The Delphi Podcast

“But it was a chart from essentially the time from an application to get to 100 million or revenue. Zero to 100 million and how many days it took. And a lot of those companies are now currently being built on Solana. And this is not just like crypto companies, this is across AI as well. The fastest zero to 100 million dollar revenue companies in the world are crypto companies. Hyperliquid was number one. I think it was 89 days for them to get to 100 million in revenue.”

“I'm actually, today, I'm pretty bearish decentralization. More so in the sense that I think we, going back, we ran all these experiments and they were fine. But I don't think people necessarily pay for decentralization. I think if people valued Ethereum, like, say, Solana had a thousand nodes and Ethereum had 10,000. Like, if people really value that, they would pay 10x the amount of costs, in my mind. I think people care about decentralization only if it makes a better product.”

“So let's use the kind of like centralized exchange example. If you're in Russia and the server for the exchange is located in New York, why would I, as a Russian person, want to trade on a server that's located many miles away from me if I'm running a quant shop? Yeah, like it's purely a latency, like physics perspective. You're disadvantaged if it's not in your home jurisdiction. If these things actually scale to where you're doing trillions of dollars a day, and this is the everything exchange, I just have a hard time seeing the United States, China, whoever, like big global powers that actually trade a lot saying, yes, I'm okay with that data center being in Japan.”

“And the most basic terms is you can have leaders in various jurisdictions around the world where you can cut down on the time to inclusion. And so, for example, instead of having someone always has kind of an information advantage, you can disadvantage everybody to call it 20 milliseconds. And so what happens is you run parallel auctions. You aggregate kind of all these transactions around the world. But if you have these world leaders and call it Japan, United States, Hong Kong, Russia, UK, wherever, the thought process is you have these 20 millisecond batches that are happening simultaneously and continuously.”

“Hyperliquid has built a great product. They deserve all the praise in the world for doing that. They're making a lot of money. But it is a different engineering designed to build this global exchange. I think purely from a fairness, so to speak, I don't think the world and its limit will say, Hyperliquid is this global exchange that I feel like as a US person or anybody kind of outside of that market, will trade on. It's a regional exchange, which is a great regional exchange and very profitable and people love the product.”

“And now within the last, call it, six months to a year, we have the proprietary AMMs, which are very unique. And these are actually allow you to update the entire pricing curve with one Oracle update. And so what has happened is these proprietary AMMs are really, I would say, the blockchain native design for order books. The quotes were sale on chain, and you say, hey, oh shit, the price moved, I need to update my price. And they couldn't update the price fast enough, and so someone would pick them off.”

“Like the popular platforms today, like Drift on Solana, to no fault of their own, it was just Solana, I don't think in my mind, was as performant as they needed to actually compete with something that was centralized, like a hyperliquid that was co-located from a data center standpoint. That gave them enough of an edge to actually win the perps, where Solana, because of the decentralized aspect, was not performant enough yet to actually build the perps product. They've done well on spot, but they have not done well on perps.”

“I think for the record, Grok is underappreciated. It's mostly just like what we've seen with GPUs, is that you throw more compute at it, and the algorithms are generally getting better, that the combination of better algorithms and more compute makes better models. I think Elon is the best set world in moving the atoms around and building things in the real world the quickest. I think he built the first Memphis data center in like 20 days. It was 100,000 interconnected GPUs, the world's largest data cluster, and now he's going to build, I think, a million.”

“The robots are actually worrying me. Because you have, in a lot of sense, I mean, you essentially get automation. And like, when you think about the jobs that a lot of people do today, like Uber Eats or just Uber Drivers, like, that has, like, the gig economy, so to speak, has become very popular. And if you remove, like, the Uber or, like, Lyft or whatever, and they're just all completely autonomous cars, and then you think about how many, like, things that humans do, from, like, baristas to, like, restocking shelves to construction. The amount of things that humans can do and that robots can probably do better, that actually, like, kind of freaks me out a little bit.”

“When I meet non-crypto people and I tell them we've been our lives, or America has been yours in 1776, we printed 80 percent of the dollars the last five years, they're like, that's just not true. And I'm like, no, no, that's true. Especially after COVID and 2020 and 2021, where they printed so much money, is that if you look on Zillow and house prices, they've essentially 2X. So unless you made 2X your net worth or 2X your salary in five years, that most likely asset that you want to buy has doubled in value.”

“Sub three, though, is sort of like... Yeah, is like the, I guess, gold standard for when you cross over to kind of doing something special in the marathon. And for me, it's like just a natural extension of a lot of the endurance training that I've done. And it's funny, like, I've done a bunch of different endurance events, but had, up until this past year, never done a marathon. And when my old boss from Slow, Sam Lessin, found out that that was the case, he publicly on his podcast, like, challenged me to run a sub three marathon. And so, yeah, I put, you know, a great training block in. And just, you know, the most important thing, honestly, is like showing up to the race healthy and getting to the starting line. Because a lot of folks overtrain, they get injured, and then, you know, wheels fall off.”

“And so that helps us triage the top part that we want to review first. And from the 2,000 or so submissions that we get each hackathon, we distill that down to about 150 that we then go and interview. Each project gets two sets of team members' eyeballs on them. We have mentors, founders, other investors in the ecosystem that represent a judging panel that evaluates those projects from that subset of 150, and then that gets cut down to about 40 that are considered for category winners down to honorable mentions for the non-dilutive prize money, and then we distill it down to about 10 to 15 that we invest in at that 250k check size and then bring into the accelerator program.”

“We recently, in the last cohort, rented an office in San Francisco for two weeks of the Accelerator. And the feedback at the end of that was, hey, guys, do you have somewhere where we can just stay longer? And we looked around, like, well, we only have this for two weeks. And so, that then precipitated, hey, we need to actually go and get an office so that, you know, not only can we have people stay longer than the program itself, but I think, too, in starting something by yourself, right, or with one other person, and especially if you two are remote at the outset, like, it's a pretty lonely endeavor.”

“The thing that's changed, at least within Coliseum, is for us to be, I think that we've like, oriented more and more towards the individuals rather than the ideas. Because at the end of the day, the ideas are pretty pliable in terms of a direction a product can take. And I think early on we may be overindexed on having to have the perfect marriage of the two. And that led to like the biggest disagreement because we were not arguing over the person, but we were arguing over the idea that they had and the merits that are associated with that.”

“It becomes very clear over the course of a conversation, whether someone is looking to us for validation that their idea has merit, versus basically not caring and being like, I'm going to will this idea into the world, whether or not you guys agree with me or not. I do think that really lobbing in these sort of grenades that pierce the veil of like someone's confidence and seeing that their ability to kind of take that in, but, you know, answer the question constructively, I think is, is, you know, the thing that we see a good bit of.”

“The other thing that they taught me that is still true with what we do day to day is they frame every investment memo in what is the 10x case and what's the 100x case. Meaning the 10x case is, hey, with these dollars, you believe there's a 10x value. And usually that's around a single particular product hypothesis. And then the 100x case is like, what is the unreasonable thing that could happen with this company that in retrospect is going to be perfectly reasonable.”

“There was a company called Orr that came through. Hardhat Chat is an Anon developer that built Orr. It was a store of value that was built on top of Solana. They basically used a proof of work mechanism, but then settled to Solana. And throughout the hackathon, it actually broke Solana or was very close to breaking Solana. And so we saw this and we were like, this is one of the coolest things we've seen in a long time. Ultimately, what Hardhat landed on was like, it actually isn't proof of work, it's gamifying the experience around mining. So he actually created a consumer application that does that. And now that's one of the top revenue generating applications on Solana.”

“MetaDao is another one that has sort of wandered through the idea of being amazed under the umbrella of these different types of companies that can be governed, but this concept of futarchy, where decision markets rather than one token or one share, one vote govern sort of the direction a corporation can take. It's, for whatever reason, just because the world is weird, aligned very closely with the founder type that we like, where in order to have a token from essentially day one with your company, you have to be essentially a public market CEO where you're communicating, hey, here's what we're doing, here's why we're doing it and kind of managing the expectations of a market, but as a like pre-seed or seed stage company.”

“I think that the impact of AI and, you know, vibe coding and all of that is like, it certainly democratizes the ability to express an idea in code, and that puts a little bit more emphasis for us, again, on the individual behind that. But also, importantly, it really creates an emphasis on the execution of that idea. Because, like, I think at least the current subset of tools that people have at their disposal is quite good at getting from, like, zero to one. But then to take that to something that is, like, maintainable production-level code over the long term and, like, has a product that is truly differentiated, that's, like, the magic.”

“By 2035? 70 to 80 percent. Maybe higher. By 2030, I would say 100 percent possible, 20 to 5 to 30 percent. There are many hundreds of millions of dollars that have been deployed with this belief in mind. So you don't really have to trust me. You have to trust, well, there's actually institutional investors out there that genuinely believe this, and we're able to convince their IC to put hundreds of millions of dollars into this concept.”

“Satoshi's coins are in an insecure format, and it's weird that Satoshi left them like that, because Satoshi was aware of this risk. But if Satoshi had really cared about the risk, Satoshi would have not left a million to two million coins laying about. It's nuts. Like in 2010, Satoshi couldn't have known that 15 years later, as it turns out, it's impossible to upgrade Bitcoin.”

“Bitcoin developers making Taproot not quantum-resistant is the craziest thing that's ever happened. I went back and I looked at the discussion. They knew. Luke Dash raised this objection at the time and said, hey guys, by the way, Taproot addresses are vulnerable to quantum. This is 2021. At that point, we did know that quantum would happen eventually. So what the hell were they thinking?”

“One out of 100. So you fail at 50 percent, right? Anything below that is an F. They mostly don't think it's real or they won't acknowledge that it's real. I read all the discussions on the mailing list about it. There's one named BIP that has to do with quantum, which is by an outsider. It's a very nice guy, Hunter Beast. He's not like part of the inner club, right?”

“I think the modal probability is on the one with the fattest part of the distribution or tallest peak is that a private firm, let's say there's no fork, right? So let's say we don't collectively steal the Satoshi coin. It would be that a private firm at the behest of the US government requisitions the coins and gives them to the US government. And they do this because they believe that China is about to do it. And so the government authorizes them to do that.”

“I spent a lot of time reading about whether people that recovered coins from ancient shipwrecks were able to claim the coins. There was a famous one in Florida in the 70s, 80s. He spent 16 years looking for this. His wife and his son died in the course of this. So it was a really like Faustian thing, like he really suffered to get, and then he got his time on the shipwreck, and he dug up all the coins, $500 million worth of Spanish gold. From the 1600s, and Spain sued him over it, and he won, but it cost him everything.”

“So when you work backwards from Q-Day, it was like Y2K with a bottom. We work backwards from Q-Day. Okay. Q-Day, let's say it's 2033. Okay. We'll be conservative, right? How long will it have taken for Bitcoin to migrate? I don't know, five years? Okay. How long will it have taken for us as a community to agree on the path forward? Three years? How long will it have taken for us to test the code? A year? That takes us back to now in the past. Yeah. But we're in the present. So the numbers are not adding up anymore, right?”

“So every six months, I get totally obsessed with something. So in 2023, it was Chokepoint, right? And then it completely dominates my focus. And before that, it was Proof of Reserves. And I don't remember what it was last year. And well, AI and data centers, that was probably my main obsession last year and the year before. And this year, at the end of last year, for the last six months of 2025, it was quantum.”

“Bitcoin is actually the only thing that can't upgrade, along with some like really old school, antiquated, like physical devices where the cryptography is flashed into the memory and can't ever be upgraded because it's physical, right? So Bitcoin is in a very unique class of things that can't upgrade easily. When I was doing it in banks and whatever, they'll upgrade the database, the government's told them.”

“So we were a nonprofit first. The fund came much later. So maybe I can like back up and say what Halcyon's worldview is. And we can share on. So super powerful AI is coming, whether you believe that means AGI or super intelligence or whatever, we can get into the distinctions there. It's going to radically reshape the world. Or you could almost think of the nonprofit as the platform team to the VC fund. You woke up tomorrow with an AI security VC fund, and then you thought, what would the platform team look like? How do we develop relationships with the labs and the researchers and the lawmakers and the biggest potential customers and the CISOs of the big companies? Well, that's just the thing that the nonprofit's been doing for the last two years, and so the fund really benefits from that.”

“You could also see AI being used to do things like create pandemic viruses or bioweapons. And we already know from OpenAI and Anthropic that the current models that are released out in the world right now do provide uplift on those fronts, which is to say makes it easier for a relative amateur to do something like create a pandemic virus that could cause a COVID or worse level pandemic, right? So this is like a present-day risk. This is not something that you have to believe, you know, that some crazy super intelligence will come into existence to sort of find this salient.”

“So people talk a lot about defense in depth, right? It's not just about securing at one point of vulnerability or one layer. You want to secure at many points. So for example, one starts in pre-training, right? What datasets are going into the models? If your dataset includes a bunch of biological information, say about like the structure of and function of viruses, then you want to take that really seriously. And so a bunch of the organizations that are building bio models are actually choosing not to include viral data in their training sets because they know that that just opens up this can of worms of risk.”

“Well, they do partner actually with big insurance companies that you would have heard of. But think of it more akin to cyber insurance, right? So thinking about how insurance can mitigate AI risk, if you just look at other industries that have come up over the years, the insurance sector within those industries was often the force that brought more safety to those industries. So for example, if you look at the history of automobiles, they didn't put seat belts in automobiles for a long time, cars for a long time. And the thing that got them to put seat belts in was the insurance industry insisting on it, right?”

“I think one thing is like, has this person made expensive choices to do this thing? Did they walk away from something else that would have just been an easy path to a prestigious career, lots of money, a great life? Yeah, I took a big pay cut to do this work for sure, and so I love seeing people who do that kind of thing. I also love to see founders who really care about how their companies are governed.”

“But I told you this story about COVID, right? About a year into COVID, I took a walk with a friend of mine who's like the world's largest biosecurity and pandemic preparedness grant maker. And so I was like, kind of like, well, but on the bright side, I'm sure now that COVID has happened to the world, it's like come to its senses. And now we're doing all this great biosecurity and pandemic prevention investment. And he's like, dude, you are so naïve. It is the opposite, right? Because first of all, everybody's sick of it, right? Everybody's sick of wearing masks, everybody's sick of the pandemic, they're sick of talking about pandemics. And now it's become super politicized.”

“Then there is what people call like, diffusion, right? So AI actually solving problems in the real world. So you see a lot of companies say like, well, we build this AI that is better than radiologists at diagnosing cancer. And for some use cases, AI has been better than radiologists at diagnosing cancer for like many years. And yet we still have a lot of human radiologists. In fact, we have a shortage of them. So why has AI not like in the real world solved this problem? Because there are all these either operational constraints or bureaucratic constraints to actually implementing the technology in the real world.”

“Sycophantic AI, right? I hate when I ask ChatGPT a question and it's like literally like how much cumin should I put in this chili? And it's like incredible question. And I'm like, it was not an incredible question. It was a super mundane question. And I really worry that we're going to AB test AI into this mode that is on some level what we prefer, but on another level, what we really wouldn't want for society.”

“Although, you know, this is sort of, there's this irony, right? Because 10 years ago, if you asked anybody, what jobs is AI going to automate first, it would have been physical jobs, right? And now we realize, but like the white collar, people will all be safe. And we've realized that that's exactly wrong, right? Because the most powerful models are language models, right? And, you know, white collar knowledge workers are roughly speaking doing language-oriented work, which are typing into a screen all day, right? So of course, that's the thing that language models are really good at.”

“If you went into the Anthropic office or the Open AI office and just pulled out 30 random people and pulled them, and you're like, what do you think our prospects are for, you know, hitting AGI and that that has really terrible catastrophic consequences for humanity? You know, you'd hear a pretty big range, but your median answer might be something like, I don't know, 20% chance that AI really has catastrophic consequences for society in the next decade. Let's say your friend was an astronomer, right? And you went over to their house and they had a telescope, and you looked up into their telescope and you saw this, like, asteroid hurtling towards Earth.”

“The reason I started Halcyon was much more about talent than money. Right. To solve any big problem, climate change, health care related stuff, education, AI stuff. The most important thing is that you have some of the world's most talented people, entrepreneurs, researchers, etc. dedicating themselves to solving that problem. And yeah, you got to raise more money and all the stuff, but who's good at raising money? Really excellent entrepreneurs and leaders. Okay, you need more rank and file talent. Well, you know who's really excellent at recruiting? Excellent entrepreneurs and leaders.”

“One of the largest DeFi launches and chains of this whole cycle, they had had an incentive mechanism for depositing funds on Ethereum, and those funds needed to be migrated from Ethereum to Berachain and then executed into all of the brand new protocols that didn't exist up until the chain was live. And somebody has handled all of that execution underneath. And these executions were not just swapping from token A to token B, we were bridging, then lend, then borrow, leverage loop, then enter into one-sided LP, so forth and so forth, right? This was handled fully by Enso. Not many people know this. They assumed this is Royco, they assumed it's Boyco. That really is just an interface. It was all of Enso's logic underneath. And we safely and securely moved over $3.1 billion into over 80 plus unique different DeFi combinations.”

“We have a project called Onplug. They built their MVP without Enso. I believe it was seven months. Then we approached the team and said, hey, you're building really cool stuff. Like you're building Zapier, drag and drop automation and crypto, and it's never been done before. Why don't you just use us and get access to all of Onchain through one tool? They were able to rebuild all their integration suites in a day and a half. They reduced the development timeline from seven months to a day and a half.”

“Enso has been around four and a half years. And we've never had one exploit. You know, not one. We've handled over $17 billion. We've done, I believe as of to date, there's 10 audits publicly. So the way that we kind of view this is our view in the future should be never write a smart contract again. A developer that is in the Web2 world that wants to interface on chain, they don't need to understand smart contracts and they do not need to write smart contracts to interact with smart contracts. If you write a smart contract that interacts with another one, that's creating another layer of security issues.”

“Brutally honest, we need products that people want. You know, the same money that cycles through DeFi, it's all the same people. Like you see money going TVL up or TVL down there, that's because this is backroom agreements of, hey, I want this APY or I want this extra yield that's not public and I'll put $500 million in your protocol. It's just the same people using it. So I think what we need to do for more people to use crypto is one, give products they actually want to use and two, make it seamless. Like they shouldn't need to care about chains.”

“Where did we all individually learn about how to build code? The reality is, it's YouTube. That's where people go and they watch it. How many Web3 companies do you see that actually have, not all, like, wait, let me rephrase. Every Web3 company has YouTube videos and they have zero distribution, right? Zero. Where do people go to learn? They go to YouTube, they go to Code Academy, they go to Udemy, they go to coding boot camps, coding tutorials and so forth. So we're fortunate that we also know people inside of the Web2 developer relations communities that have a large, large, large YouTube communities.”

“We're calling something the Enso Drop. Now the name might change actually in the next couple of weeks, but this is what we have internally so far. We're in a very unique position that we have so many protocols that use our protocol. So we have 125 plus, I believe it's even 130 plus now. You can access over 250 plus on-chain protocols. Now, we interact with their contracts and we get air drops. So we're able to now start offering to the users that secure the Enso Network by delegate staking them their ability to claim other air drops. Now, I believe this is one of the largest mechanisms in the whole space if you're able to get access to over 370 plus potential air drops.”

“I think the real value capture will be the fat engine thesis, which I've just made up on the spot is Enso. If you're able to be that connectivity glue and that one tool that abstracts all of it away, it's the superpower if you have tens of millions of apps that are building atop of you. We could in theory dictate in the future, if base doesn't give us incentives, we won't route people to base. I'm not saying that we will do that, but these abstraction layers will ultimately have to control over the chains. I think ultimately, the more higher that we go up in the stack, the less control that the chains will have, and these abstraction tools will actually have a lot more control than they will, because you can just direct it away from them.”

“Now, how to flourish and create the ideas? I think you can't. You've got to keep them far, far away from the crypto industry. You've got to keep them further away from crypto Twitter as possible because they will just recycle, recycle, recycle. They need to really come up with their own innovative ideas. I was on crypto Twitter back then, 2016 I was like, oh, let's do decentralized roulette table. Roubette nowadays pretty much to some degree. I think keep them far away from crypto Twitter for as long as possible and just have really good mentors that can support them on their developer journey.”