The Exchanges

Q Journal has a story. Senate passes Trump's mega bill bill after all night session tax cuts and Medicaid spending reductions head to the house. With a July fourth deadline reachable, what's interesting in the bill? How likely do you think it is to pass? What are the major reads on this? And then we want to get your, your takeaway from like where the dividing lines are around the bill.

A Yeah, so there's a couple trip lines. Uh, the first trip line, I would say, you guys have been tracking the moratorium, the ten-year moratorium on AI development. This was the, on AI regulation, this was the first it was in, Ted Cruz put it in, then it was out, then it was back in, there was a compromise, there was a five-year agreement instead of a ten-year agreement, and Blackburn has been sort of the key driver of getting it back out again. It's out, now it was done, 99 to one in the Senate, almost unanimous in the Senate, To get this out. This was a big win for Dario of Anthropic, probably the number one beneficiary. And by the way, somebody who certainly used a fair bit of political capital pushing for this, that's fault line. Number one, I would say fault line. Number two.

AI assessment note: “there's a couple trip lines. Uh, the first trip line, I would say”

Q the Democrats feeling about EVs these days? Because it used to, I used to feel like it was a democratic cause to be pro EV incentive. Let's save the environment. Let's get people out of the gas cars. I'm joking about the V 12 obviously, but, um, but now it's become, Maybe let's punish Elon, but is this a cut your nose to spite your face situation for the Democrats?

A I think it's a rough spot. I mean, you guys remember when Elon wasn't invited to the big EV summit under the Biden administration, right? That's part of what kicked off his political awakening is he isn't, he doesn't use unionized labor. And so he was excluded from this EV summit. There has been a huge polarization and negative polarization against EVs among not only Dems in office, but the democratic base too, which is to all the points you brought up quite ironic. You're not going to find many Dems willing to cross the aisle in support of electric vehicles to save this bill. This is a bill that will live or die by Republican whip count and Republican unity. If the whip count isn't there, it's going to be really hard to get it done before July fourth.

AI assessment note: “There has been a huge polarization and negative polarization against EVs among not only Dems”

Q Uh, break it down for us. Uh, for those who haven't watched the full one hour video, we'll obviously link it here, but, uh, what are the key announcements? What do you think it's important? What do you think is important to share?

A Sure. Yeah, so the 10 second summary of our update is, um, that Neuralink is working reliably and has already changed the lives of seven human participants that we've been working with. It's, it's really been a privilege working with, uh, all those participants individually. And that, um, you know, we sort of lay down our next set of milestones, which is most immediately going to market and enabling the scaling of this technology to thousands of people. Eventually millions. And also to just go beyond, uh, the movement and expand functionalities into speech, vision, and hopefully getting to the speed of inner thoughts. And, and then there's a little bit of a hint as to what the, the path to, uh, kind of the, the ultimate goal of Neuralink would be, which is to understand and unlock the mystery of the mind.

AI assessment note: “the 10 second summary of our update is, um, that Neuralink is working reliably”

Q Yeah. Can you talk a little bit about the N, uh, the NVIDIA announcement today? That's very exciting. Uh, seems like it might be a little bit of like a side quest or, or is this like in the critical path to, you know, mass, uh, mass production?

A Yeah, no, NVIDIA's today announcement was, was kind of a little fun thing we did. NVIDIA is an investor in us. So fundamentally, they are very interested in what we are trying to do, you know, being able to capture data from physical phenomenon and build models that can manipulate the physical world. Um, but I think we work with their artists in residence. It's actually those open AI artists in residence that we work with NVIDIA collaboratively to really turn in just the artist speaking to a system into a piece of art, right? So Alex, the artist that will work with us, Basically spoke what he wanted to build, what sculpture he wanted to build, and the full stack of generating the code, running the robots, all were done autonomously. So from speech, from intent, all the way to the physical part, without anybody ever touching anything.

AI assessment note: “NVIDIA's today announcement was, was kind of a little fun thing we did.”

Q are in the world right now? What is the, what does the supply chain look like in terms of constructing, building, R&D-ing new robots? Uh, who are the major players globally? Who are the up-and-comers? I, I, I don't even know if I could name a single robot maker, um, but I'm interested to hear kind of like what the landscape looks like, who the, who the major players are.

A Yeah. So the, the biggest robot companies in the world are all not American, right? So FANUC is Japanese. They're, they're one of the biggest ones. Uh, KUKA used to be a German company. It got acquired by a Chinese company, uh, ABB, um, uh, Yaskawa, uh, you know, Mitsubishi, you know, there's a bunch of these kind of very, very large industrial companies that make robot arms. Uh, the, the thing to note though, is that like buying a robot arms is effectively kind of useless on its own. Uh, you need a full robot work cell to make that work. So you need PLCs that often come from like Siemens or Rockwell. You need sensors that come from. What's a PLC? Uh, a programmable logic controller. So it's like the, it's the thing that controls all the peripherals and safety equipment. Um, yeah. And then, uh, you need like conveyors, you need safety scanners that are often made by like Keyence and Cognex and which other company you need grippers, which is the end of arm tool. Uh, so, uh, today, like the vast majority of people who deploy robots basically hire a consultant that goes and buys all the different components for them and then assembles it and then deploys it for them.

AI assessment note: “FANUC is Japanese. They're, they're one of the biggest ones. Uh, KUKA used to be a German company”

Q buildup, tech boom, then we get COVID, we get SVB, and we're up and down and up and down, now we're having tariffs and, and Trump pump and Trump dump, um, Does it feel like we're moving into a permanently new regime of higher volatility because of some of the, some of the, ah, factors that you mentioned, or is this temporary and maybe it's smooth sailing in the future?

A Well, I think you, what you mentioned there that's interesting is you feel like it's been like that for the last 20 years. I would say it's been like that for the last 200 years. Whenever we have market data, it's always been like that. So yes, what you just described for the last 20 years is accurate. Before that, there was the, the SNL crisis in the 19 eighties, the inflation in the 19 seventies, there was a boom, the like the sixties were a wild boom period. World War II, the Great Depression, like it's always been nuts. There's never been a period Of smooth sailing and the periods that we associate with smooth sailing, like the sixties and the nineties, we know in hindsight, we're just the precursors to a giant bust that happened. So every time in hindsight that we're like, man, why couldn't it have been like that? Like actually that time was a terrible time in hindsight. That was what you wanted to avoid. And so it's never going to be a case that it's more that it's, it's, it's smooth sailing. Like it's an inherent feature of capitalism that you have volatility and booms and busts. It's not fun. It's not enjoyable, but the, Absence of that is even worse. I wrote about this in my first book. There's a great economist guy named Hyman Minsky, who I'll say very quickly came up with this idea called the financial instability hypothesis, which was the more stable the economy is,…

AI assessment note: “it's an inherent feature of capitalism that you have volatility and booms and busts”

Q excited about? I'm sure you get pitched, uh, pretty much every company in the space. If they're not pitching you, uh, what are they even doing? What, what areas do you think are, are actually getting, um, maybe not, uh, You know, I w I would say like consumer adoption, but like real enterprise adoption, you obviously worked in this, in this problem space for, for quite a long time.

A Yeah. I mean, the, the, the take I have here is that legacy healthcare companies are actually going to be the winners during this wave. I think as you see kind of AI adoption happen in healthcare, you're starting to see widgets get adopted, right? Doctors requesting ambient scribes. You see big companies like a bridge coming out of that. But ultimately ambient scribes are the front door to the electronic medical record system. They're certainly not the entire suite of software that a physician might use. And unlike the world we're used to in tech, the person who purchased the software and the person that uses software and healthcare are not the same person, right? Physicians are using it. Hospital admins procuring it. There's a big divide there, which means they have different competing priorities. I am radically more excited by the opportunity of big legacy healthcare Software businesses taking modern AI innovation, installing it onto their user base, and scaling that out than I am any kind of novel net new startup. Like, I think we will see a thousand kind of AI enabled executive physical startups happen over the next six months. I don't think any of them are going to last. I think we'll see a thousand pill mills that will spin up over the next six months. I think they're all going to jail, right? Healthcare is just not a place that rewards this kind of net new novel distribu…

AI assessment note: “legacy healthcare companies are actually going to be the winners during this wave”

Q The Salton Sea. Uh, that, that was a really cool case study. Uh, are there other kind of like, uh, Like, long-term projects that you think, uh, are kind of on the bucket list that you want to make amazing one day?

A Yeah, the Great Salt Lake is in a deficit of about 500,000 acre feet of water per year. Um, the Bear River is the primary tributary into the Great Salt Lake, and it has mountains with lots of juicy clouds from Utah to Wyoming to Idaho, all that flow into it. Um, I think that restoring the Great Salt Lake is like an American priority and a must, and even more interestingly than that, something in the distant future for Rainmaker is, um, the Winter Olympics in 2034 are going to be held in Salt Lake, and China In 2008 had this whole big show and dance where they made it rain before the parades. So it didn't end on the Olympic games. We want to show a force a decade from now to make more fresh power for all the Olympic games.

AI assessment note: “Yeah, the Great Salt Lake is in a deficit of about 500,000 acre feet”

Q And let's not, let's not take, take away their Their desire to go political. They enjoy the culture war, right? Like those guys like duking it out on the timeline with the Taylor Lorenz's of the world. And they just enjoy it. It's like a sport to them.

A And it helps them get bigger and bigger and bigger. Like when J Cal's posting, like we're the number one podcast, he's not even talking about like the number one podcast in technology anymore. Like he will have episodes, like the all in show will have episodes. They're like the number one podcast in all the categories. And it's because it appeals to Everyone. If they were just sticking to deep dives on, oh, what AI agent is the best, or what startup got funded, like, they're not gonna, they're not gonna go broad, and they want to go bigger, and bigger, and bigger, and this is very natural, and it creates an opening for new shows to come in. This is what happened with Lex Friedman. He used to be the artificial intelligence podcast. It was all AI researchers. Then eventually he broadened out, started having political people on, started having You know, Jocko Willink on and Rogan and all sorts of different people on health stuff. And that created the opportunity for Dwarkesh who's doing fantastically. And now Dwarkesh is doing Tony Blair and eventually Dwarkesh will broaden out to all sorts of different topics that he's interested in. He won't just be the AI researcher show and then there'll be a new Dwarkesh and that's great. And it's this virtuous cycle that should not be, it shouldn't be criticized. It's just, it's a positive thing.

AI assessment note: “And it helps them get bigger and bigger and bigger.”

Q influencer, and basically the first person that came and pitched him on doing a supplement company, he was just like, sure, that sounds great. The business does like, you know, millions of dollars a year in revenue now. So you can just do things, you can just pitch things, right? And these guys, like, you know, had they waited even another few weeks, maybe the opportunity never would have happened.

A Yeah, yeah, a hundred percent. It's not like these guys were like ex-Facebook, ex-Twitter, like really established in social media. They just had the connection to Trump through the show. Clearly, like, a lot of energy, and you'll see that throughout this, this story. Like, the, the dates, we'll go over the dates, and it's, like, major milestone every, like, three months. Like, the speed of execution here is, is very remarkable. Faster than what I would think would be possible, given the scale of the operation. Um, so they go to, they go to Mar-a-Lago, they, they pitch it, and they kind of break down this idea of this, Truth Social, new media offering. They're talking about a streaming service, online payments network, a social media platform, and can you guess what the deal terms were? This is, this is fascinating to me because, like, we know a lot of companies that have done deals with celebrities, and it ranges all the way from like, oh, we just paid them a CPM deal, and they posted on their Instagram, to, oh, we gave them a couple points of equity, to like these fifty-fifty deals that happen every once in a while. I, I, I know some people that just did one of those. Um, in this case, Trump got 90% of the new company.

AI assessment note: “Yeah, yeah, a hundred percent. It's not like these guys were like ex-Facebook”

Q All right, let's get to the punishment. What, what's been your reaction to Kawhi got a 700,000 dollar fine, doesn't feel very significant, at least, you know, compared to the, the actual payments and what was received, and then, uh, you have a, a one-year ban, uh, for, for Balmer, and then, but, but, What was your reaction to the, to the, to the verdict from the NBA?

A Yeah, um, this was the biggest punishment for an owner in the history of, I think, American professional sports, and the NBA took five first-round picks, which is the most you could imagine. They took thirty million dollars, which is a rounding error, of course, for Ballmer. They made him pay the fifty million dollar legal fee to Wachtell Lipton, the NBA's outside counsel, which is also a rounding error, but still, you know, Pretty annoying, I would imagine. Um, but the real thing that I thought was meaningful was that they banned Steve Ballmer from his own building for a year. He loves basketball. He sits courtside. He measured literally the toilets at the Intuit Dome when he was designing the whole building. He was a micromanager who knew everything about every piece of the building, and they banned him from being inside his own building, which he funded for two billion dollars privately to his credit, and they also banned his president of business for a year, and they banned his GM, his president of basketball, for six months, and so Kawhi got Effectively a slap on the wrist, and that is wild on the merits. The reason they did it briefly is because when Ballmer now is rattling his saber, because he is the former CEO of Microsoft, who when presented with antitrust findings by the literal federal government, also fought, right? This is unsurprising if you know the trajectory o…

AI assessment note: “this was the biggest punishment for an owner in the history of”

Q I don't know. I think it's also very much this, this mean people love to talk about on X. Yeah. They like to kind of, you know, scapegoat. Oh, you know, it's all distillation. That's the only reason Chinese models are good. Yep. Is that actually true?

A Uh, I mean, Anthropics Head of National Security Policy, Taran Chabra, accused Zipu, Z.ai, uh, of distilling both Claude and OpenAI models for GLM-Five. Two at the Aspen Security Forum, Earlier, uh, this week, this is from Vincent Chow, senior AI reporter at SCMP. He said it's the first time that they've named Zipu specifically after previously calling out Deep Seek, Alibaba, Moonshot, and Minimax, uh, join the, join the club at this point. Um, also accused, they also accused Deep Seek of continuing its adversarial campaign of distillation. Anthropic is now shutting down distillation accounts on the order of millions accounts of per week. That is crazy scale. You have to, I, I mean, you always think about it as like, oh, there's like shut down that one company or shut down that one block of IP addresses. But when there's a really, really distributed attack, we've even heard about whole companies that just like resell Claude tokens or, or GPT's 5.6 tokens. And that looks like a reasonable business because it's just a wrapper company. Of course you want to work with them, but then you don't realize that on the other side, who are their customers? Why do they, why did they get to a hundred million run rates so quickly? Well, maybe it's a lab that's trying to distill through this pass through entity. And of course, it's hard to like watermark the tokens once they go out the API and…

AI assessment note: “Taran Chabra, accused Zipu, Z.ai, uh, of distilling both Claude and OpenAI models”

Q living environment, because, I mean, from those first YouTube videos, like, the monastic, what were you, closet? You were in a closet? Like, like, that, that was, I don't know if that was intentional, but it was, uh, it was somewhat deliberate. You picked that space over a big living room, right? Uh, and I'm wondering how you think about, Your, your living space, your workspace, your office space today.

A So I would say all of the time that I, when I work, I'm very intentional about that space. When I'm not working, it's whatever Layla wants. I don't care. Like if she, if she's 10% happier, I have like a percent to percent correlation to my happiness and me being more comfortable has no correlation back. So like whatever she's good with, I'm good with. Uh, but in terms of like, Where I work, it's almost all, how do I, how do I eliminate every sense of distraction for me personally? And so it's like, I don't have windows. Some people like having windows. I don't think anything wrong with that for me. It's like, I need no stimulation because I'm, I'm very distractible. And so I have to have like, I earplugs in no light, all artificial. I have like soundproof, like, you know, like I want nothing.

AI assessment note: “when I work, I'm very intentional about that space. When I'm not working, it's whatever Layla wants.”

Q Talk about subscriptions, bundles, uh, subscription pricing in a time when we have little Spurts of inflation here and there. Um, how important has that been? How resilient has the subscriber model been? What are you seeing there?

A You know, it's, um, it's a, it's a very important part of our revenue stream, but, you know, our digital subscriptions grew 29% last year, revenue, and, um, you know, they're growing double digit percentages this year, so it's a really important growth area of our business, and we're launching more digital subscriptions for more brands, like Pitchfork, a small brand, just launched a subscription earlier this year. Tatler, another small brand in the UK, launched it, but, you know, our big brands, The New Yorker, You know, very, very strong growth. Vogue is showing incredible growth in digital subscriptions. So that's, that's an area that's important to impress, and we think we've built up some really good capabilities, both on the technology side, but then also on just the people capabilities side, too.

AI assessment note: “our digital subscriptions grew 29% last year, revenue, and, um, you know, they're growing double digit”

Q Yeah, what was the process of getting out of the mailroom? I imagine that you have, you know, a few really iconic stories from the mailroom, but There were plenty of days that were just photocopying. Is that roughly correct?

A Yeah. I mean, we did, we, we did shopping for groceries. We, we copied scripts, we delivered scripts. So there was a whole set of kind of stories on that. And honestly, we could, we could fill a lot of podcasts with just the different things that we kind of did, but look, it re it rewarded hard work and rewarded attention to detail. Um, ultimately, and you got invited to the retreats, right? And CA did these retreats every year. And after one of them, I had some thoughts, I wrote them down, um, And I hand delivered a letter to the managing partner saying, I was at the retreat and here are my thoughts, right? I wish I still had it. I don't know what I said. And then I didn't hear anything for a long time, but when I came off, ah, what was called the runs, where you delivered scripts, you, um, then waited to be picked for a desk. And only the top five got to interview for desks to try and not keep people jammed there too long. But the desk of Brian Lord opened up, and he was the co-chairman and is now the CEO. And he asked to interview me, because he had read my memo, and he said, I want to interview the kid from the memo, even though he's not in the top five, I don't care. So I went up, and, uh, I'd never really met him or spent any time with him before, but we sat down, and, um, somehow we got talking to about John Steinbeck, which, uh, was my favorite writer, and I was reading…

AI assessment note: “when I came off, ah, what was called the runs... waited to be picked”

Q in case this is fraud, it was more than six years ago, so past the statute of limitations, can't get me. On second thought, I need to atone for my sins and repaid Chipotle the 21 dollars I stole from them by purchasing 21 dollars of their stock. I am sorry, Chipotle. Another resignation has hit the timeline. The co-founder of XAI is out. You want to read through it?

A Yeah, so he says, I resigned from XAI today. This company and the family we became will stay with me forever. I will deeply miss the people, the war rooms, and all those battles we have fought together. It's time for my next chapter. It's an era with full possibilities. A small team armed with AIs can move mountains and redefine what's possible. Thank you to the entire XAI family. And Elon, thank you for believing in the mission and for the ride of a lifetime. Some of the executives at XAI were offered cash instead of staying around, and so it's possible he just said, cool, time to, uh, time to move on. This isn't looking good, and it's the original XAI co-founders. There was one, two, three, four, five, Five, six, seven, eight. There's three remaining. Elon plus Manuel and Toby.

AI assessment note: “Yeah, so he says, I resigned from XAI today.”

Q So you didn't, you, you, you learned, uh, the things not to do from FTX. Uh, you learned, you were inspired by Satoshi. Was there any company, uh, off chain, uh, that, that was particularly inspiring?

A Yeah, a bunch. I mean, I think every big company, every big tech company, I think is like a very inspirational to me. Uh, I, I grew up in the Bay area, so it's, it's, it's hard to not kind of be sort of motivated by that. Uh, I think Amazon was a really big one that what inspired me was that They, they basically like, they're the very first principles and like driven, but also, and also very practical. And I think when sort of, I assume it was Bezos, but maybe someone else in the org realized that they had built so much of the internet stack that they, it would be a shame if they, if that just went to this like retail business, like why not sort of abstracted a bit and create the right API so that anyone can leverage this like amazing infrastructure they've built. And you know, that kind of being the birth of AWS and cloud computing, I think that's like, Such an amazing story. And that's inspired. I think like, that's kind of like how we think of hyperliquid was that like it started as, you know, a blockchain optimized just to be able to do on train for purpose trading because no other infrastructure could do that.

AI assessment note: “I think Amazon was a really big one that what inspired me”

Q So, so you guys pay, you guys pay people to use their phone less. Does that mean I start out with, uh, like, but, but break down the actual mechanics of, of how that works on a, on a month to month basis?

A Sure thing. So check it out. I'll use myself. Um, I, I know I use my phone for doom scrolling and social media a little bit too much. Um, but at the same time, I consider myself an effective in touch individual and kind of need my smartphone and, uh, you know, don't want to Count how, how much data I'm using. So the way Noble Mobile works is that you pay a 50 dollar flat fee for unlimited data, and if you use it up the wazoo, then you paid 50, and that's the end of it. Um, but if you use less than 20 gigs per month, then we'll give you cash back, uh, in the form of an end of month payment, uh, data dividend, we call it, of up to 20 bucks, and then any money we give you will then Grow at a 5.5% interest rate annualized so that your cell phone plan becomes like a mini savings vehicle without you having to think about it or do anything. And so this way you guys hold on.

AI assessment note: “you pay a 50 dollar flat fee for unlimited data, and if you use less than 20 gigs”

Q What was the, the, the most expensive Beanie Baby Transaction ever. Do you know?

A I don't know. I, I do know that prices peaked, uh, throughout just, they perfectly tracked the dot-com boom. So in two, in 1998, uh, Ty, which is, uh, Ty Warner, his company is called Ty, uh, T-Y, they make beanie babies. They made 1.4 billion dollars in sales, making it one of the fastest growing toy companies in history. Two years later, the market crashed almost overnight. Prices on the secondary market plummeted, a supply overwhelmed demand, and speculative buyers exited. So people were in the habit of, um, and yes. Interesting. The chat is correct. Uh, eBay was a major, major driver of, uh, Beanie Baby's growth because people could buy them and then the secondary market was way more liquid than ever before. Before it was like, okay, if you buy some, some rare drop or some limited edition thing, And you want to go sell it. You have to go to like a trade show that maybe happens every once in a, once a, once a year. You have to find maybe your local shop, but there's not that much liquidity. Remember when we were talking to our friend about the, uh, the, the, the, the collectible card market, Beatles memorabilia. He was saying like, I only need two buyers. It's very easy to find two buyers. If you're, if you're, if you're interacting at internet scale, it's harder if you're in some town and you have a rare Beanie Baby and you need to see Are there any other crazy Beanie Baby …

AI assessment note: “I don't know. I, I do know that prices peaked”

Q Okay. Uh, and, and Delian justify what, why, why do you like these businesses? Is that, is that even a fair characterization?

A Um, fair characterization. I think, um, you know, my sort of one line would be, I'm not sure that, you know, gross margin is actually like the right thing to sort of focus on in a business, especially, you know, sort of early on. What you want to be thinking about is obviously EBITDA margin, in particular terminal, you know, EBITDA margin. And so when I think about the, like, at least founders feeling ethos to, you know, sort of investing, we think that that terminal EBITDA margin mostly is determined by ultimately how much of a, you know, monopoly your company can be in the long term. And so if you look at, you know, sort of mag seven today, obviously there's a decent chunk of them that, you know, have Some phenomenal, you know, sort of gross margins, and those tend to be the ones that are a little more software oriented. If you look at the one that is at the biggest scale and has the best EBITDA margins, it's the one that is the most, you know, basically hardware oriented. For sure, some of it propped up by like CUDA and they're like, you know, sort of software side of the house, but like NVIDIA is the one that is performing the best of all of those. And then even if you study within those, you know, which of those, um, you know, uh, companies on the hardware side have monopolies versus, you know, sort of not, you see it's the one that with the monopoly, you know, clearly out…

AI assessment note: “terminal EBITDA margin mostly is determined by ultimately how much of a, you know, monopoly”

Q this model is going to behave weird in this weird way. Uh, like the Manchurian candidate problem. We haven't found any of that, but it certainly seems like something we'd want to keep an eye on. But being from your perspective, like what, what are the, what are the risks that we need to be aware of going into a world where China is really pushing hard into open source?

A Yeah, there's two, there's two, and you identified them, but let's, let's, let's talk about both of them. Um, so the, so the phone home thing is the, is the easy one, which is you can put up, you know, you can packet sniff, uh, you know, a network and you can tell when the thing is doing that. And you, and plus you can go in, you can go in the code and you can see what it's doing that. And so you, you can validate, you can validate that that's either happening or not happening. And I think that, you know, that's important. Um, uh, but I, you know, I think people are gonna, people are gonna, are gonna figure that out. You, you, you can kind of gate that problem practically. Yeah. Um, the, the, the bigger issue is, um, we, we have this term in the field, uh, right now called open weights. Um, and, um, open weights is a loaded term. Uh, it, it uses the open term from open source, but of course with open source, the thing is you, you can actually read the code. Um, you know, with open weights, you have, you know, just a giant file full of numbers, as you said, that you, you can't really interpret. And then what you don't, what you don't have, what, what most, what most of the open source, uh, open weights models don't have, including, you know, deep seek specifically, What they don't have is they don't have open data, right? Um, or open corpus, right? So you, you, you can't actuall…

AI assessment note: “there's two, and you identified them, but let's, let's, let's talk about both”

Q That's cool. Yeah. Do you, uh, so are, are you a columnist over there or are you a contributor? Like how does, with the actual shape, do you get assigned things? Like I've had reporters reach out to me and be like, I was assigned a story on nicotine, so I need to talk to you as an expert or whatever.

A So with GQ specifically, my editor, Dan, um, when I wrote the Zinn story and when I wrote the story about members clubs last year, both Were because I, I had written about both of those topics in my newsletter and he was like, I think it would be interesting for you to expand upon this and like a larger feature in a, in a print issue. Um, and that's how that's happened. I, I don't really, but like, I'm at the point now where if I have an idea of a great story that I want to write, like it doesn't make sense for me to pitch it to the times or to New York magazine, because it, I can get it done faster on feed me.

AI assessment note: “he was like, I think it would be interesting for you to expand upon this”

Q where I, I have to, I, I have to comply kind of at the national level, and they're kind of forcing me to comply everywhere, or can I just pull back out of that state? Because we've seen this with, with, uh, Apple intelligence not being available in China. That's right. Not being available in Europe, for example. But yeah, what, what would the impact on something like that be?

A Well, think about states in two categories, right? There are some states that are so large. And by the way, the analog that I would think about here is what happens in the textbook market. So if you're a textbook publisher and you want to write some textbook about biology or whatever, if the largest states in the country, this is category one, pass a law mandating something needs to be in a textbook, suddenly you're going to comply nationally because they are so large, they warp the market. Exactly. You have to. That's right. And so if California tomorrow says, look, We're passing a European-style bill to severely curtail, which I think would be a disaster, to severely curtail AI development. That would, in effect, be a national bill. And part of the, in my view, mistake that's happening tactically right now is this self-imposed July fourth deadline is so restrictive that it forces all of these more nuanced negotiating points to fall by the wayside, right? It becomes blanket. Either it's in or it's out. There's just not time to get to A more nuanced middle ground. So 11 Labs, if Tennessee has the Elvis Act, probably says, look, we just squirt around Tennessee. I don't, I haven't spoken to 11 Labs. But if California passes, then suddenly you have the whole country warped by this. And Gavin Newsom and the California legislature are certainly ambitious enough to try, as they did l…

AI assessment note: “if the largest states... pass a law... suddenly you're going to comply nationally”

Q biggest bill. There's all this pork. Are there any like green shoots where there's some random startup out there? That's just super excited because maybe it's not a headline grabbing, you know, ten billion dollar thing, but it's like, that's a hundred million dollars that I can go after in energy or in defense or in shipbuilding or anything. Have you seen any of these? Like they're called riders, right?

A Yeah. Yeah, it's a good question. Um, Biden was famous for stacking a lot of these. Like, the Inflation Reduction Act had a million ways you could go participate. This one's got fewer. I'll say one funny piece of pork, and I'll say one actual opportunity. The real opportunity, I would say, is for fintech folks who are looking at the Trump accounts, right, which is negatively giving huge amounts of new Americans who are underbanked or not banked before access to capital for the first time. If I'm Robin Hood, I am loving. That's a big change for me. Funniest piece of Orc in the bill, Murkowski, who was the key holdout in the Senate from Alaska, got a tax exemption for, uh, Alaskan, native Alaskan, uh, whalers. She said basically native Alaskan whalers need to be, yeah, that was honestly big delivery for, uh, her constituency of people who kill whales in Alaska, I suppose.

AI assessment note: “The real opportunity, I would say, is for fintech folks”

Q Zooming out just for a second. What is the state? Uh, how's the internet been for the last few weeks? It's been a busy, uh, month in our last month in geopolitics. I'm sure. And I know that makes your job a bit harder, or at least I would imagine, uh, things get a little bit busier.

A Keeps it interesting. I mean, it's whatever, whatever you see sort of in, in a, in the sort of kinetic side, the physical side of a conflict is, is usually replicated in the, in the cyber, um, side of the conflict. And so, um, you know, we, as the hostilities, um, ramped up between Israel and Iran, Uh, first of all, we saw very clear signals that, that something was coming ahead of time, which is, it's interesting how there's sort of a digital residue, which shows when, uh, there's going to be, uh, kinetic action, uh, uh, taken. And so, and so we were, we were watching, uh, that, that pretty carefully. Uh, big attacks, um, targeting both, um, Iran. Uh, one of Iran's largest banks was hacked by what looked like a, a group of, um, Israeli, um, hacktivists. Um, the largest cryptocurrency exchange in Iran, uh, was hacked and all its funds were drained. Uh, again, looking like it was either by the Israelis or Israeli, um, sympathetic, uh, attackers. Iran has, has tried to do the same thing against, uh, against Israel. Israel's, um, cyber, uh, defenses are much stronger. Um, but we're starting to see that Iran is targeting other, um, uh, uh, other largely U.S., Uh, financial services firms, uh, critical infrastructure firms, uh, to try and see, to sort of try and strike back from what, what is, um, you know, how, how do they strike back against the US in a way that, um, that, that is…

AI assessment note: “whatever you see sort of in, in a, in the sort of kinetic side”

Q Are, are a lot of these companies building tools like that, or is all the focus just on like, let's one shot the next Oscar film?

A Yeah, it's a great question. There's a decent amount of vertical focus. I think also, like, it's very hard for this, for startups, like, if you're not a Google to, or an OpenAI or whoever to play in the game of, like, let's train the largest one-shot best video model. Um, so a lot of them are focusing on verticals, like, um, Luma, which is a company we've invested in, did this really cool tool where you can upload, like, a, a nine by 16, like, iPhone style video, and you can just say extend, and it just Basically like out paints around the existing video and makes it look like, so you can just change the dimensions of your video really fast.

AI assessment note: “There's a decent amount of vertical focus. I think also, like, it's very hard”

Q to corporations that want to vend in talking avatars. For example, are you seeing a lot of that? Um, or like, I guess how, how, how, uh, rigorous or, uh, stringent are founders about like, Hey, we are trying to build a consumer app or we're just building a cool technology and it might land as a consumer product, but it also might land as a B to B play.

A Um, most people are not at all rigorous and often don't even know at the beginning. So what we actually saw with the first generation of AI video was it was only researchers making these like magical models and they had no idea what the use cases were going to be. They all just put like a text prompt box in front of the model and then you got an output and like a big company and an individual were using the exact same interface. Now I think we're starting to see more at like what we call the app layer, which is essentially like How do you productize this? How do you create workflow? And, and therefore, how do you go into two specific verticals? Um, maybe an example of that is all of the like standalone video ad, um, creation products. So things like Creatify or captions where, or Hey Jen has a product for this too, where, um, you can literally just like paste in a link to your Amazon or Shopify store. It will pull all of the info about your product, your logo, your brand, and it will generate a talking head avatar. Like holding your product and describing it. And that's something that like a VO three or cling, the general video model companies won't do today.

AI assessment note: “most people are not at all rigorous and often don't even know at the beginning.”

Q It is a cool name. Um, I want to talk about base. Obviously it's powering a lot of, uh, consumer crypto use cases right now. Like what, how did that come about? What has it been like incubating something like that inside Coinbase?

A Yeah. A lot of people have asked me about that because they're always surprised that like a relatively big company can still innovate or whatever, but I think the, uh, the way that it, you know, I can't take too much credit for it, by the way. The big, the biggest thing that I did was I, we have, we have a very entrepreneurial culture of different people we've brought in. And, um, Jesse Paul came to me at one point and he was like, Hey, I want to work on a new chain. And I was like, cool. Like take a small team, go run at it. You know, like, Two pizza team or whatever, like five people, and I had no idea if it was going to work, but the, the biggest thing I didn't, he deserves all the credit to be honest. The only thing I did was I kind of shielded it from the rest of the organization a little bit and just gave them time to cook. And he went through like five or six iterations and I, I saw a couple of these along the way and I was like, honestly, I have no idea if this is going to work. Um, and it turned out to actually really hit. And so that's, if there is like a secret to this, which I don't, I don't know if it's a secret, but It's like, you have to allocate some percentage of your, of your budget to these venture bets and taller and make them small bets. Cause you know, you want to have two, two, three, four people like working on it. Kind of like a YC company, not like don…

AI assessment note: “Jesse Paul came to me at one point and he was like, Hey”

Q Where are you getting the most leverage from AI internally? I can think of a bunch of different potential applications, but I haven't seen you do online virtue signaling about it yet, like other, uh, certain CEOs have. Uh, but I'm curious, you know, what, what you're excited about.

A Yeah. Okay. Well, so I think we're doing a lot of the current best practices that most people are doing with well-run companies. And then there's a couple areas on the horizon that we're, that are more crypto specific that I'll tell you too. So, I mean, the basic ones you probably know, it's like, we got, we onboarded all of our engineers to co-pilot and cursor and everything like that. So we got a hundred percent of them onboarded. That was good. Um, our customer support, I think like maybe 60% of inquiries are being answered by AI now, which is really good. We also, for a long time, been building AI models around, um, fraud prevention and, like, detecting risky transactions. Um, so, but you can, yeah, obviously, like, any process within the company that has a clearly defined input and output, you can train a model on and, like, um, generating compliance reports. Like, there's, so we're trying to integrate it everywhere. The ones that are a little more on the frontier or, like, crypto-specific, uh, for us, so, um, One is that, um, I guess people are calling it agentic commerce, right? Which is like, you know, your agent is increasingly kind of like your assistant and it might need to book you a ticket to go somewhere, or it might need to get through a paywall or something. Like if you're like, Hey, go read all the research papers from nature on this topic and give, give me a r…

AI assessment note: “we onboarded all of our engineers to co-pilot... customer support, I think like maybe 60%”

Q Do you think the conversation around economic freedom over the next few years will shift away from the U.S. and more towards emerging markets who are adopting digital assets, but may face, you know, pushback from, from their governments and You know, potentially even more aggressive way than we have in the United States over the past four or five years?

A Yeah, it's a great point because I think the unmet need is higher in some of these other countries, emerging markets, right? Where people like in Turkey, they're getting like 70% inflation a year or something. It's just like, it's devastating to their entire economy. People have very high demand for the dollar and then Bitcoin. Same thing in Nigeria, et cetera. Now, some of those markets are going to be very resistant to it. Like, especially if the government is not fully aligned with the people's interests, like the government wants to run their Fiat currency, because then they can kind of print money and abuse it, and you know, it's, um, it's one of these kind of original sins of like fiat currency, right? So, or like some of these, these markets have capital controls, right? Like in India. So I, I think the governments will be somewhat reluctant to embrace crypto in some of these places, but the, the average person will love it. Like the people are demanding it. They're rushing toward it. The governments are a little hesitant. And so in, Democratic countries. It'll be an, it'll be allowed because more people will vote for it. But in places like China, they're really cramping down on it. You know, I don't think North Korea is going to add it anytime soon. So it's a little bit like the internet, right? China has the great firewall of China. North Korea has their own private in…

AI assessment note: “I think the unmet need is higher in some of these other countries, emerging markets”