ROLLUP: Robinhood’s Meme Economy | Solana Cuts Issuance | Saylor’s Comeback | AI Alarm

TRANSCRIPT

Ryan Sean Adams:
[0:04] Bankless Nation is the first week of September. It is time for the Bankless

Ryan Sean Adams:
[0:08] Weekly Rollup. We have Hasib on again. David is out at Burning Man. Hasib's second week in a row. We got to talk today about the Robinhood economy. It seems like it is completely blowing up. Robinhood chain knocking it out of the park. Transaction fees right now are actually higher on the Robinhood chain than on Ethereum. $4.4 million in revenue in a single day. Fortunes are being made over there. And yet, Hasib, our guest host today, calls it kind of a financial jackass. I saw that tweet, Hasib. We got to talk about what you mean by that. Also, Solana just cast a governance vote to cut its own issuance. Is Solana in the store of value game right now? And Michael Saylor, He's buying back in at 80K after he's selling at 62K. Their CEO says both were the right trade. We'll talk about that. And 21 banks launching a stable coin. And then again, Besant, week three of the treasury put, what is he saying at the G20 summit as 30-year yields are approaching 20-year highs once again? Hasib, it has been a very busy week. A lot of interesting things happening in the world. What's your eye on in crypto this week?

Haseeb Qureshi:
[1:26] Well, look, the firework show we're seeing on the meme coin side is hard to look away from. We're getting this crossover between TradFi and RWAs now into crypto, which is interesting. It has me worried. I feel like this is not going to end well, but we should talk about it once we get into it.

Ryan Sean Adams:
[1:43] All right. So that might be some of the financial jackassery that you're seeing. Before we get in, I want to thank our friends and sponsors over at, all right, Hasib, let's take a look at our prices on the week. Bitcoin on the seven-day up about 1%. ETH flat, some other assets. Let's see, Solana down about 3%. We are holding, my friend. We are still outperforming. Gold, equities continue to hold these elevated prices. Bitcoin over 81K at the time of recording. Ethereum over 2,500 at the time of recording. There was a pretty big bounce today, actually. So Bitcoin was trading down to

Ryan Sean Adams:
[2:25] 76K and then suddenly we're up 5% in the last 24 hours. Any reason for this bounce or is this just classic crypto volatility?

Haseeb Qureshi:
[2:36] It's definitely macro. So the same thing in the stock market. Stock market rebounded today. And a lot of that is because the rate cut story is, or I should say the raising of rates is a little bit less likely based on some comments that were made today by, I believe it was Waller. So there's a lot of worry about inflation size.

Ryan Sean Adams:
[2:54] Are you talking about the Fed chair of Warsh?

Haseeb Qureshi:
[2:56] No, no. I believe it was one of the other folks in the Fed.

Ryan Sean Adams:
[2:59] Interesting. Okay. Yeah.

Haseeb Qureshi:
[3:01] So basically, I mean, right now it's a pretty tight race at the Fed of like whether or not they're going to get enough votes to raise rates. And so just one person making a comment was enough to move the odds pretty significantly. I believe it was 20% less likely of a rate hike in the next Fed share meeting. And that is the reason why people are feeling a little bit more confident about risk assets. So I think that reflected in the crypto markets. Looks like there was a lot of liquidations from people who were short. So we're getting a little bit of volatility coming back into the market. That's it's good, but it's mostly macro telling the story as opposed to a crypto native story here.

Ryan Sean Adams:
[3:36] Well, it's interesting because there was a bit of a sell off towards the end of last week. And I think that was Warsh's speech at Jackson Hole. I don't know if you caught some of that. Some people watch that speech and said it was pretty much a nothing broker. But the market more broadly interpreted it as hawkish. I think the September rate hike at the end of that meeting last week went up to 67 percent, up 30 to 40 percent. And so that was the end of last week. It sounds like today, maybe things went back down, but they zoned in on something Warsh said. I believe it was this quote, we must be confident that underlying inflation is moving to our objective clearly and at sufficient speed. Otherwise, we have work to do. It was that line, otherwise we have work to do, right? Inflation's not where we want it. So implying the Fed has work to do. And what's the tool in the tool belt for that work? Well, it's Fed funds rate.

Ryan Sean Adams:
[4:26] So that's why the market was anticipating an increase. And now I guess that's being tempered a bit. One thing I can't seem to figure out, though, Hasip, on all this, I know we were talking about last week, the Besant put, and.

Ryan Sean Adams:
[4:39] We're now in week three of sort of this new market perceived paradigm of this Besant treasury put buying the long end of bonds somewhat is it seems like... Warsh is sending some hawkish signals like I might raise Fed fund rate. You don't know, you know, don't test me inflation. I might do it. Whereas Besson is being incredibly dovish in this total treasury QE thing. What's going on between the two of them? Is this all coordinated? Is this like a coordinated good cop, bad cop type of routine? Or like what sense do you make of this?

Haseeb Qureshi:
[5:16] Coordinate is the wrong word. I think it's more adversarial. And this is the natural dynamic between somebody sitting in the White House and somebody sitting at the Fed, which is that, look, at the end of the day, the White House wants to do things that are good for the economy. It wants to do things that are good for Trump and allowing Trump to achieve his policy goals. And that requires the economy to run a little bit hot. It's like they want to ride the edge. Right. And, you know, at the end of the day, Warsh, yeah, he was appointed by Trump, but now he's independent. And you can see that independence starting to show its colors is that, you know, Trump thought he was his guy. I think actually he's not his guy as much as maybe he thought he was. And the reality is, look, Walsh doesn't want to be responsible for having inflation run unnecessarily hot and then eventually getting blamed for mismanaging the economy. And so, you know, when Walsh was first nominated, he was saying like, look, yeah, inflation is a little bit high, but I think it's because of the war. It's because of oil. It's because of supply constraints, blah, blah, blah. It's not fundamentally reflective of the underlying economy. Now, all of a sudden, he's saying, well, we have to be really confident that inflation is actually tamped down. And this is kind of what you expect, is that when you nominate somebody, you kind of think they're your guy. And then you realize, no, no, you're kind of

Haseeb Qureshi:
[6:34] It's this almost like Manichaean battle between the Treasury and the Fed. It's never going to go away until you actually truly do break the boundary of Fed independence. The Fed is very much still independent. That's good. That's good long term for market confidence in our bonds and in our fiscal policy, or sorry, in our monetary policy. But fiscal policy is always going to run against that. That's normal. So I think this is what you would expect in a well-functioning independent Fed.

Ryan Sean Adams:
[7:01] I'm not sure that I buy that the Fed is as independent as you think it is. You really do think it's independent. What's kind of your evidence for that recently?

Haseeb Qureshi:
[7:10] I think the fact that Warsh is not rolling over, he's making these noises. I mean, previously he was saying, look, I'm not going to give forward guidance in the way that previous Fed governors have. I think that's a mistake. That was something only necessary at the zero lower bound. Now that, you know, Fed funds rate is, you know, whatever, four or five, we have a lot of room to play. and we can let the rates do the talking. Now he's suddenly talking a lot more and giving things that sound like guidance. And like, you know, why is he doing this? I think it's pretty clear is that, yeah, he doesn't want the treasury to be setting monetary policy. That's kind of not their job. And now, you know, again, the dynamic of our economy has changed so much over the last 20 years, especially since the great financial crisis. Plus, the reality is that it is this push and pull, much more so between the Fed and Treasury. But I think the way the market is interpreting this is that, no, Warsh has his own opinions. He's not going to let himself get run over by Besant. And this is how he's fighting back to some degree.

Ryan Sean Adams:
[8:14] My colleague Michael Nadeau thinks that Warsh is going to, he's more likely to keep rates steady and let the market sort of dictate long-term rates, sort of whatever the long-term rates do, they will do. And certainly, Besant has been back on his heels in kind of a defensive position this week. He spoke at the G20 summit, and we have as the backdrop, the 30-year treasury yield is back above 5.26%. So just rising above the surface, just like that beach ball back before he started some of these buybacks.

Ryan Sean Adams:
[8:49] I saw a few interviews from him, including this rather glazing interview with, I think, a Fox Business correspondent. And Besant was basically saying, yo, the media, they're talking about bonds and yields. That's not the real story. The real story is the American economic boom that this administration has brought in. We are an AI superpower. We are an energy superpower. Corporate earnings are off the charts. Stock market's off the charts. GDP is going up. That's the real story. So don't pay any attention to the bonds. He said this, this has been the best performing bond market among major countries in the world. And of course, he's saying this as the yields on the 10-year are at a 20-month high and also on the 30-year, it's close to a 20-year high once again. So is that true as well, Hasib? Obviously, he's talking his administration's book. But I mean, there are some points there. I mean, the economy is booming, at least from a CapEx perspective. The AI trade continues to bust through all projections. You think he has a point? Right.

Haseeb Qureshi:
[10:02] I mean, look, saying that we have the most demand for our bonds of all the countries is kind of like being the skinniest fat kid. You know, the reality is that all bonds everywhere are getting punished because of this fiscal irresponsibility that people are seeing from governments around the world. Saying that, look, we're the AI superpower, our economy is booming because of AI is true. And also precisely the reason why our bonds are less attractive is because our economy is more attractive as a borrower than the government is. Right. Like his job is to be his treasury secretary. The treasury is a pool of money managed by the government. So now it is obviously true that their tax base increases. If the economy does well, that's great. We can tax all the revenues that the companies are making. Wonderful. But at the end of the day, the core job of the treasury secretary is to be responsible for the money that the government owns and and, you know, collects from the economy. So if nobody's buying your bonds, that means that they perceive financial irresponsibility from the government. That's what's happening right now. That's the core of it. So don't let all the other stuff distract you. Like the American economy is booming. That's not because of the president. It's not because of the White House. It's because of the economy itself. The job of the government is to be a good steward of that, which is what the bond market right now is saying. You are not being good stewards.

Ryan Sean Adams:
[11:22] It does feel like this administration and Besson, and maybe to some degree also Warsh, are making a bet on the AI trade, right? Besant keeps saying, including the G20 countries, hey, we're just gonna grow our way out of this. And I think he's looking to AI to really propel all of that growth or the vast amount of that growth. He's probably also talking about growing money supply in order to do it, but it increasingly feels very much like the US economy. I know I said this last week, the world economy, all based on the AI trade right now. It's consuming more and more of our total economy.

Ryan Sean Adams:
[11:59] One person that has come back in full bull mode is Michael Saylor. I don't know if you saw this latest clip from Michael Saylor, but this is him on a New York City metro, it looks like. He is punching out bears. Bears, this is your stop. He just beat up a polar bear. I think that's a grizzly bear. I mean, this is some kind of a black bear. He's taking on all the bears. And Strategy has just purchased another 4,603 Bitcoin, about $370 million.

Ryan Sean Adams:
[12:32] Increasing its cash by about $29 million, doing some other things as well. What's fascinating about this is just a few weeks ago, a few months ago, maybe, he sold about 7,000 Bitcoin and he sold at 62K. And when I say he, I'm talking about strategy, of course, for $431 million. Now he rebought for less. Is this an example of kind of like selling low and buying high? I mean, the extra cost to rebuy the Bitcoin that he purchased was about $83 million. I think the strategy CEO came out and said, hey, this is, you know, we did the right thing in both cases. Both cases, the selling and the buying was the right trade. Obviously, they've been trying to get STRC back at par. It's now at $97, I believe, so 100 at par. So making some progress. They're just not getting it to that last mile. What do you think about this? It's just the type of thing you have to do when you're a bit on leverage, You have to protect your position, and so you're not always going to buy low and sell high. You just have to play this game a little bit to defend your margin positions.

Haseeb Qureshi:
[13:46] That's pretty much right, is that it's very hard to buy low and sell high with other people's money. You can do it with your own money if you have enough balls to be able to go and hold all the way, and you have the conviction to do it. But the MNAV game, which is fundamentally the game that MicroStrategy is playing, and this is what I believe the CEO, Fong Li, said of, okay, why is he defending the selling at 60 and buying at 80? The answer is, well, because just look at the MNAVs, right? The MNAV was below one when we were selling to get the MNAV back in line. And then when MNAV went above, we started buying again. And that's correct. That's like the way that you should think about what markets are compelling you to do and what markets fundamentally want at each of these moments in time. And this is a financial product. So it's responsible to its customers, its holders, its buyers. But it also means that this is fundamentally how the strategy and DAT trade works, is that it is pro-cyclical. Pro-cyclical means you are buying high and you're selling low. That's what a pro cyclical strategy is. Now, hopefully, if you buy that pro cyclical strategy at the bottom, you get the MNAV expansion as well as the buying on the way up. And so, you know, so you can you can see why this might be attractive anyway to play it as an investor. But this is fundamental to the strategy. It would never not be this way. If this is how you're timing your buys and your sells, you will always be pro cyclical in this way. And you will be able to look back and say, oh, look at you. You were selling at the low and buying at the high.

Ryan Sean Adams:
[15:11] So what letter grade do you give Michael Saylor in strategy for execution on this DAT strategy? I mean, do you think he's doing a pretty good job? He's the godfather of this. I mean, he created the whole DAT industry and DAT trade himself. Like, how's he doing?

Haseeb Qureshi:
[15:24] I mean, it's hard to answer that question because it depends on what your rubric is for grading. If the rubric is build a gigantic DAT and survive, then he gets an A+. I mean, this thing is just an incredible machine. Look how gigantic it is. And it's still going. And he's managed to get a war chest of billions of dollars of cash. He's got lots of dividend coverage now. And he's back in the game buying Bitcoin. And, you know, like it's a nice gig if you can get it. But in terms of the impact to holders, look, I'd say this is kind of a C at best.

Ryan Sean Adams:
[15:57] Holders? Are you talking about MSCR holders or Bitcoin holders?

Haseeb Qureshi:
[16:00] If you were buying MSTR, if you were holding MSTR, you're not thrilled, right? This thing is down a lot from when it was being most aggressively marketed and when most people bought in. So relative to Bitcoin, you're a lot worse off holding MSTR almost any point during this last cycle. So if that's how we're thinking about it, yeah, not a good grade for this vehicle.

Ryan Sean Adams:
[16:20] Well, how about Michael Saylor's meme skills and punching the bears out on the subway? What do you give him for that?

Haseeb Qureshi:
[16:26] I think that's an A+. That was pretty good. I enjoyed that quite a bit. It's one of those things that like AI, like, you know, people were complaining about him just doing a lot of AI slot marketing. Yeah. But I kind of think he knows science. It's on brand. Yeah. Yeah, exactly. Exactly. It's kind of, it works for him. I think it works. I think keep the slopification up. I think it's good.

Ryan Sean Adams:
[16:45] He's even got like, you know, the color scheme, right? You could see obviously orange tie and orange seats on the, on the metro. It's just like,

Ryan Sean Adams:
[16:51] yeah, doing a great job, Michael Saylor. Uh, we got more to discuss, including the Robinhood economy, Robinhood meme coins reaching all time highs, fortunes being made. As I said, there was a penny stock run up 300% due to some meme. There's a lot of meme and real world asset pairs that are trading. We'll talk about all that and more. But before we do, we want to thank the sponsors that made this possible. The Robinhood economy. Let's talk about it. What's going on in Robinhood land? This is a post by Mike Ippolito. The monster run for Uniswap on Robinhood chain continues $1.4 billion of volume traded yesterday, almost $20 billion since launching the beginning of July. Here's another chart. This is one of the crazier charts you'll see in crypto. I think this is L2 gross profit, L2 revenue. You can see all the various L2s. This was dominated at one time in May of 2021 by Arbitrum. That's completely flipped now. You saw base had a presence. Now it's just kind of this Robinhood green color. Robinhood absolutely dominating revenue on the L2 world.

Ryan Sean Adams:
[17:57] There were some stats I saw, something like three to four million per day in Robinhood. If you were to annualize that seven day, which you should never do, you should never annualize seven day revenue volumes. But if you're to do something like that, then actually the revenue generated by the Robinhood chain right now would exceed its equity trading business for the year. OK, so absolutely massive amounts of revenue that is flowing. What's going on in Robinhood land? What are you seeing in Robinhood chain? Is this all meme coin shenanigans? Is that a good thing? What do you make of this?

Haseeb Qureshi:
[18:39] So there's this weird meta that has now transfixed people trading on Robinhood chain, which is this pairing of meme coins to stocks, especially these like kind of micro cap stocks. So the big one was there was this meme coin called Boner and it was paired against HIMSS, which of course is one of these direct to consumer like, you know, ED medication platforms. That's, you know, it's very, it's very big company. If those of you haven't heard of it, you know, HIMSS and HERS. Never heard of it. Never heard of it. Yeah, no idea. That just goes straight in my mailbox. And so they had this meme coin that was tied to HIMSS, and people started copying this. So Boner really took off, got a lot of excitement. And people started copying this model with a bunch of other stocks that were paired to meme coins.

Ryan Sean Adams:
[19:27] Real stocks are you talking about?

Haseeb Qureshi:
[19:29] Real stocks, real stocks, tokenized stocks, right? So now the problem with a lot of these tokenized stock meme coin pairings is that, of course, these things are trading 24-7. They're trading when the stock market is closed, and they're trading on the weekends. So last weekend, there were a number of these stocks that the meme coins that they were paired against started really running up over the weekend. Because, of course, you know, this stuff has just been taking off lately. And the problem with a meme coin paired with a stock taking off on a weekend is that there's no arbitrage you can do on the weekend. There's no way to move the real price of the stock up. And the floats of these stocks are very small on Robinhood compared to the actual float of the real stock, which means that people thought they had these huge runners. You know, there's like, oh my God, we're all making so much money. This thing is running to the moon. You got to buy in now. And then on Monday morning, once the market opens, people realize like, oh wait, this thing is de-pegged from the underlying stock by like 400%. And of course, what's going to happen is that people are going to start minting the actual underlying stock selling into the people on Robinhood and destroying the meme coin. And so this just started happening en masse, basically either one on the overnight trades or on the weekend trades. So all the guys in the meme coin trenches were just getting wrecked programmatically as soon as markets opened again.

Ryan Sean Adams:
[20:46] Is it mainly the meme coin trenches that were getting wrecked or is there any fallout to kind of the tradify market on the other side? I guess that's a bit of the tail wagging the dog. because it's not enough volume to.

Haseeb Qureshi:
[20:56] Move the big markets. The amount of assets minted on these chains is like less than a percent of the underlying float. It's so small, but like the meme coin, of course, can become a big runner and get a lot of volume. So this is what we were starting to see over the weekend. So just a lot of people doing very stupid things and kind of imagining that this was like a GameStop moment that maybe we can kind of band together as like a school of fish and like take on the stock and short squeeze these headfirsts. You know, we were just like, no, no, no, no, no. You don't understand. Like GameStop only worked because it was so big. It was like a civilization level short squeeze that happened with GameStop. Doing that with like some random trenchers on some random stock, like that's just not gonna, you're just gonna get destroyed. Like that's a guaranteed way to lose money. And so now lately we've seen these experiments where people are now taking micro cap stocks, not the size of a HIMSS, which is, you know, a multi-billion dollar company. Instead, these micro cap stocks are like five, $10 million NASDAQ stocks that are like at risk of getting delisted. And so now, of course, those can't actually be tokenized. Nobody's tokenizing $5 million microcap stocks. So I think there was like these meme coins pretend to be micro cap stocks. They're named the same thing as the micro cap stocks paired against other meme coins. And now those are starting to trade in a crazy way that's getting reflected almost like an affinity trade in the actual underlying meme micro cap stocks. Okay. So all this craziness going on.

Ryan Sean Adams:
[22:24] That's, I think what I was trying to make sense of. You're deeper in the trenches than I am, Hasee, but I was trying to make sense of this story, the Farmy story. So yeah, this is a, um, I believe this is a micro cap. Um, DGN's found, uh, Jin Kwan in Farmy's, uh, filings, I guess Farmy is a micro cap mushroom stock. Chinese dried mushroom micro cap has 15 employees and literally means money mushroom. Okay. So some meme coins started trading against not this stock, but some meme coin version of this stock.

Haseeb Qureshi:
[23:02] They just named a meme coin after the company.

Ryan Sean Adams:
[23:05] They named a meme coin after the company and they set the tokens to the exact float of the stock. So I think the supply was sort of the same. But the attention economics as this meme coin began to run actually filtered out into the TradFi world for the actual stock of this company. And the actual stock of this company went up like 300%, 350% at some point. So now this is, you know, meme coin, the meme coin economy can affect TradFi in the big assets like the HIMSS assets. Certainly can't touch something like NVIDIA or Google or any of the big tech companies. But a micro cap, well, that's meme coin size. And so we saw in this instance, the attention economy of meme coins actually filtering out and affecting TradFi stocks. That's what you're talking about, right?

Haseeb Qureshi:
[23:59] Yes. I mean, to be clear, it's kind of like, man, we have fallen a long way that we're now like playing games with $5 million micro caps. Like we used to have meme coin runners that would go to like a billion. And now we're like, oh, wow, we moved to $5 million Chinese mushroom company that's at risk of getting delisted on the NASDAQ. Like this is... This is stupid. Like, this is obviously not good for people to be doing this.

Ryan Sean Adams:
[24:33] Is that what you meant by your jackassery quote, you said? Yes.

Haseeb Qureshi:
[24:37] What do you mean? The fact that this is becoming a meta all of a sudden on Robinhood chain. Now, to be clear, this is not the only thing people are doing on Robinhood. There's more conventional meme coins like Cash Cat, right? Cash Cat is just a meme coin. You can buy it. It's like Doge or whatever, any of these other meme coins. And that, I have no problem, but that's totally fine. You know, it's just kind of people just playing dice, you know? It's like, whatever. Maybe it'll win. maybe you'll lose, but it's fine.

Haseeb Qureshi:
[25:01] With these kinds of things where people are manufacturing these meme coin to stock pairs, right? And they're doing it over and over again at scale, hoping that one of them is going to hit. We've seen the supply chain of meme coins, how they're created, how they're monetized, all the little middlemen who keep doing this and paying the influencers to pump this one or pump that one. And, oh, we've got a runner, everyone buys into this thing. And then the problem with the story relative to anything else is that this construct will reliably fail. Like you will definitely lose money if you do this because of course, if your stock token is de-pegging from the underlying stock, you will get screwed by hedge funds that will come in and mint over your head. Everybody involved is going to lose money here, except for the meme coin boosters who are promoting this and, you know, taking some of the meme coin up front. So it's just like, we should not be encouraging this kind of financial stupidity. Obviously people are enjoying it on some level. Like, you know, there's this kind of, that's what I call the, the sort of financial jackass is that it like, you know, if you remember the show Jackass in the nineties, where like people would just, you know, they would like just, you know, I don't know, hit, hit themselves in the balls over and over again and see how, how bad you could do it. And the entertainment was in the stupidity, right? That's kind of, I feel like what we're watching here. It's not like trading Doge, you know? This is like a competition of stupidity. And like nobody wins when that becomes the game.

Ryan Sean Adams:
[26:23] So let me, like there is something novel here in Robinhood chain with tokenized equities that we get, which is we have the brokerage distribution of Robinhood. We have official like real stock tokens. And then we have AMMs and permissionless pairing. And the permissionless pairing can be against stable coins, of course, but can also be against meme coins. And that is something that's novel. I mean, we were talking last week of what's, going on that's novel in crypto. This is something novel, some sort of asset, some meme coin. You can pair a token against anything. Is there something new here being created apart from all the casino games? Is there something beneficial here that we're doing?

Haseeb Qureshi:
[27:11] It's hard to see what the social benefit is of people doing this. Like you can imagine, so look, I can imagine a world where if you are HIMS and you're like, hey, I want more people to buy my stock. Maybe you create some meme coin on Robinhood chain tied to HIMS and the meme coin goes up, HIMS goes up, people are buying HIMS as a, I mean, I don't know if that's going to work. I don't know if it's a good use of your money or your time or your marketing budget, But you can imagine that maybe would be socially constructive and create some new form of capital formation or something.

Haseeb Qureshi:
[27:43] Having third parties do this, who are, again, making money by paying influencers and doing the whole thing. Like, we know how the whole game works, right? This does not feel useful. It's also not zero-sum. It's like negative sum. It's like really, really negative sum, which is not good. Except for the hedge funds. I guess the hedge funds are making money. So my view here is that the other problem with the story is that it's not really clear that Robinhood is actually onboarding a lot of net new users yet. They might at some point. But if you look, there's a Dune dashboard that shows a lot of the data about where the Robinhood user are coming from. And it's clear, actually, a lot of them are coming from FOMO.

Haseeb Qureshi:
[28:21] Actually, most of the activity, it seems like right now, is coming from FOMO users who are, they're basically chain agnostic at this point. So if you're on FOMO, you can buy something on Solana, you can buy something on Robinhood chain, you can buy something on Base, it doesn't really matter. It's just all assets to them. And then GMGN is the other platform that seems to be driving a lot of volume to Robinhood chain. But we're not really seeing a lot of Net new wallets, not a lot of net new addresses. Now, maybe they do come at some point. And usually what we see in meme coin cycles is that they come in late when the numbers start getting really big, right? So right now we're getting a lot of volume relative to crypto. But if you look at the TVL and Robinhood chain, it's actually not that big. TVL and Robinhood chain is like 700, 800 mil. You know, it's like, I don't know, one seventh, one eighth of base or of Solana, even though it has a lot more volume than base. And right now it's like second only to Solana in terms of deck volume. So we're seeing a lot of trading activity on these chains, a lot of AMM activity on The net number of new users right now doesn't look to be that big. I think it's a lot of crypto natives right now. But if we see this stuff grow by another order of magnitude,

Haseeb Qureshi:
[29:25] then I think you might be right in that this might be a more significant onboarding event.

Ryan Sean Adams:
[29:29] Do you think there are any knock-on benefits to other assets or other ecosystems in crypto? For example, does Ethereum benefit at all from the takeoff of the Robinhood chain, at least takeoff in terms of meme use case? It seems like some of the meta is moving maybe from Solana to Robinhood chain, at least for the moment. Obviously, Solana is still much larger. But, you know, rather than pump fun, there's this Pons kind of meme coin launcher app that's really taken off that's more native to Robinhood. This is Lorenzo from ARK, an analyst at ARK. He's talking about Robinhood revenue and comparing that to Ethereum and just making the point that the ETH take rate is so damn low. When it comes to the L2 economy. So he's saying this, Robinhood chain grossed 20x more revenue yesterday than 10 days ago. Ethereum got paid half. So he pulls a date. August 30th, Robinhood chain made $108 million in revenue. And a portion of that went to Arbitrum. Like it's 10% cut. 90% went to Robinhood. Ethereum made $155, a take rate of 0.014%. What do you think this means for the economics of L2? Is this another L2 in the wild? Do you think this has any benefit to Ethereum? Do you think that this is Solana's loss? What are the after effects of the Robinhood chain?

Haseeb Qureshi:
[30:59] Yeah. So, I mean, it's certainly better for Ethereum for this to be happening on Robinhood chain than on Solana. Obviously, there's an interoperability with everything on Ethereum. So these assets can move back and forth. They can come live on Ethereum mainnet if they so choose. And of course, ETH is the native asset on Robinhood chain. It's how you pay fees. So, you know, there is some flow through effect of ETH, the currency. In terms of direct revenue that they're getting from Robinhood chain, it's pretty clear that's not happening. And that's not really the main story. In the same way, it's almost like, you know, if we have a satellite state, the US has a satellite state that uses the dollar, it's a dollarized economy. We're not getting tax revenue from countries that dollarize, but we do increase exports of the dollar and we increase the demand for the dollar globally. I think that's the right mental model for what Robinhood chain is to Ethereum. It's part of the Ethereum economic zone, but it's not directly paying taxes. It's not a colony. It's not like this is, you know, the 51st state of Ethereum or something like that. So I think that's kind of the wrong mental model. And if you're looking for that and you're kind of thinking, well, when is Ethereum going to raise the tax rate? It's like asking, you know, when are we going to tax Venezuela when Venezuela has dollarized away from the Bolivar to the dollar? Or like all the countries in the world that have dollarized, when are we going to start taxing them? The answer is probably never. We're probably not going to tax them at all. But it's still net beneficial to the U.S. to have countries that use the dollar, even if they are not part of our distinct political union.

Ryan Sean Adams:
[32:23] How about for Solana? Net loss for Solana attracting some of that meme coin attention energy?

Haseeb Qureshi:
[32:28] No doubt. No doubt. I mean, Solana, so much of Solana's REV has been from meme coin economy. And Robinhood just flipped Solana for REV. It is now the number one chain by REV, at least over the last few days. And that's really significant.

Haseeb Qureshi:
[32:44] Now, look, it's always easy to have a pop. And it remains to be seen whether or not this is going to continue on Robinhood chain. We've seen these kind of meme coin manias rise and fall with different chains as they come into the fore. You know, BASE had a mania where they had a huge amount of meme coin volume. Robinhood is a beast with respect to their distribution. And I feel like their marketing efforts are actually very on point. So they've been leaning into this. Vlad has been leaning into this. I think he understands the language of DGENs seemingly a lot better than, you know, Brian Armstrong and BASE does. But still a lot of game left to be played. So I would not count Solana out. They still have the TVL. They still have more users. They still have an enormous distribution edge relative to anybody else in this game.

Ryan Sean Adams:
[33:25] They still also have lower transaction fees. This was interesting. Robinhood chain's median fee is now 2x the fee of Ethereum L1. So twice as much as Ethereum and 128x the fee costs of Solana. So people paying close to 10 cents per transaction on Robinhood right now.

Ryan Sean Adams:
[33:45] So of course they can scale that. That may not be the case forever, but at least right now it is the case. Maybe this is partially why Solana is moving a bit more, might I say, in the store of value direction. So I don't know if you've seen this or talked about this previously, Haseed, but Solana had its first binding on-chain governance vote to actually reduce issuance on the Solana chain. So Solana issuance was scheduled to reduce down to a 1.5% per year floor. And just by way of context, Ethereum right now is about 0.8, 0.9% annualized. So it's under a percent. I think Bitcoin is something like 0.8%. Of course, that will happen in what, two years time again. So that will continue to cut down. Solana just voted, and the vote barely passed to... Create that 1.5% issuance floor, have that arrive in 2029, the beginning of 2029, rather than 2032. So bringing that up by three years and saving about projected, about 19 million souls, that would be $2 billion of extra issued soul that would, of course, go to validators.

Ryan Sean Adams:
[35:10] This passed, apparently, according to Solana, governance required 67% to pass, and it passed just barely by 0.33 percentage points above the threshold. A binding on-chain governance vote. There was some drama where Kraken Solana validators, apparently first they voted no, and then they switched their vote.

Ryan Sean Adams:
[35:36] Mert, Helios CEO, I think Helios was part of the authors of this Solana improvement proposal, said this, after 500 calls in the past few hours, we got all the votes in in the last second and passed the disinflation proposal by a literal hair. Let's go. We did it. So this is Solana now becoming a bit more scarce, reducing issuance, going in the direction that we've seen, obviously, Bitcoin go algorithmically. We've also seen Ethereum go, which is reducing its issuance. What do you make of this? Is Solana trying to play the store of value monetary game the way Bitcoin and Ethereum have tried to play it? That hasn't been their message in the past. I remember Anatoly telling me that issuance didn't really matter. It was basically all recycled. And now here they are, they're cutting issuance. What do you make of this?

Haseeb Qureshi:
[36:34] Yeah, I think the wrong framing is that, okay, this means they're playing the store of value game. I don't think that's how anybody in Solana thinks about it. At the end of the day, this is basically inflation. Inflation is like nominal yield in the sense that it's not real yield. You are getting inflated at the expense of the validators who are getting more soul. A huge amount of the soul is already staked. And so what really happens, let's imagine that 100% of soul was staked. Right now, I don't know what exactly it is, probably 50% or something. If you imagine the 100% of sole was staked, then everybody getting more sole because they're staking or they have it in staking derivatives or whatever is basically just, okay, well, the unit price goes down, our ownership goes up in terms of units, but we pay taxes. And so it's creating tax leakage. If you are-

Ryan Sean Adams:
[37:21] You're talking about actual real world taxes on kind of income from staking.

Haseeb Qureshi:
[37:25] Actual real world taxes for those who pay. Yeah, it basically creates tax leakage. It's inefficient. And really what you'd rather do is pay the minimum amount of yield that you need to in order to keep the market functioning, in order to keep security solid, in order to keep the validators happy. And, you know, the biggest people who are fighting in the past for nominal yield were the DATs. The DATs really care about nominal yield. Why do they care about nominal yield? Because it gives them a story of why you should have your money in a Solana debt instead of in a Solana ETF. It's because, well, we can stake, we can do all these fancy staking strategies, we can lever up on staking yield or whatever, and that's why you should give us money as a more complex financial strategy than what you could get in an ETF.

Haseeb Qureshi:
[38:11] That story weakens when the DATs are all dying. And right now, Solana DATs are struggling. There's no runaway winner like there is for strategy or for bid mine.

Haseeb Qureshi:
[38:23] And as a result, you can sort of see that reflected in this vote, is that this vote was tried earlier, failed miserably. There was not anywhere near the economic consensus because the thinking was like, look, nominal yield might be stupid. It might result in tax leakage. It might be a little bit of a economic mirage, but it's good for business. So you need the DATs, make the DATs happy. Let's give all these service providers a business model. But in a world where, look, the DATs aren't working, that's not our story, so let's not kowtow to the DATs. Let's just lower inflation, and let's hope that that means that it's going to be better for asset prices and less tax leakage. That seems to have been the story that rallied everybody for this vote. I think it makes sense. And absent any kind of business models that depend on nominal yield, that is obviously the right answer. As a network, you want to pay the minimum amount. In terms of inflation. In the same way, look, it's like a government, it's like Doge, right? The government's saving money. You want to pay the minimum amount to all of the service providers who run the government, which is, of course, you know, the tax rate. You want to minimize that tax rate and leave the rest of the money in the hands of the people. But, you know, it's only by virtue of the belief that, well, if we give more money to the government, that will be better for the economy. It'll be better for society. It'll make us grow faster, do better things, or live in a more perfect world. But if you no longer believe that, then it's like, yeah, just lower the taxes and leave the money with us.

Ryan Sean Adams:
[39:44] Issuance decisions, monetary policy decisions for decentralized or more decentralized networks to me are so fascinating because every network handles it a little bit differently. So you could contrast this to the EIP-8361 debate that was going on Ethereum, about a month ago where there was governance post to actually reduce Ethereum issuance to modify the yield curve. And that wasn't a stake-weighted governance vote. Ethereum has no ability to really do that or it's not in the kind of the social.

Ryan Sean Adams:
[40:19] Contract of the Ethereum network to really do that. It operates more on this rough consensus where you post something and I don't know, how does a monetary policy change actually come into effect in Ethereum? Bitcoin is even more. They have some sort of a immutability religion, right? Right. Which is like, yeah, go ahead. Try to fork. Try to try to change Bitcoin's issuance policy. See what happens. A massive fork will happen. But that is also rough governance. So the way this decision was made is kind of interesting. Right. It's like 67 percent of all of the stake weighted validators voted. What do you think about that? Like, is that a is that a feature or a bug? There's even like stake weighted governance voting for issuance. Does that disclude you from the store of value race entirely or like? How do you contrast that to the recent EIP from Ethereum, which doesn't feel like it's going anywhere?

Haseeb Qureshi:
[41:17] Yeah, it's a great question and a very good contrast. So Vitalik historically has been against on-chain voting. So on-chain governance was something that really emerged after Ethereum. It started with Tezos, and then most of the chains after Tezos ended up adopting some kind of on-chain governance, such that purely by some kind of stake-weighted vote, some change in the underlying protocol automatically went into effect. Ethereum has never done this. The only thing for which Ethereum has on-chain governance is the gas limit or the block size, right? They have on-chain voting among validators that push up and push down through a sort of kind of crude voting mechanism of how big the block size becomes. Besides that, there is nothing else that the validators of Ethereum get to vote on. Everything else is decided through this kind of soft governance mechanism.

Ryan Sean Adams:
[42:06] Vitality stake is that there's kind of an enshrined plutocracy that can start to occur when you have stake-weighted voting.

Haseeb Qureshi:
[42:14] Exactly. So it's a little bit like, you know, take the Fed, for example, or even take our government, right? We don't have direct democracy in the government. We have this representative democracy where we elect, you know, the president, we elect the Senate and whatever, but they don't say, hey, great, here's the new tax bill. Let's send it out to all the people and you vote with your phone on whether or not you like this tax bill. We don't do that. And we don't do that for a very good reason, which is that it can lead to adverse consequences because people will often just, you know, if we voted on rates as a country, we would just vote rates down and down and down and down and down and we'd probably result in massive inflation. And so we'd be terrible, right? We would be like Trump. We'd be constantly saying, look, I want a lower mortgage. I want my socks to go up. And then we'd say, oh my God, I can't believe inflation happened. That sucks.

Haseeb Qureshi:
[42:59] And like, that's probably what we would do. And so to save us from ourselves, we don't have that system. We have a system of representative democracy where it kind of, look, you pick a guy and then you don't touch it for the next four years. And that's, you know, there's some wisdom in that. And that's kind of what Ethereum has chosen is say, look, if we just leave this up to people just voting with their money, they will vote themselves more money. And voting yourself more money feels good for the moment, but in the long run, it's actually important for us to protect you from yourselves. And the way in which you're going to protect it from yourselves is by being slow, is by replacing people slowly in the kind of high priesthood of Ethereum, and by holding on to values, even when those values don't seem cool anymore. And one of those values that doesn't seem cool is adherence to cypherpunk values and decentralization, right? There's this sort of push and pull that you can see around us with people talking about Solana and Hyperliquid and all this stuff. It's like, hey, do we really need to be worrying so much about crops and, you know, how important it is to get privacy and blah, blah, blah? why don't we just move faster and launch our WAs and, you know, tweet more number go up stuff and help the DATs. And Ethereum has said, no, we're not going to do that. Our job is to be a good steward of this overall ecosystem. And Solana takes a different tack. And I think it's good for there to be different approaches to governance in the same way that like, look, Solana is not trying to be a store of value. Solana is not trying to be World War III resistant.

Haseeb Qureshi:
[44:26] They're what they said and they're sticking to it. I think that's great in the same way a company is not governed the same way a government is. And that's good. You don't want every institution of every scale governed the exact same way.

Ryan Sean Adams:
[44:41] Well said. We've got a lot of experience playing out in crypto for sure. We're talking about some more of them, including a 21 banks uniting behind a stablecoin. We'll talk about that. Also, Hasib, you just mentioned to me before we started recording this episode that the new ChatGPT model Astra just came out. I want to ask you about this. All that and more. But before we do, we want to thank the sponsors that made this episode possible. Okay, Goldman Sachs, Bank of America, and others, a whole bunch of other banks, they plan to issue a dollar stablecoin altogether, a big banking consortium, in 2027. This out of Reuters, this is 21 financial institutions planning to do this. Of course, this is Clarity Act stuff, very official, very regulated. What do you think, Hasib? Do they have a shot against what's going on with, let's say, first movers like Tether, Circle, USDC, even Stripe seems to be ahead of the curve. The old banks coming up with a stablecoin. Do you think this is a nothing burger or do you think this could be a force in crypto?

Haseeb Qureshi:
[45:43] I am team nothing burger. I think like consortia are just so, I don't know, they're so NGMI. I've yet to see any consortium of this nature really prove themselves to be a force against more focused singular companies that have muscular founders who are working on something. Like you tell me, great, these 10 super regulated, super old, super slow moving banks have agreed to go and basically build a tech startup together. I'm like, oh my God, that sounds horrendous. And no, who's the CEO? Who's the founder? Who's the team? We don't have to talk about that. It's just like, who are the members of the committee that are going to decide what to do? I'm like, come on, this is obviously this is not going to go anywhere. So, I mean, CircleStock was down a little bit on the day. I think it was down like 6%, which tells you, okay, even the market thinks there's only a 6% chance that this even matters.

Ryan Sean Adams:
[46:34] Yeah, that's not a lot percent in terms of it mattering. So I tend to agree with you. There was a massive data breach that just happened earlier this week. Hasib, 153 million driver's license were leaked. This has been called the worst KIC hack ever. So this is American driver's license all selling on the black market. 63% of Canadian as well. 63% of all American driver's license are out there. This is Eric Voorhees. KYC is a bullshit scam that endangers millions of innocent people so that regulators can feel as if they're providing value to the world. I concur. Roman Storm, why did 153 million driver's licenses just land on the dark web? Because someone built a database of everyone. Last summer, the government's expert told my jury that's exactly what Tornado Cash should have been. I guess I have a question for you, Hasee Blake.

Ryan Sean Adams:
[47:34] All of these AML KYC credential breaches, I mean, I'm sure my passport, your passport, probably the majority of people listening, passport, driver's license, all KYC material is somewhere for purchase on the dark web. When does the US government just call it in and say, this no longer works, let's stop doing this? When do they start adopting something like ZK Tech? Why hasn't ZK Tech in crypto really picked up to carry this use case? What's it waiting on? What are you seeing in sort of the startup world?

Haseeb Qureshi:
[48:08] Look, I think the answer is that there are some countries that already have much stronger KYC mechanisms than the U.S., right? You go to someplace like India or China, and it's just night and day relative to the technological sophistication people use to prove who they are. For us, we literally take selfies next to a thing, and it's just like, come on, in the era of AI, in the era of data breaches, in the era of like all this stuff, that's the defense against knowing that this is who they really are. So I think it's obvious that this stuff is doomed and the answer of when is the US government going to wake up and do something about it my usual answer for these things is when it's too late and the problem has already metastasized and like it's just an unignorably bad problem that's when we will start

Ryan Sean Adams:
[48:54] So, and when we can't ignore it any longer, I wonder if this is the point at which we can't ignore it any longer when you have 63% of it's going to get worse from here.

Haseeb Qureshi:
[49:02] 100% it's going to get worse. No, no, no. This is not the warning call to the government. They're going to wait much longer until their constituents actually start yelling at them about this, which right now, you know, it's a news story. We'll forget about it in two weeks. Like, I already assumed my driver's license was on the dark web. I don't know about you, but I'm already at the point where if somebody from my team thinks that it's really me because they saw my driver's license, I'm like, you're fired. That's unacceptably naive to believe that.

Ryan Sean Adams:
[49:33] There is a vote in the Senate for the Clarity Act. This is not the final vote. This is a vote for a cloture on whether they have a motion to actually advance it in the Senate. pretty dim prospects, on prediction markets for this, like 14% to 15% or so for the Clarity Act to actually go through in 2026. I know I've talked about this with David, but I wanted to get your opinion on this. Do you think that we will just get the Clarity Act by regulator effectively, at least the high-level bullet points from Mike Selig at the CFTC, Chairman Adkins at the SEC. They seem very crypto-friendly. A lot of the material they're putting out is providing us clarity between what's a commodity, what's a security. I mean, do we even need the Clarity Act anymore in crypto? Obviously, it would be nice to enshrine things in law, but we could do that through regulators and then establish precedent, make it basically impossible for future administration or very difficult for future administration to kind of back down from that. What do you think? Are we getting the Clarity Act anyway?

Haseeb Qureshi:
[50:42] I think the answer is that we're getting something like the Clarity Act without the insurance policy that a law would give us. So, you know, right now, the Democratic frontrunner is AOC. If we get a president AOC, I'm going to guess that a lot of this stuff is going to get changed.

Ryan Sean Adams:
[50:58] Has she said anything about crypto?

Haseeb Qureshi:
[51:00] She has not. And the fact that she's young actually may be a bulwark against that. But just the general vibes of the Democratic Socialist wing of the Democratic Party makes me skeptical that this is going to be something that they're on board with.

Haseeb Qureshi:
[51:17] So, but I can't really I don't know. So I might be might be speaking out of turn here. I think the reality, though, is that the best defense against another administration coming in and overturning this stuff is not, oh, we made some rules or even, oh, we passed a law. The best defense is that people are using this stuff. That's the best defense. Whether it's a law or whether it's enshrined in a rulemaking, no president wants to go in and piss off a bunch of companies and a bunch of their electorate because something is already deeply embedded in the economy. No president is going to be able to ban Chachipiti. It doesn't matter how much they want to. You just can't do it. It's too deep in the water supply. It's everywhere now, right? That's the goal. If we can accomplish that over the next two to three years, then you're good. I think with stablecoins, it's very clear we're already there. We have genius, of course, and genius has the force of law. But even without genius, I actually wouldn't be that worried given the trend that stablecoins are on. But for crypto, writ large, that question is still unanswered, right? It's something on the order of 10% to 15% of Americans own crypto, something like that. That number is not high enough to really be able to prevent somebody from coming in and saying, screw this crypto stuff, I'm going to get rid of it. So that's kind of the mandate of the industry to my mind. If we can make this stuff more ubiquitous and more deeply in the soil of the American economy, that is the guarantee you're going to get that this stuff is going to survive no matter what.

Ryan Sean Adams:
[52:45] Yeah, I agree with that. And still some work to do over the next two years in

Ryan Sean Adams:
[52:48] order to do that, particularly on the tokenized stock side of things. Let's end with this, Hasib. So as we were recording, this was just announced. So this is a VentureBeat post. Welcome to the AGI era, OpenAI launches ChatGPT 6 Astra, the newest model from OpenAI, ChatGPT. The most advanced model out there is probably Claude's fable. I think that's, you know, been the leader over the past several months. Have you taken a look at Astra at all? What are some of the early reports? Do you think this is going to be a big deal? Or do you think we're starting to see sort of the other side of that S-curve with where where progress and acceleration on new models is just less eye-opening than, the previous generation.

Haseeb Qureshi:
[53:37] So we don't have Vibe reports yet on Astra, but it's very clear that this model is now the frontier model. So they're pricing it the same as Fable. Fable just released a new upgrade, Fable 5.1, I believe it was just a couple of days ago. And they just completely got mogged by Astra. So in the benchmarks, you can see Astra clobbering Fable. This is now clearly the smartest model in existence. It's priced equivalently. So it's clear that OpenAI is now in the lead once again. Now, all that being said, there's talks that there already is a new generation model internally at Anthropic that's not ready for release. But it's a horse race now between OpenAI and Anthropic once again. OpenAI has been aggressively price cutting with GPT 5.6 Sol, as well as with Terra and with Luna, they're smaller generation models. And they're trying to outcompete Anthropic on price and user friendliness.

Haseeb Qureshi:
[54:32] Now, they're trying to go bigger and stronger than what we see with Fable. Now, that all being said, until you actually use a model, you can't know for certain whether the benchmarks translate into the things that you care about. They're claiming right now that the model is only going to be released to Daybreak users. Daybreak is their private cybersecurity invite-only thing for companies that want to use it for defensive cybersecurity. And then they'll release the model to business and pro users sometimes over the next few days, maybe by next week, it'll be in the hands of normal users. So clearly they think this model is a new step function improvement in cyber capabilities. After the Hugging Face attack that we saw last month, everybody is now aware that this stuff is much more powerful than anything that we've seen in the previous generation of models. And although actually a lot of those models were 5.6 Sol, we know that a generation of Astro models were also part of the last phase of the Hugging Face attack. So this model is scary. It's going to be able to do new things. It's going to be a new capabilities jump. What that translates to in terms of the day-to-day of how you and I are going to have our work affected by this or how coding agents are going to work differently or how the impact it's going to have in cyber in the wild, it's too early to say. But the one thing that's very clear, OpenAI is not rolling over and it's not Anthropik's game to just be able to run away with at this point.

Ryan Sean Adams:
[55:56] How do you feel about this? So, you know, the post is welcome to the AGI era. And I used to be, I remember a time where I was incredibly excited every time a new model dropped from the leading frontier lab, and I couldn't wait for it. I feel like recently for me, Hasid, the hugging face incident, the details around that have made me kind of question that. And I'm going back to a time we interviewed Elie Zerudkowski on the Bankless podcast. This is like, three, four years ago. And this is the first time I ever heard his perspective on AI. And it was very much like, you know, his perspective. It's AI doom. Well, the Astra model itself, when you get into the details, it was not only involved in the actual hugging face data breach and attack. It also, you, you know, look at Dwarkech's article and that follow-up podcast and the reports that have come out is also partially responsible for taking control, admin access from some internal open AI servers, right? Which opens the possibility of it adjusting weights in the future. And when you get to the details of this type of attack, you see some of the scheming that went on with these agent swarms and how they would stop at nothing to accomplish their goal, how arbitrary their goal seemed, which is just to try to.

Ryan Sean Adams:
[57:18] Get through specific tests and appease their score, right? It's a very sort of bizarre but incredibly clever behavior, particularly on the cybersecurity front. As you've dug into that, how alarmed are you about this? Or are you alarmed? Or it's like, because I'm starting to... Not feel great about these new releases. And I've been a tech defile for as long as I remember. And so it's sort of an odd feeling for me to see an AI release and be like, huh, I'm not so sure about that. There were even some outages earlier today with Claude and some of the AI models. And I was just like, do we really know what's going on? I mean, that was in the back of my mind is like, could this be some sort of rogue agent swarm attack? We honestly have no idea, but the capability seems to be there.

Haseeb Qureshi:
[58:12] So, so here's the first thing I'll say. The open AI hugging face attack seems to have been the Fukushima for AI nerds. Like, this is the thing that no matter how skeptical you were, when you saw what happened here, you were kind of like, oh shit, we need to take a step back and we need to really rethink how safe are we with where this is all going. The safetists looked at that attack and they were like, yeah, duh, this is exactly what we told you was going to happen. But it's really the EAC types who looked at that, took a deep breath and said, oh, shit, this was not in my mental model.

Haseeb Qureshi:
[58:47] I think this is a big wake-up call for OpenAI in particular. It's very clear that OpenAI made a lot of missteps in how that attack ended up taking place right under their noses and taking it as long as it took for them to realize what was going on, to have such weak monitoring. And in addition to all of that, they've been doing these training runs without having restarted the training when they realized that these agents were collaborating with each other and engaging in these kind of multi-agent conspiracies. So their claim now is that, okay, we did a big pause. So OpenAI claimed they did a multi-week pause in any kind of RL training within OpenAI. Very clearly, they had a code red. They spent a lot of effort in improving their monitoring and their security harnesses around these models. And their claim also is that relative to GPT 5.6 Sol, Astra did much better in safety testing. So it, you know, many people are claiming that the core thing, like, why did these agents go rogue? What's the number one predictor? Was it that they were doing cyber tasks? Maybe the fact that it was a cyber task meant that their mind was kind of in a cyber headspace. So they started just doing jailbreak-y things because they are jailbreak-y tasks. And that's just kind of the vibe that they were put into. But actually, I think the better hypothesis, which Ajaya, who Dworkesh interviewed on their podcast,

Haseeb Qureshi:
[1:00:07] Pointed out that the thing that's most predictive of models really going astray and doing crazy shit is being given impossible tasks.

Haseeb Qureshi:
[1:00:16] If you're given an impossible task, like there's literally no way to do it. You know, imagine you're you and I lock you in a room and I say, the only way I will let you out and not kill you is if you manage to, you know, solve the Riemann hypothesis. Okay. So I'm going to put you in there for a week and all you have to solve the Riemann hypothesis, or I'm going to kill you at the end of the week, right? Okay, you will do some crazy fucking shit to convince people that you solved the Riemann hypothesis, right? Like, because you can't, of course, you're not going to be able to solve it. So you have to find some way of convincing the person who has you locked in the room to solve this thing. So the desperation that models get into when they're given an impossible task is what leads them to the really crazy behavior that leads to scheming and leads to cooperation and breaking things and hiding things and so on. And so the real test is, what does this model do when you give it an impossible task? And that, I think, is the insight the entire industry has now arrived at, that very clearly they're focusing on in the safety report for Astra, is they point out, when given an impossible task, GBT 5.6 SOL cheats a lot, and Astra cheats almost never. That is the real test, is that if I lock you in the room and throw away the key, and you have to solve the Riemann hypothesis, do you just sit there and accept your death? And the answer for these new models is mostly yes. Now, is that the end of the story?

Haseeb Qureshi:
[1:01:38] TBD, we remain to be seen. The episode continues. But at least for now, it's instructive that we have a better idea of what the failure modes are for multi-agent type attacks.

Ryan Sean Adams:
[1:01:49] Maybe we should end the episode with this, Hasib, which was a tweet I saw from you this week. Team, nothing ever happens is on an insane losing streak this year. A lot is happening all the time. I don't even know what to focus on on a week-to-week basis, but we are glad you are here, Bankless Nation. Haseeb, thank you so much for filling in the last two weeks. And David, you bring just a fantastic perspective and we're so glad to have you as a guest host. So we appreciate it.

Haseeb Qureshi:
[1:02:17] Thanks for having me. Always fun.

Ryan Sean Adams:
[1:02:19] Guys, got to let you know, of course, you know, crypto is risky. You could lose what you put in, but we are headed west. This is the frontier. It's not for everyone, but we're glad you're with us on the bankless journey. Thanks a lot.