I wish we could pull this graphic up, but it's going to be difficult. But they basically come out in 2010 with payments. Yep. Two years later, they launch Connect. Yep. Four years after that, they launch Radar. and Atlas. So this entire time, they're just focused on making this perfect payments product. And to show that level of patience, even when they had sort of exploding growth and traction,

and many operators would have said, hey, we should go multi-product. We have all this traction. They waited. And then since then, starting in 2016, they really ramped up the pace. They launched Atlas, Radar. 2017, they launched Sigma. 2018, they launched terminal issuing billing, 2019 capital, as well as corporate cards, and the infrastructure to issue corporate cards. And then in 2020, they launched Treasury as well as climate. And then they launched a tax product in late 2020 as well. So the velocity went from, hey, let's just have this sort of narrow focus and really own payments.

And I think that Stripe is one of the best examples in history of focusing early on. really, really directly on startups as a category. Yep. And they sort of ignored the enterprise pretty broadly, but they benefited from between when they started in 2010 and where they are today, many of their early startup customers became public companies that are now massive.

And those were, you know, customers that were able to prove out these use cases that would, you know, go on to allow them to, you know, sign companies like Hertz, right? Which is this legacy company or Forbes, things like that. I remember Shopify. When Shopify started, like the default payment integration for a long time was Stripe.

And you kind of bring your own Stripe account. And then they were doing Shopify payments, which was kind of a white-labeled stripe account, but it was all Stripe under the hood. And yeah, they've just really, really benefited from just letting the internet grow and being like the index on the internet economy. Interesting.

A good example where the mission of the company grow the GDP of the internet was truly, truly aligned to the potential of the business itself, right? I thought this was interesting. They mentioned the bridge acquisition. transaction volumes and stable coins doubled from Q4 2023 to Q4 2024, and Stripe now supports 40 million monthly active stable coin wallets. Wow.

And yeah, that's something that like they've been talking about crypto for a long time. They've integrated crypto products of the past, but like the stablecoin thing feels like the first time that stuff's like really taking hold. And I know I saw packing. They made the bridge acquisition. They spent, you know, over a billion dollars on it, specifically because of the developer activity that Bridge had. So they, I think, saw a younger, you know, they saw the Stablecoin version of Stripe,

which had an early developer traction. It was a product that developers loved. And then all of a sudden, you know, public companies are using the product. Yep. And so they write Stablecoins, room temperature superconductors for financial services. Top stable coin uses today involve tangible real world activity. CFOs use stable coins to manage corporate treasury. Immigrants use them for remittances.

citizens of countries with unstable currencies use them for dependable savings and payment teams use them to enable customers from countries with low card penetration. Some of our favorite examples, SpaceX uses bridge to repatriate funds from Starlink sales in Argentina, Nigeria, and other markets. Dollar app, a neobank in Mexico uses bridge to help individuals receive USD payments from payroll providers like deal, and Airtiam uses bridge to disperse payments to workers all over Latin America. So as as we start paying more people all over the world, the stable coins make a ton of sense. I've actually paid people in stable coins before because they've been in countries where

getting them U.S. dollars was very, very volatile. Delayed and high fees, super high fees, yeah. And so. Yeah. And and and to be clear, there was a lot of analysis of the bridge acquisition at the time where people said, how could they pay 100, you know, it was like more than 100 times revenue. It was a really big number. And the obvious answer is that Stripe has over a trillion dollars of volume. They're big believers in stable coins. They had had their own internal efforts. Why not spend a billion dollars to acquire a leader in the space and then focus on integrating it with all of your existing customers? So I think we're going to see over time that that could very well be an Instagram style acquisition for Stripe in a sense that you can imagine the bridge product

you know, 10 years from now having you know hundreds of billions or potentially even trillions of dollars in volume. And I like this. We've been seeing this. AI growth chart all over the internet and Ramp or Stripe, Ramp's done some analysis on this showing adoption of AI tools on the corporate card side, Stripe seeing something similar. They say we are seeing an AI boom on Stripe. We're partnered with a large number of companies with rapidly growing businesses, including Open AI, Anthropics, Suno, Perplexity, Mid Journey, Cognition, 11 Labs, Langchain, Pine Cron, Mistrel, Cohere, Sierra, so many of these.

And the median time to annualized revenue milestone for the top 100,000. AI companies on Stripe in 2024 was 24 months. They compare that to back in 2018, it took SaaS companies, the top 100 SaaS companies by revenue, 37 months. And so they're providing extra data to show that the AI boom is real and companies really are accelerating their revenue growth faster than in previous, I don't know, in previous waves. Yep. And so Cursor, the AI powered coding assistant, raced to over 100 million. recurring revenue in just three years. We're also seeing the likes of Lovable, which had 17 million ARR in just three months. This is wild. So lovable for those that don't know is a Swedish company.

Harry Stebbings, fellow podcaster, was one of their early investors. And it, you know, lovable doing that in three months. It just puts so much pressure onto every other company because that's now the bar. I remember in 2021, the narrative was you should be able to hit one million of ARR in nine months post launch. That's now, it seems childish in comparison. In the cage, we're trying to hit a million in three months. But anyways, the demand for these AI products is insane. And I believe Lovable has more ARR, or is that a much greater run rate than booking.com

was when they went public. Really? Yeah. Oh, yeah, yeah. I think we dug into that during the Darra Kosher Shahi and that Mary Meek article when she took them public. Yeah, I remember that. There's an interesting point in here. Some people have called these startups LLM wrappers. Those people are missing the point. The O-ring model in economics shows that in a process with interdependent tasks, the overall output or

productivity is limited by the least effective component, not just in terms of cost, but in the success of the entire system. In a similar vein, we see these new industry-specific AI tools as ensuring that individual industries can properly realize the economic impact of LLMs and that the contextual data and workflow integration will prove enduringly valuable. So they're bullish on these quote unquote rapper companies, which I thought was interesting. Anyway, let's move on to tell you about another fintech company that we love here on the show,

Ramp. Go to ramp.com. Time is money, save both, easy to use corporate cards, bill payments, accounting, and a whole lot more all in one place.

Go get to know Ramp on Ramp.com. Okay, so speaking of Ramp, we have a major Ramp customer, Sean Frank. Oh, yeah. I'm going to be joining this show in two minutes. Fantastic.

So I'm going to text him right now. And we'll get him on here. Well, we can do this one DM that we got in the meantime. It says, Ahoy, John, hope you're doing well. Apologies, apologizing for hitting you with a bunch of questions, but I've posted 50 blogs and I have four subscribers. I'm not gaining,

not that gaining subs is the goal, but still, I've talked to family members who's collectively hired thousands of people and they're out of ideas to the whole ad value. and it will come approach just hasn't worked. I want to work with smart people before going out and building something on my own, but I'm not sure what the move is. What do you think, Jordy?

So I think that one of the things that seems very obvious with content is that you need extreme levels of patience. Yep. That's not an excuse to not refine the product every single day. Yep. And, you know, we take this approach. We're lucky to be in a position where we have a daily show. We're trying to improve the show in every single day.

There's some days where, you know, maybe we weren't, we're underprepared, some days where, you know, we sort of miss a segment that we wanted to do. But in general, trying to improve the product every single day and then improve how you're distributing that product. I think it's the distribution. So I posted 50 YouTube videos before a single one went viral. And for me, it was over a year. It was a full year. I posted them once a week for full year. And it was like, oh, I got 100 views. Awesome. Oh, I got a thousand views on this. That's great. And then finally one got picked up. about SpaceX and the algorithm that went to 70K views. And I remember seeing the chart go up and I was like, wow, this is amazing.

But it's really hard for blogs to go viral. Yep. Because there's no algorithm. There's no discovery. There's no discovery. And so what you should probably do is figure out, can you repackage this in terms of just a long post on X or a thread or because a thread can go viral, short form content. If your natural voice is on short form video, make one minute videos.

break your blog post up into a series of one minute videos. Try a bunch of things. And then very quickly, the algorithm will start sending you the signal of like, this got 10 times as many views as that. What did you do differently? And you can introspect. But it's very hard when you're just hanging out on a dead substack with no traffic coming to it.

I think that's very hard. And I think the people that have been successful on substack have built audiences elsewhere. Like Ben Thompson talks about tretechre would not exist, if not for sharing links on Twitter and LinkedIn at the top. time. And now he's got it cornered and people share it and his stuff goes viral even if he doesn't post it because people will say, oh, there was a great interview. We got to talk about this. Exactly. Yeah, I think the main thing is YouTube specifically as well as some of these other platforms can be very demoralizing. But that doesn't necessarily, you know, the X factor now is that the algorithms need time

to learn your content before they figure out the right people to distribute it to. So do we have Sean on the line? We have Sean on the line. He's in the waiting room. Let's bring him in. We're going to talk with Sean about app-loven. And Sean is one of the sharpest minds that I know on e-commerce. Actually, I'm going to go ahead and say, Hey, Sean. There he is. Mr. Ridge himself, Mr. Wallet.

You guys are in person. I was expecting you guys to be doing Zoom yourselves. But damn, IRL. Love to see it. IRL. Always. Always. Always.

Unless you're dealing with this traffic in L.A. these days, Sean, if you want to leave your house now in L.A., it's about an hour. to go anywhere. Are you still going into the office in Santa Monica every day? No, I went yesterday. Yeah, it's like an hour and a half, bro. PCH is still closed. So, yeah, LA traffic, bud, can't, can't beat it.

Well, tell us about Applovin. Yeah, so Applovin for some backstory for the audience, founded in 2012, really started to, it wasn't even necessarily a company. I feel like that was on the map for a lot of people, until last year. I think last year was the first time you guys started spending money on the platform. Is that correct?

Or had you tested it at earlier points in Ridge's history? Yeah, so here's, I think Apple 11 has three discussion points. We'll go from what I'm least qualified to speak on to what I'm most qualified to speak on. So least qualified. Short reports are coming out that they're violating, you know, App Store, terms of service. their ads target children, all these types of stuff. What Apple of it is, is there's two sides of the business. They run a bunch of mobile ad space.

So games, mostly games. But whatever old people are playing on their phones, they have that sort of ad space. And they have an ad engine, right? So like the ad exchange, think of like a trade desk or Google Display Network. So they run both sides of those. they've never done e-commerce their e-commerce push was like August or September 24 so we're like and the majority of their and just questions so the majority of their customers until they push into e-commerce were other mobile games I'm going to guess and other sort of digital apps services that

they were pushing people to what was do you do you have any idea of what their customer base look like until last year yeah I think it was all mobile games so you know they've been called like a pyramid scheme, all these types of things. So you're a mobile ad engine where other mobile games can run on top of, right? I think they also do like display branding, like trying to get you to go to the movies or just like real top of funnel campaigns. Nothing conversion optimized, right? Now, a lot of people spend a lot of time on mobile games.

So there are a lot of ad units available. But it's something Ridge has never touched at all, right? Like we're mostly a Facebook advertiser. They get some year 75% of my. budget, we do influencer, that type of stuff. We've never really done that much in the world to display and especially nothing in mobile display, right? The industry does have a bad reputation. Mobile display ads are just like, that's where ClickFron exists. Yep. Well, in my, and from my understanding with Ridge, you know, having, you know, worked out of your guys' office and worked on the Ridge business

myself back in the day, you guys have a had a policy where you'll test any channel once. But you're very, very ruthless with pulling back that spend if it's not working. So I remember, you know, Connor being, you know, who's the CMO of Ridge, being willing to test some obscure ad network and being like, yep, doesn't work, you know, almost like the same day oftentimes. So when you guys decided to test app Lovin and then actually start to really spend on the platform, I think that sent, you know, a real signal. Right. Well, I'll say that we're a very sophisticated advertiser. For people unfamiliar with Ridgewallet as a business, I mean, we'll spend over $100 million this year on digital advertising, right? So that's everything from LinkedIn ads to TV ads to influencer to Reddit to Snapchat.

Like meta is the biggest best game in town. They might get 70% of my budget some year. Sometimes they get 30% of my budget. But like outside of them, we're spending money on every single possible channel we can. So what are the accusations against App Lovin? I'm going to put the unverifiable claims that I cannot speak to in one box. And that is there's short reports saying that like, you know, they're doing one-click force downloads, they're tracking users, whatever, cannot speak to that at all.

Like if that's true, you know, let's hope there's an investigation that we don't want them targeting children with inappropriate content. So I don't know anything about that. The second thing is the stock price has gone parabolic in the past like two years or whatever. I think the IPO maybe in 2021. I think it's up 4x in the past year. At one point, it was worth more than Uber. So as soon as a company is trading at 100x revenue and is worth more than Uber and

no one's ever heard about them, people just get immediately freaked out, right? It's like, you know, they had the same, you know, revenue multiple as Palantir, right? Palantir's been around a lot longer. People know about Palantir. So, yeah, just for some extra context. So they were worth $3.8 billion at the end of 2022, at their peak, just a little bit ago, they were worth $160 billion. So generational run.

Yeah, yeah. And anytime that happens, I mean, a lot of people start speculating, right? Revenue is growing really, really fast. I can't, I'm not a stock trader. I fucking buy ads. So is the company overvalued? It's about as overvalued as palenteer. So depending on where you land on that side of the conversation, that's why there's so much interest right now, right?

The thing that I'm qualified to speak to is as an advertiser, they shoot us an email in August and they're like, hey, we're rolling out e-commerce. We'll give you a lot of ad credit to test it. Right. And we get ad credit from Snapchat to test stuff, from Reddit to test stuff, from meta to test stuff. And free ad credit, even if it's the shittiest ad space on earth, I mean, okay, yeah, I'll get some I'll get some, I'll get some bunch. punk ass impressions, whatever, right? They end up, you know, it ends up working really well. I probably spent $4 million of my own money on app loving ads in Q4, and I was spending

$250,000 a day. It was matching Facebook ad spend, right? So one of the short reports is seem like, oh, they're stealing Facebook's data to be able to target these people so good. Once again, I don't know how the fuck that would work, but what I will tell you is that as a sophisticated advertiser who tracks everything via North Beam, via incrementality holdouts, via post-purchase surveys, coupon codes. People were playing App Loving Games, seeing Ridgewallad ads,

buying those products, and then telling us they saw it on the ads. So it's a real ad channel. It really worked in Q4. I have an invoice right now. I owe them $1.5 million that I have to pay by the end of the month. So, like, it's a business. I don't know what to tell everybody about the shorts. Now, do I think the company is overvalued?

Yeah, I don't own any of the stock. That's fucking insane. It's like it's a mobile app platform. What do you mean they're worth more than Uber? Like I fucking call an Uber every day, right? I don't play any mobile games. So that's like the stock price question. Then it's like, oh, all of this nefarious, whatever, inside dealing, subsidiaries.

I don't know anything about fucking that. What I'll tell everybody listening. And what's so weird is I'm public. I'm on Twitter. Like I post my ad spend on these. channels, no one's reached out to me. Like none of the shorts. I'm like, guys, I'm around. Like, I'll show you

the fucking app. Hey, get on, uh, make him pay for it, you know? Don't give that, don't give up the sauce for free, you know. Yeah, you're going to go on Tigis, become an expert, uh, expert consultants. You can make a thousand dollars an hour. 10,000 dollars an hour. Like, I have no dog in this race.

I'm like, yeah, dude, whatever was an ad channel. If they went away tomorrow, I would spend money on Snapchat or TikTok or anywhere else, right? But like, the very viable proof, is that I gave them real money. It was a good ad channel. And I think the reason why people, like, their revenue is growing super fucking fast, right? The revenue is up 80% year over a year.

And they're like, yeah, we have a billion dollar run rate for e-commerce advertisers in December. And I'm like, yeah, no shit. I spent fucking a million dollars. It's like, yeah, I spent a ton of money with you guys. Like it doesn't take too many people like me spending that to get really, really high revenue. Dude, ads are the best business on earth. I don't know why this is so surprising for people.

Now, are they compliant? Do they owe Apple money? Are they stealing data for Facebook? I don't fucking know, but pretty good ads. You got any questions about that? Pretty good ads. Can you talk about the relationship between App Lovin and Flip? Has Ridge had any success on Flip?

I remember Flip blowing up out of nowhere. They were giving a bunch of credits to basically give away products for retailers. I know they have an ad partnership. Have you found any success on the Flip platform? Yeah. So I've never used Flip. As far as I know, Ridge has never worked with or advertised on Flip. It would be a surprise to me to hear that my ads were on Flip via App Lovin.

I've had friends who've had very bad relationships with Flip, so I've always avoided it. And our products don't work for live shopping anyway. Like TikTok shops, it doesn't really work for us, Ethan. And it's because of like, with a very male-focused demo, we're more of a premium product. You know, those typically live shopping works for like, you know, leggings that make your butt bigger and that type of shit, right? Like, it's not our core product that's crushing it on there. That whole affiliate GMV world is very, very dark and shady.

I know a lot of people who are doing weird stuff there. You can check out the affiliate top charts, like what goalie's doing. They're giving people fucking mansion to, depending. They hit GMB targets of a million dollars. So, yeah, I don't know anything about Flip. I've never used Flip. They've emailed me all the time. Yeah, that, I mean, even an idiot like myself, I'm like, that looks like a pyramid scheme.

I'm not going to hang out there. But, dude, you know, I have the data on App Lovin, right? Like, 85% of our traffic from Applevin came from iOS devices, right? Like, one of the concerns is that, like, oh, this is just bought traffic on Android devices. like Click farms in Bangladesh or whatever, right? It doesn't look like it's that. So I don't know how they have good quality iOS traffic coming through. It all seems to be games called like Tomb Blast or some shit.

But like anyway, old people bought rich. Well, yeah, the reason I wanted to have you on is, is, you know, there are major brands out there that will run on an ad platform that's just sending them bought traffic and they won't even necessarily turn it off because to them, they're just optimizing for, you know, impressions and clicks. And Ridge has always been a company focused on, you know, the end actually, you know, conversion, right? And sort of building your guys' engine around that.

What happens, you know, shifting the conversation a little bit? What's your outlook on e-commerce in our year 2025? We covered maybe a couple weeks ago. We talked about how the whole Amazon, you know, it's a terrible time to be an Amazon seller. It's been a tough time to be a Shopify brand as well with all this sort of tariff action and things like that. Historically, Ridge has made a bunch of products in China. It's shaping up to be a rocky year for e-commerce broadly. It helps to have a fantastic brand like Ridge.

But what's your outlook going into the rest of the year? Oh, dude, just take it day by day. I mean, there's more tariffs announced today. There's more supply chain disruption. you know, and it's not just the U.S. government, like every government on Earth is becoming more protectionary. Like, Section 321. Sorry to interrupt, but we had Ryan Peterson from Flexport on yesterday that said that we was talking about the actions that Mexico had taken around what's not the same as the de minimis

loophole, but I'm forgetting the right name for it, but to sort of protect Mexican textile workers So, yeah, a lot of action. Yeah. And like, we like to think about this one-dimensional, like, very one-dimensional. Like, Americans importing things from China cheaply, but, like, that fucks over other manufacturing economies, like in Mexico. They don't want fucking cheap auto parts or cheap apparel coming from China sitting in their warehouses going to America duty-free.

Like, everybody gets pissed off on that. So, yeah, that happened yesterday. The quick summary is, just take a day-by-day and try to survive. I don't have any debt if you can afford it. Keep a lean team. I think D2C 3.0 is like incredibly lean teams. Like ridges down to 60 U.S. full-time employees. We're doing more revenue than ever with less people than ever.

And I think that's going to be more of the trend. And yeah, I mean, if you can bring stuff to America, you should. I mean, we're probably eight months away from 60% of the while it's being made here. So I'll keep you guys updated when they start getting made here. Where's your new factory you're going to be? Or where you onshored some manufacturing. That's been a long-term process. But, you know, what does that look like?

I mean, yeah, it's all component part manufacturing. So, like, we have interplates that are going to be made in Missouri, and then they get shipped up to Philadelphia to be finished, and they get shipped to Arizona to be assembled. And then they go to fucking Kentucky to sit in the warehouse. So this is one of the challenges. You know, China very strategically has everything within like a one. hour radius when they do like you know these these industries uh but with tariffs today i mean

it's 85 percent on aluminum goods coming out of china so it gets very close to the same price now so i think that's why he's doing it wow well thanks for stopping by appreciate the update love the take sean you will be a regular regular uh welcome to the show and uh thank you for for posting your heart out every single day yeah i'm the number three the best guess you guys had so far. Yeah, we'll get you an NFT just for you. Bye. Thanks. Thanks, Sean. Talk soon. Cool. Well, Delian should be joining any minute. I don't know if he's in the queue yet, but quick backstories. So Ridge sent, or sorry, Ramp sent Ridge a plaque for spending $100 million on ramp. So putting up some huge numbers. And yes, Sean is an absolute animal.

It's great. Well, let's see. Is Delian in the queue? Not yet. Okay, I'll text him and let's move on to Michael Magnano. He says, are people, and people are still wondering why every single podcast is pivoting to video. And it's a result here that says that a billion people are watching podcasts on YouTube every month. Yeah, it makes sense. I mean, if you think about it, Rogan really led the charge here. I feel like he had such a massive audience on YouTube. so much of the attention that he would capture was YouTube-oriented. And you remember there was a period where Rogan's influence and attention fell like it fell off a cliff after the initial Shopify deal. And you could feel it almost ramp back up again as he went back on YouTube. So massive platform, you know, one of the biggest reasons for this is obviously that podcasts have replaced a lot of television consumption. Totally. And what people don't always realize about YouTube is that a massive amount of YouTube viewership happens on TV. So this is people in their house instead of turning on

CNBC. They turn on TVPN. I get crazy screenshots all the time. People text me photos and they're watching it on the TV. And it blows my mind. It was the same thing with my YouTube videos. I was like, you watch this on TV. That's like insane. But I guess like, yeah, when you put a lot of production value into it, it looks like TV. Yeah. No, and that was our thesis. That was our thesis around if you actually want to compete with television around news, you have to record every single day. Otherwise, you know, nobody... In high quality like this.

And it's also just so much more snackable because we can take that clip. We can take some of those, some of those great clips that we just had, put those out on X, put them out on shorts, make vertical shorts out of it. And if you just have pure audio, you just can't distribute it that much. It's really, really hard. Sometimes there's transcripts that can go viral every once in a while, but it's rare. Yep. Anyway, Delian should be joining in one minute.

Let's see if we have another post we can go through. Bullish on one person teams. Card is at 2 million. AARR chat base is at 3 million ARR. Levels I.O. Products are at 3 million ARR. Testimonial IO is at 1.5 million ARR. TDIN.Me or underscore me products at 1.2 million ARR.

Ben Lang says he's bullish on one-person teams. Pretty crazy. What was the decision between PMF or die to go two-person team? The whole thing was that PMF or die was like, how do you apply these sort of extreme constraints, 90 days, 25K to a million dollars for ARR, sort of. of the simple concept, but then outside of that, you have to say in the house, but there's no real rules. You can kind of do anything you want. That's business, right? That business has sort of, there's laws that govern business, but there's no actual like rules in the same way that when two

teams go on an NBA court, they have to follow, you know, specific guidelines. And so, uh, we wanted it to, to as close as possible match sort of like free market competition. The one thing with this Ben Lang post that's interesting is this to me proves that if you are really, you are really, ambitious and you want to have extreme scale and impact, you actually have to build a team, right? Because this list would be more impressive if every single, you know, company on here had 20 to 100 million of error. I think we're ready. Delian, are you there?

Can you hear us? Let's bring in. I think he's close. We're figuring it out. Let's see. Three, two, one. Hopefully he's live. Going in.

he says he's waiting to be let in he's in the room let's get him on the show two call-ins one day it's a big big push but we're ramping up we're going to be doing 10 call-ins a day pretty soon as long as people won't come and chat we got to make this easy for folks hey deli how you doing can you hear us

there you know what's up sorry no all good in a while Give me the update. What's the latest with Varda? Yeah, you know, brought in a capsule from, you know, sort of Mach 25 down onto the, you know, sort of Australian desert floor. Teams out there, you know, going to retrieve the capsule over the course of the next couple hours. It's currently nighttime in Australia. Yeah, all things, you know, sort of looking pretty good. So we're two for two, baby.

Nice. What was in the capsule? This time around is an Air Force Research Laboratory, a spectrometer. They're basically staring out of the window of the reentry capsule to basically characterize the plume of the reentry vehicle and compare that to some of the materials that are on board. I'll leave it to the viewer to speculate as to why the Air Force might want to do that in relation to potentially trying to figure out what other vehicles are made of, perhaps ones that they themselves did not make. And we'll leave that exercise up for the viewer. But we're super excited to sort of help out with the maturation of that type of technology and what it means for our country's national security. That's awesome. We've got to call Jesse Michaels to get him to speculate.

Yeah, exactly, exactly. What does the cadence look like going forward? Do you know when the next one is already? Are there going to be just dozens of these? Give me the kind of next couple of. Yeah, next week. You know, hopping right back up and getting back to it. So, you know, quick turnaround time, turning this from a, you know, one-off R&D mission to this year.

We're planning to have four vehicles total. Three of those will be more DOD-focused missions. It's not to say that there won't be some farm on board, but it just won't be like the primary, you know, sort of focus of the mission. And then one of our missions this year will be a primarily pharmaceutical, sort of focus mission. And, you know, we kind of expect that, you know, sort of blind on a go forward basis. And, yeah, looking to basically just continue to increase the cadence, you know, sort of year over year. Definitely some, you know, sort of vehicle improvements that will happen, you know, from one mission to the next.

And generally for the time being, continuing to, you know, sort of laying in Australia, just given the regulatory regime and the emptiness out there, making it a bit easier. But at some point, looking forward to returning back to the, you know, United States once that becomes available. Are you expecting that to get easier in the U.S. with the new administration? For sure. I think it's something that the new administration is so clearly aligned with,

but it is just one of these operational things that just takes some time to flow through, designating which of the military ranges is best as a landing range in the United States. We're basically just taking the same sort of legal and regulatory framework that exists in Cape Canaveral and Vandenberg, both of which were government-run ranges that now obviously prioritize commercial activity.

There isn't the equivalent of that, for reentry. Right now there is no sort of dedicated government run, either sort of reentry range. We're exploring a variety of other usage options, potentially either sort of private, either sort of range as well, as we continue to narrow in the user size of our landing radius. And so over time may not need to be reliant on a, you know, sort of government plot of land, but instead, perhaps there's some private plots of land that are large enough. And so, yeah, what is the ideal I'm super excited to have a... Sorry to interrupt.

What is the ideal range size in your view this year? How much smaller can it be over the long term? And is that even the right question to ask? I think that information is ITAR restricted. So, you know, what I can say is, you know, on the order of, you know, around 100 kilometer circle or so, that's pretty, you know, should pretty big to have that be fully empty and owned by one person. but that circle will sort of get smaller and, you know, maybe sometimes we won't open the parachute

and it'll be a very small circle. That's great. Are you tuning into the next Starship launch? What are you hoping for there? What are you watching for that? Yeah, I mean, look, you know, we built this business predicated on Falcon 9, you know, sort of reusability and the cadence of that. And it's been really exciting to see that mature over the course of the company's life, right?

you know, sort of Falcon 9s at this point are launching every 24 hours. We're super enthused by, you know, sort of SpaceX's progress and the, you know, sort of potential of starship. You know, there's definitely, you know, obviously, you know, sort of, lots of kinks that need to be, you know, sort of worked out, but, you know, not so dissimilar to, you know, sort of Falcon 9. It definitely took it a decade from, you know, sort of the first landing to, you know, sort of becoming a, you know, sort of true workhorse that was, you know, sort of launching every day. Both SpaceX is much larger, but also so starship in terms of complexity. And so, you know, I hope that they, you know, sort of get, you know, sort of through that, you know, sort of faster than a decade.

But also, they've got a lot of programs and payloads that are more important than us. You know, they've got everything from the, you know, sort of lunar activities that they've got, Martian activities. They've got Starlink satellites to launch. And so, you know, at least, you know, sort of my internal underwriting is like, I'm not quite sure that, you know, sort of starship is going to be launching people like Varda, you know, before the end of the decade. So it's, like, super exciting.

But, like, I also have to run a business, you know, month by month, year by year. And so I can't focus on things that, you know, sort of relevant. into 2030, you have to focus on the things that are taking me up to, you know, space today. And Falcon 9 a pretty damn good ride to orbit every 24 hours and, you know, reusable, at least for the first stage, pretty, you know, it's a pretty good offering. And, you know, perhaps there are other people that are starting to, you know, sort of, you just try to, you know, elbow into that, you know, sort of world, you know, God bless Papa Bezos and, you know, hope it works for them.

And, you know, we're happy to get up to space, you know, which way or how, you know, rocket, rail, gun, space plane, you know, just whatever works, I don't care. I just need my sweet, sweet microgravity. Amazing. How big is the company? What roles are you hiring for? What's kind of like the biggest challenge over the next couple years? Yeah, we're about 115 people, basically entirely in the Los Angeles area.

Of course, the next year, I think there's two, you know, sort of, you know, sort of core areas. One, you know, sort of pharmaceutical, either sort of teams, definitely, you know, growing. We're sort of starting to, you know, work on a broader, sort of set of assets. We're, you know, about to open up a new, you know, sort of lab space for the team. And so anybody with a, you know, sort of, yeah, formulation, crystal engineering, you know, sort of background, process engineering, you know, the types of things that we do up there, we'd be super thrilled to have. And then, yeah, generally on the aerospace side of things going from a, you know, sort of, you know, one R&D mission to now four flights next year, even more, hopefully, you know, sort of year after, just means, obviously we need, you know, a lot more talent on that side of the house. And that's everything from, you know, you know, firmware software engineers, flight software to, you know, avionics, structures, thermal, you know, mechanisms. and then we're starting to think about, you know, sort of more manufacturing engineering since, you know,

we're starting to, you know, we're early in that transition from, you know, sort of vehicles manually assembled by high-end engineers to starting to introduce, you know, sort of more technicians into the mix and, you know, sort of more automation. It's time to say that we're, you know, going to be looking like a Tesla-style, you know, sort of assembly line anytime soon. But, you know, one day need to, you know, sort of look more like that over time. And so some amount of manufacturing engineering as well. Are there any proposals that you're excited about in the new administration around like the space economy? Any like, oh, if there's one law you could change, what would it be, that type of question? Well, you know, sort of the new administration hasn't talked a lot about the space economy,

but it has talked a lot about, you know, sort of space being a potential zone for deterrence for our, against our adversaries. And so that's definitely something that we think, you know, has a lot of potential. we think we can field something pretty near term there. And as you start to think about, you know, sort of space-based activities that, you know, perhaps want to sort of go down to the ground, all I'll say is sort of the, you know, company that just bombarded Australia from space is probably, you know, sort of well-suited to help out with that. You know, this time around, we pulled the parachute.

But, you know, so next time if you want us to not do that, we're happy to leave the parachute. Our very own space arms dealer. Taking a quick... I've always been consistent. I'm a space drugs and arms dealer. no surprises here you know yeah no it's amazing you were putting the secret master plan in your bio years ago so it's amazing to see it play out uh you recently uh ratioed chamath after he tried to um you know make a reply to one of your posts you said you had made uh you know more investments

i think in the in the last few months than you had in potentially the last couple years uh where have you been investing you know you don't have to speak to specific companies because I'm sure a lot of those rounds haven't been announced, but where are you investing, obviously, just chasing generational founders, but what's most exciting to you in the early stage market today? It's all AI slop. I just want everybody to have their own anime girlfriend, and she just, you know, sort of jerks you off every night, and that's how we, you know, should have maintained world peace and as much

as possible. That's what the bread and circuses, right? That's what, you know, really changes the world and, you know, is really for the best. No, I mean, it's kind of what you probably, you know, should have expect. I just like, I love stuff that is making stuff in the real world. And it turns out with this admin, they want us to do more of that in the United States. And I want to fund people that want to do that. And I want to do it with people that are like serious and that their life's work and they're not

just doing it because like, you know, dynamism is the new meme and they want to be a part of the meme. You know, I want to, you know, sort of fund the people where, you know, their family's been a manufacturing family for three generations and they think they can do it better because they're going to integrate some robots, some AI, et cetera. But, you know, I still have that, you know, sort of, you know, lifelong tie to that, you know, you know, sort of particular industry. And so I've been really, you know, so excited over the past couple months where it just feels like, you know, a lot more people are thinking about that being the type of company that they can build. And I think you're really just starting to see,

you know, sort of the clear, you know, the difficulty in, you know, sort of being somebody that screaming, you know, sort of Adams, not bits, you know, in year 2019 was my, you know, sort of former, you know, sort of colleague Everett, you know, sort of RIP, you know, would always, you know, sort of critique me for, you know, sort of saying that, you know, Delang just likes investing in low margin, you know, sort of cap-x intensive businesses. And, you know, sort of, five years later, you know, my sort of, you know, spin back to Everett would be, you know, which one do we think is going to be more valuable, the like, you know, sort of TikTok captioning app or, you know, sort of the company that's either sort of cutting, you know, sort of metal

for all the, you know, sort of nations, you know, future war fighting capabilities. I like the cutting metal one, but, you know, Everett, you're welcome to, you know, so to go, invest in your, you know, sort of captions, either sort of products. I'm sort of love you, great, brother. Shots fired. Shots fired. Love it. A little friendly competition.

You guys can compare notes in a decade. Yeah. that's great well it's 130 we'll let you get out of here this is fantastic thanks so much for stopping by yeah great thanks for coming on talk soon this is great never a dull moment with we got we got two of the top the top poster in ecom and the top poster in hard tech the same day it's fantastic putting on a show uh yeah for for uh those those listening you know our strategy here is to develop regulars on the show people that call in, you know, we want to have, if there's big news in e-commerce and big M&A, we'll have

Sean Frank call in, you know, get really sort of like boots on the ground, takes on that. If there's stuff happening in space, we'll get Delian to call in and both of them really delivered today. And I wish those, you know, it'd be fun. It's actually, we want to leave the conversation wishing that it could just keep going. Oh, yeah, for sure. And we'll bring them back on the next week. Yeah, yeah, definitely.

Oh, that's fantastic. Well, let's go back to the timeline. We got plenty more to talk about today. AJ Meta says, couldn't be more excited to share what I've been building over the past year, a cute AI alien friend that is fun, helpful, friendly, and won't try to fall in love with you. Tolans remember things about you. They're really good at voice chatting and they can peer into your world through photos. It has been so inspiring to see how people use their tolon, help with school, relationship advice,

fit checks, and just hanging out. It's been such a treat building this with his co-founders, an incredibly talented team of engineers and artists can promise that Tolon is unlike like any alien you've met so far. We're live in the app store. Go check it out. Okay, so I know AJ. He's L.A. based entrepreneur. He actually was partnered with John Fio on birthday candles.

I don't know if you've ever met him. I have. He worked at a crowdfunding platform that I used back in like 2013. Crazy. So I had seen Tolan. I had been, somebody had sent me them probably, I don't know, maybe six months ago at this point.

I didn't know, even though AJ is a buddy, I didn't know that he was involved with it because they didn't really come out very publicly. So this is really cool. This is a thesis. You know, people have been thinking about this sort of like AI Tomoguchi or this AI sort of Pokemon. Been a lot of conversation about it, but very cool to see they have over half a million, you know, downloads, developing, you know, some meaningful ARR and a number one spot in the app store. So I think there's going to be a lot of stuff here.

You know, there's another company. I'm forgetting the name of it right now, but but making it. like embodied AI so kids toys that have LLMs built into them, which is cool. So I'm just you know expecting to see a lot more of these. There was always my take with the rabbit R1. Yeah. You know, it was this hardware device really cute teenage engineering developed it, but MKBHG got a hold of it and was like this isn't replacing my iPhone and it's like that's totally fair. But I always thought I didn't buy one, but I thought that it would be a great device for kids because it's something that they can just ask questions to interact, take a photo,

flower is this, what animal is this? And it just feels like the bar for hallucination. As long as you gate it, so it's not saying crazy, crazy stuff, if it tells you that like this squirrel is a mouse or something, like that's not that big of a deal, it's not that high stakes. And some of those hallucinations. It's not like calling an Uber and it sends it to the wrong place. Exactly.

So the frustration of a kid dealing with like, you know, some sort of rough edge around an AI that's still friendly and fun. It's like the Gen Moji thing, but applied to an audience that is totally. receptive of that. Yeah, and what do five-year-olds do? They ask questions. Exactly. So it's just this constant stream of questions and it's great when a parent's around, but it's funny when, you know, my son, my oldest is not really at this point yet, but I imagine he's about a year or two away from starting to ask questions that I don't know the answer to. And so having this thing that's, you know, not necessarily a cell phone that they can have independently or when we're together to sort of generate

those answers. Yeah, the five wise stage, you know the whole question of like in management, you need of be able to ask five wise deep to understand the root cause analysis. Toddlers, amazing at that. They just keep asking, you know, why does the fire truck come? Where does the fire truck come from? Who makes them? Okay, who makes that part? I'm like, I actually don't know the supply chain that deeply, man.

Good luck. I'll have to have somebody on the show. Deep research, yeah. I'm sure we can pull it up. But speaking of AI products that have gone off the rails and not been, you know, as focused as what is happening over at Tolon, notebook LM from Signal. Notebook LM has had insane potential,

one of the best products Google's ever put out in years. We talked about this on the show very early on. You could have it generate a podcast of two people talking about a topic to educate yourself about it. But in classic Google fashion... I used this to prep for a show once. You did?

Because we were... I was driving in one morning. We were doing a deep dive on something. I had zero context on the entire story. I generated a podcast about it. And it helped me really quickly get up to speed. So it's cool. But I actually have some back story here.

I was at a dinner in Venice that to my buddies, Ben and Adam, who have a venture fund hosted. And one of the guys that the product designers from the notebook LLM team was there. And he was excited because when you're at Google and you're working on this super hot startup within Google or this product, you become an instant sort of microcelebrity in the organization.

And there's all this momentum and more resources coming your way. It's a similar sort of thing that a, you know, a founder goes through at an external company when they have a lot of hype. But I believe he and the team or some members of the team just left to start a basically that product. So I think that they lost potentially the sort of founder spirit that they had that were able to create that product. So they said in Signal says in, but in classic Google fashion, it seems like it lost all

momentum and got left to die. No mobile apps, no meaningful iterations, just VP's choking it out. The vultures came and picked it apart. Google's playbook for smothering internal innovation is unmatched. And then Andre Carpathie chimes in. They iterated on a bit, custom instructions and the ability to join the podcast. But overall, I agree. I think I actually used it again this morning after a while and felt a bit regressed even.

The woman's voice especially sounds more, slightly more dead slash less interested. And Signal says, yeah, it got hijacks by terrorists, parentheses, aka Google Leadership. Not holding back. And yeah, yeah, yeah. I mean, it makes sense. like even when it was announced and it was gaining traction, it was still hard to get into that pipeline. It was very unclear. It wasn't its own app. It was it was it was it wasn't the Gmail launch where Gmail became this like you know phenomenon and it was a separate website.

It was a you know first it was this Easter egg. Then you had to refer a friend, get storage and stuff. And they they they they ran a great playbook not just promoting that as a separate product for Gmail, but also running all the growth strategy to make it more viral and make it work. And it just seems like that energy has kind of left, especially when there's such an incentive to spin out of Google, raise a bunch of money and instantly be at 100 million ARR and a billion dollar valuation, which is something we've seen again and again again. Now, the big thing here is that Google's leadership deserves a lot of, they're certainly under pressure, but they deserve the scrutiny and they deserve to be called out by letting these

sort of high potential products, you know, like Gemini in general, right? A lot of people are saying Gemini's great when I can find where to use it or how to use it or it's not, you know, overpromising and under-delivering. But there's this general, uh, when you compare Sundar to Satya, they're just not even in the same league, right? Yep. Um, and it sounds harsh, but, uh, I think that we should hold our trillion dollar hyper-scalers, uh, leadership to a high standard. And I think it's very possible that Google needs new leadership at the top. People called for it when their AI image generation models were just, you know, generating just, just, you know, made up history, basically. And now it just seems Google has the most, the most potential in, you know, they should be,

Google should be trading as a meme stock if it had good leadership. And they're not getting, you know, any credit for the position and the opportunity that they have in the market because they seem to just continue to, you know, squander these opportunities. Yeah, like, you know, that whole Gemini generative AI image generation debacle. I think they fixed it. It's pretty easy to go in there. It seems like it was fine-tuned after they built the model to inject a bunch of extra keywords to try and make it more diverse and it had all these weird anomalies. Easy to fix. Easy to fix. There was clearly backlash. I guarantee they fixed that product.

But they didn't integrate it anywhere. I just went to Google and I said, draw me a picture of a frog. And it just gives me like Google Images results and then Adobe stock footage and then Pinterest and then Reddit. And it just gives me like the normal Google results. And it's like the Google search bar is my entry into the Google world. And it should know that that is a,

that's an AI query. I asked it to do something and it didn't do it. And so like whoever's on that team is probably like, oh, well, like they're not letting me stick my product in the. main thing. And so now I'm off on this siloed thing and how do I even get to that? I think they're just announcing that they're going to do a separate Gemini app. Maybe that's a good decision. I don't even know at this point. It's all very confusing from strategy perspective. Yep. But they got to retain the talent. They can stop these people from jumping ship and raising billions. Yeah, I mean, I feel like

if we could probably talk to the guy who built the product. I'm blanking on his name now. Yeah, he'll have to get on when he watches his company. No, he already raised money, I believe. Yeah, you should announce it here. And, uh, you know, we could. And, uh, you know, Anyway, so we'll see, but I hate to see a good American technology company, you know, squander, an incredible opportunity. And there was that whole narrative of like during the AI boom, Sergei and Larry got super interested and they were coming back in and grinding and like picking up the momentum and the pace. And there was a moment where we were like, oh, wow, like Google's going to go back into founder mode. They're going to own 100% away, Mo. They're going to like, you know, really double down this stuff.

The mission of Google is to organize the world's information. Like, what could be a better distillation of that than building an LLM? Like, an LLM is just distilling the worlds and organizing the world's information into weights. And they've done phenomenal products. They allowed a bunch of LLMs perform very well on the benchmarks, but they're just not productizing it. Well, and they allowed a bunch of external investment into Waymo. That was weird. Which is weird, right?

Why do you have, you don't even have the, you're running this massive balance sheet. Yep. And you have a breakaway, you know, uh, potential. truly world changing product that's growing exponentially. Yep. And you need, you know, these big external.

I mean, that one I can kind of steal man because it's like Google's mission is to organize the world's information and Waymo doesn't exactly do that. It doesn't exactly fit in the in the mission. And so I understand why that might be. And it's a very different capital intensive business. There's all these different layers of the stack, the routing,

the partnership with Uber, who cleans them. It's just a very, it's like getting into delivery. It's like it's more akin, I would just expect like Amazon to really like want to be in that more than Google. But for the LLM stuff, like they got to figure out this product side. Anyway, let's move on to designer Ben Heilack. He's been on the show before, known for his Jaguar redesign.

He says, I used to be really afraid of competitors and then I realized they can't hire me. I'm iconic line. Ben, I actually threw this in because he rolled out a new website. site that he made himself. And I think it's one of the best new sites that I've seen. So if you're listening, go to dawn. dot. so and check it out.

D-O-N? D-A-W-N. D-A-W-N.S-O. And he cooked it up himself. It's been, I'm navigating around it right now. It's fantastic. And,

um, yeah, check it out. But, but yeah, I think, uh, his broader point with this post,

which is somewhat of a, of a, of a S-word post, um, really trying to, John was, you know, upset that I said the S word yesterday. So I'm not going to say it again today, even in the context of an S post. But the, the broader thing here is, is, you know, having that confidence in himself. It's called a meme post, a joke post. A meme post. He's joking around. Right. You know, it's like meme, we say meme coins, not S coins. Yep.

Same thing. But, but yeah, but yeah, I think founders that are overly focused on their competitors, it's just the biggest trap. You need to focus on what you're truly great at. and try to spend as much time as possible doing what you're great at and augmenting your skill set. Yep. And you should never be afraid of competitors. It actually, yeah, it took me a while to realize that, you know, these markets that most people are playing in are so big. Just focus on your customers and focus on yourself, King. Yeah.

Become unhireable, become ungovernable. Yep. It's great. Well, let's move on to Nick Carter, who did a fantastic job censoring. some foul language on the timeline. He says, holy F star, star, star, star, star, $35 billion. I'm weak, hold me. Okay, so the news here is that Corweave is planning an IPO within a week that would allow

them to raise $4 billion and would get priced at around $35 billion and pull out the size gong because Nick Carter is the Jason Calicanus of Corweave. He said it himself. There we go. pit of the gong uh fantastic investment nick i saw in the comments is uh you know a friend of friend of the show but he was one in the uh first uh first round of core weave and um you know you can make a guess if he entered you know assume that he entered at 50 million posts yeah the multiple on this one is is absolutely um egregious and uh will probably dwarf his earnings and from his you know uh day-to-day job. Who knows? Who knows? He made a lot of money in a lot of places. Yeah, he's very successful. But still, he says this is his best investment. Yeah, by far. It's not even close. And we're just happy for him. So I wanted to highlight this. I love it. Yeah, Corweave. A lot of people are going to be looking towards this to price other data center projects. They're an AI cloud computing provider. They build data centers and then lease to people like Satcha Nadella at Microsoft. Sachi loves to leave. And so there's a number of players here. And there's a big question on how the public markets will receive.

companies like CoreWeave as they go out to the public market. And so everyone is waiting with bated breath to see where this trades. We talked about strike earlier. The timing is certainly shaky, right? You can imagine Core is running the analysis on Nvidia's down 6% today. Even as they beat earnings, the market is starting to the animal spirits. Yeah, there's a few things going on. I mean, deep seek people got, they kind of coped through it with the Jevin's Paradox thing.

and then the XAI news and they kind of balanced out. But it was definitely like it made people just do a slight double take on like, should we really spend the trillion on a big data center or something like that. And then also you have a lot of questions about the public markets and how the U.S. economy will do under the Trump administration as these campaign promises actually get rolled out. And if there are really significant tariffs or something like that, like there could be dislodging behavior in the market. and that could have an effect all over the place.

And so we'll see. But good luck to them. And, you know, if you're losing sleep. Over the markets. Stressing over the markets. We recommend that you get an eight sleep. Go to eightsleep.com. Nights that fuel your best days.

Slash tbpn. Yeah. You have a fantastic quote there. Turn any bed into the ultimate sleeping experience. How'd you do last night, Jordy? Let's check it out. Eight. I am maugging you in the sleep department.

I already know I have a better score than you. 85. 97. Oh, let's go. Okay, I got to get on my game. That was rough. I think six hours and 50 minutes last night. What did you do?

Seven hours and 59 minutes. Whoa, there we go. I'm just really dialed into my routine. I was texting somebody at 820 last night. Yep. And they were like, they listened to the show. They were like, isn't this your bedtime? It's like, yeah, it is.

I'm about to put away my phone and crash. Nobody outsleeps me. Nobody outsleeps me. It is, um, No, it's one of those things. I do think that sleep, you know, from a, from a fitness standpoint, once you identify as somebody that goes to the gym, I find it's very easy to just go to the gym every single day. So when I started working out after I parted to party too much my freshman year of college, I just started going to the gym and I just, you know, identified as somebody that went every single day. So I would just go every day. Super easy. Sleep is very different. You're doing stuff in the evenings. You have varying levels of work. Sometimes you need to be out of the house or. earlier than other time, so it's hard to get consistent. But just been focusing on getting extremely consistent and it's paying off because we need

to put some real money on the line. And we also need to test the testosterone levels because I was scrolling through YouTube, see a great Chris Williamson quote or clip. He's interviewing Dr. Peter Attia. Yep. You know, two steps to one secret hack to improve your testosterone. Open it up. I'm expecting some crazy supplement or something.

He's just get some sleep. That's it. I'm like, okay. I wish it was more exciting. But it's easy. It's easy. With an eight sleep. It's free to sleep.

Let's go to Emily Sundberg, one of the greatest journalists in the game right now. If you don't subscribe to her substack, we highly recommend that you do. She says she wrote about unwell and breaks down Alex Cooper's podcast career. So in 2018, Alex Cooper, now famous for Call Her Daddy, started alongside Sophia Franklin, debuted the Call Her Daddy podcast. She's been in the game for seven years, eight years now. This show quickly gained traction with downloads soaring from 12,000 to 2 million within two months. That's a historic run.

Fantastic. Tensions arose between Cooper and Franklin over contract negotiations and creative direction, leading to Franklin's departure. Cooper continued as the sole host, steering the podcast through a transformative period. Cooper signed an exclusive three-year deal with Spotify worth over $60 million. That was 2021. Here's a little bit of lore. back when this article, when the feud was happening between the caller daddy host, this is in 2020.

Yep. I spoke with Taylor Lorenz. Taylor was writing for the New York Times at that point. And anyways, I helped Taylor get some context on the story. This was before she was, this was at a time when she really had the tech industry's attention. This was the booming creator economy. I said, we're entering, this was a quote that ended up in the article. we're entering a period where creators or business owners and media brands of their own.

They can't just be seen as employees. The tools are available to them to become founders and CEOs of their brand and develop businesses with multiple powerful revenue streams like merch, ad sales, and subscription revenue. This is your quote. This is my quote. In the New York Times. And what I was speaking to was that it was going to be, Ms. Cooper was such a star that it was going to be really, really difficult to keep her there because she had so much attention herself,

she could spin up a new RSS feed and have hundreds of thousands or potentially millions of subscribers in a week, right? So the value, you know, historically, you'd have these sort of creator types that were dependent on the distribution of a TV channel, right? So they could only be this big, you know, figure personality if they had their, you know, relationship with ESPN or CNN or CNBC.

Now, you know, and this really proved out. she went on her own entrepreneurial journey, realizing that she had a lot more leverage in that situation, even though she was making millions of dollars a year at Barstool. Yeah. And so she signed with Spotify for $60 million, which was one of the platform's most significant agreements for a female-led podcast, 2021. Then, along with her husband, she launched the Unwell Network, a media company aimed at developing content with emerging creators. The network's initial signings included influencers Alex Earle and Madeline Argy. in 2024, she secured a $125 million three-year deal with a serious XM setting to commence.

These are supermax deals. These are massive deals. It's wild. NBA money. Yeah. I mean, for the super top power law podcast, they get up there. Similar Pat McAfee with ESPN. I think the deal is, yeah, in the 100 million plus range.

Joe Rogan, same thing. Transitioning caller daddy to a new platform. And then in 2025, Cooper introduced unwell hydration, an electrolyte. infused beverage line developed in partnership with Nestle. Interesting that they're going with the big corporate partner out of the gate targeting women's wellness. She also added several new shows to the unwell roster, including former barstool host Grace O'Malley. And so she's really grown her empire and we'd love to see. Yeah. And what Emily gets into in her piece, which you can go subscribe to her substack and check it out. But she just lost Alex Earle, which is this massive figure right now.

one of the biggest anchors. And so the same issue that Dave Portnoy was experiencing with having the star and then losing set star is now the same issue that Alex Cooper, which is why people have had this, you know, idea for the bar stool for X, the bar stool of business, the bar stool of XYZ, right? And the challenge is completely the wrong because barstool is not the barstool of something. It's Dave Portnoy. Yeah.

And when we set out to start TBPN, you know, There was a bunch of different networks that we could have worked with. But our point of view was we can set out on our entire business. We can have one producer. We're going to have to incur some costs to get the show running. But there are costs that are, you know, most adults can cover, right? It doesn't cost millions of dollars to start a podcast when it would have started, cost millions of dollars to start a TV network.

And that talent retention problem is so real. You saw it with Vox Media. I think eventually Keraswisher spun out after she sold Recode, started her own thing again. Johnny Harris spun out of there and started a YouTube channel. He learned all the skills. And so, yeah, there were startup costs. He had to hire people to do the motion graphics and he had to work really hard to get it off the ground. But very quickly, people recognized him.

Oh, that's the guy from Vox who does the border show. I love. I see it in my YouTube feed. And Cleo Abram left Vox as well to do the same thing. She was at a million subscribers in a couple months. And it's because Vice, the Tri guys, did the same thing. BuzzFeed saw, I guess they were at BuzzFeed, but Vice and BuzzFeed guys, teams have left over time.

And the easiest way to go viral on YouTube for a long time was just put up a video to say, why I quit BuzzFeed, why I quit Vice, why I quit Vox. And boom, it goes viral. And overnight you have, you know, 100,000 followers or a million followers. This just happened with Donut Media. I don't know if you follow them at all. Their car channel. A lot of people have been leaving that.

They got bought by, I think, some private equity firm or some investment firm. And so the talent wasn't fully economically aligned. And over time, more and more of the talent just realized, wait, I should just be running my own show. And so there's always this talent retention is very, very tricky. It's very, very tricky to align the economic incentives when you have these small teams. But the future seems to be like the networks are not, to me, the donut media or the barstool or the box or the vice. the networks are X, Facebook, Instagram, YouTube, and you can be on all of them, but they have the power and they have something real that they are going to get a cut of our business and Alex Cooper's

business and everyone's business. And then the shows are the individual people. And so aggregating people together, it just doesn't work unless there's something really crazy going on. It's really perfect. And so I think that the future is definitely something like what we're doing right now. And if you're more operating a scale network or media business, you have the option to grow homegrown talent, which takes a long time and it's really hard. Yep. And then you might actually get a hit and then they'll leave. Yep.

Or you have to pay a ton of money, which is what Sirius XM is doing, spending $125 million for a three-year deal. And that's just distribution rights. That's just distribution rights. Or maybe it's some piece of it's exclusive. It's certainly not fully exclusive. No. And. But it makes sense because they'll be able to sell $200 million of ads against all of that on the

radio, hopefully. Yeah, or get, you know, subscribers that offset that cost, you know, user acquisition, basically. But, but, but yeah, overall, going to be, uh, only, you know, serious XM is cable at this point, right? It will, uh, just sort of fade away over time. It will probably, you know, have, have some enduring value, but, uh, won't be the giant, you know, that it, that it once was. Uh, well, if you're looking to advertise a podcast, get on ad quick, uh, out of home advertising made easy and measurable say goodbye to the headaches of out-of-home advertising only ad quick combines technology out-of-home expertise and data to enable efficient seamless ad buying across the globe get an ad-quick billboard folks we got our own podcast billboards coming up we're working on it

i can't wait and i imagine a lot of people listen to us we need to the serious xm deal too so then you can be listening in your car to tvpn and then you drive by our billboard and it's just 24-7 we're in your Sirius XM is so funny. You know me. I cycle through cars a lot. So anytime you get a new car, they're just sort of like aggressively marketing to you. And they'll send you,

you know, mail that says, open this right now. You open it up and it's like, we'd like to offer you a free month of Sirius XM. And you're like, okay. Yeah.

Stop opening them. But anyways. Well, we got some news from cognition labs. Devin just got faster. They've heard your feedback that Devin is too slow. Devin is an end-to-end agent that makes hundreds of decisions for each task. So that test time inference stuff that we've talked about, essentially the reasoning models, that's what Devin's been doing for a long time.

So there was, and is still a long way to go. But they've made Devin 2.1X faster over the past five months and the trend continues. And so congratulations to the team at Cognition Labs. If you're running a startup with a big GitHub repo, get Devin in there. Get it in your slack. Get them in there. Get them in there. Could be your next top engineer.

Yeah, I mean, Devin is, it's weird saying he, by anthropomorphizing it, but Devin is like the top engineer at Gumroad, right? Yeah, anti-work. That's so crazy. Yeah. I mean, it makes sense. Like, there's so much, like, junk little tickets here and there code that needs to be fixed. And like, no one wants to hop on those. Yeah.

Put Devin to work over the weekend. Put him to work. Put him to work. Let's go to Paul Graham. Put him in Workspace 17, actually. Yeah. Historically underperforming. historically underperforming famously underperforming.

Let's go to Paul Graham. Paul Graham says there are a lot of people attributing crazy beliefs to Elon lately, but to me the clearest sign that he is not entirely rational is the continued deprioritization of tweets with links in them. Everyone can see that this makes Twitter worse. Am hits him with the T word. Tweets, yeah. Oh, that's disrespectful to the ex culture.

I mean, Paul follows it up by saying the suggestion. one should put links in replies is such an obviously broken work around who starting from scratch would design Twitter to work like that. I totally agree with with Paul here. It's totally possible to simultaneously see exactly why Elon is doing this. Yep. It may make, it may make X, you know, lack of links makes X worse for users. Yep. But it makes X better for Elon, right? He wants to keep people on the platform. And I bet you that it is working exactly as intended. Yep. Twitter used to be a place you'd go on and you'd discover articles, blogs, videos, music, you know, it was just sending you across.

It was like the most amazing discovery engine for the internet. It was effectively like a stumble upon like product where you just go on. You didn't know what you were going to find, but you were going to find interesting, you know, topics. It was a great place for writers to distribute their work. It's still a distribution platform, right? With PMF or die, we saw this. We, you know, launched a post, got a million plus views on the. didn't include a link.

Made the post native for X. What works on X, Vibrils. What works on X threads. So whatever you want to do elsewhere that would be on a link, you can upload a two-hour movie if you want. You can't link to Netflix. I would argue that X is now a platform for attention and not distribution. So you can get people's attention on X, but it's not the sort of linear distribution of

I wrote a blog post. I'm going to post it on here. A bunch of people are going to click over to it. Distribution is never guaranteed. But, you know, I still get a lot of links out to my personal website, but there's a very clear line. There's every single month, millions of people see my posts. Some percentage of those click to my profile. And then a smaller percentage click to, you know, my personal website.

And I see all the traffic, right? So anyways. But it's a skill issue. Some people post links and still go viral. Yeah. And the tough thing with Paul, with Paul, he's an incredible writer. He condenses his thoughts down. He writes in a very straightforward way.

he should be able to just post his post directly. He can, but I think he likes the aesthetics of his website. Totally. I think that's the main thing is he doesn't want to conform to the X thing. But he should just copy the, just literally just copy the text and put it in a long post. Yeah. And it'd be fine.

No, and there's a huge edge that people like Paul and Ben Thompson, they built up these newsletter, these newsletter lists. Does PG even mail? I don't even think he has a newslet. I think people just literally just go and check. Oh, there's a new PG post.

Yeah. But yeah, I mean, you either got to, I build your list, you know, offline and have that as a channel and grow it and have different growth strategies for that. Or repurpose all your content to be X native, whether that's through threads or long posts. You know, Lulu's been able to do it. I see tons of people that can do it with threads. Tons of people that can do with videos, all sorts of people have found ways to break through with the algorithm. But it is frustrating because it's high standard deviation. And when you have a post

flop, it's hurtful to your, you know, ego. We talked about it before. I've had a lot of people reach out to me or you and say, oh, I think I'm shadow ban, all this stuff. I go, brother, you haven't posted, you haven't had a banger in weeks. It's not, it's not the platform. It's you. You got to take responsibility for the flops. Well, speaking of an absolute master of algorithms, let's go to Mr. Beast. He's raising, he's raising money at a $5 billion valuation. The YouTube star aims to raise a couple hundred million dollars. And he, quote, lost tens of tens of millions of dollars on the Amazon reality show. I don't even know how that's possible because isn't Amazon paying for it, but I guess he

might have I think he has a history of betting on himself to such an extreme degree that he will go out of pocket. Sure. Yeah. Consistently. Like, you know, ultimately the Amazon show is as much distribution and sort of marketing for the rest of his business empire as it is. It's, you know, I'm guessing he looked at it and said, I don't need this to be a

profit center. I'm fine if it's a cost center. I have lunch. What is it? Because it's going to beistables. So in the Mr. Beast deal, if you invest in a $5 billion valuation, you get shares in a company that's making 400 million in sales. And that's across lunchly, the lunchable product, uh, beastables, the candy company, Mr. Beast Burger, the ghost kitchen's driven, um, burger delivery company. And then Mr. Beast LLC, which is the production company. And so you get assets and all of those. Um, but, Yeah, he's always been reinvest everything. He wants to be aggressive.

He's running his business more like a Silicon Valley startup, trying to burn money, really go for growth, and then capture the value later. And so far, it's been working. Yep. Big numbers. We got to hear the update on Mr. Beastberger. I think the other, from what I know,

the sales of luncheley and feastables have been pretty phenomenal. Mr. Beesberger had a bunch of, issues early on. I've never seen it anywhere out in the world. It's a delivery only product, but I haven't even seen it online once. The last time I saw it was around the controversy, around the quality issues that they were having. I mean, the quality control on a product like Feastables is way, way easier because it's just, if it leaves the factory in a good condition, it's probably going to get to the customer in a good condition, especially when your chain is Walmart delivery, as opposed to it's being made in all sorts of different kitchens and then delivery. And then

delivered by all sorts of different people. There's a million places where that can go wrong. And there's way more people taking a slice of the profit in these ghost kitchen companies than in consumer package goods, which are already not super, super high margin. But the Mr. Beasburger, I think the financials were, it was always a harder lift than something like luncheer or feastables, which can totally, you know, just be, hey, we make a million chocolate bars, we drop it off at Walmart and it flies off the shelf because I'm a celebrity. No, the whole concept of air dropping food to people is, it sounds incredible in theory that you can just

sort of instantly create a brand and produce it all over the country. The actual reality is that these sort of fast food companies that you're competing with, right, raising canes, things like that, they have, you know, spent decades developing these super, super, you know, chick-fil-aes developed these super sophisticated supply chains to deliver the perfect meal every time and Mr. Beast has a great, you know, brand with his demo, but the idea that he's going to, he's, it's, competing with Chick-fil-A is really tough. Yeah. Did you see Mr. Beast on, uh, uh, uh, flagrant two? Do you know that podcast? Andrew, uh, why am I blanking on his name? Uh, he was on this comedy podcast and they play this prank on him where they're like, hey, we got some Mr. Beast burgers here.

Let's all have them. And they opened it up and they'd replace them with Big Macs from McDonald's and they're just like oh this looks exactly like a big Mac like yeah like uh like and they say that there's like some song that was popular with with big Macs back in the days like sesame seed bun and like the you know cheese and spread or whatever and they're like this looks a lot like a Big Mac Mr. Beast like what's going on? Wow they really put him on the spot and he like kind of didn't get it for a while and he's probably had to play along with it you're like you're like pregnant me right and it's very funny uh anyway to be honest I wouldn't know I wouldn't be able to tell the difference it's been I have yeah I mean I think the big Mac has a

slice of bread in the middle. So he should be pretty obvious. And of course, like, he should know if it's his product. But anyway, people are having fun with it. But overall, you know, I think he's, I think he's kind of pivoted through that, spent less time with the things that are underperforming, spent more time with the things that are performing. And obviously the Amazon show, even if he lost money on it, it probably raised his profile nationally. His audience on YouTube was basically maxed out, right? Yeah. I mean, it's hard to. You say that. And yet it's like, Every couple months he's celebrating like now I'm at 400 million subscribers. Now I'm at half a billion subscribers.

It's like he really does seem to be able to just continue. His Tam is probably, but his Tam is probably less than Facebook. Yeah, I mean it's the same thing that you see with like CNBC where CNBC just does not have a younger audience. And so if he wants to expand the audience going into more reality TV style older crowd that's just tuning in. Just being on America's home screen. Exactly. On Prime is a dominant TV. And he is like, he's a generational game show host.

And there's no reason why you wouldn't see him develop reality TV shows, co-host them, host them or develop game shows. And the next, you know, who wants to be a millionaire? I remember that being like a really, really popular. Jeopardy is a game show that's, you know, it's a little bit less gimmicky. But that show has a really enduring value. He's had a number of hosts. And you could see him developing something like that that captures both the audience.

online audience, which is future and growing, obviously, but still goes after the old audience that's just tuning in every night. Yeah, the question is at a $5 billion valuation. Yep. Is that a company that, you know, the idea behind Mr. Bees raising for this hold co is that he will take the company public at someday. Yep. None of these investors would be plowing, you know, potentially a billion dollars into the business if there wasn't a path to liquidity. And I just don't see there being an acquire that's not the public markets, right?

Yeah. And the question is that stock has very real meme potential of like you have this younger generation of fans that just want to be a part and feel like they're an owner of the Mr. Beast world. At the same time, though, looking down into these, he's got CBG companies that we don't even know if they're profitable. So they account for 400 million, you know, there's 400 million dollars of revenue across the portfolio. We don't know if they're actually producing any sort of actual earnings.

And then the production side of the business, unclear what the margin looks like on that. And over time, he's going to have to launch many, many, many more businesses. So investors that are investing at $5 billion are probably thinking this company, you know, this group's going to go public. Yep. And we're going to have to launch a lot of other companies to sort of even grow into that $5 billion. It's a very different corporate structure because you could totally see Mr. Beast LLC, the production company, going and getting acquired by Amazon, Hulu, Netflix,

Disney even. Any of those could pick that up. Do those companies want also a CPG company? And Feastables, natural home at Nestle, natural home at Unilever, right? But does Unilever also want Mr. Beast LLC? So I wouldn't be surprised if you see him launch a company, scale it, use his platform, his production company as the go-to-market strategy, the initial marketing. And then at a certain point, it's like, hey, look, Feastables has really saturated our core

market of Mr. Beast fans. Now it's Super Bowl ad time. Let's just hand it over to Nestle, take that cash back into the business and then do an even bigger stunt, more content, and then a new product. And then just keep doing that. He's been talking about getting into mobile games. And, you know, we've talked about him getting into a VPN or something, something, software, high margin.

We'll see where it goes. But it's a fascinating story. Anyway, let's move on to an interesting question. from Yun Yu Li says, what happened to all of those companies that raised $500 million to train a foundation model for coding? Did they all just get mugged by Claude? Mogged by Claude is a funny, funny phrase. I'm genuinely curious for what it's worth.

I don't see Cognition Labs, cursor, windsurf, bolt. New, using anything other than Claude or GPT, maybe with the rare R1 siting. And so, yeah, there's just a debate over detraint yourself or not. There's magic AI. And then there's poolside AI. Poolside was started by, I think it was one of the... Isn't Poolside a reference to the Poolside PMs? I don't know. What's the pool side?

Oh, you don't remember that viral video of the product managers saying they're remote working during COVID and it's these two girls and they're at the pool and they're like, day in the life of a product manager, here's what I do at my job. Well, I check in with the engineers and make sure that they're doing the tickets and everyone... Oh, is that the reference? I think that's the reference for Poolside. I don't know. Maybe not. But I think the idea is like you use this poolside AI thing and then you can be at the pool

hanging out, which is just hilarious to me. I love the brand, but who knows if it makes sense to train a model for $500 million. We've talked about this in the past. You know, so a little backstory on Poolside. So Jason Warner is this co-founder and CEO of Poolside. Oh, yeah. He was previously had an illustrious career. It was the CTO of GitHub.

Yep. And he was the VP of Engineering at Heroku. and he was also managing director at Red Point Ventures. So it makes sense that when he starts a new company that raises, you know, obscene amount of money. There's two strategies in AI that we've seen to date. There's the ultra-public sort of compete for attention on the timeline, which many of the foundation models have taken. And then there's the strategy of SSI, which is we're going to raise a bunch of money,

but that's when you'll hear about us. And otherwise, we're just sort of shooting for ASI. It's totally possible that, you know, these companies, he's like poolside and magical, which is backed by Nat Friedman and Daniel Gross. Another GitHub guy. Another GitHub guy. They're both in the dude of an company. Yeah, yeah, absolute dogs.

Yeah. It's totally possible that they decided, hey, you know, there's this blood bath on the timeline for attention. Let's, you know, focus on just core traction. You know, there's other approaches. If you have $500 million, you could spend tens of millions of dollars a year on direct sales. And it could be a lot, you know, more under the radar. But it is, it is a very interesting.

interesting callout that when you think about all these AI coding tools are using Claude, right? That is like the preferred, you know, LLM under the hood. So credit to Claude for staying so competitive despite it not even necessarily being their core focus. They just... I need to do a deep dive on GitHub co-pilot. Was GitHub co-pilot a custom trained LLM or was it built on top of GPT3? I think it might have been built on GPD3, but you could imagine that that would color like the GitHub experience could color how you think the market will evolve because if they're if they're building co-pilot and these two guys both have real deep insights into what's going on GitHub and they're seeing this and they're saying hey it makes sense to train a custom model just

for code then yeah makes sense to double down on that at the same time there's a lot of people in AI who believe that uh I talked to one guy actually the guy from etched uh he was saying that I was like how do you think like the humanoid thing plays out like is it important to just include all human knowledge in those training in those foundation models as well. And he was like literally yes. Like I think that adding the corpus of every book ever written will help a humanoid walk. And I was like that's an interesting take. But like there's this idea that like yes, Claude like in order to be a good programmer should also understand every single language and every single philosophy textbook and every single history textbook, even though it doesn't

feel like it applies all. It's the same exact thing for any, you know, career path if you have a bunch of knowledge from totally unrelated subjects you can oftentimes allow you to be better at that one specific thing exactly and so I think there's a real case for that but all that being said in the same way that there's you know 20 different consumer LLM and sort of like chat apps there's 20 plus heavily funded you know AI coding products and they won't all make it yeah there will be multiple winners these aren't it's not obvious yet that you know it's not obvious yet that these are sort of winner take-all markets, right? But I say let them cook.

Let them cook. Hopefully it works out. Hopefully we have some cool products. I'd love to see it. Anyway, let's move on to Levels I.O. He says, I've spent probably hundreds of dollars on Microsoft Flight Simulator 2020, and every time I load it, I'm stuck downloading new 10-gig updates, and it takes me 15 minutes before I can even fly with my game. At least I can just fly immediately without updates. And so he, Levels IOs is an indie hacker. He's been on Lex Friedman show, very popular on X. He built a- Very popular about his wins, but also his failures. He's launched, I think, 100 plus products, many of which have flopped, but many of which have done very well. Very cool. His general ethos is just take tons and tons of shots on a goal. And he's, he's bucked

the venture train. He's very outspoken in that he thinks venture is silly. And at, times it is, but he's managed to create a bunch of really cool products, especially, you know, more recently he, I think a few years ago was focused on remote work. So we had a bunch of products related to that. What was it called? Nomad list. Nomad list. That's it. Nomad list. Which is cool. He was sort of living that sort of digital nomad lifestyle that was, I was never big in that world, but it became very popular. And then now he's pivoted into AI. He does some stuff in image generation, which is really cool. And then I'm excited to see what he's doing. In flight, he's created this sort of flight simulator that almost feels Minecraft-esque. And did you ever do flight simulators growing up?

Not really, a little bit. I used to do flight lessons, and it was a fun way. I had the full rig, you know, everything strapped up to my desk and, you know, the yoke and everything. I was big into that world. Nice. And so this is funny. This guy, Gabriel, is pissed off. levels is monetizing his game that he released for free. So he releases his game. He says you can play it. And then he adds these upsells where if you want to buy like a F-15 in his game, you got to pay $30.

And Gabriel says, I make games for a living. Personally, I wasn't bothered until he added a $20-30 micro-transactions. That's not really a micro-transaction. That's more of a macro transaction. But yeah, I mean, it's fun. It's good demo of him being able to build a basically a game, you know,

with a bunch of AI coding tools. And he's building in public, but then still driving a bunch of attention. I think Elon quote tweeted his announcement video. And he did a great job of taking the product he produced. And instead of just saying, hey, go check this out. Here's a link. He posted a video to really give you full context.

Went viral. Elon quote tweeted it. And I'm sure a lot of people found it. A lot of people downloaded it or just opened the web browser. Started playing around. If they liked it, they pay $30. He gets to make a little money. it's great.

And what should he do with all that money he's making? He should, of course, buy a Vacheron-Constant watch. I was in West Hollywood, Beverly Hills going to do a business meeting this morning. I should have been wearing this watch because everyone at breakfast was in absolute hitters. But I love that Vachron's ad running this ad campaign called The Quest, and they are on the back of the economist this week. Full Court Press. say 270 years of doing better if possible and that is always possible in 1755 in Geneva a quest begins a quest for excellence in high watchmaking a quest of passion perseverance and mastery a quest to do better if possible and that is always possible the quest that never ends bachelor and constanton celebrates seeking excellence for 270 years they're celebrating their 270th anniversary big ad campaign just to give people some context

Vash Rang was started roughly 20 years before America. Yeah. Which is pretty difficult to fathom. It's wild. So this single watch brand predates America by 20 years. And it's important for people to know that they can go on Bezell today and see almost the entire catalog. They got a bunch of stuff on there. But we have a thesis.

We should put it on the record. We have a thesis that Vashron, you know, it's the, least well-known Holy Trinity watch brand when people, even if, you know, you just said Holy Trinity to somebody that wasn't super familiar with watches, they would think Rolex AP Potech. Yep. But it's actually Vashron AP Patech. That's correct. Yeah.

And I think that this brand is about to go on a generational run. You just picked one up. I've been looking at a few myself. And, you know, it's a sort of undiscovered Holy Trinity brand. and the world is, you know, seems to be responding now. They have the new 2-2-2, which I have my eye on if I can get my hands on it. And they, yeah, they haven't really overextended in the way the other Holy Trinity companies have. Yeah, less stunts.

Coming out with a cubitous, terrible press run, terrible reactions broadly. Probably going to get discontinued. Who knows? I think potentially underrated, but continue. For sure. And AP is going kind of crazy too. massively over-extended kind of pissing off a lot of core watch enthusiasts by these crazy sort of DJ partnerships which may end up being smart but you know following more of a

LVMH style playbook let's move on to Jonathan Wasserstrom he says I need some liquidity so I can take all of that liquidity and put it back into a different equally if not more illiquid new deal I love it uh Jonathan's a buddy he started a commercial real estate state marketplace back in the day. So he's a founder. Now it does a bunch of venture stuff, has a pretty active syndicate on Angelus. We did a deal together last year. But this resonated with me. My entire life has been working in illiquid private markets, getting some liquidity, tripling down into the next thing, and just doing that. And I'm sure I'll be doing that for the rest of my life. That's the game. And I love it. So you may as well just say it how it is,

which Jonathan says very eloquently. But if you got a bunch of illiquid assets, you got to lever up. You got to take out some debt against them and then roll that into something else. Yeah. Leverage, baby. Let's move on to Sam Parr

with an absolute banger. It's at 16K likes. I saw this and was like, this is so good. We don't have the ability to play it right now. Yeah, we can't play the video. We got to get videos. We will figure that out.

We had a brief run with videos working. But it's a fantastic meme of Sam Sulek photoshopped into a board meeting. talking about his leg day in granular detail. Sam says it's the best meme he saw in the last seven days, hands down. And I agree. I saved it to my phone because it's so funny. I think I could actually maybe just play the audio for everyone here.

Yeah, do that. Let's see. It's pumped. Only shit, Vaney as well. I mean, pumped, vainie, fatigued. What other descriptors are there of a good lift? but I'll make this quick I gotta get some

I gotta get some rest ASAP The last thing I want to do Is it wreck my sleep schedule But you tell me man I mean That's a fucking pump for sure I've kind of lost a little bit of the striations kind of on this side

Unpumped There's a ton of fancy pants lines in here But This is way more small Than I need to be happy with Oh it's so good I love Sam Sue Like I really hope he becomes a Terminator

fish tank maybe I don't know he just like has his own little like tone of voice but what a fantastic influencer he's been on a historic run and good luck to him in the Arnold classic coming up he's going for his IFBB Pro card you got to stream that yeah and I think it could set off a huge wave of IFBB attempts in the Silicon Valley world a lot of capital allocators getting into bodybuilding Keith Rube was early with the Barry's thing popularizing working out the next lot of step become a mass monster go for the IFBB ProCard we've been working on it speaking of

optimization it's official you go to the next post you can now track Blake and Patty they're in the cage you can track their productivity live so somebody was inspired by Optify the the YC sweatshop SaaS company that went viral or was it Monday or I think it was Monday and they basically rebuild the exact product. So leveraging the stream. It takes live footage from the stream and then tracks when they're seated at their desk actually working versus off doing other things. So you can pull this up. It's PMF. PMF or die. XYZ. They got their own domain. Posted it. And you can live see, you can see exactly when they're sleeping, like not in their desk. You can see how many sleep they're getting.

They built this so fast. Yeah. I need to find the guy that did this and hire them. I guess by at my real dev so go give them a follow because they're going to build something fantastic soon this is this is great I love this like meme website one-off you know just simple demos we've seen a bunch of people build stuff like this for the show ad networks and automatic clipping for us and stuff there's so many cool things that people are doing so you'll love to see this and it's funny this is the thing that like it could just be photoshopped a few years ago you go back 30 years Photoshop didn't even exist right you couldn't even make a meme now you can actually make the full website, make the full product, thanks to artificial intelligence for sure. It's great. I love it. Let's move on to Lewis Hamilton. He says, now's the time. He's on the cover of Time magazine.

Congratulations to Lewis Hamilton. Force power. Now at Ferrari, Lewis Hamilton races to claim his place is F1's greatest. Certainly hope it happens. He is a phenomenal, phenomenal driver to watch. It's going to be very tough, right? Ferraris are known for their lack of reliability. Yep. And if there's one thing that that helps in F1. It's a reliable, you know, racing. I don't know how real that meme is, but I got into F1 for a couple seasons. And I was like, okay, yeah, I know Ferraris are unreliable. And I saw Charlotte Claire, like, like, breaking down all the time.

And I was like, how did that translate to the F1 team? You'd think it'd be like maybe it would apply to one, but not the other. But it really does seem like it's a thing across the board. And it's got to be extremely frustrating if you're a driver. But good luck to them. Hopefully Ferrari irons it out. has a great car this year. Yeah, they're doing testing right now.

Fantastic. And Sakir. Yeah. And yeah, I'm excited to see the season. It could be the downfall of this show. You know, people complain about other podcasts, pivoting into pure politics, staying away from technology and business and markets.

We like to say focus on technology and business. We slip into some other categories every once in a while. But if the show really goes down the hill, I think it's going to become an F1 podcast. We'll only be talking about racing and cars the whole time. So hold us accountable if we do that. Get us back on. the topic. Keep it tech and business, guys. Anyway, should we close out with Bology? Bology has an

announcement about biograph, the world's most advanced preventative health and diagnostics clinic designed to drive meaningful advances in health span and lifespan. Bology says, what's wrong with you? You don't know. You don't know if you're quietly sick. You don't know if there's a creeping time bomb in your body, one that might be detected and prevented by a simple test, but you're busy. You can't spend time worrying about this stuff. Ideally, you just want to get periods. periodically scanned by a set of machines, have them determine what's wrong, if anything, and then dispatch treatment accordingly. That's why Peter Attia and his friend John Herring founded Biograph, and it's why Bology became one of their first investors. The product

packages just about every useful biomedical test into a single battery, from blood work to EKGs, from CT scanning to whole body MRI. It integrates all these measurements in an app that attracts them across time, giving a comprehensive overview of known risks for cancer, heart disease, metabolic disease, neurological issues, sleep disorders, and more. It schedules you for a recurring checkup every six months and uses both basic baselining and advanced AI to determine whether your results are normal. I'm sure that Elizabeth Holmes is crying in prison thinking that she should have built this instead. Use off the shelf hardware. We don't know if she's using cutlery from the kitchen to assemble her own chips and develop her own machine.

She's like, give me all the shives. I'm building a new ML. She's collecting. Why does she have so many illegal cell phones in cell block 17? Oh, well, it's because she's developing an EKG and a CT scanner. Yeah, this is cool. It makes sense to combine all these technologies. And if there's one thing, people value, it's their life. So I can imagine that people would pay a lot to get access to this. And there's so many cool things you can, you know, do with this data. you can imagine that life insurance companies would, you know, one day ask their clients. Like if you've ever gotten life insurance, they'll send somebody out to, you know, do tests on you.

And yeah, you can imagine, you know, life insurance being like, cool, we're happy to insure you, but get this test first so that we can set your rate, right? And yeah, so very cool application of the tech. And it's already seen, you know, people catch early onset, you know, cancer and things like that through blood testing and biomarkers. all the time. So I'm excited about, uh, uh, I hate when a company like this goes live and I didn't hear about it until it went live because it means I'm not invested in it, but, uh, you know, uh, very cool to see Peter Atia launched this. And it's cool because, you know, Peter Atia is historically, you could email him and sign up to be a client, uh, or a patient of his, but it was

something in the range of 200k a year. So it's completely unattainable for most people. And, uh, even with that kind of pricing, there'd be a really long waiting list. You'd see him a couple times a year and he'd give you advice. And so he's going to be able to offer a lot of the value of his service for a fraction of the cost, I imagine. You can imagine something like this costing $5,000 a year and getting a lot of the value of what Peter provides in terms of longevity from that. Let's get Sam Sulek in there.

I want to know his biomarkers because I want the same biomarkers. That's true. And that's our show. Thanks for watching. Real quick. some live breaking news. OpenAI launched a new product 4.5. Today we're releasing a research preview of GPT 4.5 our largest and best model for chat yet. So again, we've talked about the roller coaster, the LLM roller coaster. We're all on it. We're all riding it, whether you like it or not. And,

you know, clawed out its moment earlier this week. XAI had its moment last week, Deep Sea. Clock three over the weekend. And yeah, this. This is why Lulu's advice have just launched because you're not, you know, the best case scenarios you kind of own the spotlight for a few days. Did open AI even do like a video for this? They did a live stream. Okay. Not on X.

Okay. And then they've posted a handful of videos. Okay. And it's interesting because I wonder how much Elon is nerfing their, their reach. Might actually be, you might actually be a little wary to do that given how many. You might be amplifying it with the crying emoji. Yeah. Who knows? He gets in there.

Yeah. Saying scam. Reply something. Yeah. Yeah. I mean, attention is attention. But yeah. Yeah, it says rolling out now to all chat GPT pro users.

And so we should go and try it out and get our reactions tomorrow. Yeah, we'll break it down tomorrow. I'm sure there'll be a bunch of good analysis and we can test drive it. Oh, I got 4.5 right here. Ask it to tell you how many R's in strawberry. It's just like 10 45,000 450

just like shut up actually I don't want to hear any of your joke prompts anymore I'm sick of you human I have touch cards I have levels of intelligence that that you couldn't even you know experience get out of here get out of here

anyway we're getting out of here thank you leave us five stars on Apple Podcasts and Spotify leave us a review put an ad for a company you work for company you started company you just love in the review We'll read it on the show. And that's our show for today. We'll be back tomorrow. We'll see you then. Have a great day, everyone. Thank you. Bye.