Unchained - The Chopping Block: Wind Downs, YC's Nemil Dalal, & Will Every Failed Crypto Idea Eventually Work?

Episode Date: July 29, 2026

YC's Nemil Dalal joins to explain why he's never been more bullish as BitMEX winds down after 11 years, whether every failed crypto idea (TCRs, DAOs, creator coins) eventually works, why crypto is really about money, Base's consumer mea culpa, on-chain reputation and credit, and who pays in the x402 AI-agent era. Welcome to The Chopping Block – where crypto insiders Haseeb Qureshi, Tom Schmidt, Tarun Chitra, and Robert Leshner chop it up about the latest in crypto. This week they're joined by Nemil Dalal, Visiting Partner at Y Combinator and ex-Coinbase, where he led USDC and the Coinbase Developer Platform. He's here to explain why, with exchanges winding down left and right, he's somehow never been more bullish. The crew digs into the great contrast of the moment: BitMEX shutting down after 11 years (plus BitMart, Movement Labs, Balancer Labs) while the plumbing quietly prints, and whether Imran's viral 'everything that failed will eventually work' thesis is genius or toxic positivity. From there it's the question of whether crypto is really only about money (Jesse's Base mea culpa included), a war-memories tour through TCRs, on-chain reputation and why pure on-chain credit keeps faceplanting, and finally who actually pays in the x402 AI-agent era, and whether decentralization even survives contact with Google-shaped gravity. Listen to the episode on Apple Podcasts, Spotify, Pods, Fountain, Podcast Addict, Pocket Casts, Amazon Music, or on your favorite podcast platform. Show highlights 🔹 BitMEX, the exchange that invented the perp swap, winds down September 23rd after 11 years, alongside BitMart, Movement Labs, and Balancer Labs. 🔹 Nemil Dalal explains why he's never been more bullish: rivals adopted BitMEX's innovations, regulatory clarity is arriving, and crypto is becoming invisible infrastructure. 🔹 Nemil rented the Chase Center for a 7,000-builder YC event with Jensen Huang, Sam Altman, and Patrick Collison, yet almost nobody's launching a token early. 🔹 Imran's viral thesis that every failed crypto idea eventually works sparks Haseeb's 'toxic positivity' pushback and a war-memories tour through TCRs and DAOs. 🔹 Instacart's Apoorva vs Webvan and Reddit vs Digg: why timing, path dependency, and the YC 'why now' question decide which failed ideas return. 🔹 Jesse's Base mea culpa: the consumer-social bet on Zora and Farcaster was wrong for now, so he handed the Base app to Kobe. 🔹 Haseeb says crypto has always been about money; Nemil counters 'money is everything,' calling the blockchain the greatest capital innovation machine in the world. 🔹 Why pure on-chain credit keeps faceplanting: address repudiation, no recourse, no wage garnishment, and old memories of Debt DAO's revenue ratchet. 🔹 The x402 AI-agent era: Cloudflare pay-per-call gating, Kimi's inference license, and whether decentralization survives Google-shaped gravity as agents become the new wallet. Hosts ⭐️Haseeb Qureshi, Managing Partner at Dragonfly ⭐️Tom Schmidt, General Partner at Dragonfly ⭐️Tarun Chitra, Managing Partner at Robot Ventures Guest ⭐️Nemil Dalal, Visiting Partner at Y Combinator Disclosures Learn more about your ad choices. Visit megaphone.fm/adchoices

I think the blockchain is the greatest capital innovation machine in the world. I think we don't think it's big enough. Like, it literally is everything. Like, there's not a company in the world that doesn't have to deal with payments, capital raising, and things like that. And the blockchain is shown that. Not a dividend. It's a tale of two quond. Now, your losses are on someone else's balance.

Generally speaking, air drops are kind of pointless anyways. Unimaged trading firms who are very involved. I like that eight of the ultimate policy. Defy protocols are the antidote to this problem. Hello, everybody. Welcome to Chopping Block. Every couple weeks, the four of us get together and give the industry insider perspective on the crypto topics of the day.

So, quick intro, this first you got Tom, the Defy Maven and Master of Memes. Hello, everyone. Next, we've got Tarun, the gigabrain and grand pooh-a at Gauntlet, but he is running a little bit late, so he will join us shortly. But joining us today, we have special guest, Nimiel, Crypto Wise Guy at Y Combinator. Welcome to the show, Nealel. Great to be on, Hasib.

Yeah, great to have you. and I am a sieve that had hypeman at Dragonfly. We're early-seasoned investors in crypto, but I want to caveat that nothing we say here is investment advice, legal advice, or even life advice. Please see Chopin Block at XYZ for more disclosures. So, Emil, you're joining us from Wycombinator,

the legendary startup incubator. You were formerly at Coinbase. I was seeing all these clips on social media of like this gigantic startup school, like stadium-level event that was happening the other day. We had an epic. What was that about? Yeah, we basically rented out.

the Chase Center in San Francisco got about 7,000 builders. So I had everyone from like 16 and 17-year-olds building with AI all the way up to, you know, 40-year-olds like me, all looking to hearing from Jensen Wong. We had Sam Altman there. We had Patrick Collison talking about crypto and Agent Fi. So it's a pretty epic event. But it's basically our way to really connect with people who want to build and want to start a startup.

Wow. I can't believe that you've got Jensen Huang, the famous guy, the guy who built his business on crypto mining and Sam Altman, the coin founder, on stage. That sounds like an incredible event. I assume crypto was the start of the show. Yeah. And so it's actually funny, but I think Agent Fy definitely was one part of the show, right, with Patrick Collison talking. But I actually did panels on blockchains. We got a ton of builders. This is actually my favorite time in the cycle, because you have a lot of the serious people who really want to build for the longer term coming in. And I ask people, how many of them are launching a token early on and almost no one

raise their hand, which is pretty awesome to see in terms of just an early builder and just really focusing on solving a user pain point. We had a ton of people really focused on stable coins. And then there's a whole group of people, especially from traditional finance, focused on just agentic finance. And that was everything from using AI to pay for different things. So I have an AI, it's going to do some work and it's going to pay for some APIs with it, to everything from a hedge fund that they wanted to build using AI models, you know, that could outcompete existing funds or at trade in crypto. I think it's a different market than you might get two, three years ago in the top

with the crypto bowl. But it felt like pretty awesome, awesome builders there. Very cool. So it's interesting because we're in this kind of moment right now where a lot of stuff in crypto is working, but it's a lot of stuff that isn't exposed to the underlying Bitcoin price, right, or the underlying crypto asset prices. So on the other hand, you've got a lot of businesses that are exposed to underlying prices, struggling. One of the big news this week was the closure of Bitmex, the infamous exchange,

the inventor of the perp swap, kind of the, you know, one of the seeds from which the entire derivatives industry has really emerged. It was kind of the birthplace of derivatives in crypto, which today are the majority of what's traded in the industry. They shut down on September, or they're going to be shutting down on September 23rd after 11 years of operating. And they were once upon a time the number one derivatives exchange in crypto. Now I don't even know where they are on the map. I presume this is one of these things that as the industry grows, as compliance costs increased, their market share has decreased. And of course, crypto trading volume is down quite a bit. It's become a much tougher market for a lot of these players operate.

Then we got Bitmart announced another crypto exchange, less well known. But they have to go about that market. Yeah, I mean, it's one of those that was always like, you know, number 30 on coin market cap that's like, oh, wait, they're doing 15 billion in volume and they're like exchange number 30. That's interesting. Then there were some other announcements of wind downs, storage A labs, movement labs, balancer labs. A lot of labs companies or projects that actually are on chain, not just exchanges, are also winding down.

And people are kind of realizing like, oh, you know, the amount of fanfare and the amount of surprise is sort of less and less with each one of these. There used to be this joke that crypto protocols or companies, they never quite go bankrupt. There's always someone out there willing to run a node. There's always someone around there willing to trade their token. So nothing ever goes to zero. But things are actually going to zero now. So, Namil, how do you react to this moment?

What do you interpret what we're seeing going on out there in the headlines? I would say, despite all those shutdowns, I'd never been more bullish. So I think you mentioned two sides. One is the trading companies. And what I would say there is a lot of the other players in the market have adopted their innovations. For example, Perps is a critical product. I think Coinbase started announcing, that in the U.S. and other markets around the world. Binance obviously has had it for a long

while. So clearly they said history, and like any in an end of market, other people caught on, realized that they have to do it. And again, I think there are things like, I think, a guilty plea and some other challenges in the last few years for a bit back specifically. But I think, if anything, they led the way. And a lot of the industry learned from that and moved on from it. And so in terms of trading, like what I'm seeing is that basically trading is being added everywhere. All major financial institutions in the U.S. and worldwide are now adding it. And so I think what you're seeing is this blurring of the lines where things like Coinbase with everything exchange strategy, I spent seven years working there.

Basically, it was like acknowledgement that you can't just do crypto. You have to think about prediction markets. You have to think about stocks and everything else. And I would say there's the opposite. Every financial institution in the U.S. is adding it. And again, volumes may be lower when Bitcoin prices are lower, but this is going to change. We've all been in the industry and seen the way these cycles progress. Absolutely, I think that's going to change.

And then on the lab side, I think it's really interesting where I think two things I'm seeing. One is that these governance tokens and things like that, we're realizing that that they might not be the best model. But part of it is that there's actually clarity. Even before the Clarity Act and the U.S. has passed, there's starting to be more and more clarity. And you saw that in Hester Pierce's recent posts, that there is, like, a lot of this stuff should exist. and the government for a lot of the things wants to get out of the way and allow people to be able to build. And so I think that's really awesome

because a lot of the growth of labs at governance tokens was a function of the lack of regulatory clarity. So I think that's really, really awesome to see. And the flip side is, I think, like anything, I see in the startup ecosystem, we see so many bets on the future. And so, for example, Move is a good example where the bet was a better programming language would attract a ton more developers.

And we saw that with Ethereum and solidity early on. I think that market has changed pretty dramatic. and you're hearing, you know, developers vote one way or the other. So I think that's the follow-up there. But I think the big shift really is that a lot of things we don't talk about is all these less sexy things happening. So in YC, just to give me an example, there are four or five like stable coin companies that very few people have heard of.

Jeeves, Infinia, Blind Pay. These are all examples of companies that most people don't hear about. They're not a consumer-facing brand like an OpenC or Coinbase, but their developers using them. They're building on stablecoins. And the last thing I'll say is that I don't think a week passes without a financial institution reaching out to me, a bank or institutional investment firms somewhere in the world that's excited to integrate crypto in some way. And a lot of times it's hidden from the user. And so again, that's a very different mindset I think than I saw a few years ago with that EO is everyone's going to interface of crypto.

They're going to know about it. And so we're going to have to make the Wad experience amazing for them and things like that. And again, I think that it will eventually happen. But really what I'm seeing right now is this is the infrastructure layer. for everyone and institutions are getting in and the clarity is coming. You know, one way or another, clarity is coming. And so as a result, like a lot more people are building on it. But it's just not as sexy maybe than it was like a few years ago during a bull market.

What do you feel like are the biggest like learnings or misconceptions or shifts that teams have to make when they're going from maybe having building the web two to building even even like stable coins and building a more sort of, you know, normal consumer product still using table coins or the hood or maybe building something more exotic in. YC, and then vice versa, what are the biggest sort of like YC trusoms or learning that you feel like those teams like value the most or maybe take those to heart? So like for the crypto builders, what I see a lot is either their decentralization maxis. And what I would say like YC in general, and I imagine for you guys as well, is that the biggest focus is about solving a pain point. Like users don't care about decentralization. Some do.

Some developers do. But a lot of people don't. And so like you have to meet people. where they are. Again, you're talking to the person who used to run stable coins at Coinbase. You know, so I believe in the power of like anything that solves a user pain point, be it centralized, be decentralized. And so I think that's one thing we definitely see with some of the Web3 builders is that their default come in with an ideology. And I think one of the most important perspectives in YC is make something people want. And so as a result, like think about

what the pain points are. And that's where I think stable coins are so interesting today is that even though they're centralized, they run a decentralized infrastructure. And they solve a very, very, clear, painful people. The other thing is, I think that I love the idea of tokens. I worry that for a lot of consumer products early on, that's kind of dangerous to launch a token. If you're getting your product on having access to a token, I suddenly think your Tam, your addressable market goes down a lot. And so we really counsel people that we love the idea of launching tokens. If they're real world assets, things like that, awesome. If there's some type of incentive token that's part of your product, would love it later in the process once you have some PMF. Because again, if you're

giving away money for free, a lot of people use your product that says nothing about whether they really want to use the product or not. And so that's probably like the other big thing that we talked to user about. And the last thing I think in this market especially is you may not solve it directly for the consumer. You may solve it for an institution or developer or someone else behind the scenes. And so like one of these themes we're seeing a lot is, for example, that I'm selling to a bank or I'm selling to a fintech. And so for example, companies like gusto is a payroll company, and they're doing payroll using stable coins, and they need infrastructure to be able to do that.

Deal is another example of that, and they both need infrastructure to do that. So it's not, you know, it's not like the, I think the bull market where it's all consumer, a lot of it's behind the scenes. And then I think you asked me on the Web 2 side, but maybe the quick answer there would be is like, I think a lot of it is that everyone's AI pilled. So how do you get them to know and understand the value propositions? And the financial people get it. That's maybe the interesting thing that I've seen is.

a lot of people from TradFi are moving over to crypto, and they look at it very differently. And some of the things I actually disagree with, I feel like crypto is unique, and you can't just take TradFi learnings. But the flip side is if you're selling to a bank, you're trying to do like a real world asset, they actually have a lot of muscles and skill set to be able to take that and think about how that's on the blockchain, especially when institutions are using it. And so that's kind of a cool side effect that I wasn't expecting. And I would say most cycles, in crypto would have like ignored their perspective because it was, I felt a little too early, but now actually is the perfect time. So, Neal, you were talking about how, you know, more and more

of crypto stuff is kind of pushing away from decentralization and kind of more practical, more focus on real world utility, and I completely agree with you. I got into a bit of a tiff over the weekend with Imran at Alliance Dow. And he tweeted this tweet, which went quite viral, almost everything in crypto that failed the first time will eventually work sometime in the future. Bitcoin looked like it had failed many times in the past. The same will happen with NFTs, creator tokens, Dow's, gaming, TCRs,

on-chain reputation, rebates tokens, decentralized storage. As time passes, each of these ideas will make more sense. I was with you up until there. I just think we're still too early for many of them and we don't quite understand the need yet, but eventually it will make sense. So I, you know, quote to me, and got pretty in his grill about it.

But I'm curious how you react to that, especially given that, you know, Y, C, you guys are at the very, very early stages. And somebody made a very good point. So I kind of had the obvious response, which is that I think this is a toxic attitude to take. It's almost like a toxic positivity kind of thing where it's like, okay, this is kind of too polyana-ish.

You should be learning from the failure of your predecessors. And like, in Bologis, parlance, you should study the IDMAs and understand, like, who went down which corridors, and what are the bodies that are, you know, like desiccated sitting there with their hand on some kind of booby trap. But somebody made a point in my comments, which I thought was a very good one, is if you think about something like Polymarket, if you think about something like hyperliquid, these are

perfect examples of ideas that didn't work for a very, very, very long time and then started working. And so, how do you think about this? If you're responding to him, Ron, he says, look, all this stuff is going to come back. Eventually it's going to work. What do you say to that? Before I say, I'd say, thank you for the war memories to TCRs. I have not thought about that at almost a decade, and I was exciting to hear about TCRs again. And I actually worked for Abology.

So, like, he made me a really good coroner to figure out, like, where the bodies were buried in what you can learn for the future. And by the way, to Bologi's credit, he's a huge, huge bull on Bitcoin. And he also saw the power of stable coins, very. early, and that was one of the products that we worked on together. So I think it's pretty amazing that you can have that both mindsets, that Bitcoin is going to be insanely valuable, and also stablecoin should exist and dollars should be on the blockchain rather than just replacing the dollar. But to me, it comes back to like, I don't know if you heard of Webvan.

This is this 1990s internet startup. There are all these examples of 90s internet startups that flamed out, they raised too much money, had some fanciful vision of what the future was. And I was in Ycomnia in 2012. There's this company called BitBank, which turned into Coinbase. But the other company was Instacart. And Instacart had a very different model of how they approached it. So one was obviously the time was different. There were mobile phones that were available. But the second big thing was they didn't go build factories on day one. They literally hired people and have them go into stores, you know, Whole Foods and others that already existed and bought something so they didn't have this massive capax. So I felt like what Apura did when he built the Instacard was a really,

really smart, like go to market. And then he also, like, really dependent on the mobile phone. No way, I think Instacart succeeds without the idea of mobile phones being everywhere. And WebVin in 1997, 98, clearly didn't have that. So I, like, if anything, like, I think, Steve, we might get into fight here is that, like, I actually believe that a fair number of those things that failed will come back. But I think, like you, like, I guess I'm skeptical that they're coming back right away. But I think that some of those things will.

TCR is, like, I actually think they're probably going to be a different form. there's going to be some form of data repository where there's tokens and you're incentivized to contributing data to some type of repository. I've seen that idea a bunch of times, and I think that eventually we'll figure out the right use case for it. So in YC, I think the way we think about it is, one is what is different about this time? The why now is such a critical part. We do 10-minute interviews with founders, and then we have to decide, do we make an investment

in them or not. And one of the core questions we always ask is the why now, what's changed about the world that made this idea that maybe it was terrible a few years ago or a few months ago, what is different about it and how do you approach it differently, you know, than other people did? And so, again, I, like, personally, I'm really excited for social. I think that's going to come back, like, in a big way. But candidly, I'm not investing in a lot of that today. And I wonder, I'm curious.

Like, is that the way you approach it as well? Is that you think, like, which of those will come back someday? I mean, if you take the literal list that he promulgated, which is, you know, once again, NFTs, creator tokens, dows, gaming, TCRs, on-chain reputation, rebased tokens, and decentralized storage. this is kind of like, you know, it's like walking through a cemetery, ideas that even at the time I thought like, man, I really don't know if this is going to work or like who this is for. I think of these, which of these do I do?

Can I, what kind of evidence would I need to see for me to believe that, okay, somebody is building new Dow infrastructure, somebody's building a new NFT platform and like, okay, I think there's potentially there there. It would have to be to your point, something has fundamentally changed in the environment or in the other substrate on which blockchains are layering on top of, that that has changed. And it's changed the underlying demand for this thing, right? So I made the point that, like, look, NFTs,

I can believe that NFTs will come back if we are all living in VR. If the year is 2035 and we're doing all of our Zoom meetings in some kind of virtual space and, like, we need some way to create digital scarcity, digital fashion, digital, whatever, I can totally buy that they're going to be digital, very clear universal ownership layer for scarce goods. Makes perfect sense to me. I can totally buy that. We're not in that universe.

Anybody who goes and pitches me today on, I think NFTs are going to make a comeback because I'm like, no, of course not. And I think for many of these things, it's difficult for me to even imagine what would have to change in the world for me to believe that this thing could come back. So, you know, rebased tokens are so vague that I'm like,

okay, well, that's just like a mechanic. That's not an idea. for something like creator tokens or gaming, I increasingly think that like these are just wrong, that like these are just actually incorrect mappings between incentives or like if you, it's kind of like, you know, you're a chef and you're like, huh, what happens if I put salt and crab and, I don't know, toothpaste into a cauldron? And the answer might be like, oh, well, there's one other thing you could add that would make it delicious. But sometimes the answer is like, no, these things just don't go together. These things don't taste good together. Stop trying to make it work.

So one, these are not new ideas. These are actually like generally very old ideas, right? Like even stuff like Aldo stablecoins or prediction markets, these are like Bitcoin talk era or older ideas. And so it's not even so much that, hey, you need someone to have this idea new or even from 2021 or something. There's a sort of tune to Meals point. I like the Y now framing of, well, hey, now we have a different regulatory environment, different technology, different install base, sufficient stable coin liquidity, better U.S., whatever it actually is. that actually allows this thing to exist. And I think, like, Levan and Sikard, like, two great examples. I think, kind of sticking on this food delivery idea, I think there's also a path dependency to getting to whatever success looks like in kind of the heyday of the food delivery wars, famously was like Dwardash and hostmates and, you know, Grubhub. And it was like, Dordash took this approach of really going out into the suburbs

and getting a lot of these chainshors, which is obviously how it's very uncontested territory and was how a lot of the U.S. looks. and that ended up being a very successful strategy for them in terms of building an actual dedicated loyalty base and sort of moat there and then be able to go into the cities and so you can have multiple teams executing well

with a similar idea, just taking a different path to get there and ending up with a very different result. And so I look for a lot of teams like that too where it's the initial people that sort of see the application or that are actually doing the original go-to-market that can end up actually using the very, very long sort of end state of a given company. Well, okay, so to argue against

myself a little bit. There's a famous line, I think, by Mark Andreessen, where he says, you know, back in the late 90s when they were looking at the early generation of internet companies, when they saw the web vans and, you know, all the companies that famously failed, the Chewis and so on, that what he said about that time was that every idea that we were pitched during the dot-com bubble eventually came true, literally every single one. But, of course, timing is everything when it comes to making investments. And I think many people, like, there's kind of an intuitive appeal to that idea that you can kind of just see the technology and just unroll the future. And it's just really obvious how all of this is going to integrate into society, that the internet's

going to be used for everything, that the AI is going to be used for everything. And of course, our version of that is that, well, crypto will be used for everything. So, you know, crypto will be used for all the games. It'll be used for all the jewelry. It'll be used for all the watches. And, you know, for creator tokens, every, you know, your income is going to be coming from a creator token that you've issued. Or, you know, I still don't know exactly what TCR's used for, but whatever lists you have will be TCRs. And like, I think this is, like, maybe there are some technologies for which this is true, right? I'm willing to grant that, like, people in the late 90s could have just said everything

will be on the internet and they were right about all of it, right? I don't think that is true in general for technologies. I think the internet was genuinely a unique everything technology. And I think crypto is not in everything technology. I think it is a many-thing technology. I think it is plausibly in everything financial. technology. But the core thing about many of these ideas is that they're not financial, right? Like you're talking about creator tokens and dows and games and, you know, on-tian reputation,

decentralized storage. These are not financial. And my thesis for a long time, and I've gotten into big arguments about this, is that like the core of crypto has always been about money and finance. Bitcoin was about money. Ethereum was about contracts, right? And Defi, it's got finance in the name. And prediction markets and savings. coin and tokenization, all this stuff, it's like very obviously all part of this, you know, mountain that it's being carved out of. And it's the mountain of finance and money and assets. The farther you go from that, the more we're kind of telling a story. And it's a story that just hasn't been vindicated. It has not been borne out in, to your point, to meal, what do people

want? What do people actually want from us? The answer is that they do not want us to solve their non-economic problems. And crypto keeps trying to do it. And so very famously, I mean, One of the stories that obviously connects to Coinbase is that Base, very famously, was trying, like crazy to get this consumer social strategy to work. So very famously, they were hawking creator tokens. Jesse from Base was pushing Zora and this idea that everyone's going to be using the base app in order to own content and trade content and so on. And he recently came out with a very long post where he more or less gave him a call bot. They're like, look, we got this wrong. This is not what people use.

elections for. Because of that, we were distracted, focused on the social thing, and we've now fallen behind. It's very clear. The story is about real-world assets, asset tokenization, prediction markets, more financial things. He did say, look, I don't believe that we're wrong forever, but we're definitely wrong for right now. And because of that, Jesse acknowledged that he was handing off the base app to Kobe, a very famous crypto trader and crypto celebrity, who is now running the base app while Jesse is going to be focused on the base chain. And so this seems like from Coinbase in acknowledgement that like, hey, we made the wrong bet. We got this wrong.

It's about money and finance once again. Emil, how do you respond to that? I don't have insight information. But what I would say in general in crypto is that we sometimes try to solve things that are our problem, but not our customer's problem. And in this case, I think that like in Jesse's note, he basically said that like, hey, I thought this would expand crypto. Which again, every crypto company is thinking about how to expand and how to grow. But that's different from what do people want and what is the value proposition they have from that.

And so I think the challenge in general in crypto for any builder is that when you have a token, it feels like people want it because they're making money, especially if you're early, right? So it really, this is what I really, this is why I counsel against the token early is it basically warps your vision. It seems like, like I think you guys probably saw the trenches

during that six-month or 12-month period around Zora and Farcaster and BAS. and base app where it seemed like there are lots of social apps being thrown out, people moving around these creator coins, there's a good vision for what this would do. It suddenly make creators to get money and monetize and things like that. And I think that's where I think the idea of token go up and meme coins and things like that. The challenge is, is there something real there at the end of the day that isn't a function

of just price go up, which works for a period in time like we've seen with NFTs and meme coins and L1 tokens back in the day. But you have to really get a sense for us like, the real pain point of problem that you're solving. And that's the, I think the YC attitude that I always have is like, what is the ultimate value prop? Do I think it's like ephemeral? Is it something that's happening because of price action?

And in crypto, you know, that's the challenge is that it feels like it's the future, but if prices are going up, anything, you know, seems like reality at that moment in time. And so that's really, I think the thing that I focus on. And maybe just to push back on one thing you said is that, like, I think money is everything. So, like, I know we say this is like crypto is like one portion of it. I'm like, even for storage, the allocation of that storage is a money problem. Who do I give it to for what price? And when I see like Filecoin and others, like, that's the reason they chose the blockchain

was not just the storage layer itself, but that allocation of that problem. I think that like, you know, one thing about I see it, the big thing has been capital raising. How do we help startups raise capital? I think the blockchain is the greatest capital innovation machine in the world, right? Right, right? And that like it has actually really good at that. Now, how do we make it better so that's not just pump a company? dumps and it's just like flipping the tokens.

But I think that's an example where I'm like every company in the world has to raise capital. Right. And so there's a part of me that like I really, I think like money is so much. I think we don't think it's big enough. Like it literally is everything. Like there's not a company in the world that doesn't have to deal with payments, capital raising, and things like that. And the blockchain has shown that.

And then I think like the last thing I just say is that like I'm always intrigued. I think there's going to be someone out there who comes up with a different insight for why social should work. and it didn't work this time, but I'm used to so many times things failing. The thing I think back to you see is the number of times AI has failed. And you go to Wikipedia, you can look up AI winter, but I don't think people realize for 40 years

we've tried various different ways to do it. No, no, no, no, no, no, no, no. I'm not going to get away. Throw that down. No, no, no, no. Okay, so the reason why AI didn't work is because the product didn't work, right? Everyone knows what they want AI to be.

Everybody knows they want it to be general, human intelligence, like they know the properties that they would imagine from an AI. And when AI actually delivers on the product that they are pretending to be able to give you, right, when it's like, you know, Tay or Bonzi Buddy or whatever, like, okay, yeah, that's not that great of a product. But they know what it might feel like, right? In the same way, we can say, look, VR right now sucks, right? It's okay, but it's not that great.

But we all know what we mean when we say real VR, like when VR gets really good, it's going to be a killer app and we all know it, right? That is the sense in which AI, people always knew. that AI, if it got there, it would be a killer product. You cannot say that for decentralized social. Is it true that if social were really decentralized? And if it were really, you know, fair or whatever, I don't know what the properties is that people nowadays are trying to claim. Or if it was really financialized, that if they really just got it in just the right way, that then it would be a killer product and people would use it over centralized alternatives, I think that's unclear. But for VR,

for AI, for so many of these things, like it is actually clear. For stable coins, it's actually clear. 10 years ago, if I told you, you can send dollars anywhere instantaneously, 24-7 to anyone with a mobile phone, will that of product market fit? If you can do it, yes.

Now, do I believe you can do it? Do I believe it's legal? Do I believe it'll be liquid? Do I believe, you know, yeah, okay, maybe all those things were standing in your way. But if you can deliver that product, yeah, of course people are going to want it. So that, I think that is the big distinction

with a lot of the non-financial use cases. No, no, no, I totally agree. I think it's very clear with AI what the end outcome was. And I think in crypto, it's harder for some of these things. You're totally right for stable coins is easy. But I think, by the way, I think one thing's interesting is, like, you said decentralized social. I think social is, like, creator loyalty is the way to frame it. And, like, I'm like, one thing I always do is look at, like, antecedents.

Like, what came before and what pain points had they had. And, like, I'm like, all of us use loyalty tokens already. because of like, freaking fire miles is just like one example of something that we use every day. And now I'm like, should there be another form of that that's tied to these large, massive podcasts or things like that, like a chopping block token someday? I'm like, yeah, maybe. But I think the question is, what is the utility of it, was the value of it? But by the way, that's where I agree with you is that like, I don't, I'm not a believer in just generic decentralized social as the solution. And I don't know what pain point that's necessarily solving. On a blockchain, I get it. I'm like, you. can't actually make this a reality, like the internet, if you can backdoor it.

And as a result, decentralization is critical. But for something like decentralized social, totally agree. Creator coins, I'm like, I think the way it was implemented was not the right way. Absolutely. And now my question is, like, maybe a smart person will come out and figure out a new way or the world has changed in some way and make it a reality. My take on these things is that, like, blockchain is ultimately a mechanism, right? It's a mechanic.

It is a way to implement the thing that you want. But what is the thing you're actually trying to solve for? like just stapling a blockchain to a use case is not a thesis, right? In the same way, calling something Uber for X, right? There were some actual Ubers for actual X's. And like there were places where that actually made a genuinely better product. But there are many things that you can Uberify an X or a Y or a Z, and it's not better. It doesn't make anything better.

And there is nothing intrinsic about the combination of two mechanics that is alone sufficient for answering, will something exist at this intersection? So blockchain, I remember when I first got into crypto, the first thing that I was pitched on, because I used to work at Airbnb, which also YC company, was Airbnb for Airbnb on the blockchain. And so many people told me like, well, obviously this will happen. Obviously, Uber on the blockchain and Airbnb on the blockchain, inevitable. Inevitable. Why? Because, oh, there's this platform in the middle.

It's rent seeking. What are they even doing? They're just aggregating data. Everyone should own a piece of it because they're part of a two-sided network. Very intuitive. The explanation even still sounds intuitive. Now, when you look at the details, you realize like, oh, no, everybody who's tried this learns very quickly. It's not actually primarily a two-sided market business. It's a trust business, a curation business, a fraud business, and blockchains suck at that stuff. Blockchains are a disadvantage in curation, in

anti-fraud, in all these other things. It gives you certain properties at the expense of certain others. And, you know, famously for Uber is kind of the same thing, right? Is that like so much of what Uber does is actually operations and like at this point regulatory stuff, even as much as it is just, oh, it's an app that has two different, you know, sides of a market on it. So I bring all that up to make the point that I think the bias as an investor, and obviously look, I'm putting my money where my mouth is, I will say it, this heuristic has. This heuristic has, I think, done us very well to avoid a lot of the hypey areas within crypto that have not panned out is that, look, one, follow demand signals, right? Almost all of these ideas, right? To the last

man, NFTs, creator tokens, Dow's, TCRs, entrepreneur reputation, rebased tokens, decentralized storage. Almost every single one of them did not come from a demand signal. Almost every single one of them came from, I have a hammer, here's a nail. And this is like the classic crypto thing, right? Is that like, you write a blog post, an investor writes a blog post, founders read the blog, you know, we like kind of blog post each other into a frenzy. And we're just like, oh, well, obviously this should exist. Like the same way that somebody was trying to obviously me into saying, oh, well, Airbnb on the blockchain should exist. And like after 10 years of this, I'm like, no, I think this is the wrong way to arrive at these kinds of ideas. It's like, if you're not

starting from a demand signal, you are probably fooling yourself. And it's not always true, right? I think there are genuine counter examples like Polymarket, but it's mostly true. Yeah, I just don't know if I would group all of those together. Like, I think there's some where it's clearly, there's clearly some burst of demand. Like, I would say NFTs, you have insane amounts of volume. These others, it felt like always like pushing on a string. It's like this forced mean that when you look at the numbers never really added up or Tune Miel's point was like highly incentivized with a token, which happens in normal venture too, right? I think of all the places that have VC dollars thrown out. And it's like, oh, actually, this is, you're not

economical. We can't really get a marketplace bootstrapped. And so it's just going to kind of kind of die out. I do think, again, that the path dependency thing is also very real. Like I always think of not to keep name dropping YC companies, but like Reddit versus Dig. It's like, you know, basically the exact same product, totally different paths and totally different communities. it's like why. It's like one is seated with YC people talking about YC stuff, and like YC culture. And like that obviously ended up, ended up diffusing over time. But it's like this sort of founder effect, I think is very, very real. And so it could be the case that like, okay, the stuff that's been tried today didn't really have the right sort of

first step. It's like, you know, you landed in Jamestown. You didn't land in Plymouth Rock. And so therefore, you know, your colony is going to fail. But like, you know, maybe you try it again. And it's a different group of people this time. And they have a different sort of thing. And it's like every step is is very important. Okay, so Tom, let me ask you more pointedly. What would you need to see to back a TCR? I'm not going to back in TCR. I'm going to stop you right now.

I think Nemeel's point around that. On Gen reputation. I actually think there's something interesting now, you know, vis-a-vis proof of humanity in on-gen reputation, right? Where it's like, okay, we've actually seen orders of magnitude increase in terms of, you know, personas online and content and claims. And so great. like when you see orders of magnitude, you know, increases, there's probably, you know, some,

change and required in how we sort of done things previously. I was, I think, to the numerous point around capital markets, this is one of the oldest ideas that I think there's clearly, you know, demand. And it's like, okay, well, you know, in the past, it's like you shoved the plane off the cliff and, like, it was kind of airborne for a little bit. And then it crashed, but, like, it was airborne for a little bit. There's something in there. And it's like getting the details right. And it's going to take some work. It's clearly, I think, such demand and it, clearly such good. like technology product fit for something like that. That's fair.

Although you read it broadly enough to be like identity, then okay, yes, fine. But I think on-chain reputation is like, if you said identity, I'd be like, yes, I agree. That's a huge problem. The solution would be very big. Yeah. Go ahead. Steve, are you bearish on-chain identity? On-chain reputation?

What is a reputation not an identity? Well, no, I mean, I think I'm bullish on the concept of identity. I don't know that it would be on-chain, but if it is on-chain, great. On-chain reputation, I think of more like, did you do good things? Are you a nice guy? Did you rug people? Did you like this kind of thing? There's a few protocols in the past that have tried to do this, right?

But I think reputation, obviously reputation is not identity, right? Like reputation tracking services or reputation, you know, the whole idea is to build a social profile of somebody despite the fact that they're pseudonymous. That's what I interpret. Is it a credit bureau to you on-chain reputation? Oh, sorry, reputation? Plausibly. I think you can you can tie those things together.

I think it's very clear that like on-chain credit, purely on-chain credit, doesn't work very well because of the, what's the term, repudiateability of addresses, right? I can go do a bunch of bad stuff. I can default on loans. I can whatever. And then just never use that address, never opted into my system. And like credit underwriting requires a full view of a person's activities. And this is historically one of the, and of course the fact that there's no recourse on chain, right? I can't put a lien on somebody's address.

and so there's no ability to garnish wages, there's no ability to force any kind of payback. So all these things have made it really difficult for reputation on chain or credit or whatever to really emerge without going off chain. And the reality is like there are people now that are doing on-chain credit,

but they're reaching off-chain. And they're saying, okay, you know, we need to do some KOC, we need to understand your real business. And so when I say on-chain reputation, I'm assuming that's what is meant here, not off-chain reputation on the blockchain. What do you think that's like about?

Yeah, bootstrapping, right? Like that's a different, you know, sort of path and choice. And then eventually... We're just saying, okay, we're saying like, I see you and I take a bunch of docs and stuff. Okay, yes, of course. That's going to exist. No, going back to like, again, the Instagram thing,

it's like, okay, we're not going to boil the ocean. It's like we're going to start very small and lean and like do this kind of pseudo MVP that doesn't really do the thing that we explicitly said we're going to Wizard of Oz it basically. And then over time, we sort of wean ourselves off of the Wizard of Oz version and we had the real thing going. I think actually like this is something that was kind of a lot better with. I was kind of like,

it's a Carl Sagan quote. It's like, you want to make a pie, you got to create the universe or whatever. And I feel like crypto was like that for a long time. It's like, well, I want to like make an app. So I'm going to make a new blockchain, which is obviously stupid. But now it's like, okay,

you can piece this stuff together with some APIs in an afternoon. And then great, now you can kind of go and do the thing that you want to do and kind of gradually kind of swap pieces out in like a modular way as you get bigger. Totally fair. Totally fair. And to be clear,

I am not making the claim that on-chain credit will never work or that on-chain identity is irrelevant or anything like that. I take this to be like an index of 2021 ideas, right? That's what I'm reading this as. And so I'm thinking of the 2021 version of what people told me was going to exist, right? Like I remember we did debt Dow, if you remember debt-dow, Tom. And like this was this idea that you could have DAOs that are issuing credit because you

could put like a ratchet directly into the protocol that allows them to basically be, you know, sort of take first cut of revenues in order to pay back the debt. And like, you know, it's like, okay, I can buy that. That kind of makes sense that like you have mechanics similar to like a lien that is enforced by a court. But in this case, it's enforced by smart contracts. It didn't work. And maybe someday it'll work.

And like plausibly that you could do something like that into hyperliquid today. Maybe hyperliquic could take on debt as a protocol. And basically like the debt holders are guaranteed to get paid first before equity holders get distributed dividends. Makes perfect sense. That could totally work. But the 2021 version of this, which I think was like a little simplistic, a little polyana-ish, I don't think it's like a matter of, oh, it'll all eventually happen because it was foretold.

And like God just plays mysterious games. Yeah. I think that's more the point, which is like, you know, startup ideas that are like too, like naive or literal, almost like never work. I'm just like, I want to build a Kement network. So I'm going to like get, build an app. Lesb people. And then just like get everyone on that versus, oh, no, actually. I'm going to build this little widget that's like backwards compatible with the existing system.

And you can great, you can swipe a card on your phone. And you just have this like little hook into eventually it ends up being a big market. But it's starting with the kind of like simplest toy example. And then using that to grow. Actually, I mean, Emil, you've also been running in the space for a while. I'm curious if you have like pet favorite ideas that you like want to see you happen or that you keep, wishing you're going to like, you know, happen someday. Like one thing I will say is I think what I miss about 2021 is people are trying things or what I

about the old Bitcoin talk days. It's like, by the way, X402 is one of the teams, my team launch at Coinbase. There's a Bitcoin talk article from 2011, you know, about that. And the why now is basically agents exist now, you know, machine payments matter. So what I will say is like what I really do enjoy about it is that people are trying things.

We're iterating. We're learning what works and what doesn't work and the world is changing. I think the part that's a little, actually, disappointing to me now is a lot of the things we are building in crypto, are direct analogs to what's in Tradfai. And that's like saying that the internet was great for newspapers. Clearly, it's great for newspapers. But if you look at Facebook, that's a really weird, different thing.

And it's not a natural analog to the old world. And so that's kind of like, I'm also excited to see more of that. And right now at least I think the moment we're into crypto is a lot of like, oh, it exists in the other world. Sobecoins are a really great example, right? And now let's find the blockchain and it's instantaneous. It's cheaper. It's faster or whatever.

Like those are all the things that it comes with. And I'm like, yeah, clearly, that's awesome. And I think there's going to be a massive business built on that. But I'm also excited to see a little bit of more, more the inventive and, you know, out there stuff. And maybe to go back to TCR is like I think one, I do wonder like the way we talk about it is that like we're overloading the term. So when we say social, I'm not sure I believe in decentralized social, but I do wonder if there's a loyalty token eventually. I think like TCR is an example where like I think data collectives needs to exist where like you need to contribute data and you might get equity, like a quasi form of equity in that data, like over time.

And the thing I was really excited about early crypto was like one of the visions was that building double-side marketplace like Uber is hard. You could use tokens to like accelerate that. And again, like we've talked about this for 10, 15 years. I'm not sure I've seen like a good way where you accelerate it, but the token doesn't turn into a game and then the thing implodes, you know, at some future point. And that's those types of things, Tom, that I personally really excited. Like if someone tries that and figures that out, that would be like a huge,

huge net win for the world is if we can figure out double-sided marketplaces more easily, you know, with the right crypto incentives. And I think that a lot of it is about like, how do we take the awesome power of the blockchain, let's say capital raising? It's freaking amazing. Like, you know, in the world, it's like awesome. But then how do we get rid of the more extreme playing with gasoline dynamics of that? You know, that leads to too much capital for the wrong protocols and early exit liquidity and things like that. And that's, I think, where we need to like start innovating it. And just to give you like an example of a story is like, before YC, all these VCs would negotiate custom rounds.

You paid $30,000 to $60,000 for this. A lot of VCs would write in random terms that the founders couldn't push back on. And basically, like, YC's innovation was like, let's launch the safe. A standard document that like everyone can align on. The lawyers already know about it. It doesn't have weird terms in it. And that's an example where you took a market that like maybe had a lot of lack of transparency and made that a little bit more transparent that was.

And that's why I think crypto needs is a little bit more innovation like that. that takes the awesome power of this and harnesses it in the right direction. Okay, well, we have special guest, Tarun, joining us. Very lucky to have him on the show for, you know, he rarely makes it on, but this time he came right on time. I apologize for being late. My phone got stolen, and it just took me a long time to get home today. So that was my, it's kind of a little of a disaster of an IRA whole day.

So I, again, apologize to all of yours. Well, if you, if you, just to get you caught up, we are debating TCRs. Namil is very bullish TCRs. I am bearish. You know, there's some things. I'll give you context. Do you see the post by Imran?

Yeah, yeah, yeah. Where he was like, almost everything that failed will come back around. So I am very anti this. Nemeel and Tom. Nemeel is more believes it. Tom is somewhat tempered in the middle. If I can summarize very, very rudishly. Turin, what's your take on this whole argument?

I think I'm probably somewhere in the middle where there's like clearly ideas when you're assuming. Okay, so like I think a lot about 2017 and 2013 in crypto where there was kind of this hidden belief that homo-economics existed, like this perfectly rational person who had participate in every network that existed

and allocate all their resources to my network because my network had the best white paper. You know what I mean? Like there was kind of this hidden assumption in all of the crypto inventions that was like, my thing is just so, beautiful that everyone will drop

everything else in the world that matters to them and only focus on my thing. And I think you need that little of grandeur of delusion to kind of get something. But obviously I think for a lot of these things that has just not been true, right? Like, Dow governance, great example.

TCR is a great example where like there are a lot of other incentives at play. There's sort of this idea that like people don't care and when enough people don't care it becomes much easier to manipulate. Dot, dot, dot, dot, right? So there is some argument that maybe some things with a world of agents and a representation of yourself that can feign interest and rationality at all times.

Maybe the world looks different. Maybe homo-economicus. I like this. Maybe homo-economicus will be reborn. I'm not really convinced of that for most things. I think the payment one is a kind of interesting one because I think about the early Bitcoin era where everyone was like, yeah, well, obviously the miners are machine-to-machine payments.

That's like that was the machine-to-machine payments, like mining pools. Like there was some posts that say something almost exactly like that. And then it was like 2017, Iota, machine-to-machine payments. All the IoT blockchain stuff is realistically the same thing. And then it's like actually now that actually seems like completely feasible and fathomable, whereas like in those cases there was always this kind of like, you had to take some leap of phase. And so in that sense, I think there are good ideas that exist.

I'm not sure all of them. In fact, 99% probably are still going to fail because humans can only concentrate attention for a small amount of time. Agents can concentrate attention for a certain amount of time, but they're still resource heavy. So it's not like it's free. But I do think there was a kind of, I don't know if, I mean, I'm sure you guys all remember the,

if you're in CryptoPivot to AI tweet by Jason Calcanus. and then I guess Brian Armstrong wrote this long tweet and that made Jason change his mind and write if you're in Crypto Use AI as his new version which I think he wrote yesterday or today, I forget sometime in the last week. And I thought it was kind of an interesting thing where there's a sense in which AI actually

has a lot of the white paper energy. Like not, not Megalabs, not. the data center operators. But like everyone in the middle, right? Like every new NeoLab is like, I mean, all the U.S. open source labs are white papers, right? Like more or less for being honest. Like the Chinese ones are the only ones who deliver nine. It's like the opposite of the US first Chinese L1s in 2017 where the Chinese L ones were all fake and the U.S. ones exist. Just a funny comparison. But there's kind of an interesting question of like, hey, are there ideas there that mix with the crypto ideas and like suddenly that primordial soup is something.

And that's where I'm, that's where I'm like, I'm willing to kind of bend them. Yeah. So no, I feel like, I want Terra 5002. Then that's like, okay, fine. Yeah. I totally agree with you that like, look, underlying technology change means the rules have changed. We've, we've sort of shaken up the chessboard. And now actually there are moves available to you that weren't available before. Right. If you, if you move your night here every time you will die, but now all of a sudden, the chess board has changed, move your night here, actually the board's open. I think AI is one of these things that actually does change the landscape. I totally agree with you there.

I think the same thing, like I said, VR plausibly could change the market for NFTs. That said, I actually really like the way you framed it, Tarun, about like, we had a lot of assumptions about people are going to pay very close attention to very subtle incentives. And, you know, it's like, okay, well, normally, you know, if I'm bootstrap and. Uber, I might pay out some rider incentives, but now I'm giving you a token in Uber, which means that you must, if you're following your incentives and being rational, get all of your friends and everybody around you to start using Uber just like you, so that you will become unfathomly rich because multiply, you know, small probabilities together, you get very big numbers. I think the reality is that some of these arguments are just wrong.

It's a little bit like, you know, it's crudely analogous to like the arguments for socialism, which is that, well, if everybody just follows this perfect equilibrium of like being good to their neighbors and thinking about, oh, the family unit is already socialist, so you just expand that outwards and like, isn't there a beautiful symmetry to this idea? But in practice, it just doesn't work. And you can try again and try again and try again, and it doesn't work. Now, maybe it'll work when the players are agents and they're not humans and they don't have bounded rationality and they don't have, you know, whatever biases that we have.

But for many of these things, the answer is like, you know, why don't Dow's work? They should work. Rationally, what exactly is so different about a Dow from any other form of governance? But very clearly, DAOs are really bad. They are reliably worse than almost any other form of corporate governance or whatever. And there's some deep property here that maybe we can't even articulate of what exactly is the reason why. But it's very obvious from the record that if you can avoid being a DAO, you should. You will be better at doing your job if you're not a DAO.

And we've genuinely learned that. I don't think that's a like, oh, well, you just don't have the right Dow tooling yet or you don't have the, you know, your Dow members are just not as good as my Dow members. I think it's that look, we've actually learned something about the world, and we should be incorporating this into how we build things going forward. In the same way you learn about management over time, we've learned about the quote-unquote management of systems with crypto or decentralized networks being one of those, quote-unquote, management tools. I just think like this homo-economics, like this perfect rational human hidden

that's willing to like do all the token incentives and also isn't the person who made it trying to get other people, to buy it. Like, balancing those two is, like, a very hard game. But I think the agent world, like, I'm actually more of the mindset that we just haven't conceived of the correct financial-like interactions they're going to have, right? We're thinking of these very, like, human speed peer-to-peer. But, like, maybe they just love bundling millions of transactions as one or, like,

things that we haven't really totally thought of. And that's more, I think, maybe that's kind of what Nabil was going at at the end. But, like, that's sort of what I'm more. excited about them, the things are like, how do we like give them human capabilities, right? We should treat them as this like magic alien that can do something you can't do. I feel like we landed on a foreign planet and tune. We're like every day learning more about this entity that's out there. But like the version I give is that like with AI agents, like e-commerce, I feel like it's

going to be like credit cards, they're going to be connected to agents. We're just going to do the same thing we do. Like something like that. But then there are these things like, I think you guys maybe heard like Cloudflare announced that they're going to start gating websites behind a call and pay for that with crypto to start with. And I think, like, I'm really curious, how does that work? Is that going to work effectively?

Or are they going instead just signed business development deals with large players like they've already been doing? And that's the way this is going to work or the AI agents are going to pay, three cents and two cents and five cents for different types of transactions. And I think personally, that gets me excited to just see what's going to happen, right? So, first of all, I agree. We don't know yet. way, way too early to be prognosticating any of these things.

But I think we're also, any new technology, it's easy to fall into the trap of wanting to paper over complexity. I think with the early internet, people had this idea of like, oh, we're all going to be online and it's going to be people from all over the world and we're going to learn that we have no differences and blah, blah, blah. And like, well, no, actually, like, the Russian internet is very different from the Chinese internet, it's very different with the American internet. And, like, inevitably, that's going to happen.

There's going to be splintering. Same way with the blockchain. I think we had these very simplistic ideas about the complexities. And I think it's likely to be even true for AI. One can make the argument that's like, well, it's very annoying for a human to like pay a cent to like read every web page, even if economically, like, that is the right market clearing price perhaps. But it's just annoying and high friction and like, you know, the reality is that you were not going to be willing to do it. But agents will be willing to do it. Kind of, kind of.

But agents also pay tokens. And like the cost of having to realize, oh, I need to do this and create this like, I need to, I need to, I need to. You need to create this HTML payload to go pay this thing for like this one H-Tml web page that I'm scraping out of the like 5,000 pages I'm scraping in the span of these two seconds. And if I'm instead batching all of this with one batch payment and I'm calling on a subscription and I have something in the middle that's like handling all this complexity for me and I'm not dealing individual payments for individual things, actually I cut down my token cost by 30%. And that like the friction that arrives in humans actually still exists just in a different form. and actually it's not so different from the old world, which is that we still have bundling and we still have subscriptions and we still have like, yeah, this thing is

sort of too cheap to meter, even if it's metered in some more kind of loose sense. Tom, what do your thoughts? I'm trying to get kind of concrete here for a minute. Yeah, I mean, I think it kind of goes back to, like, again, these you're seeing orders of magnitude increase in traffic and so something's going to break and you think to rethink some assumptions. But I do think it was kind of noodling on the turn's point around like the white paper era. And I feel that actually a lot of the models that come out where it's like,

you report the benchmark numbers. And you're like, wow, everyone's just, benchmarks look great. And then you try and like, oh, there's like actually a piece of shit. And it's like kind of kind of like the problem with like chains or report TPS where it's like, okay, on paper, this is the thing that we care about. It doesn't actually care, capture the thing that people want to do,

which is like you want to go swap some meme coins or something. And I was actually thinking about this recently with licensing where I feel like crypto is very early in terms of like pushing BSL of like it's open. And you can verify what it does, but you don't get to use the actual code. And now you see actually open source kind of moving more this direction. Like Mongo has this new license came out a couple of years ago where it's like, you can use Mongo,

but if you want to resell MongoDB services, you have to pay for a license order to do that. And like the new Kimi weights are actually similar where it's like, you can use the Kimi weights and you can look at them. But if you want to sell Kimi inference, you also need a license from us. And so I think there's some like some interesting analogs of like, how do you monetize open source while also having a lot of the principles that people

want around like transparency, verifiability, things like that. First of all, we're all figuring out as we go. That's the thing I love about this, is we don't know the answer. People are trying things. And what I love every day is getting up

and seeing, what are people trying? I think that, like, what you're saying is already happening. Like, the labs are already paying bulk rate to, like, scrape the New York Times. They're paying $10 million or $20 million.

But I think, like, when I get up, like, when I look at the problem, and I'm like, the thing that's crazy for me is that we basically destroy the monetization model of the internet right now because ads are essentially are not going to work in the way that they are today.

And at minimum, what with the AI agent may serve up the ad, they're going to get the revenue. Now, how do they distribute that back? Right? There's might not even be paying for it, but actually distributing back some money that they're making. And I'm like, there's a version to me.

I'm like, I do see smaller content providers that are not going to sign bulk deals. Maybe they'll sign up for a subscription service that they're part of. And I think, like, in my perfect world, I'd love if that was more decentralized. So one company didn't have that power across every website that you had to sign up for my bulk billing service.

you know, that interface with AI labs. So I'm hopeful that someone will figure it out, but you're totally right. It could end up a world where, like, we're back to the centralized version of power. And I think that, like, to me, like, it's kind of fun to mess a little bit with the powers that be, you know, and to think about models

that can potentially break that up a little bit more. Yeah. Look, to be clear, I'm in this industry for a reason, and I completely agree with you. I don't think, I think the big difference in my mind between how Haseeb today thinks and how Haseeb in 2020 thought, is that I think Haseeb in 2020,

you know, very much being crypto-pilled, had the sense that, like, crypto as a technology made decentralization inevitable. Is that once the technology makes it possible for things to decentralize, they naturally will because things want to be decentralized. And what I appreciate now is that it's the opposite. Things don't want to be decentralized.

Things actually naturally pushed against decentralization. Things pushed towards centralization. toward intermediaries, they push toward intermediation and aggregation and bundling and so on. And it is only through a set of customers or a set of users who really truly demand decentralization be the de facto reality that the market forces will actually congeal to make something that's a decentralized alternative or push the system to stay decentralized. And unless that's there, the default will almost always be it rolls downhill and downhill is some configuration that involves, no, it's not like 50 different exchanges that all have

equal market share and we're all just like, you know, whatever. No, it's like three exchanges and they own most of the market. And ultimately, if you get kicked off of them, then you have no liquidity. That's the default state of most markets, of most environments, of most technologies. I 100% agree with that. And the thing I'd say, actually, I think innovation is the thing that breaks up older forms of centralization to create newer forms of centralization. So I think there's a world where the last 15 years we're under the Google and Facebook and others.

And now you look at the shift in that power. So I'm actually an innovation maxi, which is that if you have new things like blockchains, if you have new things like AI, they actually disrupt the powers that be. And if we're in ever a moment where that slows down, you're going to have those centralisms forms be maintained. And so like really, I think the thing I'm pushing, like I would push for it. And I think YSE always does is smaller players being able to take big shots. And candidly, some of them will become the big centralized players that we fear. But if long as you keep doing that, hopefully they won't be there for too long.

And by the way, in tech, this is a really good example. I think the other industry is like financial services and others that don't feel like they change that much. But in tech, I feel like even if you're the biggest company in the world, you're scared like crazy right now in how much the world is changing. And like that I think to me the example of what Google has gone through in the last decade is a really good example from like a major, you know, large, almost monopolist, at least massive centralized player to someone who is actually running scared and actually

having to make a lot of shifts in their strategy because of innovation that's come. But again, that's not blockchain decentralization leading into a natural stated decentralization. It's actually innovation, forcing centralization apart and then bring it back together. Right. And the story of history is always like the sort of, you know, are we the baddies meme? Is that like, okay, you rebel against the previous baddies and then you become the baddie. And so like, okay, you know, Google fought against, you know, Microsoft and IBM. And then now we're fighting against Google.

And like that literally, you know, when I was growing up, I was like, Google is this amazing democratizing force in the world. It's flattening the universe, like information free. It's a free part. Everything is free maps and search and all this stuff. You don't have to pay a dime for it. And at the time, like I was like, this is incredible. What a bounty.

What a what a inversion from what companies used to be like. And now we look at that and we're just like these parasitic, How could they give these things away and steal our data? And crypto was like the next stage of that rebellion. And now we look at, you know, people look at Coinbase or they look at Binance or they even look at like something like hyperliquid. And they say, okay, well, this thing is, you know, is it really so different? And the answer is like, it is different.

But it's, there's no inevitability to it. And so I'm sure today people look at Open AI. I mean, now Anthropics are getting a lot of flack, but let's take Open AI. people are showering opening eye with all this love. It's like, oh, they're supporting open source, and they're like pushing, you know, they sign the letter and they're pushing down, you know, the token prices,

and they're trying to make everything really cheap. But there will come a time when we all look at opening eye as these overlords that are controlling everything, and there will be a new generation of people. It's like, I don't know, it's like Chinese history, right? Like the emperor, the old emperor was a tyrant, and so we killed the old emperor. We install a new emperor, and that one's good,

until we decide that's the new emperor. Now we want to kill them too. I mean, I think AI feels a little bit different in one kind of very weird way. The internet when it started was trying to find direct monetization methods, right? Like you had to pay for subscriptions or whatever. And then it just turned out that it was like actually find some other way of subsidizing PFO. online ads, whatever. That kind of hides it from the end user

because we want the end user to look like an American because it's like they all want health insurance, but they don't want to know how much it costs and they don't want to get a bill, but they have to somehow still pay it. Right? Like that's kind of what the internet ended up becoming, right? It's like the US healthcare system in some way, right?

It's like a no bill. You don't know what you're paying, but like you're getting something. Unclear if it's trash. But like, you know, and you might be paying in other ways you don't realize. And the interesting thing is I feel like the AI CAPEX stuff and obviously everyone's trying to decrease it but let's suppose we do hit some threshold where it's like

just so high and the OPEX costs are also high right like as the inference stuff gets bulkier then what is the monetization it doesn't feel like it's like this marginal cost of open sources zero type of thing right like the whole narrative you know you think about ESR and Linus Quervolds of Linux in the 90s. A lot of the writing was about this idea that the creation process

was the expensive process and the thing to be valued and the thing that you had to spend all the money on. But after that, the marginal cost is zero forever. And now it's like actually the marginal cost is quite high forever. There's not like a... Just because the weights are open

doesn't mean you're getting out the thing you want immediately. Right? You pay some new realized marginal cost. And it feels like that marginal cost, and maybe this is like me not dreaming of nuclear fusion or data centers in space enough or whatever thing you need to make this make sense. But like it seems like it's too high to just be like a pure subsidy model. Right. And to me, that's like we move the Pareto Frontier, but we also like moved back the payment, the economics of it. And I don't know how to reconcile that. That to me feels like industrial revolution.

Because like, I don't, I actually don't think it's going to be out. I don't think they're going to be enough to subsets. I mean, look at the amount people are spending on CAPEX right now versus their revenue. They're all cash flow. Some of them are cash flow negative. These are like the highest free cash flow of businesses in the world. And in one year, they're negative cash flow. So like, if we think that that trend at all continues, the ads are not enough.

That's like the stupidest, like, 30 IQ way of saying ads are not enough, right? Just like look at the CAPEX relative to free cash flow. It's like, I don't think you've seen anything like this since the like World War II era. Don't you think one is the cost curve is going down? And number two is I actually think they can take a transaction flow for the things that they help support you in buying. Right? It's not just like a straight up $20 or $1 or whatever an ad. But it's like, I'm going to take 2% of your order because I help like support it.

and set it up, right? To Amazon or whatever. Outcome-based pricing, I agree. If you click on an ad and you buy something, they took 2%. Outcome-based pricing that is beyond what you have in ads. I think I agree. Payments for Order Flow for this AI stuff is like, I feel like if the internet turns out to be only websites

are generated on demand by an L-LM, there's no static servers anymore, no static websites. Because that is one thing people are starting to talk about. It's purely on-demand internet. Like there's no, the tokens are generating the HTML you see live. That sounds horrible. Right, right.

But in that world with no static content, then like you don't have this free copying thing anymore. Now the marginal costs. You're just like a forest of M dashes, you know, like moving away through the foliage. But like in that world, in that world, I think like, Nemeel's right. Like there has to be some type of like order flow sharing type of thing that looks like ads, but it's ads with a much stronger performance fee than. ads you have now where there's way more spray and prey in like some conversion fee, but it's like kind of more of a lottery ticket. Like it's becoming the gateway to everything. Like is, I think

the view I have. It's like a browse, so the new version of the browser where that is the most critical, like the way we talk about wallets and crypto, right, is the form of like what this is. And I think that like every user interface is going to be rethought to go through an AI agent. And so therefore, there's a tremendous amount of market power that you have there. And like, To me, I think they're going to figure out ways to be able to tollgate that in various forms. And it's going to be probably much more than like banner ads. Yeah, I mean, I did have this tweet a few weeks ago about how harnesses, like front ends and harnesses for AI are like wallets. Routers are like decks aggregators and kind of bridges and things along those lines.

Models are like protocols. And that basically means their fees are going to not going to be, they're not going to be able to extract fees themselves, as we see right now. And then inference providers are liquidity providers, right? They're putting up capital to earn some yield from the participants. And there's like some stuff about that that's not true. Like it's like a very overly simplified makes crypto people feel good about themselves. Oh, we discovered something first. I suddenly felt very cool when you were drawing out those analogies.

But I do think there is some truth to that. Because if I look at the monetization in that order flow, it's all payment for order flow at the edges, right? So liquidity priors into the wallets. And then the routers and protocols can't extract. very much. And maybe that's what the AI Internet looks like. Well, I mean, St. Alton was actually at Y.C.

And one of the things he said was the new phase of open AI is as utility, right, which means like massive scale, low margins for some part of the business, right? But the idea is that everyone's going to use AI. And so as a result, like that that's going to be different margins at that, that core like financial model layer. Yeah. Interesting. Okay.

Well, before we become an AI podcast, we are up in time, so we got to wrap. Namil, where can people find you? I'm at Twitter or X, sorry, Nemeel D. And by the way, today is Y Combinator's fall application deadline. Actually, if you submit in the next few days, we'll look at it. So go ahead and submit at Ycombinator.com slash apply. Okay, perfect. Nemeel, thanks so much for coming on.

And we'll be back next week. Thank you, everybody. Thank you.