The Exchanges

Q What about the Super Bowl ad? Was that a similar one person green lights it, and who green lit it, and what was the, what was the story with the QR code?

A Yeah. Well, as, as usual, like success has many fathers and mothers, but, um, you know, it, it partially came out of a rushed, uh, like a short time, you know, time constraints breed creativity. Right. So we actually didn't, I think by the time we bought the ad, we didn't have enough time to like go film something for months and months. And so the team just got really creative and, and, um, I don't even remember who it was in the meeting, like our, our CMO and there was an agency involved and it was a collaborative effort where someone was like, You know, well, what if we did kind of like a direct, uh, direct marketing type campaign where we want people just to install the app, but you can never measure the result of most of these brand ads, right? You're always like, I don't know, did it drive revenue or not? Nobody really knows. So we're like, well, we could actually measure how many people just like scan it and sign up if we, if we just put a QR code on the screen. So I don't remember who said it actually in the room, but it obviously it worked. It was cool.

AI assessment note: “it was a collaborative effort where someone was like, You know, well, what if”

Q Yeah, it makes so much sense as a mini-app, because in many ways, people would just go to GoFundMe, Create this thing and then take it to social to promote it. And why doesn't it just exist natively in the product?

A Yeah. And I think the other thing that you, you have is like the liquidity in the sense that I'm not fragmenting. I don't fund my GoFundMe wallet over here, or I have to like connect this payment method for the seventh time. And again, anytime you're using like a card based payment method, Visa and MasterCard and the banks are getting their three percent. Right. So I think it will be slow. In many cases, but this stuff will just become increasingly pervasive, and people will choose it, If the kind of like platforms or creators are giving an incentive, right? Like if I can not have to pay three percent to a card network by having you as a stable coin, I'm going to make that the default payment off it. And maybe I give you a two percent cash back or you get some loyalty points or an NFT or whatever reason to, to get you to switch your, your payment method.

AI assessment note: “I don't fund my GoFundMe wallet over here, or I have to like connect”

Q say, well, we don't actually want to release this because we should just use it ourselves, but it, it sort of could potentially tie into this dynamic between closed source AI and, and open source AI. And is that, Do you think part of Coinbase's opportunity is to, you know, make sure these types of tools are available to as many people as possible versus just a handful of traders?

A Yeah, it's kind of, I mean, so we're not doing prop trading on our own balance sheet. So we're not like trying to be a hedge fund or something with a proprietary algorithm. So our goal would be to like make this available to as many consumers as possible. And yeah, I mean, it might, it's kind of adjacent to financial literacy, right? Like many people might come in and be like, yeah, I want to make like crazy returns. And this thing can kind of tell you, well, okay, that's pretty high risk. Like, why don't you like dollar cost average into this, um, Diversified portfolio over time, and like that might be more in line with your risk appetite, but if you want to take 10% of it and try something more high risk, like, you know, so it can, it can teach people about these concepts, um, and while also doing what they ultimately want to do, what they would have done anyway in the app, but it's just making it easier for them.

AI assessment note: “So our goal would be to like make this available to as many consumers as possible.”

Q kind of an, is kind of, You know, indexed, or at least somewhat correlated with, like, Sam Altman's career, and, and even people that are in more, like, serious, less, like, meme token, meme vibe trading career paths have found themselves indexed to a particular meme coin through one way or another. Do you think that's something that's, uh, like, a temporary phenomenon, or something that will continue to grow?

A I think it'll continue to grow. I mean, this is how, like, every company sort of works, right, in a way. Uh, every executive is tied to their business, regardless of the fundamentals. Like, if, you know, I mean, you could apply this for anything, but if some CEO of a publicly traded company showed up on a TV show slurring his speech drunk, I'd imagine it'd affect the stock price, even though nothing has changed with the underlying asset, right? I mean, Tesla commands, what is it, like, a hundred X earnings because of Elon Musk. Um, so I think it'll be a growing phenomenon. I, I don't know If meme coins specifically are the mechanism, but I think that like pricing in people, it's definitely a growing trend. Um, yeah.

AI assessment note: “I think it'll continue to grow. I mean, this is how, like, every company”

Q the FDA decides to approve drugs, whether the, whether the benchmark is better than a placebo, better than nothing, or better than the standard of care. Um, how do you, how do you see the FDA's mandate In drug development, what is their goal? Are they trying to protect certain, certain IP structures by limiting the access to drugs, or is it more purely based on safety and efficacy broadly?

A Yeah. Um, I think, you know, so we've had conversations with the FDA and at the end of the day, the FDA is really a group of people trying their best to approve drugs that work. Um, and then, uh, you know, rejects drugs that might have safety issues or are the same as placebo. Um, so for us working in rare disease, a lot of the times we are up against placebo or even historical controls, um, because there's usually no approved treatment. Um, in these specific diseases. So it's much easier to get a drug approved there versus if you work in, you know, cancer, which is very common, um, like breast cancer. Um, there's already so many drugs that are approved and effective that you have to run really long trials against drugs that already work. Um, and that ends up being extremely costly, extremely expensive. And there's some debate about, you know, is there a way to get around this five, six year trial?

AI assessment note: “the FDA is really a group of people trying their best to approve drugs that work”

Q Well, anything in voice agents, or just improving the customer experience broadly, like surprise and delight, right?

A Well, you know, it's one of those things that I think is funny, like people talk about hallucinations as this bad thing, but in a lot of ways, for like, kind of this speech-to-speech interaction model, where you're talking to an AI, Hallucination is a feature. So if you remember like your Alexa or Siri, like when you were talking about this, you know, 20, 10, 20, 12, you go and you ask it a question and it's just going through a bunch of like a decision tree. And if it hits, you know, a dead end and it can't answer the question, it just doesn't do anything or it says I can't answer the question. And so the, the second that it, it can't do something you expect it to do all of a sudden you, you know, it's a, it's such a punishing user experience. You just don't want to use the thing anymore and you stop using the thing, but. With kind of these new models that actually, you know, hallucinate answers into existence, even if it doesn't know the answer, it tries, um, you always get a response, uh, that comes back from the model. And for, for like voice interfaces that actually can be more of a feature than a bug, um, it also helps that these models.

AI assessment note: “for like kind of this speech-to-speech interaction model... Hallucination is a feature.”

Q Oh, well, well, anything in voice agents, or just, just, just improving the customer experience broadly, like, surprise and delight, right?

A Well, you know, it's one of those things that I think is funny, like, people talk about hallucinations as this bad thing, um, but in a lot of ways for, like, kind of this speech-to-speech interaction model where you're talking to an AI, Hallucination is a feature. So if you remember like your Alexa or Siri, like when you were talking about this, you know, 2010, 20 12, you go and you ask it a question and it's just going through a bunch of like a decision tree. And if it hits, you know, a dead end and it can't answer the question, it just doesn't do anything or it says I can't answer the question. And so the second that it can't do something you expect it to do, all of a sudden you, you know, it's a, it's such a punishing user experience. You just don't want to use the thing anymore and you stop using the thing. But With kind of these new models that actually, you know, hallucinate answers into existence, even if it doesn't know the answer, it tries. Um, you always get a response, uh, that comes back from the model and for, for like voice interfaces that actually can be more of a feature than a bug. Um, it also helps that these models. Yeah.

AI assessment note: “for like voice interfaces that actually can be more of a feature than a bug”

Q then getting, like, a sculpture out, like, that's, that sounds really awesome and futuristic. Um, yeah, yeah, go. How, uh, a lot of investment in humanoid robots lately, many of those promising to revolutionize manufacturing, replace human labor, automate the production of lots of things. As somebody who's been doing it with robots since How do, how do you think of that form factor in the context of, um, manufacturing?

A Yeah, I'm actually super bullish on here. That's right. I think the question is, to your point, is it going to be, is the first application going to be in, um, in manufacturing? I don't think so. Right. Um, I think the biggest problem, if you want to like think of a startup as like, what do you need to do risk first? The first thing you gave an example of like Nike figuring out how to do the shoe manufacturing. The first thing is actually intelligence. We have kinematic frameworks for a long time. You know, we could do what humans does in terms of kinematic freedom with industrial robots. We can actually do it more precisely with higher force with industrial robots, which was what we need in manufacturing setting. So the missing piece really was intelligence. So we're kind of a little bit intelligence first, right? We don't need to solve the joints. We don't need to solve People walking around, like robots walking around and, you know, having the human four factor. If you solve the intelligence, um, you have enough kinematic frameworks, which is industrial robots, to do what you need to do. But that being said, I think, you know, humanoid is a huge opportunity, maybe not in manufacturing, but, but downstream in homes and all other places that, that, that, that we can use humanoid.

AI assessment note: “is the first application going to be in, um, in manufacturing? I don't think so.”

Q actual machine side, are you, are, you know, you and the industry seeing challenges of like, you know, how, how much of like the actual robots that you guys are leveraging are, are sourced, uh, outside of the country? And Chris and Hadrian was saying like, yeah, my capex goes up by 10%, but The demand has been way, way higher. So what's been your experience? I mean, that's true.

A I mean, like, what we can, what we do here, the challenge is, for a lot of work that we do, and especially in the defense, aerospace defense, There is no industrial base, so we have to do it. So the demand is always there. Now, 70% of our bomb is off the shelf, and we intentionally tried to do that so that we can actually finance it easily, right? Use, you know, multi-push slippers, off-the-shelf equipment so that we can easily finance, um, the full hardware stack. That being said, you know, a lot of the hardware that we use is either produced in America or it's in very, like, allied countries. For example, uh, robots were produced in, in, um, Japan. Um, so the tariffs are a little bit more digestible there, but that being said, you know, we are one of the very few manufacturing companies that are like doing almost suffer like margins. So there is enough room for us to, to be able to absorb that cost and still, you know, have a very high value, high margin list.

AI assessment note: “a lot of the hardware that we use is either produced in America or it's in very, like, allied countries.”

Q software investors kind of poke fun at, at VCs that are investing in manufacturing, you know, expecting, uh, software-like margins, and, and there's sort of this sense that, uh, well, manufacturing hasn't, for most things, hasn't historically had software-like margins. You're seeing it today. Um, how, how do you think about the margin profile for advanced manufacturing over time? Even, you know, even in things that are like non-chips, right?

A Yeah. Yeah. So. Um, so, so it's, it's interesting arbitrage. I think there's a lot of people thinking about manufacturing solutions in different ways. I think you can think of it as automating what traditionally has been done. And I think that's usually end up being very low margin, right? But if you're creating something new that, you know, has some kind of arbitrage on either neighbor, Or has some kind of arbitrage on equipment, right? In our case, whereas you don't have to make dyes, right? And a dye, a single dye for, um, you know, a car door can be up to a million dollars, right? So for us, we're faster, but we get rid of that asset. Um, so it allows us to have the, you know, at the same cost parity, have higher margin. But I think, yeah, down the road, um, the margins could erode, and that's why we're thinking about it as a platform, right? It's a robotic system that can do forming today. Tomorrow it's going to be your next operation. It's going to do bending. It's going to do hemming. It's going to do forging, right? So you're constantly expanding the capabilities of the system, which allows you to sustain a very, uh, larger volume business, um, uh, as some of the margins of the older processes kind of erode, right?

AI assessment note: “down the road, um, the margins could erode, and that's why we're thinking about it as a platform”

Q Chatbots. Yeah. Wait, yeah, what, uh, wasn't there, this was sort of before my time, but there was a whole chatbot era, right? Wasn't there, like, pre-AI was like, it was supposed to be the thing. It was sort of right. It was just didn't have the, the underlying tech trend. Is that right?

A I feel like the tech wasn't there, but also the branding of chat bot to, uh, chat GPT isn't that different, um, but magically different from a outcome and, uh, sort of adoption curve. If you look like the automated chat bots, I think is what they were called for that moment in time. Uh, they all fell on their face because They weren't interesting enough, good enough. They were like pre-programmed answers. And then, uh, open AI releases in essence, just an advanced version of that in many ways from a consumer positioning. Uh, and here we are. And so you are like, the idea of chatbot was right in how you're interacting with the tech. The tech, uh, to your point, just wasn't good enough.

AI assessment note: “The tech, uh, to your point, just wasn't good enough.”

Q Yeah, what do you think about the far future of, uh, of verification?

A No, I think on, on, on identity, I think we're laughing about it, but I think in, in person is going to become important when you want to want to have the best assurance, um, because online will be able to be faked so much easier and cheaper. So I think there's going to be actually more, it depends, it depends on the use case, but if you have some very, very important transaction, like getting a passport or, you know, like very important points, you still are going to need the biometrics in person. It's going to be. It's gonna be hard to fake that. Um, so I think we'll see more in person use cases and probably in the identity stack, more solutions. Um, the, the, the other thing that I'm bullish on is, um, electronic, uh, driver licenses and wallets, you know, like Apple and Google are starting to have your driver license in your wallet, like Apple pay. And I think, uh, you know, and the government, especially in Europe and other countries, they're moving to. Mobile driver licenses that connect to the government. And if Apple and Google are, are opening their SDKs and continuing to build them right, I think that's going to be a more elegant solution, um, than scanning your, you know, picture of your driver license, for example.

AI assessment note: “in person is going to become important when you want to have the best assurance”

Q Uh, how do you, how do you guys think about, uh, international long-term, uh, get, you know, do you see a world in the future where you have millions of users outside of the country? Uh, I know U S has obviously been the, been the focus today.

A Absolutely. But the way that starts is we actually launched, uh, an app called alpha, which is a little bit of a spin out of our AI product. Uh, launched that in international markets, uh, late last year as a separate kind of product. So that's all alive and growing well in a bunch of European markets and a few other places around the world. And so that's basically like Apple stocks with AI powers. That's how you think of it. You literally come in, you take a photo of your watch list, wherever you have it. And the AI, the auto follows everything and tells you why things are moving in real time, not just they're moving and by how much, and then you can definitely ask questions and do research and all that stuff. So that's like, I would say the first step in our international strategy in building that audience sort of abroad. And then obviously, as you can imagine, investing services could come later because that's the aspect that has a regulatory component to it, which totally.

AI assessment note: “Absolutely. But the way that starts is we actually launched, uh, an app called alpha”

Q of the underlying technology to go from a Roomba. We're pulling up the video here on the stream of, uh, what you've actually built. Uh, what were the foundational turning points in terms of the, like the different models and different breakthroughs? I imagine the transformer was really important, but there's probably a ton of other, uh, developments that excited you. Now is the time, like we're ready to go.

A Yeah, there's been a lot of things that we are, we are building on top of, um, things like transformers, things like vision language models, the concept of pre-training and post-training. A lot of those things transfer to the robotics world, but they're, they're not as well understood. We are still in the process of figuring out what that recipe should be like. We, we kind of have to rediscover this, some of the steps that, that language people had to do initially and see how we can map them onto the robotics world. We don't have the privilege of having an open internet full of data. We need to collect the data ourselves, which on one hand is a big challenge. Like the data isn't there. You can't e-trade nearly as fast. On the other hand, it also gives you more freedom in figuring out what kind of data is the most important and what data to collect. For, for this particular PI-O-V advancement, what we have to do is, one, collect very diverse data set, large diverse data set that involves not only model manipulators in homes, but also static robots in the office or data of the internet. And it turns out if you collect very diverse data across many different tasks from many different form factors, they all contribute to each other. And that they contribute to a better understanding for the model of what actually is happening and how to utilize all of the data to figure out what to…

AI assessment note: “things like transformers, things like vision language models, the concept of pre-training and post-training”

Q Is it purely economics or, or is there something else?

A Yeah, the food ties, again, I think this is one of the most no-brainer, unimpeachable things they could start with, because it's actually not a cost thing. Uh, we've done a lot of engagement with the food industry on this. Um, actually, they've done a lot of research, and they actually cite this as a reason for why they should keep it. They say, well, the kids in focus groups lunged for the neon bright colors, and it's like, well, I'm sure they do, but you're feeding them crude oil, So, so it's really actually, they say it's consumer preference and actually when pushed before the food industry said, well, we believe in consumer choice and kids prefer, uh, these bright colors. So that's one reason I think, uh, the administration and secretary started with the food.

AI assessment note: “it's actually not a cost thing... it's really actually, they say it's consumer preference”

Q like booking a flight and only 50% of the time it like books the flight, it's like, well, I'm just going to do it myself. Right. Cause like, it's only going to take me a minute. Whereas like the dishes could take 20 minutes. Right. So how, how do you think about kind of, uh, the sort of threshold of reliability in order to like really deliver value for consumers?

A Yeah. I think that, that Like, we don't think right now about delivering value for consumers, and it's kind of why we structured the company the way we did. We're a research lab. We're trying to solve this problem of physical intelligence. We really like these consumer-oriented tasks, and I think people tend to as well, like, when they think about laundry being folded for them. I think there'll be a point at which it gets good enough that we can deploy it to consumers, but it's not going to be, like, 50%. It's going to be closer to 98, 99%. And there, I think we can harp on, um, self-driving cars, where there will be a period of, Of interventions, right? Like if it doesn't work, it's not that it just will stop and do nothing for a while. We can have a human teleoperator intervene and finish the task. But I think the other cool thing about the home and consumer use case is there's so much that could also just happen overnight. Um, there's like while you're sleeping, your laundry is folded, your, your, your meals are cooked for like the, you know, they're prepped for the week ahead. Um, your house is tidied. Uh, so consumers still a little far away, but making a lot of progress.

AI assessment note: “it's not going to be, like, 50%. It's going to be closer to 98, 99%.”

Q really, really simple terms, like maybe like I'm a venture capitalist or something. Um, something like, uh, how do we communicate with the Hubble telescope? Is this like a big satellite dish, like what I saw in contact with Jodie Foster? Is that how we communicate with the Hubble, or is there a different network of ground stations? What, what, what was kind of the gold standard 1020 years ago?

A Yeah. Um, I have to just say like us ground nerds in space, we do not often get asked these questions. So thank you very much. Like the ground is just generally like the not sexy part of space. Um, so it's very fun. So yeah, I mean, what you're doing is, uh, you know, generally using RF to contact a satellite that is like hundreds of kilometers away. Um, you need to concentrate enough power to be able to do that. So that's why you see like the big parabolic dishes. It's concentrating power. Got it. Um, and so when you're, you know, further away, that requires more tower, more power. So actually, um, if you're, you know, in the Palo Alto area, you go by the, um, the Stanford dish is kind of a well-known one. I imagine like some folks, uh, if they're watching might know of that. Um, it's massive, uh, a giant dish that's used to make that contact. And so it's interesting because like the legacy of space, it's more exploratory. It's more research-based where like, Booking a Nintendo is kind of more like booking a telescope, you know, where it's like a, an individual piece of equipment where, um, you know, they're located at these different locations around the world. You book the time, you're kind of in control of how that functions and how it operates. Uh, but as the space industry has been scaling, um, it's not really a sustainable model to think of, like, as you're needing to co…

AI assessment note: “That's why you see like the big parabolic dishes. It's concentrating power.”

Q Uh, question. For now, I think there's an obvious opportunity to serve existing space companies. What kind of companies do you think, ah, are potentially enabled by your technology and network that, ah, may not have been smart to start five years ago if you didn't have, kind of, the resources of, of SpaceX?

A Yeah. Uh, there's a company that we were recently talking to that I get really excited about. Um, you know, in LA, we had the wildfires a couple of months ago and, uh, absolutely devastating. Um, you know, really difficult to figure out where to route resources with a really fast moving fire. Uh, if you're trying to get a sense of like the scale and, and the direction of that fire with a helis, uh, a helicopter, um, it's often like not safe or not even permitted to go into those regions because there's just so much debris. Um, and so satellites are a really interesting application where, um, if you're able to have, uh, enough revisit rate, which is what, you know, in the space industry call like being able to go over a region again, if you're like a low earth orbit satellite, where you basically just need to have like a bunch of satellites that pass over and take turns because it takes time to orbit the earth. Um, so having enough revisit rate to where you can actually like regularly track the movement of a fire is pretty revolutionary. You can, you could stop fires much more rapidly and be able to, um, detect the movement. Um, the challenge with that is if you don't have, um, your, your latency down low enough to, um, to be able to give the information, it's pretty much useless, right? Like if you deliver information like an hour later, then you can't, um, you can't deliver an…

AI assessment note: “having enough revisit rate to where you can actually like regularly track the movement”

Q Uh, do you think it affects tech? Do you think it affects any of the, I mean, like immediately the reaction was like all the, all the NASDAQ was way down, but at the same time, like even semiconductors were, you know, on day one kind of excluded. Uh, was there, is this something that all founders should just tune out or is there something here to pay attention to?

A I think, I think it's, it's probably both, right? Like I think, um, I think Mike McDonald from, from Lightspeed had a good tweet, um, or post today or yesterday where, uh, he was like, I'm getting a lot of questions from maybe it was LPs about the tariffs. Um, and they have to, to realize, you know, I invest in companies that aren't going to exit for seven to 10 years. And so there's like something about the like duration of the asset class where, um, you know, something like economic turmoil that happens today If you're investing in the venture asset class, isn't going to necessarily impact like your, your exit because ultimately the founders and the board and the management team have some control over when they exit. Um, and so you, you can kind of theoretically wait a little bit for, for a better environment to, to either IPO or sell or whatever you want to do to get liquidity. Uh, I think at the same time, um, the, the like underpinning, like the undercurrents of what's going on, um, uh, especially politically. Are immensely relevant for, for startups and tech, like obviously everything around Taiwan and just like semiconductors generally. Um, and, and if that escalated, how bad it would be for not only technology companies, especially in AI, but like just our day-to-day lives. Like I think our day-to-day lives would, would pretty much come to a halt. Um, and, and I don't t…

AI assessment note: “I think it's, it's probably both, right?”

Q We're not going to touch it. But what's your, what's your kind of read on, on that situation?

A Yeah, I think, I mean, obviously before, before this all kind of blew up, obviously it was unfortunate that there was like, you know, four or five IPOs kind of on the shelf, um, ready to go right when this blew up. And so, so it seemed like, ah, you know, we were, we were opening things back up and then, uh, immediately some black Swan event happened and it kind of shut it back down. But for, for, you know, uh, like a full almost nine to 12 months before that, it was a solid environment. It wasn't, you know, IPO-ing in 20, 21, um, where, where you could go and, you know, SPAC for 50 times ARR or something, but it was not a bad environment, and I, like, I kind of have a slightly orthogonal take on the IPO window, which is, there's this kind of, like, rock and hard place situation for the IPO market, which sums up to, like, one, Do the IPO markets want you? Um, but then also just as important, like, do you want the IPO markets? Um, so on the first one, I went and counted this morning the number of public software companies on, um, you know, the Meritech kind of software comparables index that they run, um, that currently have over five hundred million dollars in ARR, um, which is like an incredible achievement, and right now there are over 80 public SaaS companies that have more than five hundred million dollars of ARR.

AI assessment note: “I kind of have a slightly orthogonal take on the IPO window”

Q and then went back up. Uh, and yeah, obviously you want to avoid that. Um, what, what about overnight trading? You always hear, oh, you're watching the futures market and, uh, it seems like Somebody has an edge here that they can trade before the market opens. Uh, how is 24 seven trading coming to America? We've heard some rumors. Is there a way to get in on that action?

A I mean, generally speaking, I think it's definitely coming. Like, it will also come to public at some point. Um, right now we essentially have, you know, four a.m. to eight p.m. Um, but the thing that you have to think of is that each trading window Has its own participants and its own liquidity and execution venues. And so think of it as like, there's the regular opening between your nine 34 PM. That's the most liquidity. It's when like most people participate. Um, you could call that the healthiest time in the market in theory. Then you have essentially pre-market and post-market, which is, you know, four AM to 9:30 and, uh, you know, four 30 to eight PM. And then you have overnight, which is like the new thing. Overnight right now, there's essentially only like one major player who drives the liquidity for that. And what happens there is that because you have, um, only one player, you have a lot of, um, like all I give, you have not as many platforms participating yet. And so you can have these moments where there's a lot of kind of unilateral flow happening. And that's why in the, in the, in the overnight markets, you often see certain stocks just suddenly rally or so on. And that is a little bit, I don't want to call it fake, but it's a little bit like it has these wild swings because the, the types of people that trade in those times of the markets and it's a concentrated…

AI assessment note: “I think it's definitely coming. Like, it will also come to public at some point.”

Q for investing too? You hear about these legendary deals. Uh, they had drinks, they did some napkin math and it was the best investment ever. The, the deal that they were in the DD room for sending doc sends back and forth, like it was just a mess. Um, is there something more broadly applicable to this idea of like simplicity breeds genius or, or like, you know, um, inspiration?

A I think what's true, especially for very smart, educated people is that the harder you work at something, the more opportunities you have to fool yourself. So if you're very smart and you have a one 40 IQ and you went to Harvard and you work at Goldman Sachs and you spend six months doing due diligence on this deal, you are going to convince yourself of whatever you want to convince yourself of, because you have so much mental horsepower that you can create any model, come up with any theory that justifies what you want to see. And so it's like the harder you work, The more likely that your final result is going to be fooling yourself. I've, I've seen that.

AI assessment note: “the harder you work at something, the more opportunities you have to fool yourself.”

Q your mid twenties, like I, I, you know, anybody who's had a post go viral or sort of break containment knows what happens. Like your, your message requests on every platform are probably like, you know, very dark. Um, and, uh, you know, are you, are you, uh, you know, you, you seem like you're handling it incredibly well, but, uh, does it ever, does it ever get to you?

A That's really, that's really caring. Thanks for asking, man. Um, you know, like, I, uh, I view Rainmaker as the best means by which I can serve God in my lifetime, um, and I owe everything in my life to Jesus Christ because of all that he's done for me in this life, and then hopefully in the next. Um, so this This situation that I'm going through, ah, no big deal at all. Like, if, if it ends up on the other side with You know, ideally me having, um, gotten to educate and convince some people and bring them more water in places where they need it. I am happy to take the punches and the death threats. And, um, even if it didn't work out, uh, you know, I'm going to give it my all. And, uh, I was a high school debate kid, so I live for this kind of dorky.

AI assessment note: “This situation that I'm going through, ah, no big deal at all.”

Q boards sort of broadly, you've sat on a ton of different boards, even board director at a number of important companies. Like, do you have a playbook and a system now that you come in and you say like, you know, you know, Kind of grab the reins or like, what does that even look like? What is, what is partnering with you look like at, at the board level?

A Every, um, you know, every situation is just like, you know, every good board is, you know, is, you know, the same and every bad board is bad and so ununique, um, and unhappy way. Uh, no, I think like the, uh, in general, A paint by numbers approach to anything in technology and anything in investing is bad. Um, And I think the worst board members tend to try to be the people who come in and take the reins and run their play, run that company through their playbook because playbooks age very quickly, especially in a world of accelerating change. So I think the first thing you have to do when you join a board is just take the time to learn what business you're actually in. And I think this is one thing that, um, I've found is you, you can get to a pretty good sense of a company and its market, um, You know, at the end of your diligence process when you're writing the memo, but, uh, you haven't seen the people in that company respond to adverse events. You haven't seen what the real bottlenecks in the business are and you can have opinions, but you sort of, um, uh, it's much more easy to feel those opinions when you have dollars at risk. So I think in general, the first thing I try to do is like, just figure out what business we're actually in and where the bottlenecks actually are. Um, I think one thing I like to do pretty soon after leading an investment is, uh, do an executive…

AI assessment note: “A paint by numbers approach to anything in technology and anything in investing is bad.”

Q Can you talk a little bit about, uh, paths to venture? You mentioned that in your analyst class, there is the high startups maxed out on startups versus maxed out on finance. Are both of those paths still viable today or have they shifted over time? How do you think about hiring new people?

A I think with venture, um, and it's funny, I read, and you know, I do, I help lead Sequoia's growth stage investing business, so like venture, like true venture, you know, seed, it's a different thing. Um, from my perspective, Uh, it sort of doesn't matter at all what you did before, because nothing you do before joining a venture capital firm prepares you for how truly, like, multifaceted long-term success is in venture. Like, you can divide the job into, you know, sourcing, picking, winning, and company building. Like, those are probably the core, the four core competencies. What we try to do, or at least what I try to do is I want to find somebody who can really hit the ground running on one of those things. So with finance, you'll have people who come in and, you know, they can be really good on the, you know, the picking, like the investment, uh, the investment process, due diligence, have a real point of view on why a company should succeed. But maybe you've never taken a meeting before one-on-one with anybody in your life. You know, for me, that was me. Like when I joined Sequoia at Goldman, I was the, you know, headphones on behind the Excel sheet, you know, like that was, I literally had never done a one-on-one meeting before in my life. And I got to Sequoia, I turned to, uh, Doug and I said, Doug, you know, like, is there like a sector you want me to cover? And Doug lo…

AI assessment note: “it sort of doesn't matter at all what you did before”

Q Yeah, I mean, you think about, uh, uh, Palmer with Oculus VR coming from Snow Crash. Uh, Delian's favorite book for a long time was Seven Eves, another Neil Stevenson book. Uh, can you tell me about the actual diamond market? I believe you invested in a company that It makes diamonds. And that's so interesting to go from sci-fi to actual investing thesis.

A Yeah. So I invested in diamond foundry early and the thesis behind that was, um, when everybody thought that they were just going to make fancy diamonds that we wear for wedding rings. I thought, what is the secret here? The secret is that we're actually going to need compute and diamond wafers are going to be a thing. And also industrial diamonds, which are actually the more expensive diamonds and they're not pretty. We're going to need that for cutting for manufacturing and for, you know, all sorts of applications, saw blades are made out of diamonds. And I went around to other investors and I told them about this future, going back to magically where they're like, you're crazy. Like this is just going to end up being a three carat diamond. And this market is, they kept comparing it to the beers and Karen, you know, comparing it to, you know, the, the diamond market. And I said, no, that's not, you're not seeing the bigger picture here. And that was completely informed by the Diamond Age, because in order to have this You know, post-scarcity future of abundance. Diamonds were a critical part of that, and I don't know if you remember in the book, but there are things that are made of diamonds in there. You know, there are buildings made of diamonds. They're just like, all of the technology was enabled because of that unlock, that we needed this, this material that was virtuall…

AI assessment note: “I invested in diamond foundry early and the thesis behind that was”

Q bunch of fast casual restaurants. They're partnering with Nvidia and like buying chips and like actually building, effectively building their own, um, AI in house. Has there been any sort of, you're, you're obviously working with all the big sort of foundation, uh, labs, or I would say most of them would be my guess, but has there been any customers that you're working with that you've been surprised by?

A Not too surprising yet. Like most of our customer base, we work with all of the top five labs in the U S and we're starting to see, uh, maybe actually one interesting thing is we're starting to see a lot more customization at the application layer. And I think a big reason for this is that it's now much more data efficient to customize models with RL environments. And a lot of this, like New kind of data versus fine tuning data that people would do historically. And so it's now becoming, you know, this like gold rush for all these application layer companies without too much capex, they're able to get these incredible results. And so we're seeing some of that, but at least for the most part, the like legacy fortune 500, uh, haven't really caught on to this yet. And I imagine that that might take a year or two.

AI assessment note: “Not too surprising yet. Like most of our customer base, we work with all”

Q Uh, can you do some backstory on the company? Can you break down what you're building, who you are, little career highlights? I mean, yeah, I'm sure a lot of people know, but give us the basics.

A Yeah, so, um, I started off my career in like this weird corner of operational counter-terrorism in the U.S. military. Um, that sort of branched into high technology and high technology strategy. Um, again, when I was still in uniform and got a lot of exposure to just how much, um, you know, the world is a very interconnected place, and there are things that we can do in the digital world that absolutely affect the physical world. People talk about Stuxnet, they talk about, you know, these, these attacks that have happened against, uh, oil facilities. So, um, I get out of the military, I go to Silicon Valley, and then two years later, Just like a strange series of coincidences. I get asked to come back to Washington and essentially be like the most senior cybersecurity person in the US government. Not like I'm going to reset your password, but like I'm going to remake our entire practice of running cyber operations for like the DOD and the intelligence community and Homeland Security and all that stuff. So I did that job for four years, and then afterwards, um, pulled together two buddies of mine, including Brandon Park, who helped stand up Amazon's global, what we call operational technology OT cybersecurity program at Amazon, and then Felix Plushinsky, who I'd recruited to come join me in the admin, who was a hedge fund trader, and we started looking at this problem. It's a h…

AI assessment note: “we just wanted to start a company to, to solve that.”

Q I'd love to know, like, obviously when you're the CEO, if you're selling your company, you're probably operating at level five superstar, but then you go into Twitch, you said it was a very different environment. Did you feel yourself moving downward towards like level four or level three?

A Yeah, there's gravity. Gravity is moving you downwards, right? Because you're supposed to, you know, like you, if you go to these big companies, you're not the CEO of the founder and more, not all the problems don't roll up to you. All the opportunities aren't your decision, whether you're going to do them or not. And so gravity is kind of pulling you down to get to like, kind of like a level three, which is like a bit of a, you're just a sort of a task doer, or you can, or maybe you can manage a team of people to do, to do a set of tasks. But I was just like trying to fight that. So like, I had a great conversation with my buddy Furkan when we got acquired and he goes, he was pulled up the company metrics, like the company dashboard. And he's like the CEO view basically. And he's like, which one of these numbers do you think our project will move? And I was like, I, like, I was just thinking of they bought our company to do X. We should come here and do X. And what he was arguing was like, cool. That made sense when we were trying to sell this company to them. Of course we wanted to say how big of a deal it was. And now that we're here, we're on the same team. We should just be more objective. Like, is our thing gonna move any of these metrics? And he goes, don't you just think like whatever we do while we're here, we should just be able to move one of these KPIs like five or …

AI assessment note: “Yeah, there's gravity. Gravity is moving you downwards, right?”

Q on what's happening right now. What the current thing is. We obviously focus on X as a platform. But if you're only consuming content on X, then you're not injecting anything sort of new. Um, and you're just like part of the hive mind. How do you think about your information diet and kind of where, where do you spend your time, uh, kind of coming up with, with ideas?

A Yeah, it's a great question. I, I think that, uh, the, the information diet is probably like today, something that only nerds like us talk about, but it's going to become a very popular concept as people realize that it's garbage in garbage out. So, um, you know, I think the problem that the tech guys have is you, uh, we will all go consume, not necessarily garbage content, but just the same content, right? So we're all on X, We're all reading the same things. We're listening to all in and listen to the same podcast. And then guess what? You have the same thoughts as everybody else. And like, while calling yourself a contrarian, right? It doesn't really make sense. And so, you know, for me, uh, I do a couple of things. So I have, I have a phrase that I like, which is, um, kids, dogs, and dead people. And I want to spend as much of my time as I can with kids, dogs, and dead people. And, uh, you know, so what does that mean? Kids will teach you a certain set of things. So kids, kids are highly, uh, playful. They, they work with their imagination a lot. They're completely delusional. They have no idea of what's realistic and what's not. And so the more time you spend with kids, the more playful you become, the more imaginative you become, the more delusional you become. So I take that from them. Uh, dogs are sort of these unconditional lovers, right? The dogs are, are always, alwa…

AI assessment note: “I have a phrase that I like, which is, um, kids, dogs, and dead people.”