The Exchanges

Q Yeah. What about, uh, the canary in the coal mine analogy? Uh, I was looking at, uh, unemployment statistics in India and the Philippines, and it doesn't seem to be doom and gloom over there. I don't know, I didn't dig in super far, but would you at least expect that the unemployment rate would spike overseas before it spikes in America, or do you think this all happens simultaneously?

A It's a tricky question. I think ultimately white collar work is a lot more of our economy than it is the economy of India and the Philippines. And they are much sort of like more immature economies that are growing through investment and things like that. Uh, but certainly I think we called it out the, the, uh, the consulting sectors in India are certainly going to be challenged, uh, in other places as well. But the reality is like the, the timing, timing is everything in the markets clearly, but the trick here is if you're a corporate and you are hard pressed to get AI Uh, into your organization today, you know, ChatGPT and OpenAI will send you a forward deployed engineer if you have billion dollars in budget, right? If you have a ten million dollar budget, they're not going to. Uh, and so who are those folks turning to? They can't usually do it themselves, and so they are going to, uh, the outsource providers, the Accentures of the world, and so I think those businesses are, are, are likely going to be in a lot of trouble over the medium term, but they probably will have a big bump, uh, from people really, uh, putting that AI into their organizations first, and so it's, it's a bit of a tricky timeline there.

AI assessment note: “those businesses are, are, are likely going to be in a lot of trouble”

Q nothing and win? Like if you just, if you just keep the servers online and keep the platform flexible, it's not like you have to come up with the next greatest Piece of intellectual property because you are the platform. So if, if I, if I misunderstand that, what, what do you have to do day to day to actually move the ball down the field to get to that?

A I like the way you think of that, and we, you know, for those of us that are mathematicians out there, we sometimes think of the second and third derivative, not just the first derivative. And, and what you shared is exactly right. If we walked away, just kept the lights on, there's enormous native growth in the platform. This is a, a highly technical platform. Creators make experiences. They auto translate. They run anywhere in the world. We've now seen peak concurrency records for gaming all around the world. In August, Grow a Garden, one of the more recent big games on Roblox, hit twenty-five million concurrence. So I, I would say we cannot walk away. We need to support infra and support scale. But if we did walk away, I, I, I think there may be some negative growth that keeps going. What we think about when we're running the company, though, is the exact opposite of that. We think about, are we constantly layering in systems? Are we building a machine that can keep growing? And are we building a machine that as we hit 10% of the gaming space, we're, we're growing right by that. And, and the gaming, you know, the gaming market's pretty interesting. It's about 180 to two hundred billion It's arguably really, um, sitting right there for technical innovation, for new ways of distributing, for games that run really well on low-end Android phones, and on high-end PCs, for games t…

AI assessment note: “We need to support infra and support scale... We think about, are we constantly layering in systems?”

Q Are you the Jerome Powell of the Roblox ecosystem? You get DMs directly. People ask you to print, to turn on the money printer.

A Well, imagine I'm a young kid. Why can't you just make a trillion Robux? And, and so there's a lot of learning there that the value of the ecosystem doesn't change when you print more currency, the currency is just worth less. So we're very careful with the value of our currency. We, um, we've done something really right, I think, from the start, and there's a real temptation to do nonlinear functions, accelerators. If a dev gets really big, should we pay them, you know, more per hour, per minute, per robot than others? We've really stayed away from all those nonlinear systems, and I think that's added a fairness to it. We, we, I do believe we have aligned incentives of creators and users in a good way so that if the system is working, it keeps working year after year.

AI assessment note: “the value of the ecosystem doesn't change when you print more currency”

Q to get all this oil and stuff. Um, but obviously in, in a purely financial context, uh, that's not. That's not the outcome, but, uh, has this just unlocked higher frequency trading, more leverage, uh, a different shape of trader, something more quantitative, something more, uh, algorithmic driven? Like, who are the, who are the customers, or who are the traders, and, and why are they excited about the product?

A Yeah. So they're excited because, so, so basically like futures, like you said, like futures kind of sucked because of all these like random things. They kind of settle. Sometimes you don't want to get delivery. They're like all these weird things. You have to keep rolling your position if you want it open. Um, so like if you look at Robinhood, for example, like that's actually much more popular with retail users than futures are. Um, and they, they, they primarily trade options on Robinhood when, you know, when leverage is concerned and options are super cool because it's kind of like a lottery ticket. You can, like as retail, you can buy a lottery ticket and feel really good about it. You're downsized limited. Um, but the, the trade off there is like, it's actually really hard to price an option, especially if you're retail and especially if you're using like an app that doesn't give you the information that you need. And so you're, you're kind of getting fleeced because there was like very complicated, um, structures. It sounds simple. It's like a strike price expiry, but, uh, in practice it's very hard to price. And so herbs are basically like marrying these two assets into one thing. So there's, you want to trade something you want, you want to trade something that's just like the price of the underlying. You want there to be leverage and you want there to be like one thin…

AI assessment note: “So they're excited because... you want there to be leverage and... never expires.”

Q But there are so many other examples and you guys are like the, the, like the key leaders of tech history in my mind. And so I would imagine that there's, there's more to it. And you've seen in the data LVMH performed, like I think of you as a tech history podcast and the LVMH just does incredibly well. Like, was that something you predicted? Is there something there?

A So the best episodes Are the ones that have these three key ingredients. And I always thought when we started over a tech podcast, we cover tech companies, uh, then we were in this middle phase before we sort of became more mainstream, which was, uh, educating a tech audience about non-tech phenomena. Like no tech companies are good at brand. And so when we started studying the luxury companies, it's like blowing the minds of all these tech people like, whoa, that's why this is valuable, which included myself. Like I learned during the research and I'm like, Hey audience, I, I gotta share this with you. Guess what I just figured out. And so the three key ingredients.

AI assessment note: “educating a tech audience about non-tech phenomena. Like no tech companies are good at brand.”

Q Um, so I'm wondering about your thoughts on, on when you have a, you know, uh, when, when you have a platform, uh, how hard is it to resist chasing the new shiny object? Is that the right move? Or are, are there any other things that you think Apple should be, uh, you know, changing their strategy on?

A Yeah. So look, Apple's always had this, you know, very clearly defined strategy that, you know, Steve, Steve and Tim, you know, working together figured out a long time ago, which is, you know, they, they, I forget the exact term, but it's, it's something like basically they, they, they invest deeply into the core of what they do. You know, they'll basically work internally on things for many years. They all, they only actually release things when they feel like they're kind of fully baked. Um, right. And, and, and so as a consequence, they have this thing where, and Tim says this, right. Uh, you know, they're rarely first to market with new technologies, you know, They're, they're more often in the category of what, you know, Peter, Peter Thiel calls last to market. You know, they're, you know, they'll, they'll come out whatever, three years later, whatever, five years later. You know, they're, you know, there were tablets for years before the iPad. There were, you know, smartphones for years before the iPhone. Folding phones.

AI assessment note: “they're rarely first to market with new technologies”

Q Did you, did you know, did you, did you, was that, was that your guess early on? Did you, did you, did you fully see that the, that the TAM for this was not, Like, how many designers are there and just, like, add it up and multiply it? Uh, did, did you, I imagine you believe that it could be used across?

A It's, it's, it's, it's, it's similar to what I was saying with the, you know, I think the creative technologist archetype who's gonna do a lot of things and eventually, in 10 years, the way that companies run their product engineering teams will probably look very different. I think similarly, like, their worst, there were even, you know, predating our investment, there were definitely examples of teams that went all in on the Figma way, collaborative way of building products. And you had, you know, full product engineering design teams at small companies or even at midsize companies. I remember Airbnb and Square were in our portfolio and we talked to them before we invested and they were, you know, early at pioneering this, you know, we're getting off of the Dropbox links being sent around and we're, you know, we're going all in on, on, you know, working in Figma. And it turns out it was a much better way of shipping products if you wanted to ship really good products quickly and therefore the entire world seemingly, um, changed its mind. In a few years.

AI assessment note: “predating our investment, there were definitely examples of teams that went all in”

Q the Dolce Trilindy. It's a multi-hundred-thousand-dollar car. Even if we can't subsidize that, maybe we can just take away the subsidies from electric cars to tilt the, to tilt the playing field in favor of the, of the larger engines, the supercharged V-Aids, the twin turbos. The G-sixty threes. These types of cars that Americans, you know, deserve to be driving. So, what's the news on the EV credit front?

A I agree with you that the backbone of the American economy is the Ferrari, right? It's impossible to me to imagine a world. Thank you for the last track. I appreciate that. Where a blue collar factory worker cannot own a sports car is beyond me. This bill doesn't address that. I wish it did. It'd be great if it did. It's not going to solve that huge unsolved problem. What it will do, by the way, is continue to drive a wedge between Elon and Trump. Right. Then you saw Elon take the Twitter. I imagine you guys probably mentioned it earlier. You saw Elon take the Twitter and start to really crusade against people voting for this. That is in part because at least the house version of the bill was much worse on a lot of the subsidies for electric vehicles than a Senate version. The big challenge is going to be going into now the conference committee. It's going to be how do they reconcile these two challenges so that they can hopefully tilt the market back in favor, of course, of Ferraris and supercars nonetheless, while also not alienating One of the two power centers of the GOP, at least today in Elon, who could easily primary somebody who votes against.

AI assessment note: “the house version of the bill was much worse on a lot of the subsidies”

Q the new, the new paradigm of, uh, for, for engaging with With language models. I'm curious, uh, any, um, is all the stuff around every B to B SaaS player descending on this sort of like single interface, like a chat interface that generates software? Was that predictable to you? Do you think that's, do you think that that's like part of a multi-year trend, or is that just FOMO?

A I mean, like copilot was early. It was like. Like the first GPT three copilot came out and like, that was already one of the early, like LLM applications that people were interested in at all. And then it took until cursor for it to really like, I think cursor plus like three, five summit was when it became a thing that was good enough that people were excited about it. Um, and it really ushered in the trend because people were starting to find it more useful than a toy and like a thing that actually want to use day to day. Um, and so I think like that's one path is like, and then the more background agent kind of things are like starting to take off now, which I imagine like those will get reliable enough that they're like useful for cranking stuff out. Um, they already are kind of depending on what you're doing.

AI assessment note: “it really ushered in the trend because people were starting to find it more useful”

Q Are you seeing any infrastructure players trying to do, like, do anything on, like, content verification side and, like, trying to create some sort of, um, mechanism to, to prove whether something was, like, authentic, you know, actually shot on an iPhone, right? Right. You know, proving through the metadata and some type of, like, public, um, setting. Is there, is there any pitches, uh, from, from that side?

A Yeah. So, most, Largely, honestly, today that has come in two places. One is the model companies themselves will often watermark the content in some way, like the VO three generations has a little VO three, 11 labs, it's just the audio. They actually have a site where you can upload any audio and it will tell you if it was generated with 11 labs or not. That's cool. Um, which, which is pretty cool. The other, um, place we've seen development there is for like prominent individuals, um, like, you know, celebrities or someone who's There's like value behind their brands and who potentially even might want to monetize it in the age of AI. Like if you're an actor and you suddenly don't have to, you know, film, go fly back to LA when you're filming a movie in Australia to tape like five ads for some cell phone brand, and you can have your AI avatar generated to do it instead. And it looks just as good. Like you might actually want to, you know, have some licensing company that owns your AI licensing rights, whether it's your traditional talent agency or not. Um, who can manage that for you?

AI assessment note: “Largely, honestly, today that has come in two places.”

Q Like they do with cloud services, right? I don't actually know who's under the hood.

A I think they do a lot of that. And then I think a lot of it comes down to the privacy angle is that they want to make certain claims about privacy and ESG and net zero and stuff. And so if they were to just white label, uh, another company that didn't have the same reputation in privacy or, or environmentalism, Like people would just say, wait, wait, wait, but like anthropic isn't, isn't a known for their privacy rules as much as possible. So maybe they're training on my Siri results now. And that's a whole, that's a whole thing. And they have to, you know, build up anthropics brand to have the brand of apple, which is just, I'm sure anthropics doing great, but they're not known like apple is in terms of how seriously they take it. And then in terms of, uh, in terms of acquisition, obviously there's all the antitrust stuff. Uh, we'll get into this with a scale AI deal, but, uh, every major hyperscaler except for apple now, Has basically done one of these like interesting aqua hires of a foundation model lab to kind of juice up their AI efforts. And, uh, and, and it's kind of been a mixed bag. Maybe we see one from apple. They certainly have the cash, but, uh, it would be a very, very different. It would be a departure from the current strategy. Anyway, let's go to, uh, John Gruber over at daring fireball with his review of liquid glass. He says, I've got iOS. 26 installed on a …

AI assessment note: “I think they do a lot of that. And then I think a lot of it”

Q Yeah. So, so, so break down the state of the art because like embryo screening exists. I think most parents in America, at least if they have the means do some sort of screening, uh, while the embryo is growing, is this purely for IVF? Is this just going a layer deeper? And then is, I want to talk about the regulatory and FDA component as well.

A Let's talk about it. So basically if you go to an IVF clinic today, you're a couple, the vast, vast, vast majority of clinics. The first thing I should understand is that the IVF process is principally controlled today by clinicians or doctors. Honestly, couples don't have as much liberty in our perspective as they should. It's their baby. It's their embryos. They should have the right to those, that information, and they should really pick off any vertical. However, today in the clinic, what generally happens is people test embryos for very rare and severe genetic conditions. For example, like a chromosomal abnormality, like Uh, down syndrome, for example, or even a condition like cystic fibrosis or Tay-Sachs or PKU, right? These are conditions that are very rare, um, that maybe someone might have a carrier for cystic fibrosis, but again, it's, it's pretty rare. Um, then there are conditions that we've all heard of, heard about things like breast cancer, things like coronary artery disease, the things that actually kill the vast majority of people today, right? Chronic conditions kill the vast majority of people today. Those conditions are just not tested for in the clinic, even though we have very good science, actually, that can make those predictions. How do we know this as a, as a DNA company as well? That's what we do, right? We build models that predicts disease and the …

AI assessment note: “instead of just looking for really severe, like down syndrome, cystic fibrosis, why not do breast cancer?”

Q aligned people and crypto funds, and it was split almost exactly 50 50. Like the entire crypto community really did go super bipartisan with the spend. How do you bucket? How do you think about the adoption across government B to B or just direct to consumer? What are the key drivers for the different, uh, the different constituencies and then the different applications and how do they fit together?

A No, that's a great question. And I, like, I have a broader kind of framework I like to talk about, which is sort of, I, I sort of like to distinguish technologies between what I call inside out and outside in. And some inside out of things that sort of start with established institutions, like AI to some extent is like that, you know, the iPhone was like this. It came out of Apple, a very established institution. AI came out of, you know, Stanford and so forth. Um, whereas crypto very much, you know, Bitcoin started at the fringes, right? And sort of like open source software and there's, you know, other kind of tech movements that started at the fringes and it sort of worked its way in, right? So stablecoin started for, The main use case was settling, you know, crypto trades, you know, seven years ago or something. And then over time, you started to see more and more like, you know, payment providers in Argentina, for example, and that, you know, sort of more, more kind of moving to the center. Now Stripe, you know, I think of Stripe is, is very much probably the, I think the smartest, a lot of people think the smartest, one of the, if not the smartest fintech company is all in on it. You know, they did a billion dollar acquisition, a bridge, you know, to ramp up their efforts. Um, I think the main, the gaining factor, like I We speak to a lot of like, like what I would love t…

AI assessment note: “I sort of like to distinguish technologies between what I call inside out and outside in.”

Q That's fantastic. Uh, talk to me about the tech stack. I imagine that you mentioned command lines. Are you doing post training on a, on a foundation model? Are you using llama open source stuff? Like what, what, how do you actually, uh, Solve the technical problem. What, what is deeper in the supply chain of your tech stack?

A So there's, as I mentioned earlier, like seemingly everyone is focused on vertically integrating control policy. All the big companies are doing this and you get big horizontal companies that are awesome, like physical intelligence. And I really don't want to compete with these people. In fact, I want to work with them. And so when I initially put my mind to how do I construct this architecture? I, I had to say like, okay, I absolutely need to be an augmentation atop the transformer layer. So after, uh, you know, a robot is told what to do by its decision transformer, I run this interaction layer. Now I want you guys, uh, you guys have both heard the phrase, time flies when you're having fun.

AI assessment note: “I absolutely need to be an augmentation atop the transformer layer.”

Q that once mid-journey gets good enough, uh, or kind of hits some, some peak, that they would just bake it down into silicon, and we would just have image generation models, you know, ASICs, essentially, like what happened with Bitcoin. Uh, is that the future here? Not that you would pivot into hardware, but maybe you would, like, vend your software into a hardware provider at that, at that level?

A Um, well, so we, like, I mean, we partner with, like, folks like Cerebris and Drock. Um, and so we allow you to kind of plug in their models and plug in effectively their hardware accelerated inference. Um, and, and so we're, we're compatible with that world. I think on the inference side, you can definitely, I think it's going to continue to push as low as it can get. Um, there, there's obviously some limits, you know, it's a trade-off between, Uh, kind of capabilities and level of knowledge and how fast you can kind of, kind of, you know, run that, that pass through, uh, the model to get the result out. So there are trade-offs, of course, that follow the laws of physics, but There's also kind of diminishing returns after a while. To give you an example, I once built this, uh, this Cerebris demo. I used like a Lama seven B or eight B, uh, Lama eight B have to remember these numbers, uh, on the primary accounts, but, uh, Lama eight B hooked up to Cerebris. And, uh, I got a bunch of feedback on that voice demo that the model was responding too fast and can you slow it down? And it's kind of going off the rails a little bit.

AI assessment note: “we partner with, like, folks like Cerebris and Drock... we're compatible with that world.”

Q Does that mean like a flourishing of RL Uh, big transformers for different tasks, or are we still searching for like the God model that can do everything all at once?

A I mean, so, okay, I think there's a, A couple leaps we need to have models that can do everything all at once for like a super long amount of time. Um, like my, I tweeted something about this, like my, I'm happy to call O three, 10 minute AGI. And I think like framing AGI in terms of like length of time, it takes a human to do a task is like more reasonable than like a global framing. Like, sure. There's a bar of like drop and replace for a human that we are like definitely not at yet, like for general jobs. But most things a human can do in 10 minutes, you can, like, get O three to do that pretty well.

AI assessment note: “I think there's a, A couple leaps we need to have models that can”

Q It's all about DCF. I mean, you, you could, you could justify the cash flows, right?

A You could comp to the public market, the problem, but the problem with the, first of all, the company is about the basically nothing at late stage is trading comp to the public market really. Right. And we can get, we can get into like why, how I think, I think that look, the DCF, Comp to the public market. That way, that is the factory model, right? It's basically saying like, Hey, I have a late stage thing. I put money in. It's going to triple. The DCF looks like this. It's has this much profit margin. Like this is the story package sell to the public market. The public market buys on that same story. Like that, that was the mentality that persisted for a long time. And it was a great system for a lot of, for money making for a lot of people. Right. I actually think that again, the public market Now is like, well, if I kind of just want those types of metrics, why don't I just buy more of the mag seven? Like, I don't, I don't even want to dick around with your subscale offering. Like, I don't care, right? Like I think, and there's reasons for that. It's because the big LPs are bigger. It's because of meme stocks, like a whole bunch of stuff going on there, right? That like kind of makes that happen. But the net outcome is there is no off ramp. So then the question is when you're underwriting at a late stage, you're not, you also have to underwrite to someone Buying from you. …

AI assessment note: “basically nothing at late stage is trading comp to the public market really”

Q No, you, you explained that to me, uh, a while back. I had fun with that. Um, so, so let's talk about LamaFour. How should, uh, how should startups be thinking about, uh, LamaFour as a tool in the toolkit against the other options that they have?

A Yeah, I mean, I think like, you know, Twitter is equipped to, and research in general, right, is equipped to, um, sort of view state of the art as the only thing that matters. Um, and I think that actually in many cases being first is overrated. You know, we've seen, you know, going all the way back to sort of the Slack in teams charts, where you've seen the famous chart, Slack versus Teams, right? Distribution is incredibly important as long as, you know, sort of the incumbents can wake up and can catch up. Um, you know, I would not bet against Zuck And a hundred billion dollars of profit per year. Um, you know, I think that, uh, you know, Zuck also is in some sense playing a different game. Like he's not trying to build like the, the sort of very best open source, like chat experience for consumers. Um, what Zuck sees, I think rightly so, is that, you know, having an open source model, which is really good, is good for the ecosystem and is good for meta. Um, and you know, most businesses don't love using closed source models. They want to use open source models. Um, for all kinds of reasons, you know, privacy, security, continuity, cost, you know, uh, you can build your startup on GPT-IV and it's amazing. Um, and then, you know, there's a new version out and, uh, opening eye deprecates, you know, the old version, right? And all of a sudden all of your prompts don't work the s…

AI assessment note: “most businesses don't love using closed source models. They want to use open source”

Q about, uh, entrepreneurial storytelling? It seems like it's an incredibly valuable skill, but then some founders get maybe lost in the sauce and wind up just focusing purely on storytelling, and then that needs to be handicapped. Um, what, what advice do you have for founders, uh, when, when, when they're, when they are trying to tell kind of that definitive optimist vision of the future without seeming not credible?

A I think, I think we've actually gotten much better in the last couple of years at separating storytelling from just charlatan bullshit. And like, there was a period when we weren't like, we were not very good at that. Probably, you know, late, late twenties, early twenties, we were, we were not very good at that. And there were a lot of people who got away with a lot of things that they should not have. But I think our, our threshold for BS has dropped as, as an industry and in a very good way. And people can see through things very quickly. And so it's always going to be the case that, you know, Steve jobs was the best storyteller and he was the opposite of a charlatan, the, the, the polar opposite of a charlatan. And I, I, I definitely feel like there is a world where yes, you have to be a good storyteller, but there's a much stronger sense of put up or shut up. Like you have to show me the numbers of what you're doing. You can't just keep the story going forever. So that's, that's a great thing.

AI assessment note: “put up or shut up. Like you have to show me the numbers”

Q they're speaking freely on the ski lifts together. They know no one else is listening. They share stock tips and then the market crashes. And so if you look at like Theranos, it happened during ski season. FTX happened during ski season. Enron was discovered during ski season. Uh, what do you think of that? Or why do you think the market's going down right now? What, what's gone wrong?

A Um, you know, there is a, like, uh, historical seasonality to, like, the stock market. I remember, you know, sort of seeing something similar where it was basically the past, like, you know, sort of 50 years of the steepest drawdowns of the S&P 500, something like 80% of them happened between, like, October and February. There's a component of that that I'm sure is also, like, tax season related, where, like, people want a lot, you know, sort of gains versus losses at the end of the year, but not too closely, you know, sort of coupled to it. I'm sure there's a component of it is, you know, people sharing, you know, stock tips in Aspen. There is also the presidential transfer, whether it's, you know, sort of this year or any prior year always happens at this time of year, which also introduces volatility and people don't like volatility. The one bull sign that I'll say is I was just looking at this. I don't know if you guys know this website, Truflation, uh, which basically tries to do like a week to week, um, analysis of inflation, just tracking grocery store prices that are, and as it is actually showing like a pretty steep drop. That's great. Yeah. So I'm, I'm a little bit of a believer of like the Scott Besant, um, you know, There is going to be a detox period.

AI assessment note: “there is a, like, uh, historical seasonality to, like, the stock market.”

Q a lot of sense in the, in the age of AI. Do you know more about, um, kind of like early use cases, obviously like the frontier insane, like, you know, gigawatt scale that's going to come in the future. What are, what are some of the more exciting use cases in like Like the near term, is it just like fine tuning stuff? Do you have any insight there?

A Yeah, yeah. So I mean, they're, right now, they're basically working on just like scaling this up to like state of the art level models. Like they did the 1.1 billion parameter, you know, training run. Like a few months later, they completed like a ten billion parameter training run. And then, you know, basically just serving as like the, the peer to peer marketplace between the supply and demand side. So like on the demand side, you have like AI startups that need extra compute, like labs, like, you know, random independent developers, that sort of thing. And then on a supply side, it's like data centers and then individuals and also startups with like idle compute. Like you have like hugging face, like semi-analysis that want to just like earn extra for, for their idle compute. So it's basically just like, like the way we think about it is just enabling like the AI market to like progress at a much faster rate because it's just like enabling all this idle compute to be To be put to work. And then on, yeah, on their end, it's like basically what they did is Google deep mind released a paper called like D loco, um, which stands for like distributed low communications framework. Um, and they implemented that, but in like an open source way, like they called it open D loco, um, such that like you can train across like many different continents. And that was like a huge step funct…

AI assessment note: “they're basically working on just like scaling this up to like state of the art level models”

Q like, well, that's like a stupid technical feature. Like, why can't you just make a, uh, a combined checking and savings account or a, a yield bearing checking account? But it's like, well, there's so much cruft in the legal system around the American financial system. So it's like, maybe, yeah, you do just need to like, Have some sort of greenfield fresh start and that's enough. I don't know.

A Yeah. So at Capital we had, uh, we had a high yield checking account. So it's not that, um, it is a good, I mean, I think the, There's so much, um, platform and infrastructure risk in all of this stuff, right? Like one of the reasons that they maybe had to buy Bridge, I'm just speculating, is that Stripe's existing bank partners had no interest in taking on that level of risk, right? We, at Capital, we were rolling out a Stablecoin custody product when we were informed by our bank partner that we, we were gonna have to jettison a lot of our Digital asset oriented customers. And so our entire product strategy like fell apart because we couldn't do the thing that our customers wanted. And in fact had to be like, oh, you guys should like leave basically. Um, so there's so much, um, that, that is, I feel like the whole bridge story is such a good example of risk taking, right? There's so much that bridge was trying to do that when they started was unclear If it was even going to be possible. And so oftentimes Opportunities like that attract unsophisticated founders that don't know what they're getting themselves into, and then founders that are sophisticated and already successful and comfortable with the risk, right?

AI assessment note: “at Capital we had, uh, we had a high yield checking account.”

Q How do you get a job as a venture capitalist in twenty-twenty-five?

A Um, so I, I mean, look, the, the best way, the best way to do it is to have a track record early as somebody who is like in the loop specifically on new product development. Um, and so somebody who, you know, be, be like deeply in the trenches, um, at one of these new companies in one of these spaces, um, you know, participate in the creation of, of, of a great new product, uh, and, and, and a great new company and, you know, really demonstrate that you know how to do that. Um, you know, there's, there, you know, there, there are great VCs who have not done that, but, you know, I think that is sort of a foundational skillset. Uh, you know, for working with the kinds of founders that, that you want to work with, who are going to, who, you know, are going to want you to have, you know, kind of very interesting things to say on that. Um, as I think that, you know, still the, the best way to do it.

AI assessment note: “the best way to do it is to have a track record early”

Q How do you get a job as a venture capitalist in twenty-twenty-five?

A Um, so I, I mean, look, the, the best way, the best way to do it is to have a track record early as somebody who is like in the loop specifically on new product development. Um, and so somebody who, you know, be, be like deeply in the trenches, um, at one of these new companies in one of these spaces, um, you know, participate in the creation of, of, of a great new product, uh, and, and, and a great new company and, you know, really demonstrate that you know how to do that. Um, you know, there's, there, you know, there, there are great VCs who have not done that, but, you know, I think that is sort of a foundational skillset Uh, you know, for working with the kinds of founders that, that you want to work with, who are going to, who, you know, are going to want you to have, you know, kind of very interesting things to say on that. Um, as I think that, you know, still the, the best way to do it.

AI assessment note: “the best way to do it is to have a track record early”

Q continue to, to run, uh, and scale the business. You guys had a lot of momentum, but, uh, what was going through your head Friday night? Because I, I imagine at the time you were also balancing, uh, you, you were, Uh, all of a sudden managing a team of hundreds of people that, uh, I probably all, uh, you know, wanted, wanted, uh, your time and attention as well.

A My, my immediate priority was just to get a lot of options on the table and to have a lot of pass forward. Uh, I had to tell the team, like, this is the path forward immediately because this is, this is happening now. Uh, and then outside of the meeting, you know, I, I was thinking like, okay, who do I need to talk to? What, what is the best use of my time? I can tell you, I was on the phone for pretty much 24 hours nonstop on my phone, uh, after, after the all hands and, uh, me and Scott, we, we worked really fast. Uh, we met at our office the next day. Um, he even brought in everything on a piece of paper to sign. Uh, he even brought, he took the, there you go. Uh, he had to talk through a lot. Uh, but that was, uh, it was really, it was pretty dramatic, uh, Scott, but it was cool.

AI assessment note: “My, my immediate priority was just to get a lot of options on the table”

Q applied to. Do you think you could have pulled this off with Microsoft, Google, Amazon and split three times? Or do they have defenses in place that would make this difficult? There's a variety of memes around maybe big tech companies. It's easier work. Maybe startups are harder. I'm just kind of curious about, it seems like you went for a lot of startups. What was the reasoning behind that?

A Yeah. So it feels counterintuitive. Like if I was in need of money, like why would I work for, you know, startups as opposed to big tech because they clearly pay well, they have a nine to five schedule for the most part. Like I would say, you know, people don't really care about like what you are essentially like doing. Um, the thing with me is like, if you're spending, you know, Multiple hours a week, like working on something, you would have to at least decently be passionate about it because otherwise you'll just like burn out. Um, like I said, you know, each of these companies that I've worked for, and again, uh, you know, founders, some of the founders have spent meaningful time with me and vouch, um, you know, would, would say that I actually cared about these companies. So it wasn't like cold email without context. Like I deeply read into what a company was doing, what the business model was, you know, who their customers were, I had great ideas about, like, you know, what to build for them, what not to build for them, you know, what the platform is like. Um, you know, I did deliver, like, beyond engineering for a lot of these companies, and, um, yeah, like, it was more kind of like, hey, if I'm spending, like, 1:40 hours, like, I want to do something that I actually care about. Like, I don't want to, like, you know, do nine to five and kind of, like, send out a div in, …

AI assessment note: “I want to do something that I actually care about. Like, I don't want”

Q trajectory with Varda. He was talking about ZB Land at one point, then it was Pharma. Now there's some DoD mixed in there, some government contracting. It feels like a lot of these companies that are doing stuff in space or doing stuff in hard tech, it's dual use. Um, is there a government angle here at some point, or is that just something you're thinking about in the future?

A No, it's, it's very near term. It's very real term. Um, very real term. Um, very, very real. Yeah. Across a number of different applications. I mean, they're dealing with the same challenges, like if not even more so, um, where like they have aging assets that, um, are kind of infrequent. Like there's, you know, certain networks that just don't have a lot of assets and they're old and they're vulnerable to outages, whether it's like an intentional outage. You know, by somebody targeting that site or not. Um, and so there's been a lot of interest in how that they can leverage commercial to get, um, sites deployed quickly. Like for us in the conversations we're having, we're really emphasizing like we can deploy capability quickly, um, and we can serve up capability that's like quite scalable. So if one of those outages happens, you'll have that, that backup and that resiliency. Um, and so as you know, government use cases, Um, like so much of our world runs on space in a way that I think people don't really realize. And so, um, it was, you know, that's been a refrain that you're hearing more and more through government stakeholders where there's this concern on, you know, if anything goes down in space or, or through the ground connectivity, it has ripple effects through like a lot, a lot of our critical infrastructure. Um, and so for us to be able to, uh, you know, deploy capab…

AI assessment note: “No, it's, it's very near term. It's very real term.”

Q so much distractions in life. So maybe everybody just moves like all the time that was spent writing emails, just moves on to, you know, tick tock or, or X. And it's just like, you know, you're auto completing and then there's, there's There's brain rot. That's, that's sort of like a, a dark take on it, but are you worried about what humanity loses by not thinking through writing?

A I'm absolutely worried. I mean, I see major white pills and major black pills here. Like, I think that if you're not seeing the pros and the cons, and I mean major pros and cons, you're, you're, you're missing out in a fundamental way of what's really going on. Um, I agree with you. I think a lot of our information is Or a lot of the way that we communicate is like, did this person actually write it? And I think that what's going to end up happening is you sort of see this in the differences. Like, have you noticed like how big of a separation there is between the vibe of like your private group chats versus like what you feel in public? I feel like the private group chat vibe is just going to be like, go even more. We're going to have to learn to write with voice and really show off our distinctiveness in writing as almost a Adaptive way to say, hey, this isn't written by LLMs. And at the same time, I completely agree. I mean, I think that, you know, people are super busy. And the one thing that we've learned time and again about technology is that people value convenience. You know, like I listen to music all the time on my crappy iPhone speakers, because I don't want to get up and walk to the other side of the room and like hook in my USB cord. And I think you should never bet against humans going for convenience. In the aggregate. Um, so that's how I'm thinking about it.

AI assessment note: “I'm absolutely worried. I mean, I see major white pills and major black pills”

Q I, I didn't understand why, why is the jury just, like, giving, like, an advisory decision? What is the history of, like, why do you, why do you have a jury when the judge is ultimately going to make the final call? It feels like just kind of putting on, like, a show, because, like, theoretically, the judge could just Sit through a bunch of depositions and make a call.

A Sure. No, I, I think, um, so I do think they want More often than not, want a jury of these CEOs and company peers to be the deciding factor in what they feel like is good for a civil claim. Like, that, I think, is fairly standard, but to your point, um, the judge can throw out their verdict, which is, like, is, and judges, I don't think, tend to want to do that because, like, they want to have reliance on, this is the public, the public should have a say in what goes or whatever, But, uh, the judge can do that. I will say, also, she is responsible for, uh, if they're, if Elon, or sorry, if OpenAI is found liable, judge decides on remedies, damages, and things like that, so she still has an active role in that regard, and in steering the case. But I really do think that, uh, courts often prefer, or often appreciate A peer, a jury of your peers making some of these decisions. So I don't think it's going to be, like, completely disregarded is what I would say.

AI assessment note: “courts often prefer, or often appreciate A peer, a jury of your peers making some of these decisions.”

Q Which is like, I mean, he's 54, he's got a while, right?

A Yeah, if he's a radical life extension guy, he could be like 300 years from now, who knows? Uh, that's a very vague timeline. Where, whereas in the past, with Point to Point, with Roadster, with Cybertruck, with Space Data Center, all the other Elon pitches have been very focused on Uh, we think we can do this in five years. We think we can do this in two years. We think, and, and he always gets criticized for being Elon math, too aggressive, but then history's shown that, you know, he, he over promises, under delivers, but he typically does deliver eventually. And, and if the technology is completely new, no one else was going to build it, and he's the first one to build it, like, it's fine, because Starlink works. It was the first Leo constellation for satellite internet had turned into a huge business. Wound up being a great technology, and so, uh, even if, even if it was late, I don't even remember the, the, the Starlink predictions, but, uh, if, even if that was late, it's fine. It was still good business. Uh, now the mass driver's a lot crazier. I was, uh, going back and forth with Tyler about this, trying to get him to, uh, nail down a prediction. We started with 100 years. Do you think it's possible to put a mass driver on the moon in the, in the next 100 years? And you said yes. So it's possible.

AI assessment note: “Yeah, if he's a radical life extension guy, he could be like 300 years”