Tesla Robotaxis, Saudi Arabia, AI in Big Tech, Art Market Update + Guests

28:11Why not? It's not that expensive. And it's not that expensive. Obviously, it doesn't work over the very long term, but it's fine as a path to market. And so the Texas Department of Public Safety, which oversees the highway, has not yet received an emergency plan from Tesla but said it's prepared for the launch. Such a plan would help emergency responders engage with cars if they fail to pull over or if they block traffic, such as other robo-taxi services have done. So it seems like the government's actually catching up to this stuff and is putting a plan in place to say, hey, we want this and people want this, but we got to figure out what to do if this car freaks out basically.

28:47I saw a video, somewhat unrelated, but I saw a video of somebody figured out how to trap a Waymo. A Waymo had - I thought you just put a cone on it and it stops. Yeah, so apparently a Waymo was in front of this guy's house. And so he went and just took tape and just put four pieces of tape around the main sensors. And so the car's like on and it's like running. But like, and I was like, I don't support abuse of any type of robot. That person's first to go when the salvation day comes. They're cooked. Yeah. The Terminator's not going to look kindly on that behavior. Anyway. Respect robots. Full self-driving could take over most of city driving on city streets.

29:30FSD launched to the public in 2022. By last year, more than 2.4 million cars were using it. That's a lot of data. Several autonomy researchers and engineers said Musk's confidence in Tesla's ability to quickly expand its autonomous vehicles to new markets fails to account for the logistics needed to operate ride-hailing fleets at scale. Yeah, whether they're driven by a human or a computer, it also glosses over some of the regional differences, such as traffic light and lane design that have slowed other companies that they try to scale. And this is the key thing we'll go into, but this is like the really, really important.

30:04This is a very, very important underlying strategy to the way Tesla is approaching this versus Waymo. So Tesla receives video clips from an array of cameras mounted around its cars and processes them at a cluster of supercomputers, which it calls Dojo. A primary goal is reducing the frequency with which a human has to intervene in the car's driving. This was called on-policy learning versus off-policy learning. And so the car's driving itself, whenever the human interacts to take the steering wheel or apply the brakes, that's sent as a negative reward to the neural network to say, hey, whatever you did there, that was wrong.

30:38The human didn't like it. And so go back and do more of the driving where the human isn't taking over. And you get a lot of data on that. You build a big transformer network and eventually you get a very strong machine learning model, essentially. So the computer systems handling the data learn differently from humans and sometimes stumble over issues that would be obvious to people. The engineer said, like, this is obvious. We've seen this with hallucinations with LLMs. The development has at times relied on feedback from Musk and from YouTubers driving the semi-autonomous modes who map out parts of the roads in San Francisco and Florida where the cars struggle.

31:12Interesting. This is citizen. Citizen journalist. AI. Yeah, I guess. Training. Tesla's sole reliance on cameras rather than laser and radar-based systems in Waymo means its autonomous approach is more prone to mistakes, critics say. Cameras only is not enough and likely will never be enough, says Seton Mersili, co-founder of the autonomous trucking and logistics companies Locomation and Atlas Robotics. Tesla's approach doesn't work, period, said Missy Cummings, an autonomous vehicle researcher at George Mason University. computer vision is simply not enough for this task and i completely disagree with that betting against elon musk it's not just that it's betting against it's betting against the way humans do driving like humans don't have lidar we just have vision and we drive fine and so i've never i've never bought this this argument this line of uh reasoning that that you that you definitely need something more advanced than just a camera because humans are able to drive 18 year olds can drive 16 year olds can drive.

32:18And yes, you need, you obviously need to create something that's, you know, human level intelligence and very, very capable and you need a huge amount of data, but there's, there's nothing like this is such a refutation of the... Director of the Mason Autonomy and Robotics Center. Yeah. And I, and I understand where she's coming from. And she was a naval pilot in the eighties. Yeah. I mean, I understand where she's coming from in the sense that like right now, lidar and radar based systems and gps based systems are seeing better results right now but to just throw out the bitter lesson and disrespect scale is all you need is a really really tall order that's been wrong many many times and so it's it's a very risky position to be in and and i don't i don't think it'll pay off well and the whole theory of the scaling laws and the ideas that you want to just get a whole bunch of data and crunch it down into a huge transformer, essentially, has proven correct many, many times.

33:19And it also puts you on a different path to scaling because if you have a system that's more hand-coded, more customized based on the particular locale, every time you go into a new city, you need a new team to spin up new records and scan. and whatnot and do all these things. Yeah, and her broad point is that laser-based LIDAR systems also make errors like computer vision is making today, but it's just that layer of redundancy that allows you to avoid mistakes that pure computer vision systems are making today. It's just a very strong statement to say, camera only will never be enough. Yeah. Because it's literally enough right now in a human.

34:06and so I just don't buy this line at all. I just think the timeline to actually scale this up is long. It's going to take a long time but the idea that you will never be able to operate a car safely with just camera is ridiculous to me anyway. Her analysis last year of Waymo, Cruise and Zoox and other testers found that computer vision systems are very brittle and likely play an outsized role in crashes in part because they miss obstacles and can hallucinate imaginary obstacles This is what I was saying. If you add that LiDAR system, you can add redundancy. And so it's just another effectively check on the pure computer vision approach.

34:47It's difficult to gauge safety of either approach because Waymo's deployments are still limited to specific cities. And Tesla's system is not yet fully autonomous. Federal regulators are examining how Tesla cars handle relatively common traffic occurrences such as sun glare, fog, and airborne dust, which is legitimate. uh musk has counted counted counter doubters by saying that humans managed to drive with only their eyes got him but but acknowledged last month that tesla's robo taxi services won't work in snow prompting some followers online to offer to help train tesla by driving more in poor conditions they're just like i will do i volunteer as tribute here musk musk also said it's been difficult to train on enough incidents to make sure the cars can handle them safely because If the Teslas don't get into a lot of crashes, there's not a lot of data there.

35:34If Elon can add the Tesla camera system to a Lamborghini Cerato, I'm happy to put 5 ,000 miles on it in harsh winter conditions. At this point, summer's approaching. Probably need to take it down to Australia. But doing a road trip through the Australian winter sounds great. Yeah, yeah. I mean, this also ignores simulation learning and transfer learning, which is obviously something Tesla's invested in a ton and has a very proven track record. There was a researcher from Nvidia at Sequoia's AI Ascent talking about transfer learning and how they're using generative AI to generate essentially video from the simulated data.

36:25So they'll create a physics simulation, basically a video game of a robot picking up an orange or something like that. That's a physics simulation. Then they'll create a ton of variation in the different physics properties to try and inject a lot of noise into the system, basically. But then they'll use generative AI to convert that into video feeds that then they can feed back into the system to train on. And they've seen a lot of success with that. And so I'm skeptical that this is that big of an issue, but also like you could just put lidar on the cyber cab like it hasn't been built yet and if the price comes down you throw that on and then and then all of a sudden yeah tesla there's companies that yeah tesla could acquire that would immediately give them a massive edge there so yeah anyways uh let's move on okay uh well there's some other uh commentary around this piece uh With the RoboTaxi network, your Tesla will be able to earn money while you're not using it, essentially paying for itself.

37:26It'll go to work just like you, says Tesla. Very bullish for buying a Tesla. It's wild that big public company can tweet like this. Don't use periods anymore. Oh, yeah. I saw. Yeah, it's an interesting evolution. But periods are over. Done. And then Dan O'Dowd is another account on here, pouring some cold water on the RoboTaxi launch. You can almost hear Elon's sigh of relief. He now has his excuse for canceling Tesla's June RoboTaxi launch. It won't be long until Elon claims Tesla was totally ready, but the pesky regulator stopped him. Mighty convenient. Do you think Elon ordered NHTSA to send this letter?

38:11And so the NHTSA said - Get him a tinfoil hat. Get him a tinfoil hat. And then Cole Grind says, people say you should sell Tesla because it's hard to build a ride sharing system like Uber and Lyft. But little do they know, Tesla already has over 20 million people using their app. Once the RoboTaxi launches, they'll be able to use the existing Tesla app for their ride sharing service. This will allow Tesla to gain new customers through app downloads and cross selling of other Tesla products. Once those members come online, I would not be selling here, I would be buying. He's given some financial advice, folks, but we aren't.

38:46And then Rami, who's clearly a Tesla fan, his handle is Tesla Explored, says, please send me a video of any other company that can drive me in New York, in the rain, door to door, without any interventions. I didn't need to be in the car. I could have easily been in the backseat of my own Tesla robotaxi. And the result is that it took me to lunch autonomously. So he's a fan. The level of conviction from some of these Tesla fans is incredible. It's similar to the level of It's even beyond our little corner of tech. It's an entirely new level. But we see this with other companies. We see this with Ramp fans.

39:24They go crazy. There's all those YouTubers. People saying that they would defend Ramp their life. Exactly. It happens a lot. They know that time is money and they want to say both. They both. They create whole YouTube channels all around just easy-to-use corporate cards, bill payments, accounting, a whole lot more. all in one place. They're just obsessed. It's crazy. Seeing people get time is money saved both. Lots of tattoos. Lots of ramp tattoos out there for sure. Some people start whole live shows, whole podcast just to promote ramp. In service. In service of ramp. Yeah, they call them like, you know, there's like the Teslarati.

39:58They're like the Ramparati. The Ramparati. We're like the Ramparati. Yeah. The Illuminati. Yeah. It's great. Well, if you're looking to go long Tesla, short Tesla, whatever your decision is, go over to public.com, investing for those who take it seriously. They got multi-asset investing, industry-leading deals. They're trusted by millions. Thank you for the soundboard, Jordy. Thank you, Republic. Should we talk about the giants of Silicon Valley, how they're having a midlife crisis over AI? This was an interesting post in the Wall Street Journal because it kind of all stemmed from an interview with Sarah Guo, who's been on the show.

40:32And so she did a podcast with the Wall Journal, bold names, venture capitalist Sarah Guo's surprising bet on unsexy AI. And they did a nice little write-up here. So they say, one minute you're upending established industries as the younger disruptor. The next, you're staring into the abyss, eating glasses, as Elon Musk likes to say, watching the disruption at your door. Most, if not all of the Magnificent Seven are in that position, weirdly trying at the same time to figure out the threat of artificial intelligence to their kingdoms. That dynamic has been on display for the past few weeks. Alphabet stock dropped more than 7 % after a senior Apple executive disclosed that Google search traffic on its devices using Safari fell for the first time in 20 years.

41:17They've had Safari around for 20 years as a Google partner? That's crazy. It's so long. Google clarified it continues to see overall search growth, even from Apple devices. Putting Apple in the truth zone. Don't knock our stock down, Apple. We're still growing. For his part, Chief Executive Tim Cook is trying to buy time for his company, pushing investors during his latest earnings call to be patient with the iPhone maker's delays around AI features. Then there's Facebook. Mark Zuckerberg is attempting to paint a bright future for his ad dollar juggernaut as something of an AI buddy for the lonely.

41:54That was kind of taken out of context from a Dwarkech interview. Did you see this? Dwarkech said something and they attributed it to Zuck. It's like every time. It's so bad. Anyway. You're lonely. How about an AI buddy? Even Musk seems to be sweating things as he returns from his Doge sojourn to Tesla, seeking to counter a slide in the electric car maker's stock price with the promises of deploying driverless cars. We're not on the edge of death. Not even close, Musk told analysts recently. We love to hear it, and we saw it in the stock. Over a trillion dollars now. Not bad. His protest sounded like that Monty Python and the Holy Grail character about to be thrown on a pile of corpses.

42:30I'm not dead. I feel happy. To be fair, none of these giants are dead yet, and they have lots of reasons to feel happy. They are wildly profitable pillars of corporate America, which we love to hear from corporate America. And together, they represent around$7 trillion in market value. You have the crossroads they all stand at collectively and how they react individually look like ready-made case studies for 21st century update to the classic business school book, The Innovator's Dilemma. On the Tesla note, people that were shorting Tesla over the last couple of months, sales of the cars have been abysmal.

43:08Literally every market is down like double digits it's like so bad just getting crushed by every other ev manufacturer they're having to sell cars clearly no black pilling no black pilling no i'm not i'm not i'm no i'm just i i just have to the the this is why it never be a short seller because you could have looked at all the data for tesla and said this where is the bottom here i mean it's like where like why is tesla you know and it's like elon musk don't betting against never bet again it's okay not to invest with him but betting against him that is uh it's rough dangerous and again it's like it's like are these this is what we keep going back to with like the trump tariff things and stuff it's like are these things stuff he can roll back yeah like okay so what what what's going wrong with tesla in this theoretical bad scenario well like elon's not spending enough time there right he's spending too much time at doge well can he quit his job at doge and go back to be ceo like yes okay and is that what he did yes and that's worth about 350 billion dollars it's like oh like the lidar isn't working like could he add lidar to his cars like yeah like could he map every city yes could he you know there's like all these things that people are like oh he'll never he's not doing this impossible oh he's like he's over he's he politically he's too right wing like well could he come out and endorse gavin newsom like yeah could he live out he could yeah like anything's never too late anything's possible it's never too late uh well let's talk about the art market i wanted to have my dear friend dylan abriscato on potentially the most important topic to talk about the most important story of today and certainly uh potentially the week yes i think so it's it's shaking up silicon valley there's a lot of uh there's a lot of venture capitalists a lot of exited founders, post-exit founders, and a lot of technology journalists who have generational wealth who are at Christie's right now bidding on modern art.

45:13And so they'll want an update. So we'll bring in Dylan to the studio. Welcome to the stream. How you doing? Thank you for joining, Dylan. Dylan, it's great to have you on the show finally. Long overdue. You've given us a lot of great pointers. uh but uh anyways i wanted we were we were going to cover this story so thank you for being uh able to hop on on short notice do you want to give a quick intro and then we'll dive into the story just so people uh have some context yeah sure uh i'm dylan founder of crypto the game which is an interactive uh crypto survival game thank uh hq trivia where i worked for the rise and fall meets survivor or squid game or hunger games we've had a couple viral seasons um and before that i was uh jordy's head of marketing over at party running capital so we go way back um and way back to be here way back um awesome why don't you give some context on your your uh background uh as an amateur uh but successful art collector as well yeah yeah i've kind of um always been an obsessive collector of things you know growing up i collected beanie babies pokemon cards baseball cards sports memorabilia, always collecting.

46:29Fast forward to, I don't know, probably like 2017, 2018. My wife and I moved into our first apartment together, had very sterile white walls and decided that I guess like collecting art would be a good way to kind of scratch that itch of collecting and be something that we could do together. So kind of just started as something we could do together. I kind of browsed Instagram. Back then, there was like a much more intuitive kind of like explore tab. You resurfaced kind of like art if you followed artists. And pretty much just cold DM this artist also coincidentally named Jordi, Jordi Kerwick, whose art immediately spoke to me, sent him a DM.

47:13He didn't have any gallery representation. He made, you know, buying a piece and commissioning a painting as easy as like texting with a friend. you know I think the art market historically has been so hard to get into for young collectors and there was like zero friction in the way that we went about it I think we got a little lucky in the sense that he eventually like blew up got kind of like gallery representation and tier one gallery representation he was breaking records you know in similar auctions to what we're about to discuss and kind of because of that I don't know I almost like made a name for myself in this like Instagram art community.

47:51Um, and, uh, yeah, I've kind of just like catapulted into, into the art collecting scene. How's the P and L though on paper, I'm sure the returns are fantastic. Have you ever, do you ever sell or do you just buy? I don't know. I just buy. Um, I don't know. I, you know, the, uh, everyone always asks kind of because of that, like what to buy and how to go about it. And my advice is always just to buy what you love and buy what speaks to you and it kind of sounds like you know pretty simple advice but anytime i've looked at a piece or like had a recommendation from a friend or a fellow collector on a piece that was going to be a great investment or was going to do really well has never worked out and i've always kind of regretted buying and having on my wall and kind of like looking at it um versus the stuff that we do buy and collect and kind of like do it for the right reasons and have always been kind of like rewarded on paper basically for for those early bets yeah great well let's talk about this story and then I'd love to get your update on kind of the broader art market.

48:48So we'll read through it. So the journal covered it this morning, but the Mondrian sells for$47.6 million. Christie's modern art sales kickoff, a week-long series of art auctions that will gauge the art market's strengths amid broader economic uncertainty. People are wondering, will art values soar or flop or, you know, in general, what the liquidity will be like? So this was at the kind of low end of the estimate. The 1922 work composition with large red, plain, bluish gray, yellow, black, and blue was expected by Christie's to sell for between 40 million and 50 million, but only one bidder emerged, and it sold quickly for$41 million, or 47.6, with the auction house's fees.

49:33The work still ranks as the third highest price ever paid for the artist, but the record remains 51 million, paid three years ago for Mondrian's composition number two, Both examples feature a large red square, which the auction house hoped would help because collectors tend to pay a premium for Mondrian's spare abstracts that contained cherry hues, said Alex Rotter, Christie's president. Red always helps, simple as it is, because it's a color that people like, said Rotter. There you go. That's what Dylan's talking about. Never buy brown. Never buy brown. It doesn't sell. it's also particularly small it's only about like two feet by two feet or something and so i saw on x uh born a sore a an nft profile picture says this tiny mondrian just sold for 47 and a half million dollars last night without anyone batting a batting an eye and yet people think triple a natively digital grails are overpriced at two to three million much much higher kind of saying hey hey, maybe we should pump the market of the bored apes.

50:42Yeah. Let's get bored apes in the conversation. Yeah, let's get them back in the conversation. Anyways, were you surprised at how this got priced, Dylan? Were you following it at all? Or what was your take? Yeah, I mean, there's a lot of uncertainty in the overall economy, right? So, I mean, you would expect that art kind of oscillates. But, yeah, I mean, do you follow this? Or how do you even track the overall art market these days? Yeah, I mean, it's exactly it is so similar to, you know, venture investing. I know you kind of joked about it at the top, but especially when you look at emerging markets and if you buy from their studio in the same way that I did with Jordi Kerwick, it's basically getting into a friend and family round, at least with kind of like living artists and artists that still have gallery representation.

51:25If you do, if they do basically sign with an up and coming gallery, that's the equivalent of a seed round. And then by the time they do get that kind of like tier one Gagosian level representation is the equivalent to an Andreessen or, you know, tier one fund, like backing them in like a growth stage round. But yeah, I think you referenced that kind of like Mondrian record that was set three years ago. there were emerging artists, everyone, basically like at the time of Zerk, like everything was going crazy. And we saw that both in, you know, basically across every single market, art notwithstanding.

52:03So I think the way that I personally judge it is basically just off of the information that I receive. In 2021, 2022, I was reaching out to galleries kind of cold and being told that there were wait lists and the deal flow is basically like very, very hard to come by. Fast forward to, you know, a few years from then, you know, galleries were kind of like reaching out to me, you know, hey, I know you were initially interested in this artist a few years ago. Here's all this work that's available. I'm willing to offer discounts. And that immediately signals to me that, you know, times aren't good for galleries.

52:37But I do think that the best way of predicting like how these auctions are going to go are based on basically how New York Art Week goes and that was this past week which was kind of headlined by freeze but they also had to take that as well as nada which is kind of like the more emerging art fair and from what i can tell from everyone that i've spoken to that was kind of like on the ground it was a much more optimistic new york art week than years past um gagosian sold a three million dollar sculpture on the floor um they sold that you know tons of booths that were sold out kind of like on day one I love a$3 million sculpture selling on day one.

53:19That's great news. Yeah. So I think like that usually sets the tone for how these kind of like auctions go. There's been a ton of buzz around the Bass House. I'm not sure if you guys are familiar with the Bass House, but it was a commissioned We both are, but can you explain it for the fans who might not be? Sure. Yeah. It's an architecturally significant home in Texas that these two kind of like well-known collectors commissioned a very famous architect to design this mid-century modern house in Texas, specifically designed to house their collection of Rothko's and Warhol's and, you know, pretty much every big name.

53:58And they are auctioning off the collection, which everyone in the art market is very surprised by because you would think, you know, you commission this home to specifically house this collection, it should eventually become a museum. And I think that kind of like bodes well for what they think and what a lot of people think is is going to be a lot of liquidity kind of like heading into these auctions um having said all of that i would say if we had this conversation like a week or two ago i think it would be a lot more uh pessimistic or concerned but given the very like recent kind of like positive news and kind of like market turnaround um i think there's some buzz kind of heading into heading into tonight yeah i have a follow-up question um i mean the size seems significant i feel like there's probably a sweet spot for the scale of a particular you mentioned that uh sculpture uh we're recording not too far from the brode you walk in there's this uh massive jeff coons balloon animal uh yeah i i i walked in when i was in new york once i walked into jordan wolfson uh i don't even know if it's a sculpture but it's like a performance art piece i don't know if you're familiar with jordan wolfson but it's like this crazy interactive like machine that has motors and stuff and at a certain scale it becomes like how do you even collect that unless you have like a museum to put it in uh that's who's buying it is is the institution it has to be like i would imagine you know the three million dollar sculpture that gagosian sold that i mentioned was a kuhn's and i you know the road bought their kuhn sculpture from gagosian and i would venture to guess that the bitter who or the person that bought this three million dollar sculpture was an institution looking to build out their collection for the museum yeah is there a different i mean yeah i'm curious not to not to switch gears too much but i'm curious is christie's even slotting in nfts this week at all or are they so over it that they're not even because it was a big thing like three years ago yeah i mean so the bit you got to give christie's people yeah you have to give them some credit for moving quickly and getting nfts into these auctions early but i'm curious if they're doing that now so this week is kind of like the hero week of every year and there's uh there's no way that in like a you know 21st century like masterpiece auction you'll see an nft but i do think they are still accepting crypto and selling nfts uh all the main major auction houses kind of have daily auctions that they're running um you know you probably won't see a people in something as kind of like highly publicized as this week's auctions.

56:31But I definitely think they're still selling digital art for sure. Yeah. Yeah. There's one NFT that I think belongs in the conversation. Have you heard of this Moxie Marlin Spike NFT? So he's the founder of Signal. And during the NFT boom, he wrote a very technical deep dive on kind of some of the flaws with the way NFTs were structured in that he basically designed an NFT that displayed differently on OpenSea than in a phantom wallet, than in an Ethereum wallet. And he was making this commentary around the importance of platforms and the role that a firm like OpenSea plays in the art market and how the code is not purely on chain because it can be represented differently in different places.

57:22And it was kind of a top signal at the time. He wrote it when NFTs were super hot. And obviously, he came to it from a very engineering mindset perspective. And I think that even though he never sold it, and it doesn't have any value, and it never minted in the traditional sense, it is an NFT. And it has this historical provenance and the story around it that I think, as we look back on the NFT boom, we could remember that particular story as an interesting commentary on the entire industry that could actually drive the value up so i was considering buying it from him for a while but i didn't i didn't have the guts to throw out an actual price yeah yeah well i do think similar to how there's paintings and there's sculptures and there's basically all different forms of you know there's mixed media art like i do think digital art definitely has staying power um probably just like in the more experimental uh guys that you were talking about.

58:17And then just to quickly touch on like size as being a blocker. I think once you're talking about these like, you know, eight, nine figure pieces, it's almost always someone that can house or store either in their in their museum or in their home, like a massive piece. But just for like the everyday collector or someone that might be listening to this, like, personally speaking, like I have gigantic canvases that are still rolled up that I have absolutely zero wall space for but um that is like an addiction and a problem and i do think that um you know buy for the size that feels right to your home and something that you actually want to hang well we're moving into a new studio very shortly and uh we're certainly going to have some wall space so we'll only charge you a small storage for you i'll pump my bags as a yeah family family and friends uh anyways great to have you on uh yeah come back on again soon Thanks so much.

59:09We'll talk more. And let us know how things shake out this week in the art market. Oh, great. Later. Talk to you later. I was just pulling an episode OpenSea. Let's bring in Jared because he just has five minutes. Okay. And he's all the way from Saudi Arabia. Jared Madfoss, folks. Reporting live. Reporting live. From. From Riyadh. Saudi Arabia. He's there on the ground. He better be in the traditional. In the traditional. Oh, we were expecting a suit. we were expecting more pageantry yeah no no sobe from me today boys but uh i am calling in from a non-disclosed location in saudi can you hear me yeah yeah you sound great you sound great loud and clear give us the breakdown why are you there what's going on what's the mood on the ground yeah so uh i'm here mostly just to see friends okay um but uh but beyond that there was an event today uh where there was a u.s delegation president trump was there uh the crown prince Mohammed bin Salman was obviously there.

1:00:07And a slew of top tech leaders from the U.S. also came. So we had Jensen Wong, Elon Musk, et cetera, et cetera. You know, all the big names you can look them up online. Yep. Can we hear? Yeah, I'm not this. In terms of the big headlines, look, you guys can look them up. The big ones are 600 million in new committed capital. Billion. Sorry, 600 billion in new committed capital. But I don't want to talk about that at all. What is the most interesting thing from today is the fact that sanctions on Syria have finally been lifted. Now, I know between the three of us, every time we get together, it's like, when are we finally going to be able to enter the Syrian market?

1:00:45Well, boys, that day is today. That's today. All right. So, yeah. No, it's today. Yeah, this is the day everybody's been waiting for. We should start distributing the show there, for sure. Yeah. No, 100%. A lot of people don't realize, look, Syria, after the civil war started in 2011, our companies have just not really been able to serve customers there. So the Syrians don't really know the joy of U.S. big tech yet. And, you know, they missed the whole Web 2.0 mobile development. So most of the groups that are servicing them are from the Chinese sphere versus the American sphere. and I am beyond excited for the new crop of talented founders that are going to descend on the Syrian country to spread them the joys of American big tech.

1:01:27So, you know, if anyone is watching and is going to be launching any number of companies in Syria, my DMs are open. I could not be more excited. Very cool. Wow. 23 million people in Syria. Go get them. Go get them. Future customers. 23 million people. They've had double-digit inflation for like the past decade, which means that they are a prime market for stablecoin adoption. So every stablecoin neobank needs to be descending into Syria and seeing how quickly they can dollarize the entire country. If anybody is using the Syrian pound in five years instead of U.S. dollars on blockchains, we have just massively screwed up as an industry.

1:01:59So please request for startups. We are looking to fund. That's very cool. What about artificial intelligence? NVIDIA launched some stuff. Have you been hearing any rumblings about AI stuff going on? yeah look the fact that jensen was here today on the main stage shows that uh we are likely softening some of the restrictions on being able to export gpus to the kingdom and other parts of the middle east um i might have the numbers off but i think there was an announcement that like a purchase order of 18 000 gpus has just been approved for the kingdom um if you're listening and that sounds small not enough to build you know a hyperscaler or a massive cluster you're correct, but this is clearly the first of many.

1:02:39The entire region is rapidly searching to turn their, you know, petrol-based capital into, you know, token-based capital in the future. So expect continued investment from the region trying to diversify out of their kind of one main industry into the future via AI. Just for extra context, so Humane, the AI subsidiary of the PIF, is going to be purchasing hundreds of thousands of advanced GPUs over the next five years. So starting with, I think, the number that you said, 18 ,000 and scaling. Yeah. And the first plant is, I think, half a gigawatt, right? Something like that. What's your take on, why do you think that we started calling data centers AI factories?

1:03:23Do you have a take there? Is it just sound cooler? Yeah. It lowers your cost of capital. Yeah. And that's the most important thing that anybody can be focused on at any given point. Yeah, what is the word on energy? Obviously, with AI training, it makes so much sense to put big data centers there. But at the same time, they're trying to get away from the, hey, we have a whole ton of free energy because we have so much oil. There's also a lot of sunlight, so maybe there's solar in the future. What are the hot energy sectors that people are looking at? Yeah, so a lot of people are, with regard to energy, talking about a potential nuclear deal.

1:04:04While there is an abundance of oil in the kingdom, burning oil to power plants is not necessarily the most efficient. use of oil. You know, I don't have the numbers off the top of my head, but if you look at the amount of oil that is burnt in order to air condition all the buildings in the kingdom, it's like an exorbitant expense. So some Gulf states are slightly more advantaged in that they have larger deposits of natural gas versus crude or shale, which is just, you know, it's more efficient in terms of converting into energy. And with regard to solar, solar in the desert has not really taken off as much because my understanding is there's a lot of wear and tear that comes from the sand and that you need to build a series of protections and the elements are not necessarily kind and you know 130 degree fahrenheit weather as sand rapidly batters them so we'll see what happens but yeah obviously um there's been a bunch of companies that are super interesting like crusoe i might be mispronouncing that they're doing things like capturing excess flare using that to generate energy at first to do crypto mining and now to run data centers so expect there to be a lot more, as Jordy put it, AI factories coming out of the region.

1:05:12And there's a massive focus on renewables and alternative sources as well to power these things because people would rather sell their oil versus use them to spin a water turbine. Yeah, that makes sense. Do you think you or Jensen go to more conferences? I really don't go to that many conferences, guys. I'm about to hit my quota for the year and then I will retreat to a dark room instead of my laptop for the next nine months. yeah you do it right though you do it right you go to you go to just the right ones uh we've been just like not even not even intentionally but paying attention to jensen's travel schedule and uh he's at a new conference pretty much every single day but um what's the what's the rest of the mood around crypto broadly is there a push for saudi arabia to build a position in bitcoin kind of a bitcoin national reserve from their public investment fund are they into ethereum Solana?

1:06:04Is there even a thesis from the sovereign wealth fund around crypto? Or is it kind of just general hand-waving around, hey, we want crypto to be friendly generally? Yeah. So the holdings of PIF are public or PIF, the public investment fund of the kingdom. As of now, there's no crypto holdings in that, except via funds that they're invested in downstream. In terms of if it is going to emerge as a crypto hub, that is still yet to be seen. Crypto is largely legal in Saudi. What is interesting, though, is when you look at countries that have been super friendly to crypto in the region, mainly the UAE, right?

1:06:44There's a lot in Abu Dhabi. There's quite a bit in Dubai. The customers that they tend to cater are not born Emiratis, right? The number of Emirati citizens that necessarily will make use of crypto products in terms of giving them access to banking is fairly low, right? These are nations that are mostly staffed by emigrants, expats that are choosing to live there. But when you look at Saudi, even though crypto is legal here, there are some studies that show that in terms of Gulf region, it has the highest penetration from an individual user adoption space, just how many individuals are holding the coins.

1:07:18So I think that as the regulatory framework fleshes out a little bit And as, you know, more and more voices inside of the Ministry of Finance and the central bank here grow louder, we should expect to see a softening of the rules. And then the end state of that is probably a pretty robust environment, specifically in the kingdom. But remember, the people that can benefit from crypto the most are regions that have, you know, high levels of inflation, low levels of banking, highly mobile first in terms of, you know, payment infrastructure and things like that. And that doesn't necessarily describe all the Gulf states the best.

1:07:56So, you know, jury's still out on how much and how quickly penetration will come outside of just like investment speculation, which, of course, you hear rumors about tons of family offices here building massive positions. But it's just very hard to validate that. Yeah, with the new$300 billion,$600 billion investment, is that one of those things where the investment was already going to happen and they're tagging it for this press release? Or what is the structure? Is there any clarity on the structure of that deal? Is that just investment dollars flowing into American investment funds or specific buying of American goods?

1:08:27Do you have any clarity on what we're going to expect here? I haven't really broken it down yet, but I think the right way to think about it is it's going to be a lot of infrastructure investments and a lot of purchasing of American goods. Sure. Right. So when you're already talking about the hundreds of thousands of GPUs that are going to be acquired, that's, you know, the pool of capital that that's that where that's going to be coming from. But no, I mean, they specified they want to invest in energy infrastructure. They want to invest in data centers. They want to invest in ports. Yeah. So, I mean, it could even be like more oil rigs in the short term as they, you know, they're moving away from oil, but, you know, they're obviously still producing a lot of it.

1:09:02And American companies have kind of a dominant position in the, you know, oil and gas infrastructure market. So they might be buying for that. totally so we'll see also part of that is a massive defense spending bill or a defense spending agreement i think it was 143 billion that they committed and um you know just socially talking with some of the people from the ministry of defense that you ran into at the conference um they were like oh my god have you heard of this company called seronic have you heard of this company called andriel they're really gung-ho on some of the non-prime upstarts coming out of the states and um yeah i had like three different guys from the Ministry of Defense.

1:09:40Like, hey, if you meet any founders building interesting new defense companies, refer them to us. We want to potentially pilot them. We want to potentially invest in them if we want to make a purchase contract. So yeah, it's an exciting time for sure. I would like to see somebody with a Daytona SP3 using it as a launch pad for a roadrunner. So like you're driving the Daytona, launching the roadrunner off the back. I mean, in terms of like early stage founders, like how early is too early to start engaging with someone like as senior as like a ministry of defense official i feel like you need to have a product that can be demoed but when we talked to the secretary of the army his message was very much like we want to hear from startups even at the early stage we have a proving ground you can come test here we want to work with silicon valley that seems like it's certainly a vibe shift in america but it sounds like that's happening over in saudi arabia as well absolutely i mean i i mainly do not focus on defense so i don't want to say anything too out of turn.

1:10:35But I think post the Ukraine-Russia conflict, people saw how important it is to have the most up-to-date equipment for the battlefield. So in the same way that a lot of companies immediately went over to Ukraine and Russia, immediately went over to Israel following October 7th, there's probably an increased appetite to try all new things and see where it goes. And if you have a prototype, it's probably better than not having a prototype. But if you have good contacts, it probably is a good idea to get in touch earlier rather than later. Yeah, I think oftentimes a conversation start, but no purchase activity can actually happen until the U.S.

1:11:05has actually basically approved it and said, this works, we'll buy it. Now you can export it to allies. But anyways, we don't need to get into ITAR stuff. What's the mood around the trade war generally? It feels like we're almost in like post-trade war discourse. It's a distant memory at this point after Sunday and Monday. But are people still thinking about tariffs or has the conversation just completely moved on? Yeah, the day opened with tariff conversations. So Scott Bessent started the day and it was very calming. You know, he got up, he talked about the negotiations with his Chinese counterparties.

1:11:44He made some very interesting points around how while President Trump and Xi had a very strong relationship, some of the people more junior in the CCP perhaps did not share his warm feelings towards our president. So he mentioned that a lot of the talks in Switzerland were about showcasing and building favor with some of the second and tertiary level folks within the Chinese government. And came out and was very much, we don't want to trigger an embargo on Chinese goods. We don't want to trigger a double embargo. So the tariff rates came down and talked about a number of other countries he's going to negotiate with.

1:12:24And it was very much a non-escalatory tone. And I don't think it's a surprise at all that the U.S. delegation chose to start the day out that way. So no, tariffs weren't really talked about after that because Besson kind of kept everybody at ease. We'll see how much that remains. And if we continue to lay off and have more 90-day reprieves and more agreements negotiated, or if tomorrow things ramp back up again. But for now, I think people were pretty calm. Very cool. Thank you so much for taking the time. I know it's late there very clearly. It's dark outside. Thank you for just throwing on the suit just to pop on.

1:12:55Yeah, just for us. Thanks for starting on this. Every time. Every time. Great to see you, Jared. We'll talk to you soon. See ya. Let me talk about Linear. Linear is a purpose-built tool for planning and building products. Let me get some soundboard, Jordy. Meet the system for modern software development, streamline issues, projects, and product roadmaps. It is the standard for product development. Yeah. It is the backbone of the global economy. Of the global economy. Of the global economy. Many people are calling, just considering that big tech is, you know. This really goes on for a long time.

1:13:25It's still going on. Well, we got Sam D 'Amico from Impulse Labs coming in the studios. He's talking about stoves, but also trade tariffs, manufacturing. I want to ask him about what it takes to build a BYD in America. And I want to know the history of BYD, what he's learning from that company. And I also want to get him to sign up for a numeral because I want him to put his sales tax on autopilot. I want him to spend less than five minutes per month on sales tax compliance. But Sam, welcome to the studio. How are you doing, Sam? um hopefully welcome welcome welcome sam it's great to have you good to have you how'd you uh how'd you sleep last night better than than you have for the last few weeks yeah it was it's pretty good um i think i got the notification at like midnight on uh on like was it sunday night or something like that and i was like oh sweet i didn't expect that but i did and that's why i the panic um but they and so so for clarification basically we um i guess ashley van scooped this that we do a decent amount of our manufacturing in china yeah um at impulse which is the highest performance stove you can buy and you can go check out his latest video um on core memory uh um the core memory youtube channel if you want to learn more but the fun part about that is um we probably spent a couple weeks being like hey do we have to pay this 145 tariff on a bunch of the stuff we build.

1:14:50My immediate calculation was this will stop trade between China and the United States and a bunch of businesses will not like that. And feedback will land from everyone from Brian Peterson to someone that makes Bratz dolls to, I don't know. Yeah. Everything. So, I mean, it's only a 90 day pause. Are you trying to get as much stuff in the country right now? Are you scrambling to import a year supply? Or are you thinking that this is the end of the trade war, and we're going to be more or less smooth sailing from here on out? Yeah, I mean, I as much as I would, so one big thing with hardware companies is like a year supply is like a gazillion dollars.

1:15:31So you got to be, you got to be, you got to be a little strategic about what to grab men. We made sure that our suppliers are multinational and have manufacturing sites in Southeast Asia and other places, including places that are likely scoped to have deals of, you know, coming together in terms of, in terms of like a favorable treatment on the tariff side, Vietnam, other places in Southeast Asia, Mexico, et cetera. And so basically the, the, the idea is just like, make sure that like all the customers that are already paid, we already told them, Hey, we're going to cover this crazy 145 % tariff back in the day or a couple of weeks ago, we sent them a mess.

1:16:07Then, you know, like that, obviously it's a little easier for us to do that now. I'm still, still holding, still holding to that. um but uh but yeah the i think the big thing is like now at least i've got some breathing room to think conscientiously about what the next steps are for manufacturing of not just our stuff but you know technologies that we're going to be powering other devices with as well i mean i know you're you know i know you're a relatively new company relatively small um but did you actually have to pay any tariffs or were you able to kind of like do a little bit before a little you did okay so break oh yeah i think we paid the 145 on a couple things actually a couple things it's a good we we we the customer the bulk of the customer orders that that are coming in are not subject to that but like yeah we got um we we've started mass production already um and we're a bunch of kind of the beginning of that ends up being kind of samples um and like vip influencer units and other fun things like that we've definitely paid it on a handful of those um The balance of even that tranche of units is landing pretty soon.

1:17:14So we mostly avoided all of this. We're kind of lucky that this was an April thing, not a May thing. Nice. To be blunt. But yeah, there was a lot of fun on that side where I think if our battery pack build hadn't gotten delayed like three weeks, there was actually a small chance that we would have had to decide whether or not to pay a very a very large amount of money uh to get them out of the port it's better to be lucky than good right uh that's the so the calculation so i'll give you kind of like yeah what i thought about and what was in my head as this was happening because i don't think i actually panicked at all um i was like this is actually an opportunity to talk frankly about what it's actually like running hardware startup in 2025 with a large supply chain with a pretty complicated product with with with with all these constraints right yeah and finally people care about me right like you know finally this is the main the main news story of the day so i viewed it as like let's turn the You made it up.

1:18:23Let's turn the limited lemons in the lemonade. You didn't panic. You coped and it all worked out. And I get on odd lots and a couple of things. So, so, so that was the opportunity I saw. And then the second piece was I was like, okay, so how long does it take for a boat to go from China to the U.S.? And, and then I was like, okay, so by like the first week of May, there's probably going to be some motion on, on, on something here. um and so i kind of had that in my head of like there was a shot clock going on before we're going to have to do some sort of relaxation of some sort of thing um and looking at the shapes and negotiations and kind of the various players that were you know swarming around and what i've heard privately from i probably can't say who and what but like just just just in a sense of like there were there were there were there were there were well-known folks that have complicated supply chains that were not going to want to pay 145 duties on their stuff um you saw that pretty quickly with apple right uh and apple was kind of like the largest thing they got the exemption pretty early but then there was a lot of other discussion about other companies and at a certain point you know the the noise gets so loud it's deafening the climb down from amazon on the like we won't put the tariffs in the checkout was interesting because it's like i wonder if there was it kind of felt like maybe there were there could have been some negotiation behind the scenes on this where they're like yeah don't put it in there because we're gonna just come back in two weeks kind of well i think they threatened to do it right they actually rolled out the ui and then they got slammed by the trump administration for doing that and seeing it as like hey you're you're blaming us but you shouldn't uh yeah never has the question the question i have is does the the debacle of the last month or so change your long-term planning in a meaningful way because we were talking in some we were talking on the show and off the show yesterday around this idea that okay what did we get out of this right and it didn't it didn't feel like we got a lot out of the geneva negotiations if anything we got rare earths back which was a reaction to uh the initial tariffs uh and de-escalation is it's generally good market obviously likes it uh but we just kind of like round trip back to where we were and john made a point which is that okay well if a generation of ceos like got it into their head the reality of decoupling does that happening right now but it's not happening but it could and you had a board meeting during this chaotic period and so i'm sure during that board meeting you guys thought really seriously about hey if this actually sticks what is our plan maybe that wasn't plan a but at least it's a plan in the plan document now and so uh decoupling they had concepts of a plan so so so i i actually have this like i i i i i don't want to get super political but i do think there's this like trump tariff cycle like image that that there's a meme there's a meme it's called like the trump tariff cycle it's like i add tariffs yep i like take them off art of the deal take them away art of the deal right yeah so so i i i actually there's one interesting wrinkle to this which is um the the democrats have been off in space like you can argue they've been off in space with like you know you can you can blame like wokeness or um some of the more like like some of the elizabeth warren coded economic ideas in terms of like oh we should just break up all our big tech companies because yeah well that would be and like there's a bunch of these these sort of like these these sort of ideas that have been let's call it anti-tech or or kind of on the wrong side of a culture war balance or something like that yep um what's been very interesting is or like let's legalize homeless encampments in san francisco or whatever and what's been interesting is systematically um so like grow sf has like really and daniel and all sorts of really reset San Francisco in a positive way because like they had to get the egg in their face to learn the lesson.

1:22:31Yep. But I'm actually excited about the Democrats being negatively polarized into a free trade party. Oh, interesting. That's a funny take. Yeah. Because, um, because, you know, because as a reaction, yeah, no, no, that makes sense. That's funny. Uh, I, we, we have five more minutes. I want to, uh, I want you to walk me through, I'm sure you've studied BYD. I'd love to know kind of, uh, the history of that company, what your takeaways are, um, and what it takes to build an American BYD in America and what makes BYD, what does that even mean? An American BYD? Why is that different than an American, like uh yeah exactly yeah so i spoke to i spoke to an ex-tesla person and i was like i was like would elon ever build stuff for other people and it's like i've actually asked this question for for for a number of different folks and i've actually gotten a variety of answers and it's kind of like he'll do it if it's on the main path of like getting something getting to getting to getting to the like vertically integrated goal if that makes if that makes any sense but like it's it's unless like unless tesla was forced by the government to be like you will build tanks or something like that like yeah like that they're not in that business and they're never going to be yeah um and so if you think of like what what like byd has in terms of capabilities it's like it's a car company and a car brand you can buy a car from them but like everything inside that car is also like fractally built by them um so like the infotainment system they're the manufacturer of that whereas like quanna um like whereas like quanna or pegatron or something like that may make that system for Tesla.

1:24:14Yep. And this goes down all the way into the battery cells and all this other stuff. So there's like, they're like the most vertically integrated company you could ever imagine. So interesting because you'd think in like a command and control economy, you'd be able to force a Koretsu like in Japan even more, right? You'd be able to force cooperation between a bunch of different companies. But it seems like in China, maybe that hasn't happened with BYD. Am I understanding? It's interesting because I think it's the DNA of the place more than there's like a more than more than the policy side but then the flip side of it is so byd makes stuff for other people and that's like what they do generally that's what they historically did they just happened to be a nickel cadmium battery manufacturer that just kept acquiring businesses and like bringing them into the full until they basically had a car business a lithium my own battery business and kind of like all of the all of the like ip needed to build any electric electric device so like that's everything from like a smartphone to a car to like your dreams and so and so what's really interesting about this is like i can like i could go to a firm like this and i can say hey build me soup to nuts a whole device they will commit substantial engineering resources but those engineering resources it's a little bit less like maybe it's it's there's like contract manufacturers in the west and stuff like that but this is like no they'll do it with the byd parts and the byd process it's it's like they almost have like the best lego set in in in the world and they're able to leverage all of that ip as just a really high starting point to be able to kind of develop develop all these devices yeah so i mean what does it take to build an american version of that?

1:26:03Should we be even thinking about that as a goal? Um, is that, is that a different company or is that a, uh, you know, a cultural shift in Tesla strategy? Where do we go from here? Yeah. I mean, there, there'd be one way to do it. It should be like defense production access. Hey, Tesla bill for other people. But I think that would not go over, they'll go over like a lead balloon and probably is not a good strategic choice. Um, um, I do not want to accept that idea. I don't think it's a great idea for Tesla for a bunch of reasons, but, um, but I do think that there's, I've been focused on like what parts of the value chain in kind of the electrified device universe.

1:26:39And I say that very broadly, like this is everything from EVs to stoves to phones and stuff like that. Where part of the value chain is like not made in the United States, really high level of integration, both like hardware, software, et cetera. And if you have better components, you can make better products. And this is where we realized that on the appliance industry just if you go and build appliances in the united states today you buy electronics from china or germany and all the guts inside it's kind of like you're getting um your graphics card from a taiwanese manufacturer and your motherboard from another taiwanese manufacturer have you taken any trips to germany by the way they've been asleep at the wheel in many ways i've actually never been to a german factory which is saying something that's judgmental against myself and I intend to fix that soon.

1:27:29Yeah. The, I think a lot of what's been going on is all like China has been just incredibly good at the final assembly and like that part of the value chain. And it turns out that's where a lot of the software defined technologies start intersecting is like automated assembly, computer vision, like, like building all kind of like the firmware and software to make stuff work in an integrated way. And like, that's the thing that I think people miss about the car industry. And this is probably the thing to say is the legacy car companies don't build cars like phones, whereas Tesla does and China does.

1:28:06And that's a really broad statement, but it's really like you're building an integrated electronic device, not a bunch of individual boxes designed by different firms that you've like plugged together. um and and you need manufacturers like byd like a byd or like a tesla that can actually go and span all of these disciplines with the same people being able to the people in these various groups being able to talk to each other versus how like a ford car or a like one of a major u.s appliance companies like stove is put together where in some sense there's like a corporate firewall between electronics box A and metal enclosure B, for instance.

1:28:50Can you talk about zooming out in terms of this re-industrialization, reshoring industrial capacity? What's at the top of the priority stock? I mean, there are a lot of memes about like, do we really want to be doing screwing in iPhones and making those here? We've talked a lot about DJI. It seems like there's a huge gap between Chinese capability and consumer drone manufacturing versus American. Well, guess who makes DJI? Is that BYD? I won't say. Okay. But then there's also like lower level, potentially like higher skilled activities like battery manufacturing with the, you know, mineral extraction, all sorts of stuff.

1:29:29How should we be thinking about it? Not just from a consumer perspective, but from like a industrial capacity, governmental perspective. Yeah, so actually what's really interesting is if I go take an iPhone and I like part it all out on a table in front of you and I say, who's actually doing the core engineering an ip of everything in the iphone yeah it's actually like the us and our asian allies like right because foxconn's a taiwanese company it's it's it's it's taiwan but no but even all the parts like lg in a tech makes the camera yep samsung right yeah a lot of korean uh tech in there for sure a lot of korean japanese technology in there so so so the point though is where where is china really good it's like it's like it's the assembly the integration being able to kind of span all of these disciplines.

1:30:12So, so if I was to focus, if I was saying what I think is really important is, I think, I think Noah Smith has a, he's my tenant, so I have to show him. But I think Noah Smith has a good take on this, which is basically like the US is missing out on the technology of the future. And China actually has explicitly said this, where they believe that the tech stack of the future is the electric tech stack, basically. It's like, it's like batteries, power electronics um motors that whole thing and it's like you can build a robot you can build a drone you can build a car you can build smartphones and consumer electronics that spans the gamut of like that is the technology stack yeah and in some and it's rare earth magnets stuff like that for some reason the u.s just like i think it's basically the 08 recession and like a lack of manufacturing investment afterwards um meant that like we never actually picked up that stuff with one notable exception which is tesla yeah and so and and actually spacex as well so like elon did this twice we should reverse engineer why um well that's why we're talking to elon musk of stoves we're very happy many people have been saying this the other people people have also been calling you noah smith's landlord after you dropped that he's your tenant effect out there uh thank you so much Sam for joining.

1:31:28This is great.

1:31:32Before you can even say goodbye to the sound effect board. What were you going to say? Beep. Beep. It's great to have you, Sam. I'm excited for you and the team. And keep us updated. I know there's big news coming, so drop it here first. We'd love to have you back on. Talk to you soon. Talk soon. You know, a lot of people have been saying we need an American compliance company. We need an American company that helps you manage risk. We need an American company to prove trust, continuously. Continuously. We need a trust management platform that takes your manual work out of your security and compliance process and replaces it with continuous automation, whether you're pursuing your first framework or managing a complex program.

1:32:10That's why Vanta.com exists. So head over to Vanta. Vanta makes compliance fun. They didn't say that. I said that because I've used Vanta. Fantastic. Well, we got Vincent from Prime Intellect coming in the studio next. We're talking about a bunch of stuff related to crypto and AI, GPU mining. I think this is gonna be a very interesting conversation. We talked, well, welcome to the stream, Vincent. How you doing? There he is. What's going on? A couple months ago, yeah, yeah, thanks for joining. A couple months ago, we had Bridget from Founders Fund on, gave us the high level overview, but I'd love to get it from you.

1:32:48Can you just introduce yourself and the company to kind of kick us off? Yes. So my name is Vincent. And I'm the founder, co-founder and CEO of Primatite. And basically, our goal really is to commoditize AI and compute more broadly, really also by almost like networking all the computing world together and making sure like no GPU goes idle, but rather can contribute to almost like an interplanetary compute cluster that can like fund for example, so say I models, but also the people can just like use to ultimately save on like compute. Like ultimately the goal really is realizing that like computers is like gigantic market, which will ultimately power all of those different areas.

1:33:32And it's quite inefficient and very distributed. So a lot of it is ultimately like research challenges to make sure we can like network compute. And it's like quite distributed throughout the world, but also like verify it right it's like you you can't trust like the thousands or like millions of different like data centers and like the gpu nodes um distributed across the world so there's basically like a challenge to scale trust and it's like you can trust one data center one company but it's hard to do so at scale um so basically we're attacking those two problems okay uh why don't you talk about the launch sunday yeah you guys uh i hadn't seen i hadn't seen the launch like that on a sunday since the last seemingly few trump executive orders so you guys kind of realized sunday yeah super bowl sunday why not launch a new model but but break down uh the the launch uh just a couple days ago yes so basically um we like kind of like to take a step back almost like this first uh very large scale uh distributed pre-training run into like one which was building on like deep minds distributed low communication training to train basically a model across the internet across multiple data centers in similar efficiencies in a centralized setting where we um like where all the computers network in one data center and now that we switch to this kind of like oh one like post training like reinforcement learning paradigm um it's actually kind of like a paradigm that is like perfectly set up for distributed training because most of it is like those rollouts and basically you can distribute them extremely well.

1:35:06Like there's basically less needs to communicate every training step. So what we did with like Interlec 2 was really scaling to like the larger model size, like a 30, roughly 30B model that we basically trained, distributed across the globe with different people contributing compute. And we basically built this on top of QAN. so basically the goal was also to like take the best open models out there and improve them further for like actually quite um like small amounts of compute it was honestly like more like a proof of concept where we basically proved four different pieces like on the one side basically um we proved this like verifiable compute piece that basically anyone in the world could join with their compute we don't need to trust them there's very like efficient basically ways to like um make sure that they're basically like an honest compute contributor.

1:35:57And then the other piece was really making like basically distributed reinforcement learning work. So basically decoupling those like role generations from the training itself. So you basically can train across the globe with like less frequency of communication. And what we basically prove is that like with quite a small amount of compute, like roughly $100 ,000, which is not comparable to obviously kind of like the more like close AGI labs, we've basically been able to like improve upon Quent. And it's kind of like sets the foundation for us to actually scale this much further and ultimately actually have kind of like a repeatable recipe to get like basically improve like the strongest open source models out there continuously for quite like small amounts of capital across different domains.

1:36:44So we made it much better on math and coding, but the goal is also to make it much better on like other domains. and really like how to do so is like really scaling those like reinforcement learning environments i think you had uh will brown on like he joined us to work on this yeah i was gonna ask what was what was the what was the process like recruiting uh the big dog wb absolute dog i mean i i just have to imagine he's such a good poster he's such a great guy to chat with i he could have gone anywhere but he ended up with you guys yeah i think it's it's two things like you should ask him stuff but I think like a lot of talent is gravitating more towards like open source AGI and like to basically build in the open, to publish in the open, right?

1:37:23Like to be able to also like contribute to the frontier of like the most exciting research and in a sense like I have friends at OpenAI and who's never heard because they can't publish, don't really talk about their work, right? So like I think there's this component but I think then we are basically I think like now like especially in the Western hemisphere, I think like one of the only places that is really catching up to the frontier, right? Like there's obviously like Quen and DeepSeek and others in China, but I think it's really a fumble by like America, I think, to basically be the close guy that is like not really almost like standing in for the values of like democracy and freedom to the same extent.

1:38:01But I think like he is obviously extremely passionate about like open source AGI and kind of like distributed like reinforcement learning. So like he immediately hit the ground running on basically scaling this much further. So we'll have more like, yeah, like work we are building on top of like this direct jump distributed reinforcement learning, like in the next few months already out, which should like scale this much further. I remember a couple of years ago, I was using Octane Render, and they have a distributed rendering network where you can render CGI images on GPUs all over the world.

1:38:35And they actually have a crypto layer for render tokens as well, but this idea of distributed computation makes a ton of sense in that context because each frame is unrelated to the other frame. So you can, if you're rendering a 90 frame sequence that's a few seconds long, you can just send, you know, each frame to a different machine. What have you learned from the successes and failures of the render network and what Otoy did with Octane? And is there any difference between those discrete individual frame generation and something as complex as training a whole model where you know you might be you in theory you'd need to load it all into one uh one instance every single time yeah like great question i think there's a few players like them and others that have tried this in the past and i think they've been almost like in some ways i think too narrow and like ambitious right it's like rendering like it's nice but like it's not like the general purpose like use case that we now have with AI, right?

1:39:35So I think there's basically, like with our goals, really making this much more like general purpose useful for like every single like computer AI use case. And I think that has been one big learning. It's basically being like most of our team are AI researchers and they come from the AI world, including the people that build, for example, the verification mechanisms, right? Like the verification in our case has like tiny overhead, like 1%. Like the more like crypto native verification mechanisms have more like a 1 to 10x overhead, which then make it completely like unfeasible to compete with a centralized clusters.

1:40:06So I think that has been like a big learning is basically being very pragmatic and close to the AI researchers. So that's like the main community we're like building with and engaging with and basically building for, right? So like all of our users are AI developers. And ultimately, I think the biggest lesson has been just creating value for them from day one, right? It's like we launched basically our compute platform like two or three months in with like two or three people and immediately got like a lot of users and scaled since then. Like basically that like any kind of like AI startup or developer can just like basically with our platform almost like like have like a global market on the one side.

1:40:48We aggregate all the hyperscalers, all the different data centers, even working with folks like Aaron from Hydra, and all these guys to basically aggregate and orchestrate all these different data centers and hyperscalers. And I think that has been another lesson. We are not trying to be the exclusive marketplace where supply onboards. We just aggregated all the supply out there, which gave us a huge wedge maybe to enter the market. Okay, talk about some of the hardware where this is actually running a 30 billion parameter model. At floating point 16, you're going to need 60 gigs of RAM. You're probably going to be on H100, not gaming cards.

1:41:29But are you thinking about bringing residual capacity online in the gaming market? Or are these like small data center operators that just have a stack of H100s that are not using it and they're bringing them online? Yes, so basically I think what's been exciting is like actually this run was here to genius. So H100s and A100s could join and I think it may be even like H100s and like for future ones like actually for next one we're starting like even 3090s and like gamer cards can join and I think what's interesting is you can basically with like this role generation and like basically synthetic data generation you can distribute the tasks like basically based on their like difficulty to different cards right and to different models.

1:42:11so the goal really is like to have like a fault tolerant like heterogeneous like pool of compute right so like what's also a big one that we did already in the first one but also in this one is fault tolerant so it's not only that you have like H100s joining and A100s and 3090s but they can also drop out right like they can join for a few hours when they're idle and I think that's like the border thesis is like almost like every GPU that is like idling it's like a market failure right? Like especially in a world where increasingly like my theory is like the biggest almost like commodity in the world will be compute and like it will be one of like one of the biggest slices of GDP and it's like unfortunate like half of that is like sitting idle instead of like curing diseases, generating like movies or like some like creating the basically technological progress, right?

1:42:56And I think the goal is to basically just get to like 100 % utilization of compute and we're very very far from that yeah yeah well uh what what are you seeing in like the trend in chip size and scale you know obviously nvidia is pushing pushing those like djx uh djx like rack mounted units racking everything together with nv links you can do even bigger models cerebrus is scaling up the single chip size to wafer scale uh the the the theme of the last few years seems to have been uh let's not decentralize let's actually centralize further and so uh where Where does this go? What is your take on kind of the ever bigger racks and ever bigger chips for larger compute training?

1:43:37Yeah, I think we'll see both. And I think actually, ironically, like an NVIDIA actually has an incentive to fragment the market. Like they don't want one buyer. They don't want AWS to be the only customer because AWS is building their own chips or like the big tech clouds basically, like have their own chips competing with NVIDIA, right? So NVIDIA is very happy to give a bunch of chips to Core, a bunch of chips to Cruiser, a bunch of chips to Lambda and all these other guys, right? But then, so I think we'll see both at the same time. Like, I think like all, like we'll have thousands and like millions probably over the next few years, like of data centers, there will be these gigantic clusters, right?

1:44:13Like Stargate from OpenAI. But I think at the same time, like there's increasing amount of like capacity in almost like the mid tier of like smaller data centers. But then I think like also on the other side, like there's a lot of compute, like sitting idle everywhere, right? Like soon whatever, like autonomous cars, robots, like everything else. So we actually had like some of the biggest like car companies in the world reach out to us or phone manufacturers that are like, oh, we're sitting on like hundreds of millions, like billions of like phones or cars that are like have idle compute capacity.

1:44:42And I think we're not there yet, but I think we'll have like, we'll have just much more compute everywhere, right? Like we'll have more like mega clusters, but we'll also have like more mid-tier and small clusters. Mega clusters. Mega clusters. mega clusters uh i have a question before you leave can you can you give some uh any expectations around deep seek r2 i've seen some crazy rumors flying around online uh i don't know what's what's real and what's fake news information warfare curious what you're expecting out of the next release yes like i think they'll they'll definitely catch up to the frontier like again i think to some extent I think they'll they'll probably be not as good it would be my guess as like um Gemini or OpenAI and but I think there's obviously this trend where I think like and all the closed lab CEOs have admitted it right it's like that open source has catched up like very uncomfortably close to their lead and I think that will continue I think that the close guys obviously have some um advantages but I think there's also huge advantages obviously the open source uh AGI space has so I think like the trend will continue that ultimately like the deep seeks and the clients of the world but hopefully also like players like llama will like catch up a bit more like obviously they've been a bit disappointing with their like llama for launch but I think they might also actually uh like catch up again and and then obviously I think like our goal is like to be able to basically build on all of these um on all of that progress right it's like basically we kicked off our like largest synthetic data generation run the day deep seek dropped and like generate a basically a lot of like verifiable like reasoning data that we couldn't use for example in this recent model so I think the like what I see more is like basically we can like improve open source AI progress like literally day by day week by week like it won't be a thing where we need to wait like four months yeah like some progress to come but it's more like continuous basically progress well thank you so much for joining this is a fantastic conversation yeah I really enjoyed this and congrats on scooping yeah wb man legend it's really truly truly legendary pickup deal yeah uh but uh come back on again soon i know you guys are gonna have a bunch of launches uh coming up and uh appreciated your perspective we'll talk soon appreciate it uh and if you're looking for distributed compute go to prime intellect if you're looking for distributed advertising go to adquick.com advertising made easy and measurable say goodbye to the headaches of out-of-home advertising only AdQuick combines technology, out-of-home expertise, and data to enable efficient, seamless ad buying across the globe.

1:47:17They just make it too easy, John. I want to spend every single dollar we have on AdQuick, billboards. Anyway, we got the CEO of Backbone in the studio. Let's bring him in, Manit. Excited to have you here. Welcome to the stream. How are you doing? Thanks for having me. Appreciate it. I feel like I walked into the middle of an ad read. You did. You did. You did. We are 100 % corporate sponsored. Independent media is dead. The future is corporate media, and we are leading the charge. But good to have you here. Can you introduce yourself and explain what Backbone is? I'm a fan, but some people might not be aware.

1:47:52Sure. My name is Manif. I'm the CEO and founder of Backbone. Backbone is the leading platform for the future of gaming. The way I describe it to people is that every time there's been a shift in streaming media. It's resulted in some of the biggest consumer businesses. So like music streaming, you had Spotify and Sonos. And then TV streaming, you had Netflix and Roku. And you can think of Backbone as the analog company for streaming games. We have a number of products. We sell devices that you can use to connect to your phone or any screen and play games on them. And then we also make an app that aggregates all the different streaming services like Xbox Game Pass and others.

1:48:31so you can access them all in one place. You can buy our products in any major retail store, Walmart, Target, Best Buy, et cetera. And we're backed by investors like Index and Sound Ventures. Yeah, so I got one a couple of years ago. I was not getting a lot of time in front of the PlayStation 5. And so I would drop my phone into the backbone and stream from the PS5 over the Wi-Fi to the phone with the backbone. It gives you the whole gaming experience. Very fun. what has the relationship been like with the big console providers? Can you break us down the different strategies? It seems like Sony has been very competitive, launching something that's kind of like a competitor with the PlayStation.

1:49:15I forget what they called it, but it has a screen and device and does remote play. Microsoft's taking a much more open approach, letting you even stream Xbox games on PS5 and Forza Horizon is on PlayStation now. And so what has been the relationship with the different major gaming platforms been like over the last few years? For sure. So I think, of course, each one has its own strategy. Microsoft in particular is really interesting because they're running this campaign now called Xbox Everywhere, where the idea is that you can turn any device in your life into an Xbox and run games on any screen.

1:49:54And the way they're doing this is through streaming games from the cloud. So Microsoft's subscription service Game Pass, if you subscribe to it, you can basically access the entire Xbox games catalog on any device that you own. And of course, the devices and software that we're making work great with Game Pass and make the experience work really well end to end. So you can do it on things like smartphones. So I think in Microsoft's case, their goal is really to drive more Game Pass subscriptions and just gather more Mal and Dal for their service. And it's super strategic to, you know, their overall roadmap with cloud and all that as well.

1:50:29Why wasn't Google able to pull it off? I feel like they were early to that party game streaming. I can't even remember the name of the product, but it never really hit like major takeoff or do I have it all wrong? And actually it's better than ever. I don't even remember. So you're talking about Stadia? Yeah, that's right. Stadia. Is Stadia still around? What's that? Is it still around? No, no. It was Shudder, unfortunately. Yeah. Classic Google. Classic. But I mean, you would think they have all the resources. They're a hyperscaler. They have GPUs in the cloud. They can do rendering. And they're also somewhat platform agnostic.

1:51:05So any game developer could go to Stadia. It just seems like super additive. I'm sure you had intimate interactions with them as Stadia, as you were building the company. Why do you think Stadia didn't take off? For sure. So I was at Google when they were working on it. That actually was sort of part of the basis for the idea. I went into the office one day and there was a demo that an engineer had built where you could use Google's cloud technology to stream games to like basically a two generations old Android phone. So you run GTA on like a two generations old Android phone. And I picked up this demo and I tried it and I completely lost my mind.

1:51:46I was like, we've seen disruptions in gaming across every single vertical, right? Business model, audio content. But this has got to be the single biggest disruption in the history of the games industry, at least that I'm aware of in a long time. And so that kind of became the basis for the idea behind Backbone. I basically made a slide deck on why I thought it would be like a really big opportunity for Google. And I sent it to the head of product for YouTube gaming. And then I realized it could go a bit wider. So I sent it to a bunch of the vice presidents internally. And basically, a bunch of people ended up reading it.

1:52:19And it became one of the most viral slide decks internally in the company. Like 15 % of the company read it by my last day. And so I got to meet with some of the executives that are working on Stadia and other initiatives. I think there's no doubt that the technology was incredible. But I think other players like Microsoft just had a better go-to-market strategy and more content. Yeah, that makes sense. I mean, the content's really, really key. Switching gears a little bit. But how have you seen the industry react to the news last week around the App Store sort of payment policies? I imagine you've been hearing a bunch of chatter.

1:52:57Yeah, I mean, certainly if Fortnite were to return to the App Store, it would be pretty big. I think it's been like... It was resubmitted, right? It hasn't been approved yet. But normally Apple approves in 90 % of apps get approved in 24 hours. We're on like hour 46. So everyone's saying, Apple, what's going on? Dragging your feet anyway. Yeah, I mean, I think it's huge. I mean, it means that if you're able to link out to external payment systems, you're basically going to be able to take in more margin as a game developer. And I think it'll, you know. But how do you expect users to actually respond, right?

1:53:35So let's say you download a free game and you're like, okay, I'm willing to pay for this game. And you get the option to just subscribe in the app. Maybe it's at one price or subscribe outside of the app at a lower price. You'd have to offer, I imagine have to offer some type of incentive to get somebody to want to go to a separate webpage. So I can go outside of the app and I get maybe some... Instead of 30%, you're splitting the difference. You're splitting the difference kind of, but then I imagine some users like the fact that their subscriptions are native and they can go into the app store and churn if they're not using something.

1:54:16And so how do you expect that? How do you expect developers to actually play it? And do you think it's, are they really getting back 30 % or is it more like, again, splitting the difference? So I think what we're seeing right now, and in Epic's implementation, they posted a tweet online, like the revenue cat CEO replied to Tim Sweeney's tweet. And I was asking like, how did you implement this? Did you use an app purchase? So in Fortnite's case, they show an option to do in-app purchase. And they also show, you know, the web flow through Stripe or whatever. And that one is, you know, a 20 % kind of cashback or something like that.

1:54:52You get like 20 % back in Epic credits. So there's like an incentive. People might have to incentivize customers to transact through the web. And that - But giving them back free credits, like points in the game, which they're going to value on a cash basis. Tim Sweeney, absolute dog. What a dog. I want to talk about the actual technology roadmap or tech tree to make mobile gaming even better. What's more important, better devices, faster chips, more memory, better screens, or just ubiquitous 5G, Starlink, better connectivity, lower latency. Because Call of Duty, you can stream it on your phone.

1:55:38Probably Xbox Game Pass allows you to do that. There's still a Call of Duty mobile that's optimized for the phone. And I think people probably play that one more on their phone than the streamed one. And so we're still not in the fully streaming era. What are the blockers to get us to just one installation of the game everywhere? So, I mean, both of those things are bottlenecks. And we often used to get in these like furious debates internally over like which one was going to overtake the other. And then we kind of realized after a while it was a bit aimless because we just want the overall basket to succeed.

1:56:11And I think you'll see a lot of like hybridization where you'll have, you know, games that are relatively like thin clients, but like a very significant cloud component to them. versus right now we think of this dichotomy of like fully streamed versus native. And I think maybe there'll be some games that flit across the boundary line. I mean, Fortnite is one such example of that where there's a pretty heavy cloud component to the game. What are you playing right now? Right now? So I don't know if you've heard about Expedition 33. Oh yeah, I haven't played it. I heard about it. I'm just about to get into that.

1:56:43That's been popping off on the internet. Yeah. I've been playing a lot of Fortnite. Yeah, that's good. Were you surprised about the GTA 6 release date? It seemed like the market had been pricing in a 2025 release, and then all of a sudden it was like, no, 2026. um i mean it takes these games take forever now the production costs of there's some really great articles uh online talking about like how the development cost of these titles of ballooning because of like this need to have really lifelike graphics and realistic gameplay. So not surprising, no. What about tariffs? I imagine that that was a tumultuous time for you.

1:57:27Obviously sigh of relief recently, but what's your outlook and what did it mean? Take me through the emotional journey of the last couple of weeks of the company. Yeah. I mean, definitely short-term painful. It was awkward timing given the Backbone Pro launch, which just occurred this past week. And this is a huge launch for us. I mean, we generated, I think, around 700 million media impressions, which is more than Beats and Aura and Sonos generated for recent product launches. And so it was a big launch and it was an awkward time because of the tariffs that made the pricing strategy a bit complicated.

1:58:06that being said we think in the long run anything that promotes you know free trade or reduces tariffs like globally um is beneficial to the company because then we can offer our products in more markets um than we do today at you know more of a fair price to consumers yeah i gotta pick up a backbone pro i'm looking at it right now i'm adding it to cart john has always said the biggest risk to the show is that he just gets sucked into some some it's like my one vice like You'll never find me passed out drunk, but you will find me 70 hours into Metal Gear Solid for the third time. Anyway, this is fantastic.

1:58:44Anything else, dude? No, I'm happy for you and the team that you get a little relief on the tariff. Congrats on the new launch. And good luck. I know you guys are shipping in about a week now. So I'm sure there's a lot of work to get to that point. Head over to Backbone.com. Pick one up if you're a gamer. Check it out. Anything else? That's all. Thank you. We'll talk to you soon. Great. Have a great day. Bye. Speaking of sleeping easier, post tariffs, 8sleep. Go to 8sleep.com. Five-year warranty, 30-night risk-free trial, free returns, free shipping, pod four ultra. I got an 88 last night, seven hours and 11 minutes.

1:59:19Not too bad. What'd you do, Jordy? Did I beat you or did you smoke me? A hundred. A hundred. He knows Brian Johnson's coming on the show. He's prepping. Proof of work. He's prepping. Anyway, we have our next guest coming in. Put it up. put it up 100 see it's not a screenshot i'm moving it around there we go yeah the new app is beautiful go download it uh anyway folks we got aaron ginn coming in breaking down a new uh wall street journal article he wrote aaron how you doing he's back yeah doing well man can you put on those sunglasses jordy i can't recognize you yeah yeah yeah these ones these ones the pit vipers uh unfortunately so so there's some crypto there's some crypto coin that that people are using pit vipers to promote so we got to be careful but you have to be careful but uh but they were sent to us uh from a from the ceo i think the ceo of pit viper he just likes tbpn he said have some pit vipers it's pretty cool that they like rotate up and down like this yeah i've never actually worn any of these um anyways it's great uh it's great to have you on um oh it's great the prime intellect ceo was on earlier and he said that you guys are working on some stuff together just cool yeah yeah yeah like they're handling the higher level like training layer if we're doing the underlying infrastructure distributed layer but they yeah like going back to about the op-ed that I wrote I've been more or less battling slash encouraging administration to take a more reasonable approach to the threats that Huawei has against our dominance on AI and when the I met and know the people who under the Biden administration that wrote the January 15 ruling that we don't really know, by the way, like what's going to happen.

2:01:03We just heard rumors that things are going to change and they're going to be more positive towards Jensen and NVIDIA. But generally speaking, they wrote this kind of the precursor framework that people understand is it created a G18 of AI countries, which included like Japan, South Korea, Taiwan, America, Canada, and some European countries by half of Europe. and the original biden ruling right these are bureaucrats like there's not there's no it's not it's very vague law area on their governance they're basically operating under 1980s missile technology proliferation law those are in fact whenever we had soviet union but but so they're kind of regulating or that and the biden administration you know a couple days before they're leaving decided that they would create this kind of g18 uh of countries and then a hundred-ish other countries that were going to be basically prohibited or require license and then you know absolutely no good countries car tier three the tier three countries are quite logical right they're like libya you know yeah venezuela right it's like oh okay this makes sense but it's your two countries required license was like greenland portugal you know switzerland austria all of eastern europe was on the crap list all you know the presidents right now and instance in the Gulf region, all that was on the crap list.

2:02:22And then like kind of random countries like Mexico, Mexico makes all of our servers. And the original kind of premise around this was regulating cheap power. So it was mostly anti-competitive. They didn't believe that Mexico got to be used or sent to China. They just thought, oh, they have cheap power. So we should just say, oh, that's bad for America. And so I've been advocating for this kind of idea of like this new Monroe doctrine, into thinking that uh ai is more of a forward operating base of america uh due to the neocons and neoliberals we have over 300 operating for operating bases of the world uh where our special forces are you know active and doing things that they don't tell us and so i basically have been trying to like kind of you know uh think about the ai is kind of something similar where every deployment that is an american deployment is real estate that we own and that running the world growing like growing on our infrastructure is better than growing on chinese infrastructure and it's either or and i think that under the bite administration there was this view that they couldn't do anything like oh they just make icons it's not to be overly trite but i think they had a kind of a less view of where china was and as the data came out with uh more and more advancement on the ascent line under huawei more advancement under deep seek and manis it became very much clear that this actually was a race uh and that they were significantly closer than we thought.

2:03:44And so the goal then is proliferation. And that's essentially what the Trump administration, the Commerce Department announced last week was that they are going to embrace a trusted proliferation model, which is what I've been advocating for this forward deployment of NVIDIA gear, running on our infrastructure, running on our frameworks, because it's either us or them. Either the world uses dollars or uses something else. Like there's not this kind of secular world where trade happens and it doesn't include a country, it involves us or involves somebody else. And that's becoming more socialized and more accepted.

2:04:17And so now they're trying to think of other ways to try to manage proliferation. Tom Cotton introduced a bill to GP, track GPUs. Not very bullish on that, but, you know, like there are other ways now people are trying to think about doing this rather than just saying, hey, Portugal, you have good beaches, but no GPUs for you. Saudi Arabia, we can sell you F-35s. We can sell you even, you know, ICBMs with no nukes on it. But, hey, you can't get the Blackwell series. that's too risky for us to do for you. So there's a great resetting that's happening. It's good for America. Yeah. Yeah. What is your take on the recent Saudi Arabia deal?

2:04:5018 ,000 NVIDIA Blackwell chips going in. With hundreds of thousands. With hundreds of thousands on the way. It seems like it's an affirmation of everything you've been saying. But what's your interpretation? Yeah. Like, I think it's one way Huawei's already deployed there. So I think that that was a wake-up call. and there's also a ton of Chinese cloud companies that are doing deployments in the Gulf region. But it's positive because, again, you have to realize that people are going to get this regardless. So another good framework is because this is all leading-end stuff and people get confused about what's actually happening.

2:05:27They also overly express our actual dominance on a particular area. But think of it as a MIG platform versus Lockheed or Boeing. You're not going to tell a country, hey you just can't have a fighter jet they're like we're gonna go get some mix right we're gonna go get some i forgot the ring of the french one right that that drops you know uh baguettes from the airport right whatever that french land right whatever that that one is right uh it like i forgot the name of it but the uh so it's like that it's like it's like it's an either or question so we either have a decision to engage with the country or not and then if they don't they're going to go somewhere else and with saudi arabia which is a very clear ally of ours very clearly israel the leader in the region with israel uh we have to be in a position of they understand this like like like the gulf region kind of gets real estate and put crap in the in the building they kind of like they understand that right and and so this is something they can feel they can be very successful in and as well as it gets to kind of the the you know the jedda tower birch leaf a thing of like we're ahead we're winning the line right like all these like crazy projects that won't happen until yeah but our data centers but but our data center you know and obviously saudi will have fantastic partners and nvidia and i'm sure many other uh hyperscalers uh for their new ai factories that's the new aim we got to stop saying data center it's ai factories uh yeah tokens but uh but but yeah isn't isn't this not is it is it is it correct to call it real estate today or is it something else well i think the dynamics behave like real estate like that's how the actual investors think that's how they get finance the execution is different yeah yeah so so i like the way that this blows out what i believe is what like the premise of our model why we're a part of like over about almost a dozen summary summary projects we're kind of the For NVIDIA, they advertise us to a lot of their growing data centers internationally as Hydra is the best sovereign AI go-to-market provider in the world.

2:07:32So we're really, really good at this. And our operating model is much more, I think I mentioned this last time, think of it as the airlines industry. That the commercialization of AI really resides in the platform. And that platform is Jensen. It's not anybody else. It's Jensen. And that's just like Boeing. And you have these kind of like rare birds that are coming out, which is OpenAI, but the technology mode around OpenAI, as we come knownly accept, is relatively commoditizable. So its real value is the brand. It's the accidental consumer product company, has been Thompson, like I said. So when you have the commercialization happening at the platform hardware layer, like Boeing, you create lots of billion-dollar companies, American Airlines, et cetera, Emirates, whatever.

2:08:19But you don't really have this parabolic power law outcome like you had with software-defined nodes, like Google or Airbnb or Twitter or something to that effect. So when you have a world where it works like that, Jensen becomes the play. Jensen is the platform. You can still create lots of different successful companies that are very large. But I'm very skeptical you're going to see this huge outlier into the stratosphere like we saw with other software-defined nodes. rather it looks more like the airlines industry where there's lots of very successful people lots of very successful uh you know winners but there's no clear category leader there's no clear like so far ahead that nobody nobody kind of will ever touch it because the fundamental product is a product that gets deployed in a data center that gets run and commercialized and all the value is created there and everything else kind of plugs into it it's the fastest way to become a millionaire start as a billionaire and buy an airline uh yeah this is a famous thing because Because, yeah, the value doesn't really accrue to the airlines in the way that it does to consumer tech companies, of course.

2:09:21I want to talk more about this idea that it is valuable to America to have foreign AI models running on American infrastructure. And maybe we could go back to the F-35. If we send an F-35 over to another country, they have the ability to use that however they want. We don't really have control over that. We don't necessarily have a kill switch. Um, obviously it's good economically because we sold it to them. We get some money. Uh, that's nice. But, uh, is there, is there, is there some downstream benefit to, well, now we're really allied with you because yeah, we sold you the planes, but you're still going to have to remain friendly with us as we resupply and send you more screws and bolts to, to menu, to, to maintain these planes.

2:10:08Is there a similar analogy in AI factories or data centers where sure. yeah you bought 18 000 nvidia blackwells but if you don't keep working with us those are going to die you're going to need to replace stuff and you need to keep us an ally is that the goal yeah like that's my goal because i believe america should win i don't i don't believe in a multi-polar world i believe in a new one road doctor and i believe that we should reshore to the americas and and some of that will be here some of that will be mexico some of that will be no salvador we're already you know monetizing el salvador with prisoners so like why not them make fiber optic cables why don't they make fans why don't they like for in this in the service supply chain or textiles like mexico is the cheapest place for us to manufacture it's been that way for over a decade but they don't have the technical capability to do it which is why you don't see that there but but in the new monroe doctrine uh my thesis is that we reshore critical things to the americas and then we have these we have these external allies that reinforce that just like the old manro doctrine was america's and we have britain that enforced it and with us And so the same thing goes with the Ring of Fire from South Korea down to Singapore that becomes our allies in that region.

2:11:16We ship America goods there. They become America enclaves. But they're relatively sovereign. They have their own identity and their own military. And partnerships like that not only advance – like you said, sell goods. There's an announcement with the Boeing. I mean they already bought Boeing. So it wasn't – both Riyadh Air and Sideya, they're Boeing platforms. They don't have as many Airbuses out there. because everybody says it is like just so you know like that it's a very technical thing but their engine design is not really compatible with sand that well so that's why they they prefer belly because it's more adaptable high heat um so like either way like that that's an advantage to us because it's a form of soft power that we can allow the proliferation of not only safety for americans globally but it makes us rich like trade is american it's not secular it's not something that just happens like if we define it the dollar is a platform trade is us we define we do we rule the open notions because we say that it's important for us to trade so i there's kind of like this weird thing happening on the right that there's like these two types of like uh china hawks there's like og china hawks which is like me who believe that the way you beat china is you win then there's this new china hawks which are like protectionist semi-socialist and want to withdraw from the world they believe in a multiplayer world uh and and they kind of crept in but they're actually a little they're more on the left side of the spectrum and the tariffs and uh kind of attitudes around nvidia you that's when you hear these things of like we shouldn't sell the world like kind of can wait on us or we don't need their goods we shouldn't trade um versus the original og china hawks were like no we should trade trade is peaceful it's great it decompresses issues between countries but it doesn't mean we give away the farm it just means that we rebalance and we make it fair and and it was kind of this hot minute where like these like new neo-chinox got in charge and started doing this kind of crazy things with terrorists where we had no freaking idea what was going on like believing that terrorists make us rich is one of the most profoundly illiterate things like i've ever heard from a lot of backgrounds and economists but they like it's literally stupid the places that have highest tariffs in the world are absolutely poor and the places that make external revenue service no again i'm pro that right because i don't want to pay taxes but taxes are fat right so that they but i'm willing to engage in some type of thievery if it means that i'm safe when i go to i just came back from belize but i'm safe when i go to belize i'm okay i'm okay engaging that uh but but america only wins when we have this peaceful economic transaction that occurs with countries that want to adopt more pro-western frameworks.

2:13:58We should support that because when the world becomes more west, it becomes safer, it becomes more prosperous, people have rights. We had Sam from Impulse Stoves on earlier and he was saying that free trade is almost becoming a liberal policy. It's a liberal position now to just be like free trade, just from a reactionary standpoint. I'm curious what you're expecting out of the uh qatar visit which i believe is happening in the next 48 hours yeah like i it was a cutter we we've already been talking to the world family there like they they definitely want to get in gpus they're they're have you ever been to cutter by the way it's so it's you have you have yeah briefly it's it's absolutely bizarre it's a imagine like the largest country club in the world as a country and there's like everything is clean and well done but like in the street and you're walking it's mostly just like servants there's like no people it's like it's very weird um and they have so much money it's it's they they they even build like the cutaries right there's only i think 300 000 total citizens and the total number of like the broader servant class is like around 3 million uh in the actual city uh but they're so wealthy they have they they do these um you know sand duning thing uh which i went to go do and they build these homes in the middle of the desert like these massive camps and they just like have these hundred thousand dollar uh they just go up and down sand dunes all over the place or there's like these huge camps it's like they have so much money uh and they're they're trying to get into uh power and energy as well it's the data centers but that there's a there's a tricky problem which is that they have a habit of being anti-israeli uh obviously something happened to benji and trump nobody really knows something happened uh you know and benji could not be the most friendly uh you know person to deal with so something happened there yeah but the cutaries like to think themselves as like the switzerland of the region in reality they're just more pro iran and pro pro al-qaeda um and and so So that's a, that's a, the one thing that never understood me is that the ultimate goal of every Trump, you know, conversation is a deal.

2:16:16That's what he loves. He loves deals. And, and so us being particular hawkish on any type of country, I think is kind of not who we voted for. It's we voted for somebody to make deals for America. And that includes a jet, which I don't know if it does or not. I heard it was. It was not. It seems like a rumor. Yeah. Yeah. It was a new S747 or something. Yeah. It just, it seems like a rumor at this point. like no one really knows one way or another what's happening with that but it's an interesting story certainly no i i had no problem like he went to go see kim kim jong-un like whatever no no i completely agree with this that's a great take yeah so like i have no problem meeting people and if that means more i want them to be more pro israeli and i want them to they will get the release last american that's a win for our country yep um and that that's what i voted i voted for trump and i wanted him to do deals i wanted him to break the mold reset relationships and make it more pro - And create more allies, like swing more people to our camp, regardless of the history, figure out how to do deals and build more allies in the American sphere.

2:17:17I want to talk about the learning curve, Huawei versus NVIDIA specifically. Is that a dynamic here with pushing NVIDIA chips abroad? Is there value from actually advancing the technology staying on the cutting edge just by increasing the volume of shipments and manufacturing we've obviously seen there's tsmc on the learning curve um but uh but what is your take on the huawei ascend versus nvidia hopper uh blackwell kind of uh uh fight that's unraveling right now it americans need to realize that like like lithography is not like a white technology it's not like white people like own it right it's a technology map that's available to anybody that can work on it sure and their ability yes it's not tsmc level stretch they're probably at five nanometer they're more likely at seven but probably stretch our five tsmc is at two it's about to go to like uh you know uh 1.5 i think or something to that effect um but but the thing you don't ever underestimate is that china from a from a supply chain perspective is it's almost completely sovereign not only do they dominate rare earths uh but they have the entire manufacturing infrastructure to make wallets yep and and that is a type of competitive advantage that we don't and and so now again like it doesn't mean we become central planners and like drift into you know lineas or trotsky thinking that like oh we should just essentially plan this entire thing like that's not my view i my view is comparative advantage it works like as as geordi was saying like all of a sudden liberals remember that adam smith exists and and you know that tds has benefits right is it it pushes people to is it because it's not a philosophy it's a it's a it's a psychosis so yeah yeah so it doesn't have an orientation it's just not that big yep uh but but you know like we we have the ability to to win because we still are the best but but they're just going to take a different approach like they're going to take advantage of their their power they're going to take advantage of the fact they move fast and they're cheap but a new ai belt and row initiative has already started already around the world we know this we see them at you know different data centers that we work with selling their stuff and it comes as a fully complete solution with deep seek open like a manis models combined with infrastructure combined with support manis has has that kind of distribution already no like huawei is designing for their their their open source ecosystem so it's still huawei is being the hands of heat right they're building i think what jory's asking is like have you ever run into a country that's not china that's like yeah we're we have huawei ascend and then we have deep seek and then we have manis because manis came out of nowhere just a couple months ago feels like that would be a very quick ramp to be like it's deployed in another country as a part of the ai belt and road initiative deep seek i can believe manis it is so yeah give us the update on that yeah yeah yeah it's deployed like deep seek and manis are everywhere in Europe, like they're everywhere.

2:20:16And again, it's open source, if we're open source works as a distribution strategy, they're not open source, because they give date, they try to care, they don't care about open source, they care about defeating us. And, and, and so the approach that this is why opening, I made that announcement of, we're going to do like this combined deployments of software plus like infrastructure is because that's, they're mimicking what China is doing. But but the proliferation of the open source models go look at open router like you can just see the growth in other models those other models are chinese so so it it's a it's a race that actually is a legitimately an arms race where there's not much of a uh like like what is the ceiling right like if if we have to kind of think of it as imagine your multiple countries that someone invented a new technology like the nuclear weapon like you're the reason why uk france israel uh ussr india basic proliferation that happened was because america said no we're not gonna give you any intel even though we promised by treaty actually that the uk and france would have our intel after offennheimer did what he did and then that's what started the race is like okay well we can't just have america have it we need to have it right and so what is the actual break point at which all of it stops it's like how do you it's like it's like it's a kind of different it's not an economic question right it's it's a diversity question uh and that's why the the gulf region announcements are really exciting is because fundamentally that region is going to serve europe it's probably not going to serve america it's really going to serve europe because they can say all they want they want to build this many data centers or whatever but the trying to make the eu do anything is is is you know it's trying to convince an opie's person to stop eating they'd rather just use those impact like it it's It's not going to go anywhere.

2:22:00They need some kind of injection to finally do something. And the Middle East region and the Gulf region is just really good at this stuff. They're good at real estate, building power, deploying crap. And in the airlines world, that market is dominated by Europeans. Europeans use Emirates and Qatar to fly to Asia, to fly to the region. I think it's going to be really similar. So we're able to dominate Europe through those relationships. but if we weren't going to do it my guess is like a year or two Huawei would have some big announcement of doing with probably Qatar probably the country I think of first maybe Kuwait but yeah and we still have to lead but we have to realize that Huawei is going to make inroads and places that we have decided to not and we just have to accept that we can't stop yeah I'm trying to think of another how did the how did the Manus investment from benchmarks sit with you uh a lot of people that are a lot of people have pretty much everybody has an opinion as peru hover well about it yeah everybody's got an opinion on it yeah uh but there's varying uh a lot of new comments actually i i think everyone has an opinion but not everyone's willing to share them a lot of people have opinions offline why not share it i mean dynamics and benchmark they don't necessarily yeah if you're competing another firm yeah and I think that's the right stance generally for the most part.

2:23:24Yeah. Unless you're done. I mean, I don't care. We could free exchange. If they want to do that, that's the prerogative. If the Georgia department wants to go after that investment, as I told me they're going to, that's their decision. Is this a free market, free exchange of ideas? I believe your capital is your capital. And if you want to deploy it to a Chinese model, then what I question your patriotism, probably. But it's your right to do that. and it doesn't need to go much farther than that um a vanus has serious adoption and we should think of it as i think that's a maybe benchmark we think about it is like you know one that they're going to create you know a tiktok style like relationship it's not real all this stuff is a figment of our imagination i think it actually matters to chinese it doesn't but it's that's just an excuse the reality is like they want to win they want to make money and in some respects i understand that i mean we're capitalists here uh and in other respects you have to think about it from the patriotism perspective of like we're not really interested in it's one thing to like maybe even allow gpus to be sold there which is going to happen and by the way uh gpus are all the trade deals include gpus man like to think that that we're really that serious about like we don't want them to have infrastructure it's like it it's it's total baloney like they it's such an easy gift for the administration we're the best we want to make a crap ton of money uh we like that nvidia is rebuilding here it's an easy give it's not it's not even hard to to debate over it's like oh you want that sure now give me reduction of fentanyl or give me um you know uh stop tariffing auto imports by 100 like that's the stuff they really care about so this is an easy win uh but but there will be a uh like a separation is happening to where china clearly is embracing and winning open source, they will have comparable advantage, sorry, at least a reasonable alternative on Huawei probably in like 12 months, which will basically be like an H100-ish.

2:25:25And it'll be cheap, large scale, easy to install, all open source. So the next phase, if we want to actually win China, is what the Trump administration is doing now, which is trusted proliferation. The second phase is we've got to get serious about open source. and we have to be very intentional that the closed model thing is just not going to work. It's not a serious durable moat. I think most investors would probably agree with that now. It's a myriad of models approach, and we have to be in this game. Otherwise, it doesn't matter if we have the best infrastructure in the world, if the LLMs themselves are all behind closed doors, which is never going to reach this traction, which is never going to reach this level of adoption, and there are there are so many levels of advantage china has on open source right now uh that we have to be serious what are you expecting from deep deep seek r2 we've been we just asked the the prime intellect ceo that he had his own take what are you expecting i i think you'll be uh the rumor is like mostly on huawei i haven't validated that that's the rumor i keep hearing uh it probably will be very good uh because this is accessible this is mass it's open research it's uh it's like physics it's like it's like thinking that we can control physics like if physics is available to everybody uh to explore and discover uh i think it'll be very good i think i think it's a quinn 3.5 was very good uh and manis just had an update again it shows that i unless we were heavily invest with i guess you know gemini or or or llama to really like up their game that europe's gonna continue to use that the middle east continue to use it.

2:27:08And it just is what it is. Like we can't stop people from doing this stuff. It's available to everybody. Yeah. So, I mean, walk me through this hypothetical scenario where let's just use like Switzerland as an example or something like Switzerland wants to control their own AI for their own population. So they build SwissGPT.com and every Swiss citizen is going there as the front end to their AI experience. But that in one scenario, you know, that is run on air-gapped Huawei Ascend data centers running DeepSeek and Manus to actually do the computation. But the data is not being exfiltrated. What is the risk there?

2:27:51Is it just that we're losing the economic power from selling into Swiss, the Switzerland market? or are we worried about some sort of like Manchurian candidate in the weights shifting the minds of the Swiss population or something like that? Like what is the risk of the power accruing in that scenario? I mean, I don't really worry much about like everyone knows it happens in 1989. Like you don't need your TAM score happened. It's not them when DeepSeek tries to delete itself. Like everybody knows, right? It's very obvious. like you know is President Xi you know Winnie the Pooh would possibly be like you know we don't really know right he did he did disappear for six months after COVID so maybe he's Winnie the Pooh in a little outfit so yeah like the issue with what I believe about Sovereign AI is that most of the content internet is American English the estimates are like 70 % so that's most of the training data although if you look at the proprietary data what does that mean i just did it 70 percent of the internet is uh is american let's go we're just excited about that the the uh uh so so they basically like sovereign ai what is the data that they're going to be training on yeah and because if the internet is english american english it's our slang it's our you know it's our the second is spanish and the third is mandarin yep it's it's actually government documents that is how most people are training the governments are training their own subset sovereign ai legal cases if law right and none of that crap is going to amazon none of that crap is going to corey like they're they they're going to build that to their liking so then okay so what is the end end use case i one is that most of this is mutual destruction sort of frameworks that if switzerland doesn't do it then maybe austria does and if austria does it they have a larger switzerland so they're like well i can't have that so i have to do that think of it as um uh did you ever watch show billions yeah uh love that show so yeah so axelrod right everything he thought about was like both him it's like it's like two lanes of thought it's like him winning and him not being owned by one of his competitors and that's the way g gpus and ai works it's like it's like i think from the perspective like sure it was an advantage to my my country like i was doing the investment but most of it's just like i want to screw over that other person and and in investing like traders today they're doing gpus ai stuff not necessarily because necessarily there's a return today because they're trying to figure out is someone going to screw over me somehow because when you when you put something into an llm it's immediately uh uh everywhere right like anything now like private information that is even more valuable than ever before because lms make information so accessible and can substantiate content even easier and and so private information becomes significantly more valuable And so how Switzerland would deploy that, in my view, would be basically spying on other people to providing an alternative within its own government about how to actually use LLMs internally.

2:30:57And so I think we, as Americans, need to just kind of take a step back and realize that not everything the world does has to make money. People just kind of do stuff sometimes because they think they should do it. And we, as Americans, don't really get that. It's hard for us to kind of put our hands around. So commercial. Exactly. We're so far away from Reagan where we just built crap because we're like, Soviet Union could do this to us, so we're just going to do it. We're so far away and we're so dominant for so long that we're lazy. We're self-absorbed, lazy people who think that Pope Leo is a Chicago Democrat.

2:31:42I'm not a Catholic, but he's certainly not. and like, oh, he said this about immigration. And I'm like, do you know anything about Catholic doctrine? Like, do you read anything? Like, do you know, like, it's, what I mean by that was that we define so much international dialogue by America's framework when the world doesn't work that way. Yeah. Real quick, I want to have a couple more questions and then we have a pretty hard stop here. But who is going to be the American open source AI front runner? uh xai did an open release of grok one open ai is planning to have some strategy around open source uh do you have any type of you have any expectations here i i don't think it'd be open ai uh i think it's either going to be zuck elon or somebody else that we just don't know about you yeah cool uh last one uh do you expect to be satisfied when a tiktok deal gets announced or extremely disappointed because the rumors that we've heard so far are that uh it doesn't necessarily get at the core of the inference what's wrong with that the inference is the important part right yeah that's a good question uh so do i think it's propaganda oh absolutely do i think it matters in terms of the trajectory of our culture no like like creating blaming chinese boogeyman but like on the fact that our kids are lazy and stupid is asinine it's it's classic like white guilt sort of ducking freudian frameworks and like the reason why isn't it's actually you know like your kids will actually be fine they're they're leaning very heavily libertarian so they're actually good this kind of like late millennial because i'm gonna i'm a midlife millennial so i'm more gen xer so it's kind of like late millennial early next gen uh that are very marxist in their kind of like frameworks and they've been told things about there were objectively lies about our country about how like life works and success works and yeah i feel bad for them like because they were told things that were not true uh but like saying that oh the reason why you know chinese kids are doing math because tiktok like encourages math is is it's pretty stupid like it's like it's like no maybe you're just like no my my frustration was uh you know if you're if you're in the middle of an ideological war trade war you know sort of great power conflict we're not you know multi-polar is not possible it's just sort of like the u.s or or nothing uh why would you allow your adversary to own what is effectively like the one of the largest news channels into the u.s population and why would you kind of ignore that right because even even yeah getting uh you know delivering brain brain rot like americans are going to get brain rot from somewhere right and they have an insatiable but i'd prefer to be no i'd prefer to be american brain rot no but the but the you know you know these platforms do influence public opinion and, and, but anyways, probably a longer conversation.

2:35:00We'll have you back on when the news gets announced and, um, always a pleasure. Yeah. Have a good day. God bless guys. Cheers. Jordy. I want a hat. Oh, I got you. I got, I got you. You get a jacket too. You're going to get the full set. You're going to get the full set. Thanks Aaron. Great chat. Bye. Find your happy place. Find your happy place. Book a wonder with inspiring views, hotel-grade amenities, dreamy beds, top-tier cleaning, 24-7 concierge service. It's a vacation home, but better, folks. Also, head over to GetBezel.com. Your bezel concierge is available now to source any watch on the planet.

2:35:36Seriously, any watch. And our next guest is Dan Gray. I have been following him for just a few days, but I wanted to have him on the show. He's put out some fantastic posts about venture capital market dynamics, And I thought it'd be interesting to have him on the show and pick his brain and have him break down what he's seeing in the venture capital markets. I want to talk to him about large scale up VCs, the cram down in the angel and seed market, all the fun dynamics. So welcome to the show. Dan, if you're there, can you hear me? How are you doing? Long time listener, first time caller. Glad to be here.

2:36:11Yeah, great to have you. Can you take me through your... Actually, first, let's kick it off with just the introduction. Who are you? What do you do? Break it down for everyone. Sure. Yeah, I'm, you know, my kind of day job is head of insights at Equidem, where it starts at valuation platform. And that's kind of the perspective from which I come at all of this. Yep. I think what I'm most known for is like long walls of text on Twitter about the venture industry. And what was the most recent long wall of text that you put out? Yeah, that was kind of like a summary of the last year or so of observations.

2:36:47um cut like not a very optimistic title uh like the venture capital and ailing asset class addressing a lot of the challenges that we've seen yes but it moves towards like solutions you know better practices some ideas for how the industry can evolve so everybody comes out of it better you know lps founders gps so what are the key cracks why is venture capital ailing what are the key problems that you're seeing what are the indicators what are the metrics that you're tracking because a lot of people have been saying we're so back it looks really good there's big funds everyone's doing well everyone's making money it seems like it so this is kind of an odd take to see but open to hear the argument obvious yeah i think kind of what we see at the moment uh if you look very broad big picture is like stagflation there is i would say less activity than before but prices in concentrated areas are much higher and for a lot of people that's a very bad thing you know if you're a founder's fund and you can win on merit and get into any deal you want you're doing fine at the moment that's okay if you're andreason horowitz and you have the scale to to buy into any deal you're also doing fine but a lot of my writing is is actually really aimed at like the middle of the market gps who you know could be doing better maybe if they if like better practices were more accessible to them or if the market was a bit more stable.

2:38:10And similarly, LPs who have a difficult time picking, you know, if they're not already in like the founders funds and US fees, how they can be guided towards better decisions and better returns. So yeah, is this just affecting new managers who raised funds maybe during Zerp? Or are you seeing cracks in kind of the legacy firms that have dropped out of the public eye? They're not as competitive, but they've been in the game for a decade. They have a kind of an institutional structure. Maybe they're not raising multi-billion dollar funds, but they have a model where it's like, yeah,$300 million fund every two years.

2:38:49We deploy it. It's great business. All the partners are going to retire wealthy. Is there risk for that type of mid-market VC firm? I think the ones that are struggling the most at the moment are perhaps the ones that haven't evolved as quickly as the market has shifted in the last year. So we've seen this, this much talked about bifurcation into like what I call venture banks, like Andreessen Horowitz and Thrives, etc. And the kind of traditional model of VC, which is, you know, contrarian investing, a bit more disciplined, you could say on that front. There's a few people, a few funds, kind of stuck in the middle, you know, and they're a midsize series A firm who's still trying to invest in AI and they're having a very difficult time right now so you kind of have to pick your lane a little bit these days and lean into the strengths of what those two different lanes offer did you ever read Everett Randall's post on like the Tiger Global and the and this and like the squeezing between what was he calling it he had some great phrase for it playing different games why Tiger is eating your lunch and your deals it was kind of talking about how Tiger and the crossover investors came in and they were just commodity capital that you just show up with a deck and get a term sheet the same day.

2:40:11And it was very much like venture bank, uh, versus the, uh, well, uh, how did he describe it? He said the, uh, he described, uh, one of the funds is like JC Penney funds, which is very, uh, rude. That is the dead zone. So there's Tiffany and co, which he said was like benchmark founders fund, like the really like, uh, prestige Sequoia of the world, low velocity, higher IRR, main advantage is signaling, building a board of directors, but it's expensive capital. Then you have the Walmart, which was Tiger Global, higher velocity, lower IRR, but they're built to accept a lower IRR, although that didn't play out very well for Tiger specifically.

2:40:51But in theory, these venture banks are kind of playing this Walmart category. But the dead zone was the JCPenney, where you're not super differentiated, sending a really strong signal to the market that we have picked the winner. they're going to the moon and so get in or get out but you're also not super founder friendly not taking a board seat just dumping cash in a company is that dynamic still real in your mind has it grown or has to have things changed yeah i think it's very real it's definitely real and you can see that in a lot of the liquidity issues today so you have you know the big like venture banks let's call them yeah they're finding more opportunistic liquidity through through through secondaries for like the huge winners that they invest in who have the strength of brand and the size to generate a lot of liquidity in that manner.

2:41:38Where other VCs would traditionally get liquidity, IPOs and M &A, we all know has been very slow. And that slowdown for a range of reasons, but it's part of because the industry became so bloated and confused for a while. You know, hopefully as this bifurcation is better understood, those more traditional managers are like teeing up companies for exits in a healthier way. You know, they better understand the kind of exit metrics you need. Yeah, as opposed to the venture bank model. Do we need to kind of re redefine these terms of venture capital? Because it used to be there was like angel investors, then venture capitalists that would do series days and maybe series Bs.

2:42:17And then there was growth equity, private equity. And then like, yes, a hedge fund might take a large position in a late stage firm, but they weren't seen as VCs. Now you can have a VC firm that I'm pretty sure this venture capital firms that have never invested in a company at less than a$1 billion valuation, because their whole business is maybe we buy secondary or we buy common, or we do growth stage. And we're really only trying to be along for the unicorn to$100 billion ride. And we'll do that all day long because there's so little downside. And by the time we're investing, The company has durable revenues, is growing, product market fit.

2:42:54So we're not even trying to underwrite the founder or their ability to explore the idea maze. We're just piling in. But we still call that venture capital. And maybe that's, is that a problem? Or do we need to think about the market differently? I think it's definitely helpful to think about it as a slightly separate beast. You know, venture capital, in my mind, is finding the non-consensus companies. and then you fund them through into like legibility, into growth, growth capital, and ultimately to an exit. That's how you generate your liquidity. The new game today is kind of like this financialized version of venture capital where it's not necessarily about like finding great companies and taking them to exit.

2:43:34It's about like funding through the like the incremental milestones and hitting those incremental metrics to raise bigger funds. You know, and you do that through pursuing TVPI in a way that like it's slightly disconnected to generating value in the traditional sense. You know, TVPI has slightly different drivers, you could say. Yeah. Do you think that part of the reclassification as somewhat other asset classes into venture capital is driven by a desire to mimic the fee structure of a traditional venture capital firm in other industries? So I'm basically doing public markets investing and I might be able to get one in 10.

2:44:17But if I call myself a VC firm and I play a little bit differently, I get two in 20 or something like that. Do you think that the branding matters when VC firms go out to LPs and try and raise money? That's a good question. I'm not really sure about that, to be honest. I mean, I'd hope LPs were scrutinizing that a little bit more, but perhaps not. uh what do you do you look at any data on the health of the sort of retail in the private markets right like the the everyday angel investor are you seeing more or less activity today than in 2021 early 2022 i would say quite a bit less probably like a third less at the moment because so much attention is concentrated on ai and i think a lot of angels frankly have just resigned themselves like they're not going to get in on that game yeah interesting yeah i was thinking i was uh i was even going to consider doing a poll uh to anybody within you know just a poll on x or whatever but i think if you ask the average angel investor would you sell your entire angel investing portfolio for 15 cents on the dollar today and i bet you that i bet you the honest answer is that like 90 % of, uh, of people would because angel investing is this like weird dynamic where to me, I, I legitimately feel addicted to angel investing.

2:45:43Like, like it's, it's a weird thing, but, and I, and I joke about this, right. It's like all these kids, you know, these kids that grew up in San Francisco and their, and their parents are addicted to angel investing, you know, just like refreshing their, refreshing their email on the first of the month, you know waiting waiting for waiting for the update um but it's but it's this interesting you know dynamic where it's honestly one of you know to me it's one of the most interesting ways to spend free time it's just like meeting exciting founders but at the same time liquidity timeline it's so long literally if you make money it's like well i didn't get to enjoy that in my 30s yeah yeah totally maybe it's and i have i have buddies that are that are towards you know the end of their careers or maybe even retired that are basically like somebody recently said to me, I don't want to make net new angel investments because I don't want to deal with this when I'm like potentially in a retirement, potentially like, you know, retired to Hawaii, right?

2:46:41I don't want to be worried about, you know, figuring out, oh, when, you know, when should I try to exit this position? Can you break down this idea that most venture capitalists don't understand risk management and only 10 % of venture firms are making it to fund four, what do VCs get wrong about risk management? What does risk management even mean? I thought the whole point was to be risk on, swing for the fences. That's a great question. It's a real fundamental point as well. I talk about the importance of risk management in venture capital and people do give me a slightly crazy look, like the point is risk.

2:47:17But if you properly manage risk in a systematic way, you can take more risky bets. Like that's kind of the point. Like if you understand the maths of portfolio construction, how big your checks are going to be, you know, what the outcome is likely to be, how many you have to write, what maximizes your exposure to potential unicorns, you know, what discipline across the whole portfolio looks like, you can take more risk on an individual investment basis. And yes, so what does a poorly risk managed fund look like? Is it just not building concentration in the winners, not leaving money in the fund for pro rata and follow-ons to kind of ride the winners up as they become outsized winners?

2:48:01I feel like that was something that we heard from like the last generation of managers that got a$50 million fund. Then everything got marked up in Zerp. And all of a sudden they had to make these decisions on huge pro rata checks out of their main fund where there really was no new data about the company and they were still at a seed level. but some big firm would come in and preempted a series a or series b and so all of a sudden you can't do the normal thing of wow there's so much traction on this company i need to really make sure my fund has a significant ownership because i did get a foothold in but i need to build that position is that what's going on or is there something else some another dynamic yeah that's a part of it that's quite a big part of it um i think like the fundamental thing is the having a right sized initial portfolio, the number of investments that you make to begin with, that should be diversified enough.

2:48:52And, you know, the average might be like 25 today, maybe it should be 50. How do you even calculate that? Is that based on like historical trends in like the number of unicorns or something? Like, how are you thinking about that? Yeah, exactly. There's great work done by a few people. Dave McClure from 500 Startups, he has an article on this. He talks about, you know, 100 startups should be the portfolio size. for a seed fund. Wow. Because of the small percentage that eventually become unicorns or deck corns. Yeah. And you fundamentally don't know, to begin with, which which ones are going to work out.

2:49:28That's kind of the problem is VCs over concentrate. And of course, if you're, you know, I've spoken to some great GPs about this concept, and they always push back because they have 15 companies, and they're all doing great. And that's like, fine, good for you, you're in the top 2%. It's the other 98 % that I'm kind of thinking about. But is that really a way to get to like, by definition, the don't you to be in the top 2%, don't you have to behave like the top 2 % and just actually like consider it a skill issue? And just make sure if you're trying to if you're trying to get, you know, I can see the argument if you if you're trying to get past fund four, which is like the milestone, you need to have bangers in the first one, right?

2:50:09Like, even if even if you're not necessarily even if your ownership is not amazing you need to be able to prove that you can get into winners at all to some degree and so i've seen that strategy play out fairly nicely where as long as you're in a handful you know just get showing that you can get in is kind of like key proof points even if you're not able to concentrate you know five plus percent of your of your fund in a single company yeah do you think do you think lps are more willing to forgive on mistakes of portfolio construction, not sizing into winners easier than missing winners and having like, okay, yeah, you built the portfolio perfectly.

2:50:52You got concentrated into your winners properly, but you didn't have any of those really sexy companies in the portfolio. Well, I'd like to think ultimately what matters is, you know, the end multiple, the IRR, the performance of the fund and everyone you know the majority of emerging managers come out swinging for the fences every single time they want a 10x fund and in doing so they are like are themselves boom or bust if they fail and it's less than returning capital they have no room to learn they can't improve for the next fund you know there's there's no step two if they are a bit more conservative, a bit more diversified, and they return like 2.53x, maybe they get a fund too, and they learn and they can improve and earn concentration over time.

2:51:40Isn't there something beautiful about the fact that both startups exhibit a follow a power law and venture capital firms exhibit a power law? Like there's a kinship here that it's like, hey, we're all either going to be rich or poor here. There's no median outcome. You're not just a bank to me uh you're not going to be around unless this works out you're living under the power law just as much as i am as the entrepreneur i would disagree with that to be honest with you please i i think it's a better world for everybody if the vcs are more stable okay you know they they they are very power law they are boom and bust they don't really have to be interesting interesting well yeah this is great this was great thank you for coming on and sharing uh next time you have report give us a heads up and uh love to have you back on yeah yeah we'll talk to you dan this is great cheers thank you uh let's run through some uh timeline let's give some updates whatever else is in the news i've got an idea before we jump into the timeline should we talk about figma of course figma is the design tool of my dreams it's a tool that that i've used for my entire career at this point.

2:52:50Think bigger, build faster. That's great. We should hit an air horn for that. How you design, explore your ideas freely and iterate quickly. We should make an actual old school style radio ad for this. But anyways, Figma launched a bunch of new features last week at Config, one of which is very cool. They allow you to generate marketing various types of assets based off of your brand book. It's called Figma Buzz. So exciting new feature from them alongside sites and kind of the core design tools that you probably already use and love. So go check out Figma Buzz and sites and thank you to Figma.

2:53:32Thank you to Figma. Should we talk about some timeline? Let's do it. Jason Calacanis of the All In Podcast is taking the movie industry to task. He says the movie industry is effing clueless. make a$19 a month all you can watch$10 plus$10 for each kid and$49 for a family of five watch the money pour in he wants a subscription movies would be huge if you made it a season pass like the ski industry did with the epic pass show old movies too for bonus points amc will slash prices on wednesday is the news by 50 starting in july the goal is to get more people in theaters between weekends what's interesting is that i saw a separate article about how uh the modern instantiation of a movie pass is wildly successful have you heard of this the yeah i think it's like well yeah apparently amc has a subscription product already they do i think the interesting thing that jason called out is showing old movies i think one of the challenges is that there's so few new movies coming out that people are genuinely you know wanting to get up and go see.

2:54:38I don't know if old movies would do it, but... They do re-releases all the time and they are available and I don't think they build as much hype because I think there's some weird dynamic in the movie industry where like, unless you see the billboard and the hype and like you're texting your voice. I think it comes down to scarcity. They need to pull movies off the internet entirely for a year and then just build it up and being like, we're bringing it back. you know the people some people will be like well i have a dvd i can watch it anytime good luck playing a dvd on your home entertainment system if you got it in the last five years not going to work um anyway staying with the all-in podcast chamath palihapitiya says not nearly enough people are talking about the implications of klarna rolling back some of their ai bets not knowing any of the details i can guess why replacing determinism of you or humans with probabilistic code is fraught with edge cases and require new ways of software development and process engineering that aren't well solved.

2:55:36The implications to an entire generation of AI apps will be as severe, will be severe as more companies come to terms with the difficulty in getting products to work reliably in production with AI in the loop. Customer service may be the first funeral signpost. The result will be that many startups need to pivot to simply use AI for narrow use cases and otherwise act deterministically. So what will, what will, so what have people funded then an AI company? No, not really. Just a really overpriced SaaS company at AI valuations. Interesting. What do you think? I mean, I thought the conversation with Owen yesterday from Intercom was super insightful, right?

2:56:15They are incredibly well positioned to deliver AI to CX teams and unlock the power of it, but they're not saying, he wasn't coming on and saying 99 % of tickets are are solved by AI today. He believes that we can get there over time. And his hot take was that AI was better than humans already at a lot of things, right? It stays on script, it doesn't get tired, it doesn't get short, it's not temperamental. All these things that end up creating a better customer experience once the quality is at a certain point. It's this idea that we overestimate what can be done in a year, underestimate what can be done in a decade.

2:56:57you got superhuman autocomplete for every single one of your customer service agents. You got much better ways to navigate a knowledge base effectively and just serve up a better Google search, use it as a knowledge engine. And so that should drive those, but trying to go full Klarna mode and saying, you know, we're not hiring anyone. We're going to be a one person company, probably a little aggressive. Anyway, Avi Schiffman, founder of friend.com says, I met my investors LPs and realized I'm just working for college endowments and family offices. Activate golden retriever. Yeah, you are. And it's good.

2:57:36It's an opportunity. It's an amazing blessing. I mean, we heard about this from Andrew Reed talking about when he was pitching Dylan Field at Figma. I brought him into the Brown conference room and said, hey, we work for Brown. You have an opportunity to run it up for Brown. Yeah, you have an opportunity to run it up for Brown. Exactly. And it's not just college endowments. It's teacher pension funds. It's, you know, you're supporting firefighters, all these things. So you're supporting, you're helping big oil, right? I mean, that is part of the beauty of like the intertwined financial system is that losses are born across the entire economy.

2:58:19So that any one company that blows up doesn't destroy everyone's pension. um and but but when there's big winners everyone kind of benefits if you have a pension or you go to one of these schools like their endowment is enriched and they do more research and science and all sorts of stuff and so you you have kind of a uh um uh i mean it's just it's diversification right it's just diversification and so uh like you you don't hear as many of those stories about an individual main street investor that lost all their money on a risky startup bet because angel investing is, while addictive, it is for high net worth individuals and accredited investors only, but they still have exposure to the venture markets through LPs.

2:59:02Prime Video has renewed Beast Games for not one, but two more seasons. Y 'all are not ready for the big stuff we have planned, says Mr. Beast. Very interesting. This was one that was like hotly debated about whether it was good. It's hard to track because you can't just look at the YouTube metrics to understand if Beast Games is truly successful for Amazon Prime. But we have the data now, season two and season three confirmed. They probably wouldn't have done that if it underperformed. So interesting watching a YouTuber transition to mainstream media or bigger platforms. I mean, Amazon Prime's new media, but still a huge company, major network.

2:59:37He's clearly worked out a fantastic deal. And we've seen a bunch of this. And it feels like it's more of a increasingly, social media is a proving ground for Hollywood. Talk about Pat McAfee, what he did with his podcast, took it to ESPN eventually. I think we're gonna be seeing a lot more of this as people build independently and then ultimately leverage the distribution that comes from something like Amazon Prime. So congrats to Jimmy. We should have on the show. Dalton Caldwell says, the S &P 500 now has three YC companies, DoorDash, Airbnb, and Coinbase. Let's hear it for YC. Can we get a size gong or some clapping or something?

3:00:17congratulations i can't wait to go back to the next demo day yeah me either uh it is it is actually crazy how big that yc has done this because no other incubators even close right and i'm excited for the update in in 2055 the s &p 500 is now 60 percent yc company for sure uh certainly uh if you expand that to yc founders running companies probably says 497 to go 497 to go pg i like it on a conquest good job uh and all are consumer tech yeah right everybody says build b2b sass until you see some of the biggest outcomes just wait i mean a couple of those payroll companies will be in the s &p 500 pretty soon i'm pretty sure anyway all three hopefully uh aiden says uh uber would be a lot cooler if it was called american express i i just love this guy seven likes low tam but low tam banger yeah he also had another post that we didn't get to yesterday that was like, if the airline's calling your name personally as you board the flight, it's basically like flying private.

3:01:21I thought that was funny. Anyway, shout out to Aiden. Great programmer, great hacker. Airbnb is launching, they're unveiling their next chapter. This is going live today. I don't actually have the update, but Brian Chesky has been in the founder mode cycle. So they're launching new services, businesses. Oh, interesting. There's a redesigned app. So remember they condensed a bunch of launches into one. Yep. And they were doing Airbnb experiences for a while. Yeah, expanding beyond homestays. And I guess they're going to invest a quarter of a billion dollars into this new sort of business line. Wow, big money.

3:01:57It seems like an extension of their kind of experiences products. Sure. But anyways, hopefully they can. hopefully they don't crush under the weight of competition with Wander yeah honestly that's my only hope because I love Brian Chesky the rumors of the Wander series B founder mode CEO love him but they're going to be feeling the heat they're going to be feeling the heat from Wander for sure Adam D 'Angelo said at Quora we recently tested switching from Zoom to Google Meet for a week Google Meet is better in many small ways but a bit worse in one big way audio quality particularly background audio noise cancellation echoing that kills it so we're staying on zoom those are funny just like hey i did this uh av test and we're sticking to we're sticking to zoom and uh sundar pachai responds says founder mode we'll follow up offline as we haven't experienced this as read notes below it works pretty well in my experience we will debug to understand the root cause and fix thanks for flagging thank you would love to switch if this is fixed uh and nathan a buddy of mine says i have the inverse experience zoom is well rubbish in it he's british or french maybe i don't know he's over in europe somewhere i don't really know the specifics of that rubbish anyway in it hilarious um anyway uh teddy blank had a coinage uh once your vc has uh what was it the actual coinage was something about once your once your uh company's linkedin profile picture is the founder holding a microphone at a conference you're caught with one of the yeah with one of those which I don't think is really true necessarily because a lot of the big CEOs, they wind up doing that anyway.

3:03:34Bezos, Steve Jobs. I'm sure all these guys have been pictured with lab mics, but still funny to see that, yeah, if you grew up too much. So anyway, shout out Teddy. Anything else you want to cover? Should we get out of here? It's 205. Thank you for watching. 205. Any more plugs? Got Aurora.com. Go to Lucy.com. You should always go there. Go to Aidsley. But no, I'm excited for the show tomorrow. We've got Brian Johnson calling in, as well as Mateo from 8 Sleep. That should be great. And yeah, go and leave us a five-star review on Apple Podcasts or Spotify. It helps us. I think it helps us. I don't know what it does.

3:04:14But if you do leave a review, we will talk about it on the show. And head over to Polymarket. Make sure to put an ad in it. Keep tracking the largest company at the end of May. This is the Super Bowl of big tech. Microsoft's at 60 % right now, but they dropped just recently down from 80%. This chart is actually - NVIDIA is climbing fast. Apple's climbing as well. It's a knockout, dragout fight for the largest company at the end of May. Head over to Pollock and check it out. Anyway, thank you for watching. The only market that matters. We will talk to you soon. Cheers. See you tomorrow. Goodbye.

3:04:49There's a question about whether it may not to be Thank you.