Unchained - The Chopping Block: Meme Coin Frenzy, Decentralized AI, & Uniswap's Fee Strategy - Ep. 617

Episode Date: March 9, 2024

Welcome to The Chopping Block – where crypto insiders Haseeb Qureshi, Tom Schmidt, Tarun Chitra, and Robert Leshner dive into the hot topics stirring the cryptocurrency world. Could Uniswap's fee switch revolutionize DeFi's economic model? What's the real potential of decentralized AI in blockchain, and how close are we to seeing its practical applications? And in the whirlwind of the crypto market, what makes meme coins like Jeo Boden not just viral, but influential? This episode offers a detailed analysis of Uniswap's strategic move to potentially activate fee switches, exploring its implications for liquidity providers and the broader DeFi landscape. We also navigate the promising yet challenging terrain of Crypto AI, dissecting decentralized inference and the feasibility of verifiable training. The phenomenon of meme coins, with a spotlight on Jeo Boden, is dissected for its impact on community engagement and market dynamics. Join us as we probe these intriguing questions, equipped with Robert Leshner's insights, to unravel the complexities of DeFi innovations, Crypto AI's frontier, and the unique ecosystem of meme coins. Tune in for an enlightening session that aims to demystify these trends and provide a clearer understanding of their role in shaping the future of cryptocurrency. Listen to the episode on Apple Podcasts, Spotify, Overcast, Podcast Addict, Pocket Casts, Pandora, Castbox, Google Podcasts, TuneIn, Amazon Music, or on your favorite podcast platform. Show Highlights 🔹 Uniswap's Fee Switch Debate: Unpacking the implications of Uniswap's potential fee switch activation and its impact on DeFi's revenue models. 🔹 The Intricacies of Crypto AI: Diving deep into the world of Crypto AI, focusing on decentralized inference and the hurdles of verifiable training. 🔹 Meme Coin Phenomenon: Analyzing the cultural and economic forces behind meme coins like Jeo Boden and their sway in the crypto community. 🔹 Regulatory Winds in Crypto: Discussing the changing regulatory landscape and what Uniswap's move might signal about the future of DeFi regulation. 🔹 Market Predictions and Speculations: Exploring expert predictions on Bitcoin and Ethereum's price movements and the factors driving market optimism. 🔹 Technological Developments and Security: Delving into the significance of memory-safe programming and the White House's surprising tech endorsement. 🔹 Attested Data in the Age of Synthetic Media: Examining the role of cryptographically signed data in ensuring content authenticity in the blockchain era. 🔹 Decentralized Platforms' Disruption Potential: Probing the capability of decentralized platforms to reshape traditional social media models. 🔹 The Economic Model of DeFi: Discussing the potential transformation in DeFi's economic landscape post-Uniswap's fee switch implementation. 🔹 Crypto Culture and Community Dynamics: Investigating the unique cultural aspects of the crypto community, especially around meme coins and their market influence. Hosts ⭐️Haseeb Qureshi, Managing partner at Dragonfly  ⭐️Tom Schmidt, General Partner at Dragonfly  ⭐️Tarun Chitra, Managing Partner at Robot Ventures ⭐️Robert Leshner, Founder of Compound Disclosures Links Bitcoin Is Hitting All-Time Highs Around the World (CoinDesk) https://www.coindesk.com/markets/2024/02/29/bitcoin-is-hitting-all-time-highs-around-the-world/ Learn more about your ad choices. Visit megaphone.fm/adchoices

I also think the meme coins, do they pass the Howie test? Like, literally no one has an expectation of profit. Well, none of the money is going into the development of anything. Yeah, there's no like shared fees. Yeah, yeah. Like, nothing is getting built, period. Not a dividend. It's a tale of two quons.

Now, your losses are on someone else's balance. Generally speaking, air drops are kind of pointless anyways. I'm into trading firms who are very involved. I like that eat is the ultimate pump. Defy protocols are the, antidote to this problem. Hello, everybody. Welcome to Chopin Block. Every couple weeks, the four of us get together and give the industry

insider's perspective on the crypto topics of the day. So quick intros, first you got Tom, the Defy Maven, and Master of Memes. Hello, everyone. Next, we've got Robert, the Cryptoconassur, and Tsar of Super State. GM, everybody. And we've got Tarun, the Gigabrain, and Gran Puba at Gauntlet. Aloha. And I'm a sieve the head hype man at Driving Fly.

So we are early stage investors in crypto, but I want to caveat that nothing we say here is investment advice, legal advice, or even like. for advice, please see ChopinBlock.XYZ for more disclosures. So boys, we took a week off and turned out we took the worst week to take off possible. Because Bitcoin, well, for us, best week, but for I think the listeners, they probably want to hear from our reactions to Bitcoin kissing all-time highs. Now, interestingly, so Bitcoin, it basically went right up to 69K and then retreated. I think right now as we're recording, it's like 67.5 somewhere in that ballpark. There was a great article in CoinDesk that pointed out how, in most currency,

Bitcoin has already surpassed all-time highs because the dollar has been very strong relative to most currencies. So in Korea, in India, in Europe, Bitcoin's already surpassed all-time highs. But in dollar terms, Bitcoin is slightly below all-time highs, but it has really cemented Bitcoin mania at this point. Bitcoin has pulled ahead of almost everything, except for, of course, meme coins, which are going nuts, which we'll also talk about. How does it feel now emerging on the other side of what was, I mean, people were calling it a bare for really quite late. Now I'm hearing people say it's starting to be a bull market. I'm like, dude, it's full on bull market. How do you guys feel and what are you observing? Well, I can't

speak for everybody, but speaking for myself and for a lot of crypto Twitter, I don't think it feels like prior bull markets yet in that this is like the quietest all-time high bull market rally. We've had probably in crypto history. The last couple times we said all-time highs, there was a lot more mania. There was a lot more retail. This was a slow burn up pretty much the whole way. I mean, I don't know if that's true. We had the biggest monthly candle in Bitcoin history. It went up 20K in February. Yes, but that's a notional increase. And, you know, versus prior markets, you know, that's the equivalent of like, you know, four or five K. Yeah. I think when you look at sort of the traditional like media metrics, like, you know, Coinbase ranking on the app store.

and also just like friends texting you. It does feel very muted. And I think we're going to talk about Eat, Denver later. But Eat, Denver also felt that way. Like, 2022, Eat Denver, like, immediately post-COVID was, was maniac. Like, downtown Denver was fucking packed. Like, Diplo was flying. It was like, this is too much.

This Eat, Denver, you know, it was still big, but it felt a lot more subdued, which overall was a lot nicer. Yeah. It did have one celebrity. Who are you thinking? Doelipa. Oh, yeah. Yeah.

I agree with that. My signal from the people messaging me is it's still like people I know in normal trading or in tech being like, hey, like, is this real? The last time I got messages like that was when Bollagy did the like, it will 2nd G thing, whatever. Because I think that was just like so egregious. Yeah. But I, yeah, I mean, I guess it feels like the ETF is the cause.

Although the interesting thing to me that I have a funny statistic I think about. is like the total amount of inflows from the ETF is roughly the same as the total TVL of Eginlayer which is crazy. 10 billion, 10 billion. I thought 10 billion was just BlackRock. It is. Yeah, I think the full ETF complex.

I think the net inflows is like 8. Yeah, you have to take out the grace. Oh, I see. It's attracting great. The total, yeah, sorry, I was just looking at a chart of like the total cumulative. Got it. And it, come on, you have to admit

that is a crazy statistic. Yes, but... As an obvious disclosure, Igenlayer investor, the porter, whatever. You know, like, I still think that to me is the biggest statistic. Well, but with the ETFs, that's a net inflow of new capital purchasing the asset. Whereas with eigenlayers, taking an asset that's already owned and simply staking it, and it's not necessarily... But locking it up for a long time. No, Igen layer's not locking it up for really any amount of time.

You can withdraw from Igo. No, no, but I mean, like, in theory, your point distinctions. All of the reward mechanisms are if you want to maximize your points it's a long time Yeah, but it's not a net flow of capital. Yeah, I still it's just let the man chill Let the man chill. It's still interesting to think about like the net inflow of new capital and the size of this new thing are the same which is to me wild. Yeah, they're both super impressive. Yeah, yeah. I mean, I think I hear what you guys are saying about the mutedness. That being said, I think it's like we have to remember that all this is happening with interest rates at five and a half percent, right?

The last time Bitcoin was at this price, interest rates were at 25 basis points, which sort of on a risk-adjusted basis makes Bitcoin much more valuable than it was at that time. That being said, there's also been a lot of inflation since then, so technically like it would what would apology be saying? I know, that's right. Yeah, yeah, yeah, yeah. Another reason I think it's muted is one that maybe is not visible to the West, which is like all the money that left the Chinese stock market, where do you think some, some of it had to

go into crypto? Oh, for sure. There's no, like that deflation of that index. Like, yes, there was like at least 30, 40 percent those foreign money that was leaving the Chinese equities instruments. But like, you got to admit, you got to expect that the majority of that like a capital outflow had to have gone. The foreign capital outflow or the domestic. Even the domestic capital outflow. I don't think the domestic capital outflow is explaining what's happening in crypto. Like, one, I don't know how to measure it well. That's the problem. There's been enormous wealth destruction in China over the last like year. So that's, that's for sure. And also the discretionary like, you know, the discretionary income in China is super low right now. Youth

unemployment is super high. They have a lot of problems domestically. The other thing, too, you can see when prices are moving. And prices are moving the most during Wall Street hours. So basically 9 a.m. to 4 p.m. Eastern time is when crypto prices are moving. But Asia is little run-up. I mean, like, until December, there was, like, the Salana December era run-up. That had a lot of the, like, anti-perfectly anti-correlated to, like, the Chinese stock index for, like, a month straight. The pre-ETF. Yeah, I think the pre-ETF probably more so. There was, like, there was, like, a lot of, like, funny. But right now, it's very clear. Everybody you talk to says that U.S. flows are driving every day.

Right. Coinbase prices on Bitcoin is, like, $200 ahead of finance and international prices. Like, there is. Is that right? I didn't realize that. It's a, yeah, the last couple of days, it's been a huge discrepancy. Holy crap. And that's indicative of the flows. Yeah, yeah. So it does, it does feel like right now, I think the, the interesting thing is that the mania is being driven more, if you can call it a mania, it's being driven more by institutions than by retail, right?

Retail is engaged. So I, you know, I also like tracking the where coin basis is on the app store. I think as of a week ago, it was below the top 100. Now it's within the top 50 on the app store. So it's climbing. It's climbing. It's making progress. And of course,

Quimmy has been down a couple times within the last week, which is also kind of a bull market sign of like, hey, you know, things are falling over again. It hasn't happened on any of the other exchanges. So, finance hasn't fallen over,

OKX hasn't fallen over, which is also a sign, again, this is a U.S. driven thing. Well, yes, but Coinbase has always had more site reliability issues

than the other exchanges. Like historically in other bull markets, Coinbase was, you know, having site degradation every day in some cases. Right, yeah. So they say they're working on it, but we'll see maybe that's one of the things to watch in this cycle. Yeah, once Coinbase hits top 10, that's when you have your exit signal.

And that's when you... Not financial advice? Not financial advice. Well, it kind of is. Yeah, he's giving financial advice. We're not giving financial advice. Yeah, I mean, if you remember in 2021 and in 2022, Coinbase was the number one finance app on the app store basically continuously for the entire year.

And on certain days, it was number one across the entire app store. Literally more downloads than Facebook on TikTok. Having never looked at this. What are the, what is the top five right now? On the app store for finance? In the finance. In finance, number one is cash app. Number two is Venmo?

Yeah, it's like Venmo, Zelle, like Chime. Robin Hood? Robin Hood is pretty low. Robin Hood is like slightly above Coinbase. Okay. So yeah, it's mostly like pay people type apps. So that was all that stuff got pushed down by Coinbase in the 21 cycle. So that I think is a sign that you're getting to a full-on retail mania.

But the other thing is just the qualitative thing where everybody at a party is talking about tokens, like a non-cryptop party. Has that happened to you? Not yet. I mean, but it definitely happened in the bull market, right? Or like you get into Uber and the Uber driver is giving you token tips. Oh, yeah. That kind of thing is like bullmark. Well, the funny thing is that I've seen more. Well, in Denver, in Denver doesn't count. Not just Denver. Even in New York. I feel like I see you're getting that. is having like the price chart open multiple times multiple times the last two or three weeks okay

interesting what are the trading eth and whiff really oh really whiff because the last bull market at like the very top my neighbors were asking me about doge and well she she like flipped uh avax today which was crazy yeah 17 billion fdb super volatile it's like shooting up and down like 20 30 percent a day uh so okay There's the question of, are we in a bull market? If so, what stage of the bull market we're in? Fine. How does it feel as investors to be back here so soon? Because I think most of us were not guessing.

Happy we invested a lot when everyone was like crypto sucks. That's all I got to say. Yeah. Yeah. I think prices have kind of been creeping up. Obviously, people comp privates to publics, but it doesn't feel as fraught the 2031 when every Normy VC, every random person who cuts angel checks is hitting you up for allocation.

or sending you deals, those people still feel very sidelined. And so similarly, it's like, we're in April 2021 if we go by my following metric, which is when I start getting like traditional VC analysts following me on Twitter. And I start seeing like X, X, X, X, X, X at whatever, just graduated the like Y, Y, Y, Y Finance Club. And I'm like, okay. So, like, they're like the ones who are early who have to present to their IC to like make, and like, you can see that happening again.

Okay. I mean, I think in the private markets, valuations are not going up as fast as they did in prior board markets. Wait, wait, wait, right now is crazy. I don't think it's crazy yet. I think it's going to get crazy within a couple months. Yeah, do you remember at the top of the last bull market, someone with an idea would walk into a room and be like, I need a $300 million valuation and was taking seriously?

I think we're getting close to that. Where are you seeing that? Where are you seeing that? I feel like a lot of the copy pasta infrastructure stuff of like X already exists and is live and like here's the we want to be number two. Yeah. The number two with no code and like, okay, maybe some pedigree is raising at like 10% of the, you know, the public comp. I think the public comps being so crazy. Yeah.

You know, the public and private have this leading and lagging cycle thing, right, obviously. But you're seeing the private comps like make a big jump because like, Until the end of last year, they were still super low. And I think the catch up happened in January. And now it's like. I think what I'd say is I'm seeing like the A's and B's really creep up. But the seeds are actually not that crazy yet. Well, there's no growth.

There's no real growth stage companies that are raising right now. I mean, there are plenty of plays that are trying to raise like 30, 40 million. But I mean like C, like have like kind of survived the last cycle, right? Like how many people are- Survived the last cycle? No, I'm talking about mostly green-bill projects. I think the thing you're seeing that's interesting, that I don't know if a lot of listeners kind of know.

The dynamics of crypto companies has changed a lot. It's very different than the traditional venture. Like, I raise a seed round. I meet some milestones. I raise an A. I meet some milestones. I raise a B. You know, it's sort of this thing where like, it's like, I raise a seed.

Maybe I raise an A. And I get, I launch the network. I have a token generation event. And then the B or C is like a treasury sale. Right. And those treasury sales, I think, are like, very differently priced than like a B or C. And I think that's why you're kind of seeing this market

move differently than it did last cycle. I don't think there was many like large, like there was like Solana and stuff like that. But now you see much smaller projects doing treasury sales. And I think that's sort of like a difference qualitatively also. I'm not sure if that's a sign of like we haven't hit mania. But like, yeah, it's some capital market structure change. That seems relevant.

Well, it was also, Alts have really lagged this market. like I don't know, you know, multiples from their from their lows, but there's still, you know, a good chunk off their all-time highs. And so, you know, there is summary rating, but it's not as, you know, pure euphoria as 21 in that way. Where, again, looking at comps, it's like, you know, they're not hitting all times high like Bitcoin. It's true. The memes are also aren't as violent as the 2021 memes. There's no have fun staying poor or whatever means. I think we're getting there.

I do feel like we're getting there. Really? I mean, have you seen any kind of like overly aggressive? I mean, it depends on what you mean by overly aggressive. Like there is definitely a lot of dancing in the streets and sort of like, you know, I think the, especially in the meme coin space, I feel like that's where it's the most exaggerated, I think. With Bitcoin, like, Bitcoin is, I don't know, at this point, like, what is there to say?

Yeah, Bitcoin is going to win. It's going to keep winning. Yeah. But I feel like, I mean, even like the ether people or like, I don't know, the bankless guys, they're like super agro and getting out there and getting in a fight. So, you know, Mird is constantly sub-tweeting people. No, but he seems to be giving out olive branches because now his tweets are all like praising non-salana thing. That's true. That's true. I mean, maybe he's found a different strategy in the bull market that works better. We all, we all, we all together. But, but to your point, Tom, just for a sense of scale right now, so Bitcoin, so Bitcoin, so Bitcoin, so Bitcoin. So Bitcoin, 20, 20%? It's very close to all-time highs, basically, you know, a few percent off, basically, you know, a few percent off. ETHU-800 is. ETHU-800 is. $3,800.

I think it's like 10%, 10, 20%, something like that. It's close. No, more than that. No, it's 4,600. It's at 38. No, I thought it was like 45. No. Okay, well, that's more like 20%.

So it's like 20% off all time high. At this rate, it's like three days away. Guys, you're making, you're giving this type of estimation. You're giving credence to all of the people who are like crypto's astrology for me. Well, the thing is we don't, okay, I don't actually know what the all time was for either. I know, somewhere between 4 to 5K. Okay.

I know. Watch. It's a 4,500 vibe. How about that? Yeah, yeah, yeah. Exactly. But the point is that for all, it's rising 4,500, Satter and Sattern, Sattern, Sard and 4200. Okay, let me finish this point. Bitcoin is a few percent off all-time high. Ether about 20% off. Most of the alt, especially the Alt-L-1s, are down somewhere between 60 to 80% from all-time highs.

So you see big spreads. So-Long has to be leading 50%. Near is like also 40%. Near is probably 70% from all-time high. No, no, it jumped a lot today. It's up 50%. Yeah, that's high. A tie was 20. Oh, high was 20.

So it's down 75 from all time high. Yeah, so a little bit under 70. But I think it's characteristic of most of them. Wait, it was 20. I didn't realize. That was the top. That was the top for near. Yeah.

Disclosure were investors in there. Yeah. I think everyone is. Yeah, yeah, that's right. So most of the alts are quite a bit down from all time highs. But also, I mean, again, interest rates haven't been cut yet. And I suspect that the spread between the all-time highs for Bitcoin and the is going to be really tied to risk appetite.

And on an absolute scale of risk appetite, you feel pretty tight. Well, I guess, I guess what you're also claiming. naming here, though, is that a lot of, like, new capital has no leverage in it versus, like, that is also true. Which is probably why it's also less manic. Right, right. And it's coming in through Bitcoin, right? Yeah.

It takes some time for that. You're right that. I think the majority of the net inflow has been the ETFs, which is, like, not leverage. But if you look at, like, futures open interest, future funding rates, I mean, there was like 100% APR, like on Monday. Funding rates on foreign exchanges are insane. Yeah. Yeah. So, I don't know.

There's, like, still. Well, I just mean there's not, like, people, like, I remember in the, there's no GPT. trade this time. Yeah, well, I also just remember in, you know, the pure exuberant retail manual 2029, just like people posting things on Reddit, like, I just took out a home equity loan and put it all on Bitcoin, you know, like, you know, like, I don't see any, you know what I mean, I don't see any of that type of stuff. Maybe they took it out on figure, the, the blockchain. Oh, there you go. There you go, provenance blog. Provenance. The blockchain. Yeah, it's funny because it does feel like right now everyone, everyone agrees Bitcoin's going to go up.

Yeah. Like there's nobody I can find who's like, well, it's a little scary. Right, exactly, exactly. But it is also self-fulfilling prophecy. The big event hasn't happened yet. The big event being. Avening. Is that the big event?

I don't know if it's that big event. I don't think it's a big event. Here's why it doesn't really matter. Well, I'm just saying it's a big event on the calendar. It is a big event on the calendar in so far as there's not a lot of things on the calendar. I mean, the shocking event in my mind that's driving all of this is just seeing the magnitude of the ETF flows. The ETF flows at this point are 10-X.

the newly mined Bitcoin. Even after the happening, if nothing changes, that goes from 10x to 20x, but it's still, because of the ETF flows, 10x daily issuance. You know, Bitcoin always had this daily overhang where there had to be buying demand that exceeded its issuance for it to stay flat, right? It's like walking on a treadmill. Ether, a different story. But we're already at the point where the ETF flows are so meaningful that the

happening looks unimportant. You think an ETF underwriter will buy a miner? like a E2 official or like sorry um no because I could see that being like a competitive edge for them if they have why well because like they they now don't have to like net out and get give market makers huge incentives daily to like find inventory for them oh yeah but I've like kind of wondered it's like small enough I well yeah I don't know I'm just saying like I could imagine in some state of the world that it actually makes sense yeah total total M&A speculation yeah That sounds a little too

Gig- Crazy. Yeah, it sounds too many steps ahead also like the, I think that there's all this weird reg risk with holding a minor, like all the EPA stuff that's going on. Okay, then explain the diversified income G-Ga Scale Trust. Right, so Grayscale is launching a product that basically has a bunch of assets that you can stake. So they generate staking income. And so people are excited.

I think it's like Solana, Aptos, near, what else is in there? Osmo. Osmo. Really? Yeah. No kidding. Yeah, so a bunch of, a bunch of stakeable tokens. And people got very excited about this as being, hey, maybe this is someday coming to an

ETF near you. Obviously, these are way farther down on the list of assets that are probably in line to get ETF'd. So we still got to fight the ether battle in the summer. I think May 31st is the big date for whether or not the ether ETF will be approved. And then I suspect what happens is, so right now I think Polymarket is pricing it at roughly 50-50, whether the Ether ETF gets approved by May 31st. If it doesn't get approved, almost certainly lawsuit time, people are going to see the SEC.

And I think at this point, most people expect the SEC, it's more or less the same argument as Bitcoin. So the SEC will probably lose that case. But it takes, you know, whatever, four months, six months. I mean, how long was Grace Giles case? It felt like it was pretty quick, wasn't it? I don't remember. I think this is a good point, which is it's not like the SEC has changed their mind on crypto and is now super eager to approve ETFs. They are dragged kicking and screaming.

they will approve an ETF, but it's not like they're really gung-ho about improving this ETTF. Yeah, I mean, my suspicion, though, is that the situation has materially changed since the Bitcoin ATF, right? So, like you said, this is the biggest product in ETF history. So, like, the SEC, ultimately, I mean, people who work at the SEC, they want jobs after they get kicked out if there's turnover in the administration, right? And so if you are the enemy to like the biggest product in the history of ETFs, it's just like, okay, well, great, no one of you'll ever hire you. you're an asshole, fuck you. So I think the other thing they realized is that, look, we were worried about the bad

things happening. No bad things have happened, right? It has been smooth sailing. It has been totally wonderful. And so to the extent you were worried about market manipulation or whatever boogeyman they were claiming, no. Like the CME also trades ether. Whatever happened to Bitcoin should also happen to ether.

And so it seems like I would think it's more likely, especially given that, you know, the same players are at the table of Black Rock and fidelity and so on, that if that was the sign that like, hey, the Bitcoin ETF is going to get approved. It seems to me like the same should be true of the Ether ETF. But if they get sued, yeah, I mean, it might take another four or six months. Like nothing that's too rational in an argument. Maybe so, but I have to think that the fact the ETF has gone this well does change things for the SEC. Right? I mean, if something went badly, you would think, okay, the percentage should go down. So things weren't great. That should mean, the percentage goes up. You know, I would think anyway. Obviously, I'm not a suitor for the SEC.

see. But I think what happens is that ether's first in line. Once you get ether in, then you probably get Lycoin, XRP, right, things that already have, you know, more or less either XRP's case, you know, they won at least the preliminary case before there's an appeal. And then for Lycoin, nobody gives a shit. Like coin trades enough that it's definitely not security. And then you go to, okay, Solana and then Salana is a fight because they have claimed affirmatively that Salana's security, whereas they've never claimed that Ether is security. They've never claimed that Ether is security. They just never said anything about it. So I think that's probably how it goes. And if the Solanified happens, then, you know, you get next in line, next in line, next in line, next in line. And so you get Avax and near and blah, blah, blah. You know, they're probably seventh, eighth in line. For what it's worth, the Plymarket does have a Doge ETF approval market that's sitting at 10% this year. So.

For your end? Yeah. I think that is an easy sell. Well, you know, not financial advice. Not financial advice. Yeah, yeah, yeah. Interesting. Okay. So this has also been the season not just a Bitcoin, but of meme points. So we alluded a little bit to, I think Shiba flipped Avax, every single meme coin. There's Wiff, there's Doge,

there's Shiba, there's a bunch of new ones. Pepe. Pepe. Pepe. Pong. Yeah. Okay. So tell me about, what is it, Geoboden? Yeah. Okay. Explain to me the story of Geobotin because I have not been following this. I don't even know the story of Geobotin that well. I don't own it. I wish I did at this point. Okay. Does anyone, does anyone, is anyone able to recount the story for our audience of Geobodin? I don't have had the story. I don't know the story. I think, you know, it's like a lot of these meme coins. I think it's a salonin meme coin was deployed 48, 72 hours ago. It has like a wojack button. Yeah, yeah. There's like a, but it's like a pure jojacks. No, it's like a weird. So they got that going for it. But it makes sense. It's sort of the counterplaced the hedge to the Trump mega token, which has also been been ripping. And so, you know, it's. efficient market is forming around these. But the Trump one wasn't like a demented Wojack one.

The Trump one was meant to be like a serious trying to ride his coattails into meme coin frenzy. Yeah. And like almost like a credible meme coin. Geo Bowden is not only like a demented Wojack meme coin. There was also a copy of GeoBoten within like an hour of like with Brian Armstrong's misspelled. No, like there's a whole cast of mispelling. every crypto celebrity type thing. Meme coins out there.

So where's your M-Boden is taking off? Where's your Memecoin? Yeah, I don't have one. Please don't make one. Don't make it to ruin one. Bad news. But, you know, there was a Joe Biden meme coin that followed the Trump meme coin.

And it was not as successful. There was obviously a lot of people speculating on it because of- I just now realize that Gio-Boden was a misspelling of Joe Biden. Okay, that didn't click until you said it. You had to see the meme. Okay.

Yeah, all right. They were like really demented. But Joe Biden was not as successful as Trump. Right. And so at some points, people start just gambling with like, just like completely like broken memes. And like, G.O. Bowden caught on in my opinion because someone tweeted out like, oh, look at my friend's portfolio, you know, like, you know, is he going to make it? And it was like all of these demented wojack meme coins. It was like, you know, a $600 portfolio. All right. I love how you're like, I know nothing about this.

Is the history of provenance of this? Insert provenance of this. Yeah, yeah, yeah. I know this is striking how much detail you have. So this obviously went viral. I mean, I saw multiple, you know, chat rooms where people were passing around this, like, tweet, laughing at it hysterically. And like, all of a sudden, Gio Bowden started like, its rally.

Geobotin. Okay. Yeah. Yeah. So. But it's not the only meme coin. Of course. No.

I mean, all of the meme coins are. Yeah. Yeah. It's funny. There was somebody from Delphi who was telling me that, I think I tweeted something about the polymarket prediction markets on the presidency. And he was like, yeah, I mean, polymarkets for suckers. Like, you should just buy Trump coin.

You can only get a two X on polymarket. On Trump coin, you can get 100x. You know, who knows. And you can get leverage. You know, I'm sure there's a lending market out there. And you can get leverage. It takes a Trump coin as collateral. Yeah, obviously, obviously.

So there is, there is. And the problem is that I agreed with him. It's like, you were absolutely right. It's only worth like 400 million today. So, you know, like, guys are living. Antifes have also been creeping up. Is that right? Yeah, yeah, very slowly though.

I still own one. Well, I have to check. There's too much supply. Yeah, that's probably. There's more supply. Yeah, so there's a, I guess, connected philosophy that people have been talking about, which is called financial nihilism. And financial nihilism is this idea that basically claims that when people lose any sense

of connection between financial markets and real world. outcomes. They start to think that, like, actually this idea that financial assets should be priced based on their intrinsic value is bullshit. Everything is in the eye of the beholder, and therefore, you know, it's kind of like the Wall Street Bet's view of the world as like, you know, whatever, you hedge funds are playing this stupid, boring state game. If we don't play that game, then financial markets are just a different thing. And so, so it is with Geoboden, you know, we all decide Geobotan's hilarious. And so it, you know, ends up going up whatever a cagillion percent. How much do you think that's going to define

this rally or this stage of crypto. Well, I thought it was funny. I was reading a post by a renowned crypto hater. Hater might be the wrong word. Just someone who, like, I feel like has disliked it for a long time. But then like every once in a while will like change his mind a little bit and then go back to disliking it. Who is this?

Just Joe Wisenthal from Bloomberg. I don't think he's a hater. No, he's not a hater. He's not a hater. He's a cautious appreciator. He's a realist troll, you know? He asks real questions. That's true.

Yeah. He's like the trad-fied war. Yeah, more of Bitcoin Maxi, but yeah, yeah. That's why I was like a hesitancy hate her. It's like something, it's like some, but like there is a little bit of like over-scepticism of the technology side, which, you know, whatever. Everyone has their own preferences. Yeah. But he was like, I love this financial nihilism.

He wrote some posts. It's like, first time that this industry has been real about it. And he has this long blog. Was it a blog post or was it just a thread? But yeah, he's like a long post about this. So I feel like there is some interesting thing to that. I think there is something to it. I mean, it's like especially like with NFTs.

It's just sort of like the, you know, deconstruction of this idea of like. Well, I also think the meme coins, do they pass the Howie test? Like literally no one has an expectation of profit. Well, none of the money is going into the development of anything. anything like yeah there's no like shared fees yeah yeah like nothing is getting built period like you're not like there is no reliance well but there was like wasn't there like some doge metaphors ship ship made a l2 yeah yeah yeah and they like amm oh yeah yeah yeah and doge queen was talking about doing some bridge or something weren't they there are dogenals there is like a ordnals

ecosystem on on on doge so okay got that going for you yeah yeah yeah so i mean doge coin has devs right so like I mean for some of these meme coins more than others I think it yeah we We also haven't seen that can devs do something memes yet, like the equivalent. Yeah, this goes down. Yeah, right, right, right. But that's like, you know, you start to see that. That's when you know, like, you're getting close to the precipice. But I think, but I think we've seen this total disconnect, right?

In the last market, it was like, oh, the devs need to, like, create a metaverse. And now it's like, please don't do anything. Just like, let it be a meme. The more that you try to accrete technology onto it, the more people will actually try to understand what its value is. Right, because of the financial nihilism of just like, you know. I guess the astrology thing is. I mean, it also speaks to AI tokens, right?

Like the more legible the AI project is. Yeah. Well, I mean, that is arguable. I think on AI tokens, I think it's pretty clear that the less clear what you're building is, the higher evaluation the market can ascribe to it. Although arguably, BitTensor is, despite it doing something more or less not useful, it does work and you can run a validator and a miner. Is that right?

Which is like, which is a lot of steps ahead of most AI coins. So that's bearish. Well, I'm just saying, no, but they're the highest market cap. Yeah, you can't do inference on it. Like, you know, it does some stuff. It literally has a transaction type for taking a gradient step. Let me tell you how inefficient that is. Yeah, it technically works.

It like tech, yeah, but it's like 10 to the nine times slower than like every. It's like kind of idiotic. It's like almost like, why would I run a neural net on Solana or Ethereum? But like at the same time, like it does actually exist. It does exist. Unlike, unlike, that's a bad thing. Like, the fact that you can discuss credibly, like, how bad the efficiency is of actually using it is. But that's why it's the largest cap window is first.

Yeah. Yeah. Being first was what mattered. Yeah, but you can objectively prove that it's pointless, like looking at the specifications of it. Or just like looking at their block explorer. I look at the explorer. I'm like, oh, my God, this is like very useless. Yeah.

I mean, it's, it's like my, my, my, my, my face ID. chip on my phone is more efficient in the six seconds than like, you know, the entire blockchain. Well, anything, anything, you know, in terms of computing, anything that's distributed has this, like, for sure, for sure, for sure. But I feel like, I feel like the, the AI stuff is like kind of funny because everyone who goes into it is coming from like AI land and like has no appreciation of like adversarial networks means you have to slow down bandwidth, latency, whatever. And then they like try to build something like, ah, shit, we can't do it. But like, we need the token. So I guess we're going to like, I think it's the way around. I think most of the people building in

crypto AI are not AI people. I think it's mostly no, no, no, no. But I think some of the some of that, they hire people who are transplants. Sure. Like they raise money. Then they hire people who work at. Yeah. And then those people are like, oh, we're just going to like bolt on all the shit and like run it centralized. And then they're like, oh, shit, we can't actually do this. Yeah. Decentralize. And then you get kind of either Rube Goldberg machine, like some, some of them, which are not live and honestly you read their docs. You're like, I think this is a Fendi scam. And then you have the other ones who are like,

BitTenzer where they act. I mean, they've been around for a while too, right? Yeah. And they actually did essentially do something. It's just not something. It's not very useful. But they became the meme coin first. Yes, they did.

Because they tried. Okay, so let's do a little deep time on crypto AI. Because I think it's very timely because it's one of the things besides Bitcoin and meme coins that is capturing the popular magic So with crypto at first, have you guys made an investment in crypto AI? Ritual. You did ritual. Okay.

Are you comfortable talking about ritual? Sure. Okay. So there's a few categories I think that now have kind of reified as the main things within cryptoEi. So the first and probably the oldest is GPU marketplaces. Yeah. Render Akash.

Back in the day it was Golem, which did really purely rendering. And now it's like GPU jobs. You can just, you know, SIS. Actually, you could argue live peer is also in this category. in this category. Oh, kind of, but it's not, it's not AI at this point, right? But arguably it's doing the same hardware aggregation.

Yeah, same, same core idea, but they get a shudy evaluation because no idea. Yeah, but like, okay, I don't think any like distributed compute slash distributed hardware is necessarily AI, right? Yes, but it's getting the AI. No, no, I think in the live peer case, it's like they're still buying A100s, you know,

they were still hardware, they had the same provisioning. They're buying A100s for video transcoding? Well, I think they actually allowed you to do like, I mean, remember, think about this way. When you're taking Kuda and you're compiling into a new chip, you're going to run faster. And if you're a minor on this thing and like the faster you respond, the faster you get your rewards, you would buy the new chip. Oh, it's just like a block reward thing.

It's not even that they care about the... I can't imagine that any supply finds the best way to monetize going to live here. I think every single one is... I think my point there was just more that like you could run this. You could use the remnant. You could. Yeah, yeah. But I think A100 can just get better rewards elsewhere.

I would assume. I'd be surprised if that were not true. Okay, so category number one is GP rental. And that I think is a thing. And like you're seeing Akash is doing okay. You know, render is doing okay. Well, I mean, I will say one thing. Well, something we invested in that's a centralized company that does GP rental is doing much better than any of the.

Yeah, in absolute terms, but not in token. All the, all the non-decentralized ones have like way better product markets. What did you guys invest in to? Together. Oh, together. Oh, yeah. But that's like, isn't that managed? But that's what I mean by centralized.

Right, but you don't get the GPU, right? They just run model. Yeah, yeah, yeah. But like, arguably all of the decentralized ones, the APIs, they're starting to give people look like that. So like you don't have to know where they're running it. Oh, I see. Because like people don't want to know that they have to stake the token and whatever, right?

I see. They basically are making like an intent looking like RFQ style thing. Give me dollars and I'll give you, you know. Yeah, I'll give you the element. But hey, that looks like the ones who are managed. Right, right. Yeah. Yeah.

Yeah. Yeah. I mean, it's clearly the case that very few people are using these decentralized GPU networks relative to the centralized ones. All the inventory, especially of the high-end chips, are held by centralized cloud providers. Also, the networking constraints for executing a lot of these things are like, why would I do it over the internet? Right, especially if you need multiple GPU simultaneously, then you're kind of fucked.

So, okay, so that's category one is GP rental. Category two is decentralized inference. So this is, you know, a little bit different. GPU rentals, like you get the whole GPU. Decentralized inference is just, you know, I have this model, I want to run this model, I don't want anyone to censor me, I don't want anyone to control me whether I can use this model. Maybe I have a query that normally would not be allowed on a centralized provider, whatever it is, right? Or I have an anime model and this anime stuff is, you know, trained on unlicensed data that, whatever.

Insert your use case here. So that is ritual. Ritual is decentralized inference, correct? So talk us through why you're bullish on decentralized. inference. Yeah, so I guess there are a couple things. The first thing is that, you know, I think people kind of want some notion of understanding of on-chain data that is usable by contract on-chain. And this has kind of been the Holy Dream Grail for a long time, right? Like everything from the graph of like, I have indexing that's off-chain, but I prove some portion

of it on-chain to the ZK versions like Axiom where I take previous state and I can prove that like, hey, this was what happened in a theorem historically, versus having to like index manually and like be a centralized index provider, right? Like the dunes or Nansen's and stuff, which I think you'll always have, right? They're always going to be way faster. And so there's a reason for that. But for some use cases where contracts actually need some either historical queries or models processed on historical queries to make a decision about themselves, you're kind of giving them an ability to look in their past. Right now, if you look at how blockchain state works,

either I increase the cost to the user by storing much more historical state or I'm sort of autonomous in a time one sense. Like, I only use the current state. I can't really look backwards. But like in reality, a lot of applications, most applications, you're right, need some notion of historical state.

Yeah. And I think all of these things are giving you different ways of getting at that. And it's more a question of like, how do you value it? Like, how much security do you want for that? Yeah, yeah. So it sounds like what you're describing is that you can create an AI,

have it interact in real time on blockchain-based data, and basically say, oh, looking at this data, basically non-chain a gauntlet, more or less, right? Yeah, yeah. I think that in the long run, that's, like, kind of where you want to go. So I would actually distinguish that as being verifiable inference rather than decentralized inference. Sure.

I mean, I think the hard part is you have to make sure the verification stuff can happen, right? And you have two ways of doing that. One is cryptographic, so like ZK, is going to be very slow. Like right now, that's kind of, you know, I can do a single decision tree. I can do a couple boosted trees.

It's not like a real, you can't really do like high capacity stuff. Yeah. The second thing is these like economically secured things where multiple people run it and then they come to consensus and that's more in the... This is supposedly what BitTensor does, right? Right. Which is decentralized inference. Right, right.

But BitTensor seems to be doing, you know, like almost step. by step verification of the inference versus like an aggregation. So it's just, you know, you look at the blocks and it's filled with like each step of the thing instead of just like, oh, I verified one thing. So I think that the idea is to like really find a way of like figuring how to batch these things and make it easier. Now, the hard part I think for using it for something that's like really high value like defy is that like you're kind of creating a bunch of MV, right? People can see exactly what you're like in order for the verification to work. The entire thing

is open source. So why can't I run it locally and front run your transaction? Yeah. You kind of create all this. So we're not still at the point that that's settled. But I think for these lower frequency use cases, like the farcaster, social media type of stuff, or like sort of some of the games. Did you have like a commitment to the model weights and just say, look, here's a hash of the model, here's what we're running, and like the nodes don't disclose what the model is? Well, how do they get the initial set? And how do I verify that that commitment?

And like you could make something where they're constantly committing and revealing. Yeah, yeah, yeah. But now you're just adding in another handle. It ends up being like kind of basically then multi-sigging it. Yeah. So I think we're not quite at the stage. Like I think like once things like Eigen layer are live and you can use that for bootstrapping these networks quickly for these economic rules, like flashing people when they don't reveal,

like for instance, then it's a lot easier. Right. But I sort of think like the long-term view of the decentralized information. is just like applications that are already decentralized that don't want to use an Oracle for historical data. This is a way of doing it. But why the inference part? Like you can easily do verifiable historical state.

Yeah. I think a lot of it just boils down to like the storage size and amount of data you want transfer, right? Like what am I using the query for? Like I'm going to pass it into some function on chain and like, okay, well, now I'm charging the user gas for that. Why can I make the off-chain network do that and kind of split the cost with them. I think it comes down to a cost basis thing more than a, like you could do the

inference or the computation from the queries on chain. But now you're just bloating your on-chain contract. So I feel like if you're already outsourcing some of that, then you might as well just outsource the whole thing and only get the answer. So the last category is decentralized, or sorry, verifiable training, which I think is just batch it crazy. That is the one where I read the docs and I'm convinced that the people who do it have either, never run a cluster and tried to keep it alive and up and done the DevOps. B, they've never actually trained a model on like a realistic size of data. They like, they like did the unit tests and like, oh, I trained it on five gigabytes of MNIS data. Okay, great. Let's like decentra's.

Like, that's not going to get your model anyone wants to use. Yeah. And the third thing is they, a lot of them were being like, okay, well, like actually we'll give you privacy. And it's like, there's no real. None of the privacy stuff is actually made for doing this. scaling that's extremely hard. And I feel like it's like the worst of both worlds. It's like it's taking the worst solution for each of these, dumping it together and saying like here's a token. Yeah.

It's a junk bond. Yeah, I think the decentralized training slash verifiable training, like it's just a, it's a solution looking for a problem. Like I have never heard anybody ever tell me like, you know, the big problem in my life is that I need to decentralize training this model. Right. Training should be centralized. It should be like the most professional and opinionated.

And it's the most expensive part of generating a model, right? Like if you balloon that by even 2x. Right. And some of these are claiming, oh, only 2x increase in cost for training. That's a massive, massive upfront investment. What's the advantage of I'm getting the exact same output? Yeah, I mean, I think that, exactly. The whole point is that you get the same output as training in a centralized way.

I think that the claim they would make is like, well, we can prove what data was actually gone into training this model. And so if you want to prove that like this model didn't use any, didn't use the photos from customers or whatever. It's an idea of like crowdsourcing funding for the model, right? Like, okay, we can actually have a model with what it's that are like community own versus you can contribute your your own data to this model to the. I was going to say, what do you think of decentralized data labeling?

I think it's fine. It just feels like scale. Yeah, basically decentralized scale AI. Yeah. I mean, I think that's fine. It just feels to me like there's one business that you are displacing, which is like, what does scale like, 10 billion? Yeah. And they've actually kind of pivoted more towards doing like fine-tuning.

Like if you go to their website now, it's like way less about explicit. Right, right. They actually run infrastructure now themselves. Right. Yeah. Right. Yeah. So they basically do like RLHF for, you know, all the big models.

And like that's that it's a very different kind of business. It's not something you can easily do decentralized because it's all about quality. And even a small amount of, of, hey, we kind of downshift it, or we have this, like, challenged and, you know, verifiers dilemma type game is going to end up just lowering the quality of the model. Yeah.

For how much people invest into training. I think another thing people have realized about Decentrius training, I remember, like, exuberance of the bull market. I heard a lot of ideas, like, oh, we're just going to, like, have a full ZK proof of the training, which is ridiculous. But then there were people who were doing,

like, statistical proofs of training, which are like, hey, I'm going to challenge you to give me some fraction of your weights and like you have to prove to me that like, there's the initial step, here's the end step, and I sample random. And like, you can kind of do these bisection game type of things,

but those things are like interactive. How much bandwidth are you going to put on this network, right? Like I'm not only am I saying like you have to do all this compute. I'm saying you have to stay up and like you're extremely deducible. It's like all the things you don't want in a training procedure are happening. You're like asking for it. True bit was right. They were just ahead of their time.

Yeah. It's true. Yeah. So, okay, that was a very brief crypto AI overview. There's one more topic. What do you, what do you guys feel? Are you excited about any? So I mean, there's been, there's been a lot of funding in crypto AI lately. It seems like most of it is either going to decentralized inference or GPU marketplaces. There was just Ionet that just announced, I think, billion dollar valuation that they raised a bunch of money at. I, my view on this is more or less that I think there is stuff that's investable here, but I think decentralized inference in GPU marketplaces. I think GPU marketplace is, unequivocally they're a thing. I don't know how big they will be because ultimately they're

always going to be playing at a disadvantage relative to the centralized players who just can get the economies of scale. They can get better inventory. They can like route things more efficiently than the decentralized players. Like if you look at vast AI, like so when everybody compares like the decentralized marketplaces like a kosh and render to Amazon, Amazon like is three times the cost, but they never put it next to Vast AI, which is like the centralized version of the exact same thing, which always is like 5% cheaper than the decentralized versions. And so my view here is that like, I just don't know. I think what crypto excels is where there's a lot of censorship, is where people don't

want you to trade GPUs or do whatever. I don't think there's anybody getting in the way of you wanting to SSHing into somebody as GPU and paying by the out. To be fair, though, I will say for a lot of the centralized training stuff, a lot of the revenue does come from things that people like opening eye or anthropic don't let you do. Centralized training or no, no, even the centralized training that's like fine tuning. You mean fine tuning like a chat, opening I model. Yeah, like there's a lot of stuff they kind of have restrictions.

Yeah, yeah, for sure, for sure. And like that's where the thing is like almost all the spend in AI right now is coming from large companies. Yeah. I mean, they're just doing much, much larger training runs than the hobbyists who are creating the wifu models or the, you know, the anime models. There obviously is a long tail of that stuff. stuff. But the reality is like, yeah, but if you're using VasiI, some random guy in Russia is renting

you his GPU, he doesn't give a shit what you're doing. So I think the reality is that, yes, you know, if you're on opening AI and they see that you're shoveling in like, you know, thousands of frames of Studio Ghibli movies, they're going to be like, yo, what, you know, you obviously don't have the rights to this. But I think most of the places where you can just rent a GPU, even on Amazon, you go rent a GPU, like they're not going to check what you're doing. Yeah. And I think this is one reason inference is a little bit more. easy to stomach because it's like, okay, well, it's not, the value is not actually in the persistent perpetual data, right? Like in training, I'm like making this data set of the

weights. Yeah. And then it has to continue to live forever. The value is not in the making of it. The value is just the just in time compute that is, you know, my chain doesn't want to do. I mean, how do you think about remote inference the versus edge compute though? Like the other trend has obviously been that like these open source models getting really good and just running locally on your device. And it just seems like that's going to continue. I think the client side split stuff of like I have a big model that it gives enough of a hint to a small model so the small model can run locally. That's definitely going to be a big deal. But I still think it's going to be higher compute than like most blockchains. Like I think blockchains are still going to have trouble

running those. I suspect where blockchains are going to thrive is the places where there is real censorship. And I think that also looks pretty ugly at scale. or like at equilibrium, right? Where like, I think, I think like the, you know, the porn models, the unlicensed models, I think that's probably where decentralized networks are going to really have a big advantage. I just don't know how big those are going to be at scale.

Yeah, for, yeah, for sure, for sure. But I just feel like it's much more believable they can exist versus the training ones. It's like, yeah, no, look, that I definitely agree with. Like, it's not even that, like, if someone censors in training, I don't even have a very good way of verifying it, at least in the inference of fixed, all the inputs and it's really only about the execution.

Right, right. Yeah. I think the reality is that, you know, open source AI and home brew AI are huge and they're going to remain huge and anybody who's like, I think people are conflating the two and they're saying, oh, if you're long open source and home brew, you should be long. So I have a thought experiment I always think about with this stuff, which is if you think about like 1989, like Linux and Windows were both kind of similar install base, maybe a little before 87 or something.

And then obviously Windows had this huge burst and Linux was kind of slow. But then all of a sudden all these like data center operators, like the rack spaces of the world started realizing like, oh my God, 90% of our revenue is going to pay Microsoft licenses. We got to like lower that. And they kind of started having Linux. And then eventually once the web 2.0 companies, I guess, existed, then they started eating up all that space and they needed just like tons of servers.

and it became so much more cost effective that Linux was sort of at data centers. But like, would you rather have invested in Rackspace or Google? Like, there's kind of this problem of like right now everyone you can invest in is like closer to the data center operator, which totally could also be a great investment. They're not going to be as crazy outcomes. Yeah, I mean, that's absolutely right,

is that the application layer is being more or less ignored at the juncture of crypto AI right now. I think in many ways, so if you remember during the dot-com boom, Cisco was one of the highest performing companies because Cisco was more or less an index of the internet. And you didn't have to take a directional view on any particular application winning because people didn't know what the internet was going to look like. But there was one Cisco. There was one Cisco.

There was one Cisco and everyone knew Cisco. Yeah, that's right. That is true. That is true. But I guess the way that I see it, like, yes, the network infrastructure is an easier bet to make because you don't have to take a view on what's going to win on the application side. But the other thing about it, which is also, again, what concerns me is that it's less legible.

If you see an application, you can see application metrics. If you see infrastructure, who knows, you know, maybe someday all decent an inference and da-da-da-da-da-da. And so that is what concerns me about the fact that the bid is going all to these infrastructure companies. I'm totally concerned by it, too. I mean, the fact is there's not going to be one Cisco. There's 10 new ones of these like three like month. Yeah. And none of them have anything in terms of like a supply.

advantage or remote or a... There's way too many picks and shovels. Right. And there's not nearly enough people holding those picks and shovels. Right. Right. And, you know, I think it's too early. Like, one, to know what any of the infrastructure, you know, will be successful.

Right. Most of it probably will be obsolete by whatever comes out in a year anyway. Yeah. The next company gets founded. And second, there's not enough things building on the infrastructure. We're all just speculating on infrastructure is going to be important, period, full stop. And there's almost nothing else really to invest in at this moment. And then there's WorldCoin, which at this point is more or less a meme.

Well, WorldCoin is the best crypto AI usage in the sense of like, I mean, I don't, I don't sort of like agree with it mainly because I think their token is so poorly designed. But the, I think they are that objectively the only ones who are actually like generating a running inference, generating a proof, posting it on chain and using it as a. It is really application. What's the inference? I mean, the scan stores a model that takes features of your... That's just like an image hack, right? That's not inference. No, no, but they actually use, they do, they run models on that, on the input to that.

They're just proprietary models that they don't... I mean, the reason they do all the ZKML stuff is because of this. Like, they actually care about moving in that direction. Wait, it's ZKML, but they're not revealing what the models are? No, no, that, well, that part, I think, supposed to come at some point. Okay. Orb two or whatever, because I'm pretty sure orb one did not have this. Okay, okay.

No, no. Okay. There's a new thing. There's a new thing. What I'm saying is like they're the only ones honestly working on an application where they're like dog fooding their own thing. Right. But ironically, it's like opaque and there's hardware that only one hardware manufacturer

can create that has all this software on there that nobody can actually verify because it's not open source. But I mean, it's not that different than video. Well, no, it's comparing them to an video, right? I mean that they're uniting these threads. There was actually a, Elon's been making these memes about. closed AI. Yes, yes, yes. And there was a closed AI meme token that launched this morning. Of course there is. So yeah. Of course that I don't invest in closed AI. Yeah, yeah.

Yeah, I think I think okay, so maybe they're doing some ZK something somewhere. Well, no, no. I think they're actually really honestly in earnest trying to add it and use it to make their product better to make the identity. I would put World Coin in earnest in the same sentence. I'm just saying like I think the develop, no, no, ignore what the BD people are doing. Like, I'm just saying, the developers at WorldCoin are actually. Look, I know they have brilliant people on their research. Who are, who are, totally agree.

Who I think are earnestly using this stuff. They're the only ones I think you can argue are earnestly. Okay, but that's not why. That's not why WorldCoin has the valuation that doesn't exist. Correct. The reason WorldCoin has the valuation that it does is because it has the most microscopic float relative to the total supply imaginable. Yes, so like 1%.

This is where I was saying the token is they should. Yeah, okay. So not only is there 1% float on a token that has almost no price discovery, but then second, what's been very clear is that it is kind of the Sam Altman meme coin. So when Sam Altman posts something, when OpenAI, you know, SORA came out, the video model, the World Coin ripped. More or less, I mean, it's kind of self-conscious because you see traders on Twitter basically saying, look, there's no way to invest in Open AI except Microsoft, which owns too much other stuff.

So the best way to get exposure to the Sam Alvin meme is to buy WorldCoin. And so more or less, like, the market agrees that WorldCoin is the same all of the point. Yeah, I'm just saying, I'm like ignoring all of that. Ignore the token. Let's ignore all of that. People are just bullish on swap revenues from the. As someone who, as someone who kind of was abhorrently never going to invest in it, I certainly have a lot of like other qualms about the go-to market that they're doing.

Right, right. But I will say that the things they're trying to do to actually keep, they care about the verifiability component because they want to basically do transformations of like the images they're taking of say like the eye and prove to you that they did those correctly for someone else to use without, you know, revealing things. And they actually have a real reason to do that. Okay, yeah.

Maybe eventually that will be worthy of congratulations. But right now, like I said, like everything is close to ours. They don't reveal anything. I'm just saying, like, I can't think of anyone else who's a good application at all. Well, that's the fundamental problem. Like, is that the ratio of infrastructure to application is, like, absurd. Yeah. And right now, like, it's basically infrastructure that's crying out for an application, right?

I mean, more or less, the core thing of WorldCoyne is that it's an ID system. And nobody uses the ID system right now. They're just kind of air dropping it on people in the third world. There was a story. So apparently WorldCoin recently got banned in Spain. And I was reading apparently some people, I think they were either in Spain or in Latam, who were saying that apparently there are all these shops popping up that will create posters of your iris, like these really, like awesome laminated posters you can put in your house. Because, and they do it for super cheap because they want to take your iris and then go get a whirlkoin air drop. And they're basically clipping the spread on like whatever you pay them and then the value of the air.

Wait, they take it to a foreign country? No, no, no, no. They do it there. and then they take the super close-up scan of your iris and then scan it in a WorldCoin orb without telling you. Well, why do they need to go? Because aren't there operators of the orb

and you have to put up some money to be an operator? I'm sure they can bribe. I think you could just buy. Yeah, exactly. I feel like that's the cheaper way of doing. Yeah, yeah, yeah. Well, but you need somebody's iris to actually get it. No, no, no, I get that,

but I just meant like it seems cheaper than like, just to like collude with one of the operators. So what happens when a Spanish person who's gotten a high-risk focus? of their eye taken once to sign up for WorldCoin in like 10 years. Well, I'll say they're already registered. Well, yeah, it'll say they're registered, but then also like they won't get an air drop, right? So the air drop is emptied out. So it's like kind of prehacking.

Right, prehacking, yeah. So anyway, it's, I don't know, weird situation, but I guess Bollman. Again, again, I was, I was ignoring all of the... Yeah, but why would you ignore all of it? Like, I'm just saying, I'm just saying from a technology standpoint, a tested sensor. You just said it for the tech? Yes, but I don't want the fate, I don't want the fate of that guy who was a isn't that meme. All right. All right.

Anyway, so I think surveying the landscape of crypto AI, it does feel like right now, obviously a lot of hype. I think the possibility for some real substance, and we're seeing kind of these first Do you see any applications? So we invested in one application, the sort of chatbot character AI type thing that I think is not public yet, so we can't reveal it. And we're looking at some other stuff. I think right now, like, look, I more or less assume that there's

going to be a lot more interesting stuff within the next few years. I think this first gen is kind of just, you know, very surface level, like, okay, I see this Web 2 company. Let's make a Web 3 version. Yeah, it does feel like the Uber on the blockchain. Exactly. It feels very Uber on the blockchain stage for crypto AI. And I do think some of these things will exist, but it's a little bit like, you know, I don't know, like storage, a storage A where it's like, okay, this is a plausible business. It's not going to be the biggest business of the world. It's not going to take over Amazon S3 or even Glacier or anything, but it's a thing. It's a thing that has some demand. Yeah, I always think about this like, like, you know, Gellum and like tribute and stuff, right?

It's like, what was the actual verifiable compute that we want? It's like EVM verifiable compute. It's roll-ups. That was actually like the verifiable compute market. And then what you want in terms of verifiable storage, it's, you know, fucking blobs. It's 4844. It's Ethereum. Like, you know, it's, it's right. Like, people have the right idea, but it rhymes. It's not like a one-to-one copy. And so, yeah, I don't know what that looks like for AI. Maybe it's it's smart contracts, creating models to get historical state. But I just think the historical state thing makes a lot of sense, right? Because like right now, all these contracts are paying oracles a lot.

Yeah. In fees to give them basically aggregate. It is cool, but it is kind of niche, right? Like, I don't think that the amount of protocols that require that level of value for sure, for sure, for sure. It's probably less than 20, I would guess on Ethereum Maynet. Yeah. I just think of like your addressable customer base for Gauntlet.

I mean, it's probably more like 50 to 100. But yeah, it's not like, it's not like, it's not like,000. Yeah, but ultimately, like, this will be a commodity service, right? So it's not as well-Ave is going to pay millions of dollars. They have this ad. There's this adverse selection cost problem. So you kind of, the part I would say that kneecaps the market is it needs to actually be used on not the most valuable thing.

It has to be used on the thing that like the cost of certain attacks is like cheap enough that it. Well, it also potentially opens up the design surface by. Yeah. And like two orders of magnitude, right? Sure. So like, when you start to get to the idea that instead of needing the data on chain, Or, to be honest, even the calculation or compute on-chain at all, you can suddenly design things that we don't really have in any Defypiret Protocol.

Like, theoretically, if the only thing that's on-chain is the actual, like, store of value in the assets of itself. And everything else is off-chain, but there's a verifiable process by which you know that the off-chain system is working correctly. you could theoretically build almost infinitely complex, infinitely fast systems that just use the blockchain assets for settlement. And that's really cool. And I don't think anyone's really spent much time in this like. I feel like the fully private defy type stuff, like the penumbras of the world, they of course already do this. But, you know, there's overhead for that. So, but yeah, there's something in the middle that's like somewhat off chain slash private and somewhat public and like merging.

But I don't, I just don't, I think like right now you're seeing kind of like all the, you know, the bell curve meme, all the midwit ideas that are just like, oh, copy pasta X and like it'll work. And it's like there's no like thinking a little bit about like why this might not work right now. Right. There is something to kind of on the outskirts of AI that I know I've heard you talk about many times, which is, you know, the importance of like a tested data. So the White House recently announced, I think a couple of weeks ago that they were going to start cryptographically signing all the content that they put out. So every speech, every video, every audio clip. And I think right now the White House is clearly ahead of the curve because I haven't seen this for most governments and certainly not from most companies is just in an age where video models are rapidly getting better. And already we have synthetic audio that is fantastic.

This is a great idea. They should be doing. Yeah, they should be doing this, right? And then there's a way that you can... Well, where you post those? Yeah, exactly. Where can we all post them in one commonplace? Yeah.

So it does feel like there will be something along those lines at some point. But it is in some sense a coordination problem more than a technology problem, right? The cryptographically signing stuff and hosting it, hosting where your keys are on your server and your certificate, this problem has been solved for many, many years. The coordination problem is what's not solved. And I think it requires more than just like, hey, I made a blockchain for all attested data. It's more like, hey, you need somebody to actually create the shelling point in the consensus, which requires a great founder and like a lot of work in order to make that convergence

into a single platform. But there will be something like that. I'm going to say the White House put announcement that they they hate memory of memory unsafe. Oh, I saw. Yeah, sorry, do you want to recount this? Yeah, I don't.

I mean, the White House just put an executive order about how like, or not executive order, like a post about how like memory unsafe programming languages are the source of all hacks and like we need memory safe programming languages. And of course, the crypto rust developer community went crazy. you know, I think at the end of the day, the people in crypto are a little bit myopic of one fact, which is like because crypto has to restart building everything from scratch, like, you're building basically the operating system from scratch in like these virtual machines

are building like all the libraries from scratch. You're fine using a programming language that's new. But 90% of the world is like too afraid to like jump even like make those like little jumps. You know, you see some little things, right? at like Microsoft or Google where they'll be like, oh, like our kernel team for our custom kernel on our data center is going to use Rust. It's like, okay, great.

But like I, you know, I think the, it's going to take a while for that, the memory safe stuff to disperse. So I thought that, but it was like funny that they actually cared about that. It was very sweet. It's like a lot of the other things they've done were kind of technologically incompetent, like the AI thing that was like.

Capping the model way. I mean, most of their crypto policy is also incompetent. Yeah. So I was like shocked to read something that was like legible, but. It's like, it's like your dad took a CS class on the weekends and he's like, hey, guys, I think it's time. Memory, you guys about memory safety.

Yeah, yeah. I heard from my professor that you guys shouldn't you see. Yeah. I mean, I mean, I just think like, you know, maybe you're right that this is how you, you pill the governments into thinking about blockchains of like, oh, actually, this is just a way for us to store this and we don't need to worry about any actors or not. The biggest problem though is whatever chain they use, if it is one of the existing blockchains, the second that they post onto, let's just say, Ethereum or Solana, or whatever the chain is,

you just know the response from the community. You're going to be a Geoboden role? Yeah, yeah, for sure. Yeah, what's the meme coin for this? Yeah. Good question. No, but you're probably right that it will immediately get filled with spam and transaction fees will shoot through the roof and there will be like all sorts of weird shit.

try to DDoS like the next, you know, authenticated message. Yeah. That's why we can't have anything nice. Like, we can't have anything nice because you know that they're going to like genuinely try to like have like a really good. That's why it's probably going to be like an enterprise chain. Right. Although theoretically they don't need any chain at all to cryptographically sign a message.

No, no, but don't tell them that. No, no, they have to store. They have to have persistent storage somewhere. And they're exactly. And it should be decentralized. No, they don't. Fuck off. Unbelievable.

You don't need any. How do you sell it? How do you sell us out like this? You don't need any storage at all of it. No, that's absolutely not true. He's lying. You need a blockchain. Call us.

Call us. Blockchain mandatory. No, no, no. I mean, I get this idea that like, especially in ZK land, a lot of people are like, oh, we don't need the whole chain. We can just have this like storage. The problem is like the input in order to verify, I still need some of the input. And I need some, some kind of agreement on some state.

And it's like, that's the stuff that's most likely to get censored, to your point. And I think, you know, in this case, a government might view itself as like no one can censor me. But I feel like for this type of stuff, especially given all the hacks and like how they're complaining about their inability to keep things safe, wouldn't they want the thing that's battle tested and is constantly being battle tested? I think U.S. government should start an AVAC subnet. I'd say that's the solution to their problems. Not financial advice. Well, to the government, that's just voting. That's not financial advice, right?

Anyway, technically we have a constitutional right to do that. Cool. Okay. So the other big story in the last week has been Unitoken. So Uniswap recently had a governance proposal from the Uniswap Foundation that proposed to finally turn on the fee switch. So for those of you don't know, Uniswap is an AMM. It's a biggest on-chain decks.

And Uniswap has never actually taken a fee. There is a fee on Uniswap, of course, but it goes to the people who are providing liquidity in Uniswap. The protocol itself, or UniToken. or Uni Dow gets nothing. So the assumption is more or less been that this is for regulatory reasons. Maybe this makes you a swap security. Maybe there's some other reason why they've been avoiding fighting this fight.

And people assumed that the community was going to end up having to drive this and the foundation was going to have to be hands off. And what we've seen is the opposite, which really took a lot of people by surprise, which is that the foundation was the one that proposed to turn on the fee switch. And now it seems like it's likely to pass the flying colors. everybody's very excited about this and all of Defi has pumped with unity leading the pack on this thought that do they know something we don't know? Is there a reason why they think that now is the time that they can go in and turn on the

fee switch without concerns about being targeted by the SEC or something or other? So first of all, Defi founder, any alpha or anything? What do you think is driving this? Here's my take and I don't have any alpha. It's not really just turning on the fee switch. Actually, more than that, it's fees of which is one piece of this. The second piece, which I think is actually more important, is that the fees will be diverted to uni stakers who not just stake, but also delegate their votes into the governance process.

And so you have to assume that the ones who are actually delegating in governance is a significant subset of the overall token holders. And the idea is that the fee switch will be diverted into those who are staking in participating in government. So, you know, I don't know what percent of uni is delegated as of today. Whatever the number is, it will be higher in my expectation after this system is rolled out. But let's just assume that there's 20 percent of uni tokens that are actually staked that are not sitting on exchanges, sitting in cold storage, sitting, you know, in a format that's not participating. I've seen estimates that the fee switch will lead to somewhere between $60 to $150 million

of fees, given current prices, the current market. $60 to $150 million directed to maybe 20% of token holders is per year, is incredibly significant. And it transforms Uni the protocol and the token from something that the token today is used for votes, but those votes are really only on the treasury. Like it doesn't seem that meaningful. it transforms it from feeling that's not that meaningful to one of the most important economic systems in crypto period, where the fees are going to be more than most L1 blockchains, period, full stuff.

And it suddenly catapults uni into a position where I think it'll be one of the most analyzed and most talked about systems, period. And it's- But is it not? Maybe I live in a bubble. On a day-to-day basis, not really. Like, you know, besides us on a podcast. I don't think you need to get that much discussion. I mean, even though the fees paid by users are significant,

100% of that is going to the liquidity providers. And I think it fundamentally takes Unity from being a quiet system to being like an incredibly loud one. And, you know, I think the magnitude of this really can't be understated. I think if it's done successfully, it will transform so many other projects. And, you know, I think it's like the beginning chat. of, you know, what does defy look like, you know, when it's sort of unleashed? I, that feels a little bit too broad because most defyp protocols do have revenue.

Like, Uni is kind of an exception in that it had, it had no revenue for so long. I mean, most don't. I mean, if you, if you talk about like what's going into the Dow, like Avey does, compound does. Compound doesn't? Compound doesn't? No. Isn't there money going into the insurance fund?

It's an insurance fund, but it's not being like paid out to like stake. True, but it's kind of like, you know, it's controlled by a token holder. It's controlled by token holders at the end of the day, right? So like it does seem to me that uni was kind of unique. It sort of feels like maybe the Saudi Aramco moment where it's not so much that there have been no oil companies. It's more like the scale of this DFI protocol has never actually been absorbed by by markets. And that is, I think, what's revitalizing things. It's just realizing like, oh, shit, that's a lot of money at scale that can be made by token holders.

I don't know, Tom, what's your take? I think maybe I might be wrong on this. I think the government proposal is basically just to switch the like the owner of the factory contract over to token holders versus just having it be sort of deployed as is. So the current proposal does not turn on the fee switch, but it allows token holders to add fees to new pools in the future if they want to do that. I thought they were turning on fees for a select number of pools. So it's they outline like the sort of potential fee schedule. But I think the current sort of snapshot that's up is just, like a temp check on shifting the owner of the factory contract over to like token holders basically.

I see. It's a multi-step process. Yeah, which again, from a regulatory perspective also probably makes sense, right? It's not the team turning on if you switch, it's token holders, you know, who are staking voting to turn on if you switch. Yeah, I heard there was a there's a conspiracy theory maybe a week or two ago that like secretly Coinbase, you know, had one or was winning. It's, it's case against the SEC. And so that was sort of driving up both the price of coin and sort of allowing you need to go and do this thing. I don't really know how credible it is. It seems, you know, kind of conspiracy theory-ish, but I thought it was kind of interesting.

And yeah, I mean, it's like you said, I think in terms of market re-rating, it's not just uni that has been ripping, right? It's like every defy token that's in a similar position of, hey, you know, we have some fee switch, or maybe it's, we have some, you know, revenue that's going to a treasury or going to an insurance fund or something like that, and is opening up the possibility that those might go to token holders in the future instead. So, yeah, I just feel like maybe it is the market maturing and sort of thinking about how to rate and then price these assets.

Yeah. Any of you? I mean, we did a lot of the fee analysis for the foundation for this, so I've spent a lot of time looking through which pools I want to do and how to think of this. I will say there's a lot of careful consideration over at least a year plus. And, you know, I don't think they're doing this randomly in the way that maybe the conspiracy theory suggests. Anything more you can share or you can, be enigmatic. You know, I'll leave it at that. Probably better, better. Okay, so we have one insider

who can't share too much, unfortunately. Hey, hey, read the post. We are, I read our report. Yeah, yeah, no, I'm sure there are very good sound economic reasons for choosing the pools you did. Yeah, it does seem like the market is more or less pricing in that the regulatory environment is shifting, or at least the winds are shifting, right? So obviously the SEC is expected to continue to be obstructionist, but I think the whole industry feels like it has more ammunition. And in some sense, you know, it feels like uni is kind of the battering ram that everybody else can kind of get behind and say, okay, well, if uni is doing this, we're probably good because they're going to go after uni because they're kind of the biggest, baddest guy in the

block. And so I suspect that's a little bit of the energy in Defi right now that's booing all these tokens. And if they go after uni, which, you know, we know that Unis had run-ins with the CFD at least, and the SEC almost certainly has chatted with them, that there's going to be a showdown at some point with respect to defy in the U.S. So assuming that the SEC gets around to it and they don't end up, you know, turning out after if there's a handoff in the administration. So we'll see right now, I mean, I think we got to wrap because we're getting up on time, but it's a wild market. My guess is that probably by this time next week, all-time highs,

in Bitcoin at least. I mean, we've already hit in dollar terms. We touched all time high very, very briefly. When a CoinDesk has the headline and Bloomberg has the headline. Yeah, Bloomberg and Wall Street Journal

posted a thing right after. No, I know, no, no. It is not in dispute that we have hit the all time high. We've hit the all time high. I mean, I guess the real question is when do you think Eth all time high? I think it takes more time.

I mean, at this rate, it's like we're days to weeks away. I mean, at the speed. But that's what I'm asking for a more precise prediction. Like, do you think by April? Well, okay. So last Tuesday, eight days ago, I was at a dinner. And people were talking about, you know, will Bitcoin in ultimate?

I think Bitcoin was like 54,000. I said within two months and I bet on it with a friend. Oh, nice. And, you know, we were all shocked when that prediction came true eight days late, seven days later. So are you willing to make this? Yeah, should we make another bet on the year? My prior bet was, will it happen within two months?

And like, it was even odds. I would bet anyone here that happens for either within two months. Within two months is a lot. Happening is what, a month and a half? April? I think the happening is overrated in terms of it. Yeah, yeah, but I think, I suspect what happens is that ether catches up. It's actually very close.

I need to go check options pricing on Deribet to get a sense of, which Mark's pricing. Yeah, see where the Delta is. Yeah, yeah, yeah, yeah, that's kind of cheating. It's against the spirit of it. Yeah, I do, I do think that. I can't use it because like a bet like this is like a binary touch option versus like, you know, an American. Yeah, yeah, I was gonna say. Sure.

You can't even look at the price. Well, let me ask you this. When ether, it's all-time high, what do you expect the price of Bitcoin to be? $76,000. 76. Okay. Which is more or less same ETHBGC ratio.

I think ETHP maces a little bit from here. ETH pays a little bit, okay. But not dramatically. I think that's probably right. That sounds right to me. Yeah, I think it'll be a while until ETHBTC hits the same level that it was in, but probably by the summer. Yeah.

I think it probably catches up. So anyway, obviously we are not traders. And so we have no idea what the fuck we're talking. We're all long-term private market investment. Except you. You're the meme coin trader. The rest of us are long-term investors. You weren't you the word giving us the... Most knowledgeable. Yeah, you're the most knowledgeable about Geo-Botent and all this random shit. I might be knowledgeable about it because I like, you know.

Okay. You have the oral tradition, right? Like, sit us around the fire and tell us the story of Geo-Botan. Back in the day. Yeah. Geo-Botan. Beautiful. Okay. With that, I think we got a wrap. Thank you, everybody. We'll be back next week.