The Exchanges

Q doctors better doctors, making nurses better nurses, which is what you just said with scribing, right? Like you're basically by doing AI scribing, you know, you're kind of like removing this thing that wasn't billable. Now they can do, go do more billable things, right? Yeah. Couldn't your software by being in the middle of every conversation between doctors and nurses on the planet, theoretically do this and improve billing?

A Yeah. So, so the question is, could you create a company where you're selling off of one incentive, but by once you're in, you're just kind of doing good for the world. Maybe most companies kind of revert to their incentives over time and invest in the things that make them money. So I'd be nervous about a company where you're trying to scale it off the wrong business model. But the second thing I'd say is if, if that's the case, it should probably be somebody else. I'm not a healthcare expert. I don't know, uh, how healthcare systems run intimately. And so it's not to say that that strategy couldn't be effective. It is to say that I don't know how to do that.

AI assessment note: “It is to say that I don't know how to do that.”

Q which was the tail end of the party. But it is a great question, Dave, and point, Brian, because if it turns out that sports, and especially football, is going to drive the economics of all of this, I mean, it's relatively constrained in terms of types of content. So, If Sam were here, he'd talk about the football league. He tried to back Brian, were you in that too?

A Yeah, we, we, you know, we have mutual friends who started, started these leagues and that didn't, I mean, it's very, very, very, very, very, very difficult to compete with the NFL, right? The, the product is, I mean, you don't get better than the product, and plus you have the community element with the fantasy football stuff, right? It was just the perfect marriage, like fantasy football. Fantasy does not really work well for other sports, and it works so well perfectly for football that it just gets people entertained and jazzed about the NFL who even otherwise wouldn't care about watching it, right? And I do think that, you know, the betting ecosystem, but especially fantasy ecosystem, has been just Put the NFL just so far ahead of everything else.

AI assessment note: “Yeah, we, we, you know, we have mutual friends who started, started these leagues”

Q Apple agent. I've got all access to everything on your phone. Don't worry. It will never be in the cloud. It will never be exploited by a third party. What do you want me to do for you now? And you just like go through and give it connectors and access as you need certain things. It saves your login as you need it. Why wouldn't they just win this?

A Well, I think Apple might win it, but I think the interesting thing is I don't think you want to be the one who takes on that liability. If anything, I think the goal is to like radiate out risk to other people and be like, well, we're not like to keep yourself safe and be the one that's aligned. Cause like the reality is you're like, Peter, you're saying that you think the ban is going to go up. The reality is the hacks are only going to continue to grow and scare people more and more and more, right? Not less and less and less. Everyone's going to have them, and so I just wonder if the real thing is going to be, if you really want to win in, like, the new economy, in this AI thing, you have to somehow figure out how to hoard the data and the trust and build the trust, but then when things go badly, always have a third party to point at to blame.

AI assessment note: “I don't think you want to be the one who takes on that liability.”

Q But then, and then OpenAI filing, saying it's expected, our IPO is expected within the next year is very vague to me. Do you have any more intel on that?

A I'm going to hand it to OpenAI, right? So they filed their, both Anthropoc and OpenAI announced themselves that they had filed Confidentially smart. Cause you know, all those pesky reporters are all over them all the time. And then open AI then proceeds to explain all the reasons it might not go public. Like, well, actually, I don't know what's in the press release, what versus my sources say, but, but combination of my sources in the press release is like, well, you know, we might like the development of the models might change the trajectory, or we might want flexibility to change our token economics to gain market share, or. We might want to do another unusual deal. Like the one, the information just reported this week of where they're doing a data center, but Nvidia is involved in potentially leasing the chips.

AI assessment note: “combination of my sources in the press release is like, well, you know, we might”

Q Is data moat or have we lost that too?

A Can be, can be if you can only uniquely create it. So I think this is, like, this is my new precision, right, on this, which is, like, these people, like, I was talking to some entrepreneurs, actually two today, who I felt like I was listening to, like, a 2021 pitch, where they're like, we did the user research and talked to the buyers, And the problem the buyers have is blah, blah, blah, blah, blah. So we were building software to address the buyer need. And I'm like, what universe are you living in? Right? Like, if you believe that the tendency is to like, all this shit is like commodity software that can be cranked out, which is the direction. I'm not saying we're there yet, but we're in that direction. You can't build a business doing a thing a buyer thinks they want right now, just because it's like, and it's just software. It has to be. That you're doing something where you write software, right? And you basically give it away, right? Or it's like basically free. You, you don't bank on the software, but you use it to build some sort of durable compounding data asset or trust asset or marketplace asset where like they have to keep using you because you're fundamentally structurally better at something that's defensible, right? Where you have a compounding advantage. But this idea, like anyone who, who believes That they can build a business because something is hard right …

AI assessment note: “Can be, can be if you can only uniquely create it.”

Q On that front, Sam, I was going to ask you, you had a post about You've had several posts, but one this week about, I'm going to get it wrong, but when intelligence goes to zero, what happens?

A Yeah, yeah, yeah, yeah. And blinded intelligence is bad business. It's like, everyone's like, all these computer scientists running around gleefully talking about how computers will commoditize human intelligence, and so all the humans will have no jobs, and they're kind of like, sad about it, but they're also kind of proud of their power. But what they fail to realize is that The, you know, the original narrative was someone would get to AGI and then that would be the end of the game is not how this is playing out. Everyone's gonna get to the same shit. It's all commodity. It's all getting a price to the cost of energy. And it's all worthless because if things are abundant, then they have no value. So I'm just like, guys, if you really, like, if you really believe that we're moving towards a world of abundant machine matrix multiplication intelligence, y'all in the pretty bad business. Being hard is not a moat anymore. Software is not a moat anymore. I mean, I've been, I'm, like, having fun with this because I've been saying for a while.

AI assessment note: “It's all commodity. It's all getting a price to the cost of energy.”

Q Where are we roughly now, just so we know how fast we have to get there?

A Quarter billion. Okay, that's big. Yeah, we're 12 and a half percent of the way there, or 25% of the way there, so I think it's possible we'll hit that number. I mean, there's, I think it's macro-driven if I was, I thought the industry was going to die when interest rates went to five percent because the reason why we're like, you know, things feel like they're struggling or whatever is because treasuries are soaking up every risk capital, right? Like there's, there's no point to, to risk money on anything when you're getting five percent from the Fed, but somehow it's not, right? There's this still this huge appetite to take risks and Bitcoin and all these other things. And eventually rates are going to drop, and that extra premium that everyone earned on their ten-year T-bills of five percent has to go somewhere. It's going to go into stocks and a whole bunch of other things, and you'll see, like, I think a massive kind of growth mode, you know, risk-down mode, just going to drive more stablecoins issued and stuff like that.

AI assessment note: “Quarter billion. Okay, that's big. Yeah, we're 12 and a half percent of the way there”

Q moment of reckoning for the tech industry on a lot of fronts. What do you think of that? And I'm not, and perhaps you haven't, you, you haven't been thinking about it as much as I have just with our different jobs, but But you're on the board of Snap. You're very close to a lot of tech founders. What do you think this moment means for, for Silicon Valley?

A It's a great question, and I'm definitely not in the heart of it like you are. I do think that there is enormous anxiety around, uh, the speed with which the Chinese have caught up with AI, and that sense of, oh my goodness, they're doing it cheaply, they're doing it effectively, they're doing it quietly, and, um, that is what I'm hearing from My friends in Silicon Valley. I think actually, um, the White House appointing an AI czar in David Sachs and Sriram Krishnan going, going in there, you know, arguably we should have been on this five, 10 years ago, and we've been slow, and what I'm hearing from everybody is that, um, you may not like David Sachs, you may not like Sriram Krishnan, but at least they're there, and they're trying to accelerate Um, and of course there's anxiety around, well, you know, from some people we need more regulation, and then there's anxiety that we've got too much regulation, and certainly, arguably, under the Biden government, there was too much regulation, and they didn't feel at all business friendly. I certainly felt frustrated when I was working briefly in finance, having to deal with all sorts of government agencies that, that You, you were, it was shocking how unresponsive they were, and yet they are arguably representing us. Uh, so I'm not unsympathetic to the need to speed up around, certainly around regulation, but what I'm hearing is the g…

AI assessment note: “I do think that there is enormous anxiety around, uh, the speed with which the Chinese have caught up”

Q over the shit wasn't, why wasn't I added, and just sort of that giddiness of, like, how much fun he's having, which I also felt like, I'm like, this is, this is, like, this has got to be really fun. Um, what is your analysis on, like, the severity of this? Um, and, and what's sort of playing out now in Washington with the, the grilling of these, uh, officials?

A Yeah, I mean, I would say, like, it's, I mean, Britt has a real point. Like, good that they're using Signal. These are people who have to talk to each other, and it's actually a much better way to communicate than an open phone line or text. Um, and so, like, and in fact, they had been told by the Pentagon they should be using Signal. Except that, you know, if any of their, like, there's an enormous effort by the Russians and the Chinese who are really good at this stuff, and probably the Israelis and others, to compromise the, the, the devices of these specific guys. One of them was in Russia while this is going on, with Steven Witkoff. What are the odds that the Russians are reading these as they come in? I don't know, one in three, one in four, they're not zero. Not, you don't have to compromise signal to get to that, right? You just have to get to the screen of the phone.

AI assessment note: “What are the odds that the Russians are reading these as they come in?”

Q And I think there's some marginal advantage to it. But in my mind, that's like a five to 10% problem of like more great businesses, not like the 90% problem. That doesn't mean it's not worth doing. But like, it does mean like, to me, it fits very nicely into trajectory of like, yes, it's much easier to start all sorts of businesses for all sorts of technical reasons, right?

A Yeah, I think, uh, my point of view is, Generating software is actually the killer app of AI. I was interviewing Nat Freeman today for our internal Vercel podcast today, and he was talking about how much conviction Microsoft got into OpenAI because of the effectiveness of Copilot. Like when Satya saw Copilot, when, when they were doing the initial iterations, they had actually way more ambitious applications they wanted to create with AI. And the one that ended up being the killer app is Making people more productive at producing code. And what's happening now is that because the models are getting so much better, we're finding out that we need to make the models better for producing code. And so you use something like, uh, uh, uh, Claude and, and which is actually one of the main models that powers V zero. And it became like 10 times better at designing software all of a sudden. It became 10 times better at creating React applications. And so let's just look at the numbers, right? Like there's four million React developers. I can move that from four million to fifty million trivially, because now like you just speak English to that thing. Even if like it doesn't do a hundred percent of the job, it's just like this like infinitely available assistant.

AI assessment note: “my point of view is, Generating software is actually the killer app of AI.”

Q And zealots, right? To your point. But Britt, over to you. What are, what are you holding in your head and thinking about?

A So a couple things. One, I think anytime you need a flow chart to explain how your board works and is hired, you're probably starting from a bad position. Um, the second thing is I feel like Satya is getting like all this credit for like saving open AI and like, you know, doing the right thing. And I don't think a lot of people are recognizing like he like hit a Hail Mary and he was Screwed. And like, frankly, I'm surprised that Microsoft and Sequoia and Thrive, like didn't do more board diligence when they were investing all this money to figure out like this board, this board was like not adequate for the type of company that this has become. And so now, oh God, I'm going to lose it. Loving billion dollars overnight. I better like figure out how to like, you know, hit this one out of the park and get Sam to join me or else I'm screwed. And like, Microsoft barely is holding on to its stock price right now, and we'll see what happens next, but I just feel like if that didn't happen, like, it would be a world of trouble, and I think a lot of, like, fingers pointed at Satya to be like, what the hell were you doing when you invested and not, like, doing all this, like, diligence around the board itself?

AI assessment note: “So a couple things. One, I think anytime you need a flow chart”

Q But Vinod, isn't each large model successively that's coming out costing like an order of magnitude more to train?

A It is costing an order of magnitude more to train, and that's why I don't think open source models will be viable because the training costs. But once you're trained, you want as broad a set of usage for two reasons. One, You want the most revenue from it, and the lowest cost models will get the most revenue, but more importantly, there's huge data generation for training the next generation of the model. So for lots of reasons, you want to maximize usage, and if you're playing the long game, and I think the AI model game is mostly being played sort of on a five-year time horizon, not a one-year time horizon. In that time horizon, costs will drop here today. NVIDIA extracts a fairly good tax from everybody, but every, every model's gonna run on multiple types of GPUs or compute, and they will want the most data generation. So, I'm pretty convinced The next few years, the revenue isn't the important metric. You don't want to, of course, lose so much money you can't afford it, but you're not trying to make tons of money. You're trying to get tons of usage and you're trying to get tons of data from the usage and learn to be a better model. Uh, I do think there's a lot to be gained in intelligence, whether it's reasoning, whether it's probabilistic thinking, whether it's sort of pattern matching, any of these things. And these models have lots of headroom to get better. And I think…

AI assessment note: “It is costing an order of magnitude more to train”

Q Who basically was asked about the next big AI applications. And this is a guy as in the weeds as any. And he goes, I don't know. So I guess I, I kind of put that out there as a little bit of a framing a lot over to you. What are you seeing in this moment? And what is your thesis about how and where to invest right now?

A Yeah. You know, it's an interesting question in terms of where, where adoption is happening. And I think like any technology wave, what we're seeing is a bunch of stuff get adopted really fast and then a bunch of things not working. And I think it's clear that 90, whatever, 99% of AI companies are not going to work out, but the one percent that will could be quite large. And, you know, I kind of think of it as three or four sort of layers to it. Obviously there's the semiconductor layer and that's NVIDIA and some startups. There's the foundation models like OpenAI, but also there's a bunch of other types of models that nobody ever talks about, right? We always talk about the language models, but there's models for biology, physics, materials, et cetera, et cetera. There's all the infra, and there's like five different types of infra, and then there's the application layer. And in the application layer, there are a handful of areas that enterprises are adopting really fast, and then there's a set of consumer products that are very clearly skyrocketing in ways that we haven't seen in a long time, right? That's ChatGPT, that's Perplexity. That's mid journey. You know, there's actually things that are working there and there's things that are working on the enterprise side. And I think from that is this under overhyped perspective, one could argue this is very under hyped in part b…

AI assessment note: “I kind of think of it as three or four sort of layers to it.”

Q But Vinod, isn't each large model successively that's coming out costing like an order of magnitude more to train?

A It is costing an order of magnitude more to train, and that's why I don't think open source models will be viable because of the training costs. But once you're trained, you want as broad a set of usage for two reasons. One, You want the most revenue from it, and the lowest cost models will get the most revenue, but more importantly, there's huge data generation for training the next generation of the model. So for lots of reasons, you want to maximize usage, and if you're playing the long game, and I think the AI model game is mostly being played sort of on a five-year time horizon, not a one-year time horizon. In that time horizon, costs will drop here today. NVIDIA extracts a fairly good tax from everybody, but every, every model is going to run on multiple types of GPUs or compute, and they will want the most data generation. So I'm pretty convinced The next few years, the revenue isn't the important metric. You don't want to, of course, lose so much money you can't afford it, but you're not trying to make tons of money. You're trying to get tons of usage and you're trying to get tons of data from the usage and learn to be a better model. Uh, I do think there's a lot to be gained in intelligence, whether it's reasoning, whether it's probabilistic thinking, whether it's sort of pattern matching, any of these things. And these models have lots of headroom to get better. And I…

AI assessment note: “It is costing an order of magnitude more to train”

Q in that. And so, you know, is there any signal in being like an alliance of sorts? Like, do you think the university presidents are all talking about how should we position ourselves in this together? Is that, is that a, would that be a good thing, bad thing? Is there any way that You know, there's just, like, a quicker way for the dominoes to fall into place here.

A I don't think so. I think these guys, unfortunately, don't really coordinate enough, right, in terms of how they think about or respond to these things. They think about their campuses and communities differently, and they're very hamstrung, is my honest sense, in a lot of cases. I mean, there are exceptions. Like, University of Florida has done a really, I mean, there are places that have done a good job with this, and there are places like Chicago, which have really clear principles, right? Because they have really clear principles, and they articulate them really well, The president of Chicago's letter was fabulous, being like, this is the process allowed, this is what's not, you do this, you get arrested, you do this, it's fine. And you're like, great, thank you for the clarity, right? But I think the problem is, is that a lot of these, you know, complicated organizations without enough clear direction, they're just kind of drifting. And the students, you know, they have nothing, they think they have nothing at stake, although I'll never hire any of them.

AI assessment note: “I don't think so. I think these guys, unfortunately, don't really coordinate enough”

Q Okay. So Brian, you've been thinking about immersive worlds for some time. What do you think about the, I, I can't just, Vision Pro, it's too, it doesn't roll off the tongue, this thing. What do you think of this thing?

A The Vision Pro, the Vision Quest, or whatever you want to call it, seems like it's pretty promising, because for the, what, for the reasons that Dave was saying, if it's truly an interface that you can pick up naturally, there's the VR element, which is, you know, kind of the entertainment stuff, You can do on a plane. And I think we've had that for a little while. Obviously you can do the VR stuff on a quest. We've seen that for a while. In fact, one of the reasons why I first invested in Oculus was similar to the experience that Britt had, which is that I was walking around in a fantasy world and a knight came up with a sword and I was actually scared. And so we thought we had to invest and that's similar to the dinosaur experience. And that's the kind of the entertainment VR aspect. The, the, the way this is going to actually You know, improve beyond that is if it truly does get the spatial right. Two ways I think you can improve on it. If it does actually get the spatial right, which I have not tried this yet. We do have one in the office, but I haven't tried it yet. In the same way that, you know, my two-year-old daughter can look at photos on an iPad because it's just such a natural interface for her, or they can, you know, scroll through stuff real easily. If it's that kind of an interface, you can interact with your environment and get more information from your environ…

AI assessment note: “seems like it's pretty promising, because for the, what, for the reasons that Dave”

Q Awesome. Well, what else about, before we move on to some of the other exciting topics of the week, streaming sports, anything else that we want to say? Brian, any thoughts? Or are you just chomping at the bit to invest in startups developing for the Vision Pro?

A No, not, I'm not, I'm not chomping in a bit to do that, but I am, I am excited to try it and see how it progresses. I, you know, as somebody who's been around this space for 25 years now, and my license plate actually represents that, I do think this is, I mean, the Quest has been really good for lots of people, but I do think that something like this, if it truly works, and again, as Dave was saying, he thinks it actually works interface-wise, I do think it unlocks A new, you know, kind of paradigm. Do I think it's gonna be as, like, massive mainstream as the phone? No. I mean, it's still maybe in a decade.

AI assessment note: “No, not, I'm not, I'm not chomping in a bit to do that”

Q Did you do Pernuvo? Which one did you do?

A Tell the people. I did QBio, but I also did, like, back in the day before QBio, like, actual DEXA scans, um, three 60 body, um, scans. I did my VO two max every year. And I would compare year over year to see like how my body was changing physically and then internally and all the, all the blood tests and urine tests and all those things too. And I think it's really awesome for preventative healthcare. I think AI obviously is going to help a lot. I think that it's an interest. It's like going to be interesting from a privacy and data standpoint of like, now do these companies own the data and they're analyzing it with AI. So what does that mean? And then I also think there's going to be a lot of like false positives and people are going to have some really bad Um, outcomes.

AI assessment note: “I did QBio, but I also did, like, back in the day before QBio”

Q So does that mean you've stopped? I mean, you've been the, you've been the north star for all of those of us that love, you know, real things in the real world, real experiences, you know, coffee you can taste, uh, donuts you can eat, you know, hospitality experience, right?

A I think we've paused and I think, and I understand why we've paused and we needed to pause, right? Because You know, this kind of, this kind of moment, you know, this AI revolution moment is quite, you know, I think meaningful and impactful. And so I just think for us as a firm where we're not going to invest in the kind of the infrastructure, you know, kind of layer per se, right? That's just not, you know, the, those checks were already just way still beyond. That's what we talked about. Fifty million, you know, a hundred million, you know, start these things, right? And so, like, it's just not what we do. Um, and You know, as we're missing that layer, you know, and we're waiting for the application layer, both in enterprise and the consumer to kind of catch up. Like, I think it's just required us to just say, pause for a second on all things that are non AI, because we're going to have catch up like almost every firm, by the way. Right. I mean, there are obviously firms that get to play in these big foundational models, but like, For the most part, like we're all kind of playing a little bit of catch up. And so it's going really, really, really fast. Uh, I think it's really dangerous moment in a lot of different ways, um, from valuations and things like that. But you just, you gotta play and you gotta play in this, in this space and you gotta, you just gotta, Be taking enoug…

AI assessment note: “I think we've paused and I think, and I understand why we've paused”

Q Sam, does Bitcoin hit a hundred K this month?

A It might. I mean, it's, that's within random walk at this point. I mean, the only, well, the, the only thing that I've heard contrary to that is that Look, I think there will be a lot of cell pressure into like just around a hundred K just because a lot of people have been up and down and say that's a good number and I'm out. And so I don't, I don't know. I mean, I, you know, like many of us on this call, I'm like a long term crypto holder and will continue to be and have no intentions of selling or anything. But, uh, you know, I think the hundred K the hundred K thing is kind of an emotional thing more than a practical thing.

AI assessment note: “It might. I mean, it's, that's within random walk at this point.”

Q Are there other aspects of, I know beyond AI, longevity is, is a big topic for you. Are there, are there others that are currently in your broader?

A There's a lot of stuff I've looked at over time and I'll give you an example. I started getting into defense tech when Google shut down Maven because I was like, oh my gosh, if Google isn't doing it, what a great startup opportunity, you know? So I started looking for defense companies and then I met Anduril and I was like, oh my God, this is like an amazing company. And at the time it was very unpopular. To back a company that was doing defense, right? And you're like, it doesn't matter left or right. It's just, let's protect our democracy, you know, like that was kind of the impetus. So, you know, it kind of feels like there are these moments in times where you realize there's either something missing, or there's interesting things happening, or a bunch of really smart people are all starting to collectively aggregate around something, and sometimes those are real trends, and sometimes those are fake trends, but

AI assessment note: “I started getting into defense tech when Google shut down Maven”

Q all, we're spread out across the spectrum, but I feel like Are kind of kids and their age in school, we're entering the age when this, these issues become really, really serious. So it seems it's sort of like we're entering these, these debates and discussions, but are they at the same time getting more serious? Like, do you feel like as an epidemic, this is kind of growing and.

A Well, I think there's two defining inflection points that when we all look back, we'll understand had an outsized impact on our, on the humans that we have been raising. The first is what Dave was talking about, the advent of the smartphone, and, you know, if I look back at, at pictures of Emma, who's, who's our oldest, we had just gotten our smartphone, so it was like she is the smartphone generation, and then COVID, and what, I mean, what age and what stage of life they were at when COVID hit is going to be these distinct chapters. I think we have not even We haven't even begun looking at, at the impact and the trauma just even as adults. So, but kids too, it really mattered what stage of life they were at and what kind of impact it had. And so I think that those two are, are converging into a massive mental health epidemic. And, you know, I think it's, It's also part of us having to bridge back and understand what parts of being connected all the time and living life like a micro influencer, which is what, what, um, I mean, she's going to kill me if she heard this. So what, like maybe her, Type of people do. What that's going to mean for them later in life. What is it going to, I think there's two really interesting things. One is this idea that you are projecting yourself all the time, and it's actually not what we used to do when social media first came out, where it was l…

AI assessment note: “I think that those two are, are converging into a massive mental health epidemic.”

Q gear from the side. There's no data on the screen. You have things like World Series of Poker that became extremely popular because they let the fans in on what was going on with each hand, you know, and once they did that, they went to the top. And I, esports was growing really fast for a long time. Has that changed at all? Or where, where is it now?

A I, I think that it is My guess is that it's not growing as fast as it should be, and I think a lot of this has to do with, I thought that a lot of the, it takes, you know, major production, and it takes, you know, kind of drumming up community, like the community aspects of these things is huge, and even though the viewership of things like League of Legends Worlds is still huge worldwide, in America, this hasn't fully taken off, and I'm not exactly sure why, like nobody has really established the For lack of a better term in the old days, we had the concept, the water cooler conversation, right? And that hasn't happened, even whether it's with the demographic of e-sports or something else, right? Like the talent of the people like running the shows is probably just not there and not at the same level. And we don't have that water cooler stuff with e-sports. I, I would, I would have expected it to have taken off a lot more by now. Um, especially as kids are just growing up playing these games more than they're growing up Playing, you know, basketball, football, right? Like I, I would have thought that, you know, we would have seen like, you know, Fortnite championships or Counter-Strike championships or League of Legends championships take off more in America and it just hasn't done that yet.

AI assessment note: “My guess is that it's not growing as fast as it should be”

Q Do you think that's why Taylor likes being in New York?

A It could be. Like, New Yorkers don't give a shit. Like, somebody's funny tweeted the other day, like, everybody in New York's famous. Why should I give a shit about you? And there's such beauty to that. Um, when you leave, like, we had, this is a weird story I've never told publicly, we had a little house in Connecticut Candace and I, like, in this kind of, like, very middle-class town. It was a house we bought. It was near the beach, and it was, like, 280,000 bucks, so it was a very, like, the rich people of the town. They were, like, the dentist and the guy that owned the cat, the tire shop, lived next to us. We loved it there, and when this internet fame kicked in, it just kept precipitating and precipitating, and then I remember the final straw was after my daughter was born. At 9:30 at night, there was a knock at the door. Like I got an aggressive knock and like Candace goes upstairs with the baby. I'm like, got a baseball bat and I like cautiously open the door and it's a woman standing there and she's like, can you take a picture with my son? And I was like, this is unbelievable. You woke my newborn daughter. This is so inappropriate. This is such a violation. And she was like, oh, come on Casey. We're from around here. We're from around here. We know you. And it was like, I don't blame her. Like I think that she wasn't doing it to be mean. She thought that was totally f…

AI assessment note: “It could be. Like, New Yorkers don't give a shit.”

Q So, okay, so is this even worse, Scott? Are people gonna be like, not only is AI, like, using all the water and have the wrong answers and all that stuff, but it also is gonna make most people poor in the process of, like, Nvidia stockholders being, getting well off?

A I mean, it's still a drop in, in the bucket. For overall assets for these guys, but you have to assume a double collapse, right? You have to see price collapse and you have to see demand collapse, which is basically what we saw in the mortgage crisis, right? If you see both of those collapse, then you have a systemic issue. Otherwise it's idiosyncratic and it just will compress margins. If it's price, the way Nvidia set it up was they said, look, we're going to make these kind of generic. And so anyone can use them. So if I have a default from one Off taker. I can just move it to the other off taker. And so there's some good protections in there, but it's fascinating.

AI assessment note: “If you see both of those collapse, then you have a systemic issue.”

Q what does it sound like when a tool bit is going to break and how do you, what is the path that you take? I mean, those are, there's a bits of data that I think, you know, in some ways, Sam, I think you and I saw this at Facebook. It's like, what did Facebook do? It just digitize the information that was in our head. Who's your friend?

A Yeah. A hundred percent. I mean, I used to, we used to talk about, I mean, like talk about the composer and Facebook era. It was a data pump, right? It was an efficient, free data pump to bring data online that couldn't exist anywhere else, and the reality is, if you go way back in the Facebook history, the real story was on day zero, internet's scary, so no one wants to post on it. You're never gonna put a real photo on it. Facebook created the trust and security for people to take a bunch of real-world data, photos, relationships, messages, et cetera, and bring it online, right, for the first time, and created an enormous amount of value Doing that, right? And so now we're in this weird multiverse where it's like, the internet has all been scraped to hell, right? Like good luck New York Times with their lawsuit, right? But like that has been like copied a thousand times and distilled a thousand times. You have personal private digital information, what's in Gmail, what's in da da da da, where people have some say in theory, as opposed to it as all being like dumpable effectively. And so that's where we were talking about the trust, things like that. There's then, like, the, the whole class of companies that are out there trying to, in some ways, build new data sets, the handshakes, the scale AIs, where you just pay a ton of money and you turn financial capital in some form in…

AI assessment note: “It was an efficient, free data pump to bring data online that couldn't exist anywhere else”

Q But you as a reporter, don't you feel like the need to disclose everything to anyone you're talking to? Cause you're literally having off the record conversations all the time.

A But I'm also talking about meeting. Like if I'm having a meeting about an upcoming project internally, right? So it doesn't even have to be But I bet if I had off the record conversations and said, hey, this is off the record. Same terms. You've known me for 20 years. It's always been off the record. But by the way, I'm really interested in what you have to say, and I want to make sure I understand it and get the most out of it, so I'm going to record it for myself. You know, maybe 50% of people say no, but 50% of people say yes, right? Because, and they might, I, I would give them the same courtesy, right, if I were sharing my thoughts.

AI assessment note: “if I had off the record conversations and said... I'm going to record it”

Q Why is that the case in the US? So I get the value of stablecoins outside of the US. I don't quite understand what the play inside the US is.

A It doesn't matter. Like the, the world is so big that if this, everybody wants us dollars and if the programmable us dollar is tether or circle or whatever, they'll use it. And then us has no choice but to accept them because the economy is so big right now. The global economy is so big and us is a big part of it. It'll effectively kind of Switch to it. And normal people are not going to switch from credit cards because it's just kind of entrenched. But the way Visa deals with like banks, I think it's going to change because they don't have to care about banks as much anymore. They can actually just do everything in stablecoins.

AI assessment note: “the way Visa deals with like banks, I think it's going to change”

Q to the Atlantic and what do you do? Do you have like a war room and you're like, how should we go about like launching this? What's the marketing around this? What's the business model? How do we, how do we like make this about the people in the news, but kind of drive some traffic to the Atlantic and get people to pay us? Like what's that conversation like?

A You know, it's funny. Cause I, I, I don't think, I think that like you're thinking like an entrepreneur, which, you know, fortunately maybe Jessica is in my partner is like, I would just think what an amazing story, right? And like, this is crazy. And particularly the interplay where Vance is sort of saying, like, I don't think Trump really knows what we're doing here. He's a little lost. And like Vance is really exposing something there that's really complicated and is going to play out for the next four years. And I think Jeffrey was, I actually think there's like the, the, the moment when you start to, and this is actually like a common Silicon Valley misconception that we are like obsessed with clicks as opposed to interested in information. Um, and so I think actually my, Guess is that he was just, like, fascinated by this thing that played out. Of course, also, part of that is what an amazing story. I can't wait to, like, tell everyone this story, but the marketing, I mean, the stuff markets itself.

AI assessment note: “this is actually like a common Silicon Valley misconception that we are like obsessed with clicks”

Q Like, why does this need to exist? Like, why and what are you, what are you fundamentally leveraging that all the other ways that people build websites today, um, are missing?

A Yeah, it really depends on the size of the customer. For e-commerce websites, for example, we just give them a lot of new revenue because we make their website so fast that you put it on Vercel and you get better SEO, you sell more stuff, people convert at a higher rate. So for example, Under Armour.com, Ransom Vercel, ASICS, like when you buy shoes, Vercel is behind the scenes, hopefully you have the best experiences on the internet and most visitors on the internet won't know. But they will like that website more because it feels better. So when, when think of it as like a global acceleration system, not only just the way to build it. Um, and the other reason that it's needed is that, uh, especially with, with the advent of AI and code generation, Versailles is almost becoming more like a design tool is that it's just so easy to create software these days.

AI assessment note: “think of it as like a global acceleration system, not only just the way”