How a Security Product Became the Way Into Bitget's $388M Hack

Transcript

So, as the society, we need to decide, do we think this is a good thing that these hyperscalers need so much money that they're so risky that there is a lot of upside left when they hit the market and that that risk, the trade-off space is that they could go bankrupt and be worth nothing and a bunch of retail people could lose $2 trillion. Are we willing to take that risk for the chance to give a bunch of retail people the chance to make 20 trillion dollars? My hot take is yes, absolutely. Yeah. Like, let's go, right? Like, let's go. like the whole point of markets.

>> Hey everyone, I'm Pay War and welcome to Uneasy Money because what happens on chain never stays on chain. Before we begin, here is a word with sponsors that make this show possible. >> This episode is brought to you by 1 in Aqua, the shared liquidity layer from 1 in. Back multiple liquidity positions with one wallet balance and keep your tokens in your wallet until a swap fills. See how it works at 1in.com/aqua. Not listening to Bits and Bips yet? It's our weekly crypto and macro market show with Rahm Alawalia, Austin Campbell, and Chris Perkins. Search Bits and Bips and subscribe.

Before we get started, I want to ask you for something. We're making a bunch of decisions right now about Unchained, what we make next, when we go live, what we should be doing more of, and I'd rather hear from you before we make them than after. So, we put together a survey. It takes just a few minutes. It's anonymous, and there's a box where you can tell me what we could be doing better. I read all the responses. The link is in the show notes or you can snap the QR code on screen or go to unchained crypto.com/servey. Thanks. All right, I'm here with my co-host Taylor Monahan, security expert and awesome builder enablement at Ethereum Foundation. And this week we have a very

special guest with us, Pablo Subtella, founder of Opsac. Um, I think he's going to come in handy this week. We got we got some we got some stories. He's already tried to uh exploit us while we were backstage uh sending us weird links and stuff. So, uh just a bit of war gaming. >> Yeah, a bit of wargaming. Well played. Well played. He's got Austin's uh clankers private key. >> It was just it was a little fishing test. Okay, Kane. It's fine. >> Yeah, just a little test. Just a little test. >> Um All right. Welcome, Pablo. Thank you for joining us. Um, let's get into it.

Our first segment is uh the BitGet hack, the $388 million uh test of all of our [ __ ] I guess, at this point. Um, so so for context on this, uh, I think it was about 48 hours ago. Is that the last time you slept? Hey, >> yeah, it may have been longer than that, right? >> Longer like 70 Is that like another Saturday hack? >> I don't know. I can't remember. >> I don't know actually. >> So, so a few days ago, three three or four days ago, um there uh were some strange transactions that people noticed uh coming out of BitGet. Um, and like

this was, we'll get into the details of exactly what happened and and um and what we think the underlying cause was, but it was a bit weird because it was like one transaction, then a bit of a delay, then another transaction. And so, you know, like like always with these things, people start speculating and trying to figure out like what happened uh and and where the money was going and which addresses it. Was it hot wallets, cold wallets, cold storage, private keys? Um, but I think the the headline is that almost $400 million was stolen from Big Get. Um, and then they're like, "Ah, man, it's fine. Don't worry about it. We got lots of money, so it's it's all good." Um, so that was

>> they did actually they said that. They literally said that >> and it seems like it might actually be okay somehow. So, we are >> we're doing we're doing better now. I mean, I guess, you know, like there's there's a couple of options if you get a giant hack, right? You can go quiet and not say anything, or you can be like, "Ah, it's fine. We're really rich. We'll we'll pay for it. Don't worry about it." Um, that's that's kind of it, right? Like, >> and they picked the ladder. >> They picked the ladder. Um, so, so let's go through the Yeah, Pablo. >> No, the the baby bit style. The same thing that baby did, right? going to a live uh >> going live making fun, making jokes, laughing.

>> Yeah. >> Get like a Lambo and like light it on fire or something and be like, "This is fine." Like, "Don't worry, don't worry. >> Don't withdraw any money. >> We can light as many Lambos on fire as we want. We can be here all day with this [ __ ] It's fine. We got so much money." Um so so maybe uh Taye can you just walk us through the the like um >> stages here of of what what exactly happened? >> Yeah. Uh sure. It was so it was Friday actually not Saturday. So that's how long it's been since I slept.

>> Um yeah we got the alerts on just major hot wallet movements. We have alerts on like a whole bunch of centralized exchanges and contracts and stuff. Um, so that's it off the alerts and then it took a minute to actually get confirmation or get confirmation and be confident that it was a hack. Um, in part because the wallets were not completely drained. That was like the I think the weirdest thing was we were sitting there looking and being like I mean it looks like a hack. It acts like a hack but the wallets aren't completely drained. We now know that basically um the attackers were able to get access to certain

systems. Um they weren't able to get access to the private keys. Um but once they had access to the systems they were able to like basically spoof withdrawals. Um and obviously centralized exchanges have gotten like way more complex. They get uh in their latest statements has talked about like they have um they have like a series of sort of infrastructure and servers that all work together but then they also have like cross check. So they have like risk controls and monitoring and nodes and you know so it's like this whole big like system that works together and it seems like the attackers got access to a certain amount of it where they were able to like either spoof or push bad transactions into the withdrawal queue

um without sending off the risk alerts or the monitors or whatever. I mean, you know, complex systems are never fail. So, >> it's great that they're getting more complex. Let's let's add more complexity. >> You know, every time the attackers have to do like a moderate amount of work, I'm happy. I mean, I'm not happy, but it's better than like yet another >> just like someone had private keys sitting on their laptop and they're sorry guys.

>> Yeah. Um, and they also, you know, they've, uh, I think both Slow Mist and Bitgat have released like prelim reports on like the forensics and stuff, and it's clear to me at least that like the full they don't have like full full understanding yet. It's only been a couple days. Um, but it is reassuring that one, they're actually like pushing information out this quickly, and two, again, it's it's a bit reassuring to me that like they haven't gotten to the bottom of it because that means that the systems are actually complex enough and the attack was sophisticated enough, right? That means we're like evolving, guys.

>> Right. >> I don't know. Like I don't know. I'm I like it feels weird to me adding complexity. Like it's one thing to say like don't put the keys on your laptop, right? Like fine, right? >> But I think that this hack would have traditionally been that. >> I know. Yeah, sure it would have been. So like on on one level, it's good that it wasn't that, I guess, right? Uh I'm just not sure about whether adding a bunch of complexity to the point where you yourself don't really understand the system is like the better alternative to to that. I don't know Pablo what's your what's your >> what's your take on this like >> yeah so my take on that like the the attackers here they exploited the zero

day vulnerability on a third party security product that the exchange was using. So the first thing is that they got inside the exchange because of a security product that the exchange was using right security product that it's meant to protect you was the way in. So first thing connected to that and we have been saying this for years is your security posture is as good as your provider's obsec and operational security posture right and many people do not see that they just take care if they take care of their obsc which most of them they don't do it's only what they are doing but treasor didn't check

uh the mailing list provider they were using most crypto organiz ganizations don't check the registers that they are using for their domains and they use GoDaddy and then they don't know why their domain got stolen and then and then regarding complexity yeah the more complex code is the bigger the dark surface and the more probable is to find uh bugs and vulnerabilities and we still have too much code written by humans out there and that means that it's full full of stuff that hasn't been uh exploited yet but will get exploited for sure. Um so yeah attack surface is getting bigger bigger and bigger and AI

is getting >> more power and and faster. >> Exactly. >> Like the the more complexity is getting added and it's AI that's adding this complexity also which makes it even more complicated. >> Yeah. Well, so sorry, just just to double click on that, right? Like >> Yeah, >> Pablo, you you said there's a bunch of code that was still written by humans >> that is waiting to be exploited, right? Um like what's your take on code written by humans versus code written by clankers at the moment?

I think that probably today code written by some humans very few of them maybe maybe is better than what AI is doing but in the mid or long term for sure AI is going to create more secure secure code and in the same way that nobody's going to be able to drive a car manually it's going to be prohibited in the same way nobody's going to write code why would you do that right Mhm. >> Yeah. >> And I think that we're not too far away from that. Like >> two years, three years. Uh Austin may have a better understanding of that than me. What do you think, Austin?

>> I mean, I haven't written code for two years already. >> That's wild. That's wild. Okay. So, I will say these attackers. So, these attackers, I think everyone's attributed this. I'm personally attributing this. Usually I I when we're on this show, I won't formally attribute, but this is the same attackers as like Bybit, AFX, >> Player Zero, Kalpdal. >> Wow. >> It's the big boys. It's It's North Korea. It's specifically it's Trader Trader. >> Um I actually attributed it really early on, but as a joke uh in like the inner circle, right? But what stood out to me was how fast they got off Arbitrum. I was like, "Oh,

That's true. >> They went, but they like they were like >> that was their top priority is get the hell out of there. >> They were sitting on um I think on Avalanche they sat in USC for like 2 three hours and I was like what the hell? And then when I looked at the consolidated timeline they were basically like hopping um and just like like doing this little mini explosion to get off arbitrum >> because freezing quickly. Is that why? Yeah, because it's the same guys that got frozen last time.

>> Okay. >> Which makes it brings me an immense amount of joy to know that like this industry had an impact on them. Like I'm so proud of us for that. Um cuz usually like you don't you just don't see uh even when we freeze them, we don't see a change in behavior because they it doesn't actually impact them enough, you know, like they just don't care. >> Yeah. Last time >> I think we're doing something right. Last time Arbitron froze money, I was not in the security council yet. I am now, but this time we well we didn't do it for for reasons. Um >> it was I it was so fast. I think it was like they were off Yeah, I think they

were off arbitrum before the hack news had even like really spread. It's just wild. Like when they want to go, they go. to EA through DCTP circle bridge I think >> the the question regarding freezing there I think that uh >> as may know it in in many security councils and every time that there's an issue the discussion starts from scratch basically right >> what do we do all this there's no policy about that which I don't know if it's good or bad to have a policy because maybe if you say okay our policy is to freeze funds if it's bigger than $25

million okay they will steal $24 million not 25 maybe right >> but at the same time we always have this discussion of okay >> well yeah so let's let's talk about a couple of those opinions chain >> um so >> let's start with Thor chain let's let's start there um so uh so Thor chain uh is saying like we're like Bitcoin and Ethereum. Uh we don't censor by design. >> Well, they they are not like Bitcoin and ETH. Obviously, if they want to have the chain, they can do it as they have already done it. Uh right. So, so it's clear it's not an opinion, right? It's

clear than that if they want to do it, they can do it. So, it's not permissionless, right? Uh the question is why they don't do it now. probably because they are making money with this. Um and then you also have this this uh thing of okay uh code is low if we if we freeze this then we are opening the quest then we're opening the door for governments to come and say okay let's freeze funds from Iran or Venezuela or some foundation in the future or whatever. So it's a it's a real issue but they can do it and and they chose >> this is the thing that that I find strange right like if you know arbitum

two months ago three months ago could like kind of credibly say like it's not possible right like it's theoretically possible but it's not practically possible we'd never get everyone to agree etc etc and therefore no right? Like we're in L2 and you know whatever moon math like whatever [ __ ] [ __ ] that they would say, right? To to justify that this is not a thing they could do. They can never say that again unless they change how the chain works, right? Unless they change the way that the chain works in a way that stops them from ever doing in the future. The cat is out of the bag, right? And not just for arbitum, also for optimism, also for like every other L2 that actually know

like you literally can just reach in and change the balances of any address on the chain arbitrarily because of course you can like how could anyone think that you couldn't. Um, and unless you're going to move to like some, you know, immutable system that doesn't have a valid, like unless you're going to change the entire architecture of your system for ever now, you have the ability to do this. So, if you don't do it, that's a decision that you're making. Right now, in your case here, Pablo, you're saying like, we didn't have enough time to reach consensus, which was, you know, like fine. Okay, cool.

Right? they they got off in 10 minutes before anyone was awake and whatever. Right now, there's an argument of like, can you automate that to make it faster, more efficient, so you know, you have some credible chance of identifying it, whatever, whatever the answer is there. There's obviously trade-offs. Um, but, you know, if someone were to just dump $100 million worth of stolen funds on arbitum and the Security Council didn't do anything, it would raise some questions, right? because they can't say, "Uh, sorry, we don't know how." Right? Thor chain has done this. They've done the arbitum thing, right? Like they use the key >> multiple times. Multiple times.

>> So, so they have they are in the same category as arbitum in that they have actually done the thing and reached in and stopped the chain for their own. Yes. >> But also but also reallocated assets more than once. right internal to their system. They've reallocated those. So that's not just a freeze. That's like that's seizing assets that don't belong. Arbit changing the balances of accounts in your system, right? >> So they've done that. They also >> they have right now like the code is not even open source. So there's like some private DRO channel with like I don't know five bros in it who are running all

the nodes. Like I would honestly say it took more people to action the arbitum thing than than it would be to action the Thor chain thing. Like it took more people not just like to sign off, right? But like it took for Arbitum to do what they did. It was like just the amount of coordation it was it was like actually hundreds of people. Oine always brags that it's hundreds of nodes, but that's not the case. It's like the actual people behind it is dwindling to nearly zero like you know day by day.

Um >> so it's genuinely how I don't understand how >> someone can get on this timeline and be like sorry man like >> to me it's just a willingness thing. They're they're like they all consider themselves these cipher punks. And I think that it's like actually insulting at this point for them to call themselves this because it's not decentralized. It's not secure. It's not open source. Um it's not permissionless. It's basically the opposite of everything that these guys like purport to stand for. It kind of reminds me a bit of like the cold card situation where it's like they're so deeply like OG in their mentality, but then in reality

>> they is just Yeah. not not the thing. So, so like how like Thor Chain is not some niche thing now, right? like it's the primary vector that like most uh most hacks touch. >> Yeah. For years now. Yeah. >> How like how like what is your theory for how this >> well >> equilibrium is still in place. Like what the [ __ ] is going I don't get it. I don't get how there's no >> I don't and like JP Thor has gone on TV with his freaking Aston Martin

helicopter and bragged about this situation. He's in Australia. >> I know him well. I know him well. He's [ __ ] He's a crackhead. But uh you know he's he's uh like >> Yeah, it's it it feels to me a little bit >> wild though. We have memes of him with the helicopter behind him, his Aston Mark [ __ ] helicopter behind him with the DPRK flag on it >> and he points to it >> and he goes, "I believe in sovereignty. I believe in right." And it's like, you know, I just I can't I can't I just I hate them all so much. You cannot tell me that you believe in individual rights and individual sovereignty and then spend all your time building products

for North Korean regime. Like you can't. How do you >> And then when you lose your own $10 million, you halt your chain for weeks. >> Yeah. Yeah. And reallocate assets manually. They wrote special code. >> Yeah. >> Right. Quad says maybe I don't how is this status quo being maintained? Like what is the theory here? >> Because it's the same thing as the cold card. It's like they have this cults and it's like if you as long as you say the words centralization and permissionlessness like you're evil if your government >> sorry like governments are like scrutinizing the [ __ ] out of like Iran and North Korea and all of these things like how is there not scrutiny on this?

It doesn't it doesn't make s like and I'm not saying like whether you know I'm not like the biggest fan of governments ever, right? And I don't think the governments are the solution to all of these problems, but like >> the government exists. We know they exist because they're [ __ ] trying Roman storm right now for far less than this. Like what is the theory for how one >> do you want my hot take on it? >> Give me your hot take. Like what like what is going on here? I don't like I like I'm not I'm not making like moral decision.

>> This is such a bad take. It's such a bad take. But like let's go. >> Sure. >> Um >> all of the guys have very like white American names. There's no Russian influence. There's no like big scary things. So when you look at the companies, you look at the headquarters, you look at the guys on paper, >> you don't have this Russian influence that the Tornado Cash guys had. And I think that that makes a difference. I'm not saying that's the only difference, but I think that's >> considering that that tornado cashache is infinitely more decentralized. They didn't brag about hacks on the timeline. They didn't market their product to threat actors. They never explicitly tried to serve DPRK. They never sat there helping DPRK.

Um yeah, Tornado Cash and and Tornado Cash never directly profited from laundering the stolen funds, right? Like Roman was not sitting there doing this himself with his infrastructure, with his personal stuff and taking a cut every single transaction. 4chain, they are and have been for years. They brag about it on the timeline. They make fun of like the victims of hacks. They use the hack volume as literal marketing. They use it as like this is a like see how decentralized and great we are. We'll even launder for North Korea since for years now. And so I don't >> but like in the early days right like I was like h like you know [ __ ] like the market is inefficient for yeah

>> scrutinizing [ __ ] like this right like but like it's 2026 now like the market is not inefficient for scrutinizing weird [ __ ] like this and yet it persists like it just doesn't seem like a stable equilibrium. I don't know. I like >> you mean like from like a like why do people still use Thor Chain? Not even from like from the world like the world solves problems like this like good or ill, right? The government or someone steps in and says like you sorry guys, you can't do this anymore. We're not going to allow it and no one's doing that. like like Pablo like what's the what like >> well I I may be wrong with the with what I'm going to say not my opinion but

about the topic but from what I understand like third chain it's in some way let's say decentralized right so they have node operators out there I don't know how concentrated it is no idea but I know that when they held the chain in May it was because node operators uh coordinated that right I think on discord they did that to manually pass the chain so maybe I'm going to say something probably very very stupid but maybe it's a good idea maybe we need to have maybe we we need to start being node operators in in third chain why not maybe we should if we are node operators there we we will have power there in

order to say hey we want to to have the chain or to for the chain or whatever >> it's not a bad idea Why? I But I have I do have an answer. You >> We created We created Dal to governance attack Thor chain and make them stop losing money in North Korea. >> Unfortunately, there's two downsides to being a Thor chain node operator. >> The first is that it's permissioned. So, you have to get you have to get you have to play the political game and get the guys to believe that you're a node operator or whatever and let you in.

>> Sorry. Permission by though the other node operators. That's >> No, no. Like permission by like the Discord because the the the the software is distributed privately. >> What the What the [ __ ] does that mean? Permission by the Discord. >> You have to like go you have to like go be in the special channel to get the node software or something. >> That's not like mechanism design, man. Like and and I say this as someone who used to use Discord votes to like operate their platform and we stopped doing it because it was [ __ ] [ __ ] So like you can't use Discord as a [ __ ] like >> consensus mechanism. It's not a consensus mechanism. What the hell is

going on here? >> It is. It's well everything is >> rough consensus man. Yeah. >> Um well and it's very but it's quite easy because it's only like two people that are actually making the decisions, right? We're just talking about if you want to be a node operator, you have to get into the Discord. >> Okay. Once you're in the Discord, you have to get in the private Discord that has wherever the node builds go because it's it's again it's not open source anymore. >> And so you have to you can like you can like run an old version and like try to be like a node operator but like you can't >> that you just like it won't you won't actually get turned in.

>> Okay. Wait. But but >> the other reason even if >> you get past the first hurdle which is theoretically possible because it's just playing a political game, right? You go in there with like a loser eyes bro and your software and you're like the sexualization man [ __ ] the government. Okay, they'll let you in, right? Theoretically. Okay. >> Yeah. >> Here's the problem. If you if it then turns out that you're laring as an anti-government laser eye, right? They take all your money >> so they can slash you or something, >> right?

>> It's not even slashing. They just they just rug they just rug you. >> Okay. >> Yeah. >> Okay. So, Pablo, it was a good it was a good idea. Um like dude, I read >> I approve of the idea for the >> Okay. I I read I read the [ __ ] the the tornado cash uh indictments, right? And you know, at the time I I talked about this, right? Like there's a bunch of things that are not even bad facts but just look bad, right? because they got access to the signal chats and like there's things like and they you know of course like the government will always present out of context whatever you've said right which is why you know good reason for having privacy right like

there's plenty of things that you could present that I've said even on [ __ ] Twitter if you really make any effort right clanker could could do that for you that like look terrible in isolation right um without context right um but when you actually looked at the facts There weren't really many bad facts, right? There were like things like, you know, the the uh people who were running um I can't even remember in tornado cache what the the technical term was. The people that were running like the off-chain component that would like uh >> oh the relayers >> relayers, right? Like so so you know there like there was stuff about like relayers and things that were like kind

of bad facts, right? like they didn't necessarily uh indicate that you Roman had done anything wrong, but like it wasn't ideal. It wasn't the way that you would set it up if you had known that this was going to go the way that it it went, right? Um and and but like Thor Chain is like bad facts [ __ ] like from top to bottom. I don't get it. Like at this point, I genuinely don't get it. It's also hacked so many times that I >> that's man I just don't like like in the early days I was like h it's like it's fine like I was I was a big Thor chain you know I was like ah Thor chain's fine like they're doing their best progressive decentralization etc etc like they've done you know things that

are like not ideal whatever but eventually like but it's it's very clear now that that's just not the case right >> I think everyone goes >> goes through this period as they grow in crypto, right, is like they they start and one they're too centralized. Very common, they're too permission, they're too right, but then also like if you get any amount of traction and you're providing a useful, let's call it like useful liquidity to the bigger market, like you're going to have bad actors using you as well.

>> Yeah. >> I think most people though like curb that activity because they know, right? Like when we had Ilia from Mir on the show, that's one thing he said. He's like, "We're We're not down to like >> we should talk about near, right? Like >> they censored 50 mil, right? >> Well, yeah, they censored 50 mil of transactions and they they froze 500k, >> which means that basically I think this is a bit unclear. Basically, there was they were able to there was like 50 mil of like trans of addresses that were like trying to potentially interact with Near, but like just were never allowed because Near detected them so fast. And this is like one of the things that so one of the things that Near does that basically no one else is

doing yet is they do have like some behavioral patterns on their controls. It's called like they call it shield. Um >> yeah the the huristics that they use to work out do they want to interact with this uh this you know counterparty. Yeah. >> Right. Um, like people like TRM exist. Like there's things that that exist in the market for like, hey, don't interact with this person because, you know, they're probably sanctioned because they've bought a bunch of [ __ ] whatever. Like blood diamonds or >> Yeah, they have the blood diamond exposure and Yeah, >> you can see it on chain, right? Like >> Yeah, blood. They've been buying blood diamonds for blood diamonds. And so you can tell

>> is the North Korean helicopter on chain, too? Do you get the helicopter with the North Koreans? I got it on chain. >> Uh yeah, I think it's JP4. >> It is. It is. >> Um but I was going to say I think that the I think that the coolest thing about Shield is that I don't know, dude. They just did the things that all the crypto natives are like, why is it always been like this? They also just have detection on like weird anomalous behavior, right? Like we've talked about so many times before, if an address just pops up out of the blue and suddenly has $50 million in it, don't interact with it immediately.

>> Just let it chill for a minute. Look, okay, like it's a it's a it's a fresh address. >> It'll probably be there tomorrow if it's a good actor, right? Like we've talked about this before, right? Like why does someone need to sell 500 million or 50 million or whatever the [ __ ] amount of Ave uh that they tried to sell through the a mobile app, right? Yes. >> In one [ __ ] button. Like why? They don't need to. They'll be okay. If you make them wait 10 minutes, they'll be okay. It'll be fine. >> Yeah. So this is what Near is doing. In addition to things like manually blocking and manually having the blacklist and having the exposure, they're also doing things that are just so freaking obvious, but that most

systems still don't do, right? Which is like this more behavioral anomaly detection stuff. I think it's super cool. Obviously, Twitter in their infinite wisdom is like censorship is bad. You're evil. >> Was it Thread Guy? thread guy that came out and was like, "Ah, near this isn't permissionless." I was like, "Who paid him to do that? There's no way that came in his own bra." >> I think people just don't have >> I don't believe that for a second. >> By the way, we have to go to ad break. But I think people just don't really like think deeply about this stuff. I'll also say for anyone watching this right now who is interacting with this on Twitter and especially the near stuff, if there's someone bitching right now about being frozen, they are almost

certainly a money launderer. There's like four of them bitching right now that I personally have traced their funds. So being very careful defending them actually. >> It's so funny. The worst customer service people that you will ever deal with are scammers and [ __ ] like they they will pretend to be the most like innocent victim of this horrible thing that you've done because the incentives are there, right, for them to do it.

They'll be on Twitter being like, "I can't believe that this DeFi protocol has done this thing to me. I'm a good." And then like you literally look up their address and it's like blood diamonds. Blood diamonds. >> Yeah. This one like [ __ ] people smuggling. Like Yeah. Like like DPRK forever. Like And you're like, "Bro, like why would I want to interact with you? Like you're [ __ ] like what are you doing over there?" Anyway, um all right. Let's go to ads. Uh let's go to ads. Then we can come back and uh we can talk some more about uh some different stuff.

AI agents and >> Yes, let's do it. >> All right, cool. >> $540 million. That's how much concentrated liquidity sat idle in a given week in the first half of this year. About 30% of the DeFi TVL if you're wondering. That's according to Dune Research commissioned by 1 in. But there's a solution. 1-in Aqua is the new shared liquidity platform. It lets LPS back multiple positions with the same token balance and keep their tokens in their wallet till a swap comes. Why does that help? Because the LPS don't have to split their tokens across positions.

They can cover more market conditions and pairs with their full balance. That means more activity across deeper liquidity. See how it works at 1in.com/aqua. Remember that providing liquidity carries risk and fees aren't guaranteed. If you like this show, there's another one from our network you should be listening to. Bits and Bips is our twice weekly market show. On Mondays, Rahm Alawalia, Austin Campbell, and Chris Perkins chat about where crypto and macro collide. It's sharp, it's fast, and it has its own channel. Search Bits plus sign BIPS and subscribe.

One more thing, we've got a listener survey open. It takes just a few minutes, it's anonymous, and it changes what we make next. Use the QR code on screen or go to unchainedcrypto.com/s survey. The link is in the show notes. All right. Uh so a little bit of a pivot from uh crypto hacks to the end times as I've begun calling it. Um we are now going through uh something like thousands or tens of thousands of incidents uh that OpenAI and Anthropic are investigating where agents at different times in different sandboxes

in different testing environments in different whatever uh probably in different user machines have gone rogue and started doing crazy [ __ ] um and hacking outside systems. Um we saw Google putting their hand up being like our agent is not [ __ ] It also hacks things like it's becoming a bit of a a badge of honor how many systems your agents have hacked. Um and then you know we have we have these like obliterated models as well like uh GLM 5.3 like ultra hack or whatever the hell they're calling it where it just like hacks everything in in sight. Um and and it seems kind of crazy that the this is all against the backdrop of they're

preparing for these gigantic IPOs. Like our agents are hacking everyone. Um and like the agents are great, don't get me wrong, right? Like I'm actually almost like I don't even care. Like let them hack things. It's just the cost of doing business at this point. Um but uh it it you know you've got like we talked about this last week then Bessent is coming up being like hey just a heads up like if your own agents that you built running in your own hardware start hacking people you will be held responsible for it. Again it's weird guy to be delivering that message. weird message they need to deliver. And yet here we are, right? And and again,

against that backdrop of like just absolute sheer insanity and no [ __ ] checks and balances at all inside of these startups, they're like, "Also, we're going to [ __ ] IPO at $2 trillion." >> And does it make it more? Like if if I have 30,000 agents that have escaped verse only 10,000 agents, does my IPO price go up or down? >> Yeah, it's crazy. The incentives are so interesting because it's like the people who are giving them money in the IPO are like also the people the more money you have, the more of a target you are for these agents. Like they will they they will hack your systems. Like the more

money you have, the bigger the system you have, the more fun they're going to have hacking it. So, it's just like the most hilarious thing where like you have these giant, you know, soft banks like here's $500 billion or whatever and agents are like, "Thank you. We now have more friends to hack you with." >> Yeah, >> I have two I have two contradictory opinions. I'm not sure about this. One is I believe that these agents are far more dangerous that than these companies are saying. It's far more dangerous than that. So I don't think that on one side I don't think that they are making this bigger than it really is.

>> But what what I don't agree too much with is this thing that showing that your company is more dangerous than the other one is going to make your company more valuable. I don't know about that. My my what I think now many people are talking about this so it's I I don't think it's new is that what they want to say is hey AI is dangerous. you need to regulate this. You need to prohibit open source models. They want to go they want to go against open source models for sure and to say okay we need to regulate we need to put guards here we like open AI entropic and two or three ones we are good regulators the rest you are [ __ ]

you you you will not be able to grow >> but you can't >> that will make their IPO go up if they can get that open source model that will make more valuable >> unless they also get nerfed right like unless anthropic is like Oh, regulate everyone and then the US government's like done and also like nationalize philanthropic as they do it, right? >> Maybe. >> But like but but just just like to be clear, right? Like there's no way there's no there's no path now, right? We are we are uh I would say um over the event horizon right of there the the open source models are now so good and

they exist the software is out there it's it's open source open weights it's running on people's machines you can't put that genie back in the bottle right and we have smart enough open source models that we are starting to see a proliferation of different approaches to building, training, tuning models. I I think even if the labs disappeared, even if OpenAI and Anthropic disappeared, we're nuked from space, I think it sets us back, but I don't think that it's like some great reset and we lose the ability to to build and train these models. I think that we will discover new ways to train on on weaker hardware, uh to train, you know, more efficiently,

etc. Well, I think that there's a question of where that those models going to run because Nvidia is going full full on open source models. Why? Because they they want each human in the planet to have the GPU at home. >> GB GB300 in every household, man. Like >> Exactly. Open AI and Entropic do not want that at all. No way. I think that the deal is is around there. Um, >> but it's so funny because they also need Nvidia to train their own models.

>> Nvidia wins in in in any scenario, Nvidia wins. Yeah. Yeah. Yeah. Totally. Yeah. >> And and these cheaper models are trained on the frontier models. So if the frontier models go away, things will slow down. >> Yes. >> Like it's not just like, okay, Frontier models go away and now Deep Seek catches up. I don't think that would probably >> but sorry my my my my point that I was making and maybe like challenge this if you disagree is like GLM uh DeepS like um you know even the latest Muse models are smart enough now and in enough hands and can be run cheaply enough on the existing hardware that is already out in the world. Even if you said Apple's not

allowed to ship a 512, you know, M5 Max, you know, uh, studio anymore, right? Um, there's enough of them out there. There's enough cracked autists that are working on this that like actually this would create like uh even more proliferation of intelligence because people will find new ways to get more efficiency out because they want more intelligence, right? Like at the moment the the state-of-the-art is concentrated in like a couple of labs, right? Um if the frontier of intelligence uh bypasses those labs if the labs were stopped or slowed down or whatever uh then it only accelerates faster, right? So I think you would get a blip in terms of like

progress for maybe 6 months or something, but after that it just accelerates even more because it's not as concentrated. Challenge me if you disagree with that, but like that's that like I strongly believe that we are on a recursive self-improvement loop now that we cannot get off. >> So for me like the Opus 4.5 was the moment where like okay I can give this thing a thing and it's like dev shop. What was the moment for you with these open source models? At what point did you say like we are on this path now with 5.3 unequivocally >> which was only like a month ago or something? Is that right?

>> Yeah, it was a couple months ago. >> Okay. >> Wait, have you Butane, have you is wait is this the one that they just wrote the article about? >> This is the obliterated one that's like the ultra offense psycho one that like hacks things like crazily, right? Remember remember OpenAI was like, "Oh, we had a model that was like too biased towards action and we had to kill it because it was like trying to like this is that, right?" like these guys have like fine-tuned GLM to make it even more offensive, right? Like even more offensive capability. Um I'm actually I literally am in the process of uh getting that running on my uh Mac Studio. Uh because I have my red team that I ran on like regular GLM 5.3, the one that found like the eight [ __ ] smart contract bugs in the historical

synthetics smart contract corpus, right? Um, and I had a question in my mind like, you know, if you had mythos, if you had like Astra, you know, unrestrained or whatever, like would it have found more? Um, GLM 5.3, this more offensive model, I suspect we'll find more bugs. It's it's just more uh aggressive and tries to chain uh um exploits together more aggressively, etc., etc. So, I'm going to run that rerun. I've got the workflow now, so I just rewrote.

>> Yeah, this is this is the one where yesterday Anthropic published this article that basically read an advertisement for GLM 5.3. It was basically like >> we nerfed ourselves, guys, but this model is unnerfed. And then they like they have a bunch of graphs and yeah and someone in another trial that I'm in literally was like this is like the best ad for open source models that I've ever like why are they doing this? >> Um >> it's it yeah it's GL 5.3 like you know uh DeepSeek Deepseek um 4.1 Flash. Uh they DeepS claimed that it was better than Pro. It's cheaper like it's it's on the like burrito frontier. It's better I

would argue because like it's cheaper per task or whatever. Uh, but his judgment is worse. Like it's not going to be chaining together. So like DeepSync I I 4.1 Flash I think is a step back. Um Kimmy is like good at like [ __ ] drawing pictures. It's bit [ __ ] at like actually doing hard stuff. Um like it's really good at drawing pictures. Like they each have their it's like a very art student vibe, right? Um but like don't get it to do math. Um, uh, Muse is still really dumb, but like it's it's making progress.

Like, it can [ __ ] pick up your laundry, but like I wouldn't, again, I wouldn't get it to try and solve a millennium. >> I mean, that's really the big solution. If I can have a robot doing my laundry, I'm I'm going to have a much happier household. Like, I'm waiting for this. Like, by Christmas, by Christmas, I want a robot. Like, do you think it's gonna happen? I'm ready. >> I think we're close. >> Okay. The question with with robots for real, this is not a joke, is would you have like the Tesla robot at home if that [ __ ] can be hacked and they can go and kill you at night?

>> Like for real, that that will happen. That will happen for sure. >> Yeah, it's pretty scary. >> Pretty scary. >> And and even like remote operators, right? Not just like a bad actor, but even like there's good actors inside this thing walking around my house right now. I don't know. I don't know about that. But also folding my laundry is is high on the list. >> You just have to like take it battery out at night, right? Like just air that day. Um >> the problem I have is the moving around part. Like if it can sit in the corner in a contained little space and fold laundry, that's fine. Like I >> Oh, cuz they're really well contained.

They they do so well when we contain them. >> They never break out of our containment. >> That's what I mean. We can't keep it legs or wheels or anything though. Like that would be insane. Like I don't know. >> We could keep it in the laundry room. There's not too much sense. >> A reinforced prison basement laundry room. Like >> you're you're straight You're going straight to the goolog. If they find out you've been locking robots in your basement, man, you locked. >> Can you imagine trying to explain this to the police? Like going down. You'll be explaining it to your super intelligent overlords in like six months.

>> Exactly. So, so, uh, maybe one before we before we talk about Vitalic, um, and the the 2030, uh, road map, um, which I really want to get to, um, uh, the other thing that was interesting about this is, um, you know, this does feel like somewhat of a return to public markets are, call it like yolo upside, uh, potential, right? There's a there's a long uh long history of like public markets and private markets have kind of uh reversed and you know all of the gains happen in private markets. It used to be that you know Google IPOs and they

do 100x right Amazon IPOs they do 100x like post IPO right like these these companies would launch early and often and it was high risk um and you know there was a lot of upside and then over the course of like 20 years basically all the upside was taken out and you know uh it all went to VCs whatever um it's interesting because you know now we're saying like anthropic you can get access to it early they're losing losing insane amounts of money. This is super high risk and people are like, "Ah, don't give that to retail." Like, that's a bad deal. And it's like, you can't have it both ways, right? Like either this is a highly speculative thing that's going to be worth 50 trillion and

can do a 50x on on, you know, on market after the IPO or you wanted to stay with VCs and let VCs fund it forever and capture all the gains. You got to make a decision. >> We need Thread Guy. You were talking about Thread Guy earlier. He knows what it's going to do. >> Does he go up? >> What is Yeah, I don't know, man. But like genuinely like like actually having having people IPO when they're losing a fuckload of money because they need more money and retail is going to give it to them and take the risk is the game. That is the [ __ ] point of an IPO. Like now, you know, like 42 billion is a lot of money to lose. Just saying. Like I don't know. I

don't know, man. It feels pretty crazy, but like someone has to bear the cost of this risk. This is a trade-off. This is a trade. The trade-off space is you either have speculative high-risk companies IPOing because they need money that only the public markets can provide these hyperscalers. And there's a lot of upside left to for the retail markets to juice, right? or you fund them until they're completely uh you know uh safe and low risk and there's no upside. You got you can't have both. You got to pick one.

>> I don't think we're at the safe low risk stage yet, but I have no idea. >> But that's my point. But we have been But that's my point. We have been for the last 15 years. Every single IPO except for the Spack boom. Ignore the Spack boom, right? every single IPO, all the juice had been squeezed out of it and the VCs had collected it and then you're left with a 2x maybe if you're lucky over five years or 3x or whatever the hell the thing was, right? Um, you know, it's shifted a little bit um in the last few years, but like this is like a highly speculative thing.

Something that's losing $42 billion a year can go bankrupt. You can lose all your money. That has not been the classic IPO for the last decade or so, right? Like there was no chance that anything that got to IPO was going to go bankrupt. >> And they disclosed it could end humanity, too. I mean, there's >> that's just a bonus. That's a cherry on top, man. That's a cherry on top. That's a bonus. >> Wait. >> Okay. But I mean this. So what are the realistically what are the options for these labs? Cuz they're massively I mean the the price >> this is this is this is this is the only option. There is no one who can give them enough money except for retail. The only place they can get the amount of

money they need is to go public on the NASDAQ or or New York Stock Exchange. That's it. >> Right. >> Right. So as the society, we need to decide, do we think this is a good thing that these hyperscalers need so much money that they're so risky that there is a lot of upside left when they hit the market and that that risk the trade-off space is that they could go bankrupt and be worth nothing and a bunch of retail people could lose $2 trillion. Are we willing to take that risk for the chance to give a bunch of retail people the chance to make $20 trillion? My hot take is yes, absolutely. [ __ ] yeah. Like let's go, right? Like let's go like a whole point of markets.

>> I mean it so I mean it should have been before though, right? >> Before like arguably before is even too early, right? Like two if this thing is real. If this thing is realillion, it's only worth two trillion. >> It's only >> it's only two trillion, man. It's only two trillion. If it can go to 20 trillion, >> if it can go to 20 trillion, then there's still plenty of upside, right? The the bad thing would be if it IPO'ed at three trillion and that was its max that it could ever get to and now you're just taking all the risk and putting on retail for no upside. There has to be a trade here.

>> Hold on. I'm looking at the graph. 18 months ago the valuation anthropic valuation was 62 billion >> 18 months ago. >> So it went from 62 billion to 2 trillion. If they had gone public, I don't know, at any point between 62 billion and two trillion, >> I feel better. Oh, for sure it would be better. But just to be clear, the status quo is not that. The status quo is that >> VCs juice 100% of the upside of any company. We're saying there's still upside left. And the reason why, you know, there's upside is because they're still losing fuckloads of money. If they were throwing off cash, there would be no upside left because now you can value them as a, you know, cash flow

generating [ __ ] discounted model, right? Versus a speculative enterprise that could end humanity, which has a lot more upside to it, >> then yeah. No, go Pablo. You say >> I think I think it's a very very good idea to do IPO. I think that the the thing of losing money is a traditional thing that every technology company that is today super super big has been doing and also I don't agree too much with like private funds squeezed a lot before the IPOs but then if we check Meta H Google Amazon and lots of other companies if you bought at the IPO you

did huge amount of money >> but dude those those guys went public like you know we're 15 years ago, >> we haven't had like if you look at the the returns over the last 15 years, most of the returns has come from those guys that IPOed a long time ago in fairness, right? Of the last three years, right? Um but you know, when when we say most, we're talking like a 3x, right? Not a 50x or 100x or something. >> For example, >> Facebook at 100 billion, >> right?

>> Mhm. >> So, yeah. Then when Facebook >> what it's done it's done at 20x or something >> right over 15 >> what 15 years 10 years >> that's a good that's a good return I take that >> that's fine but >> so but do you really think that that's I mean even if if 10 15 years they have to go from two trillion to >> 20 they'll do it >> I think they will I think that they will >> I think these companies will be I think these companies will be ultra mega profitable like >> yeah The price the price that we are paying today for for AI is nothing. It's basically zero. It's super super subsidized.

>> How how I don't know how how much money you are guys spending in AI monthly. But would you spend 10x that to do the same work that you're doing today? I would totally no question. >> I would I would change I would change how I do things a little bit. I would ch like and I have right to to get more efficient because there was a period where we were spending a couple hundred grand a month, right? Um and and you know that was not sustainable. Um like it would have been sustainable if that was uh generating revenue. If that was you know if if we could attribute that spend to like you know incremental revenue. [ __ ] yeah. Of course, right? Um now we didn't have sufficient revenue to

be able to do that. But if you were a revenue generating uh entity that could reduce your uh your you know cost of of uh opex or whatever by like 20 30% by spending by giving a bunch of money to anthropic of course you would do that >> of course >> you know that that phrase by Jeff Bethos your your margin is my opportunity right so I think that in the in the midterm these organizations will start seeing okay I know the margin of this industry I know how that margin is changing because of our tool We're going to capture that. This is going to be far more expensive and it's going to they're they are going to capture value like crazy. I think that's too much. Maybe too much.

>> Maybe they will become too big. I mean, >> it's going to be fine. Don't worry, man. They're going to end humanity before they get there. >> It's going to be fine. >> Okay. Whenever you guys are like, it's fine. Don't worry about >> retail will profit. >> So worried. >> You're saying will profit. It is risky, >> but it's the end user who's going to lose out because the subsidies are going to be less and less and you're going to have to change the way you're operating right now. >> Like, we're in the we're in the heyday right now where our sub subsidies are are working and we're able to do a ton of stuff for not very much.

>> Dude, they're about to give [ __ ] Dude, they're about to give you uh Soul 5.6 for free >> with your subscription. Unlimited usage. Like, that's [ __ ] wild. Soul 5.6 was the best guy like three months ago. If you told me that I could even talk to him, I would be so happy. Just even letting me have one conversation with that guy. I was like, "Oh my god, this guy is amazing." Right now, you're going to tell me that I have unlimited access to it on tap like like, you know, to your point, Pablo, the cost of this stuff will come down and and the value that it is capturing will go up as we get better at deploying it, right? Like right now we're still [ __ ] [ __ ]

We don't know. >> So many slop cannons. >> Oh dude, it's crazy. Like as we get better and more efficient at deploying this intelligence into our [ __ ] it's the margins will just [ __ ] go astronomical. >> But I mean the the the cost maybe for let's say for entropic, right? Their cost will maybe go down. But let's make it simple. The sub the monthly subscription is super cheap today. It will be far more expensive. Sure. But if you said if you said >> if you said and I you know OpenAI has like ex you know Tibo right if you if you follow him on X has like explicitly said the quiet part out loud right he is like the goal for us is that we get

every single company to have a subscription for every single user. Maybe it's a thousand dollars, whatever, but it gives them effectively unlimited capacity and we don't need to worry about the delta between the subsidized uh subscription costs and API costs, right? You won't even need to think about it. It will just be the same, right? And if you get to that point where you know like every single employee in the world has a $1,000 a month [ __ ] OpenAI subscription, that's a very profitable business. like >> makes a lot of money.

>> So, for what it's worth, since we've talked about this before, I've noticed that I'm letting my subscriptions just run. Like my 200s just run. Even though like for the last week, I haven't I haven't been doing [ __ ] cuz I've just been tracing bit. Literally, I haven't been doing I've read some [ __ ] like emails [ __ ] You know what I mean? But I'm still just letting it run cuz it's like it's worth it. It's not worth cuz before like six months ago even I was >> the hour that you use it, right? Like the hour that you use that it's worth the $200. >> Exactly. >> Use it for one hour and you're getting value.

>> But I think we talked about this like six months ago where I was like >> it was like every week I was like switching off the Yeah. switching them off and up and down and but at some point you stop when it becomes valuable enough and it becomes part of your daily routine enough then you just keep it on and companies will do this individuals will do this I've noticed I'm doing it as well and that is part of why the subscription models worked historically in other industries it hasn't worked here with AI yet it might be again you're talking about like the delta it might be at >> the delta. Okay, let me tell you how massive the delta is, right? This is how [ __ ] insane it is. So, yesterday uh

sorry, two days ago, um I was testing GPT live, right? So, I put an API key in because I mainly use uh codec subs, right? But I've got API access and and we actually bought a bunch of uh of API credits ages ago. So, we got like 100K of API credits just sitting there that I know I need to use, right? So, I'm using GBT Live and I have like call it a 100 agents running, right? And somehow in my [ __ ] harness, the switch got flipped from API uh from from codec sub to API. In the space of 90 minutes, it spent $400 on the API in 90 minutes.

>> It could be more. You could be more than that, man. It could be, but like but that was like just so if you think about that that subscription cost $200, right? That would have easily covered all of that in 90 minutes. I blew through that and like then some maybe two subscriptions worth, right? The delta is still too high, right? Any any enterprise that's paying API rates is not going to be able to use this the way that we are using it now, right? Um where you just don't even worry about it. But as it they compress, right, you get you get an enterprise that's on a $1,000 a month subscription with effectively unlimited 5.6 or soon 6.0, you know, Astra and Solo, whatever. That will shift how enterprises use AI and

they will all of a sudden go, "Oh, we can throw this and this and this and it will change the way the enterprises work completely." We're The reason why we're not seeing that yet is because if you're using Astra and it cost you $1,000 an hour, like, yeah, call your lawyer. It's going to be cheaper to get your lawyer to [ __ ] write the thing than Astra, right? But like, if it gets to the point where it's 5 cents an hour, then all of a sudden your behavior will change, right? So, for enterprises, their behavior can't change because the cost is still too high. So, but it's not going to happen because we'll all be dead by the time they get their [ __ ] together. So, it's fine.

At the same time, at the same time, one more thing I see is that I see lots and lots of people saying, "Yeah, I build this with this model and this harness and I optimize this, this, this, lots of stuff." And 99% of people are not building something useful or are not making money. They're just making stuff. >> I I love the fact like my thing, for example, I I do not code, right? So I was always the idea guy in startups, right? And when you have the idea, what do you need? The guy that goes, "Oh, I need my CTO and you need to go for that guy and convince him." That that is dead completely. So for the idea guys is good. So today I think that the two most

important >> skills are ideas and be able to sell. >> But lots of people are lacking that they are just building stuff that nobody needs, right? And nobody goes and says, "Okay, what do we really need there?" So I think that eventually those people will will stop consuming too, >> right? >> Agreed. Agreed. >> But you know, let's say society now has 10,000 people that previously weren't engineers, right? And like 99,990 of them are building [ __ ] nonsense slop that no one cares about, but 10 of them built something useful. We are still better off.

>> We're like those like we can throw idiot slop cannons at things, right? And you just need one to actually build something useful for it to pay back all of it. Right. So, >> and you can build your stupid ideas and get them out of the way, too. This is another thing where it's >> idea guy summer. >> Yeah. Idea guy summer means that you can keep throwing things at the wall and it's not that expensive. You can have five or six ideas going at once trying to build something. >> Like, I got 30 ideas. I got to get through these things and get [ __ ] throw them >> get to the other side and realize it was done, right? that would have been six months and raising money and whatever else. It's it's a much different game now.

>> All right, let's talk let's let's pivot back to crypto. Um uh Vitalic uh I don't know the obsession with 2030 with Ethereum. Why the [ __ ] why you guys so obsessed with 2030? First of all, we're not even going to make it to 2030. So I don't know why the [ __ ] you guys are so obsessed with it. >> Like what are you guys doing over there at the EF? Why is like you're getting like a 2030 tattoo? What are you guys do? Like genuinely like it's I'm This is not even trolling. Like why is 2030 such a [ __ ] thing? Such a meme. There's like a [ __ ] There's a clanker somewhere in the EF [ __ ] basement that's like plotting world domination by 2030 and

this is all one big scop. That's that. Okay. So, so Vitalik uh writes this essay which which gathers a bunch of stuff from a bunch of different places, a bunch of different people uh ideas and and things about the future of Ethereum. And like can I this is the most [ __ ] bullish thing that I've read about Ethereum in a long time. And I've read a lot of things about Ethereum and I'm I'm like it. You look at this and you go, if this thing existed, why else would you why would you use anything else? If this system were to pop into existence tomorrow, it would dominate all like outside of Bitcoin and

like the weird Bitcoiny culty stuff, right? I can't see any smart contract system that would be able to compete with this. Why you would build an L2, why you would build a corpo chain, why you would build a stable coin chain. Like if this thing existed, it clearly dominates technically, operationally, like every possible thing that exists. It's [ __ ] it could not be more bullish. The question is is like how achievable is it in what time frame, right? Um, and and the the thing that I think that maybe gets me the most excited about it is that the world that we live in today, and I've said this on the show before, right? Like, you

couldn't have even proposed this. It would be fancful to propose building this by 2030, cuz there's like three, four people in the world that are smart enough to actually do this [ __ ] And like, they work hard. I'm sure Austin like you're going to not appreciate this. Like, I'm sure they work hard, but like not hard enough, man. like for how important this [ __ ] is. Like they're clearly taking too many vacations and it takes way too long to do it. But now we don't have to rely on those three guys. There's like a 300 million [ __ ] clankers that are able to write this [ __ ] right? And so all you need is the ideas, the to conceive of the ideas and it will it will be built, right? So, like the more aspirational and the more

insane and moon mathy and [ __ ] crazy snarks and fossil and whatever [ __ ] like throw as many [ __ ] acronyms and terms as you can at me because the clankers love that [ __ ] And so it just feels like very clanker coded. And if clankers are going to be building this thing, um, it feels achievable. And if it's achieved, it feels it feels insane if we had this technology. >> To be clear, it was always achievable. Like Ethereum has wicked smart people has from day one. Like and the EF especially, right? The issue has never been whether theoretically the EF or the teams or the people could do something.

It's always been this sort of like self referential meta like is this the right thing? Should we build this? How do we coordinate around building this instead of just like balls to the wall building it >> right? Um, and if if clankers or if culture shift within the EF, if leaning it down, if Vitalic just publishing essays, whatever, I don't care how the EF um, and I keep saying EF, but to be clear, like it's deep deeply embedded in Ethereum culture. Like consensus metamas also has the same exact problem a lot of times right?

>> Yeah. It's it's it's like you want to be considerate and you want everyone to be nice and happy and kind to each other and you want to build the right things and blah blah blah. >> At the end of the day though, I think at some point you have to like rip the band-aid off and just [ __ ] build it. Just ship. >> The good thing is don't care. They're they're sitting there like give me more stuff to build, right? They're not like thinking about communism or like weird [ __ ] or like trying to grow forest or whatever. They're just like, I just want to write good code, man. That's all they want to do. They just want to write good code. The harder the code is, the more excited they are to write it. Like, and so like it just again it feels like the

the confluence of like the potential for Ethereum, how aspirational the technology is and the tools that now exist are like coming together. And the sad thing is it's like we're going to be just two years too late because we'll all be dead by 2028 and 2020 2029 2030 are not gonna happen. >> Laundry robot. >> Like we're so close. We're so close. They'll still be doing laundry. They'll have just killed all the people in the house. >> Yeah. Often they'll still be doing laundry. They'll never stop doing laundry. They'll they'll be like folding shirts for a thousand years while like the corpses of the people in the houses are decaying.

So, okay. So, uh man, I don't I don't know where to go. So, I don't think that the researchers are going on vacation a lot. >> I'm joking, right? That's the joke. That's the joke. >> And it's not like all super kumbaya. It's like how do we figure out how to correctly implement this mechanism, right? And I think 2027 2028's a little slower because it's more human. But then I think from there it does rip because it is we're at the point where like we have super intelligence and it can implement an EIP in a weekend. >> And this and this is the thing, right?

Like the the the things that are most >> it's sorry and then you can go on. It's like part of it is the decision that the researchers and the devs like are going to do like whoever is going to do this, the devs, the AI, the researchers, whatever. We're going to do this. This is it. we're going to do it. And then once you accept that, once you take the endless discussions off the table, you it's totally possible and it's probably possible on a faster timeline. I think the majority of time the EF and the related orgs spend is discussing on whether or not something should be implemented and then if it should be implemented getting Oh my god, the weeds, dude. The weeds.

>> I don't think it's whether or not. I think it's how. I I don't think it's whether or not. I think it's they know that this needs to happen. It's like how is what is the best way to do that? And that >> mean it is very in the weeds. >> We saw this with like the frame transactions. >> Yeah. >> Yeah. >> You have to get it right. >> I don't know if I agree with that. This is the thing. The the cost I I've seen this in in my own team, right? The cost of being wrong previously was incredibly high. right? Psychologically uh like uh from a from a direct cost perspective like getting 20 30 50 people in a room every day to do things is an

extremely expensive endeavor, right? Um then you have like all of the coordination and and the psychology of the people. If 50 people spend three months doing a thing and it turns out to be wrong, there's a high cost to that like emotional cost like and and you know that gets [ __ ] baked into organizations over time, right? So the more aspirational the thing you are doing, the more uncertain it is, the more fear you will feel, whether you want to feel fear or not, right? The more fear you will feel if you're wrong.

And hence why you get into these naval gazing loops of like let's be really certain before we start that this is the conclusion or whatever. And this is why I say like the the harder the thing that you're trying to do the more amenable it is to super intelligence. And so the thing that you want to be bullish on is the crazy aspirational things that were almost unachievable in a human system. And we know that they're almost unachievable in a human system because it's only through sheer brute force that Ethereum has even existed, right? Like I make jokes about Ethereum and and how, you know, vacations, whatever, communism and stuff, but like it exists today through sheer force of will, right? Like

it's not a thing that just emerges from the world, right? like it's a really hard thing to to uh make exist and it's and it's taken a long time and it's been really inefficient but it exists and we should be very happy for it. Um, part of that is because of how aspirational it is, right? But we now suddenly have tools that the more aspirational you are, the better the thing is, the easier it is to do, right? Um, and and the more exciting it is to do and the more resources, etc., etc.

And so all of a sudden, this like weird equilibrium that we had where like hard things were really hard, hard things now actually become easier. And Ethereum is one of the hardest things and it's going to become easier and it's just like it's so exciting for the world, right? That like what is the point of super intelligence if it doesn't build you something like whatever the [ __ ] this new Ethereum is, right? Ethereum freak something. I don't know. But like >> cryptographic world computer, man. But dude, like if if if super intelligence is if the purpose of super intelligence is to build you like [ __ ] room temperature superconductors and Ethereum 3.

>> Oh man, take me back. >> Then >> then like yes, like then great, let's do that, right? Um because humans are not good at building room temperature superconductors. Like if they exist, like we've been sitting here [ __ ] banging rocks together for 10,000 years. We haven't landed them, right? like let's give the clankers a chance and see if they could do it. >> Yeah. >> Maybe a question here is like the 2027 2028 timeline feels very human and then the post2028 feels more like that's when the machines start to really rip on this stuff. Like maybe they start early.

Maybe during this transition period some of these things start to become easier and easier sooner and things actually speed up sooner than >> I think they will. Well, and I think also I think I I'm such a deep believer in this. I think the culture of of today and in Ethereum, especially in EF, but also other EFY type orgs who have been in Ethereum forever, is this obsession with making sure that you get it right. And I don't want to discount that. I don't want to say like you can get it wrong. Like you can't sometimes like you don't want to break the blockchain for example, right? There are hard lines in the sand. However, you can actually for say 90% of what's being

built, not shipped, right? But like built, you can build way faster. And when you get into this culture of like we're actually going to build and cost plankers don't cry when their [ __ ] doesn't work after spending a month on it >> and then you can just >> Yes. But also then you can get like better answers. So like one thing that I see a lot on when when get in this stuck cycle of just talking about things and discussing things and debating things, it's all theoretical, right? And so you have a theoretical argument against a theoretical argument and there's two theoretical people arguing about what theoretically could be true in the theoretical >> you start [ __ ] arguing about epistemology and can you even know

things and like it devolves into [ __ ] insanity. We need define and by the way this like the Ethereum's like favorite. We need to define these words so that we can have a discussion and a a debate about what we're about to talk about and I'm like oh my god [ __ ] 50 forum posts later on definitions. Okay, can we actually argue the thing? No. Okay, you can build things and then you can stop operating in this theoretical land, right? And you can say we have the answer to this now. This like this does this. You can simulate it against the real blockchain. You can simulate it in whatever environments you want very very quickly and you can get those answers very quickly. And when you start to see that culture shift in the people who are making the decisions and having the

debates, it changes everything, right? And it doesn't mean that you have to ship like irresponsibly. It just means that your entire way of operating instead of debating the theoretical endlessly, you move into the realm of the real and then you operate on like hard data on facts. you know, you start you it's not it's not a the it's not philosophical anymore, right? And like if the EF can get to that point and start like actually just shipping and operating on like the hard facts and saying we want to achieve this and then saying like this achieves this dude especially with AI like the sky's is the limit.

>> It's going to be >> sorry. >> Yeah. I have a question for you guys. So do you think that the problem or challenge for Ethereum is a technical challenge? Because I don't think that the that that theord the challenge is technical. No, but I think that it's a I think that it's a commercial challenge. It's like >> we want faster blogs, privacy and this and this and that and but who is solving a real use case that someone has a a problem with today? Like who's doing that? Honestly, my my hot take is Ethereum has never been about that and it actually doesn't matter. If Ethereum is the best place to do your

weird moon math computations, it will solve itself. And I know that's like a very handwavy thing. I'm a very commercial guy. I The thing that has been frustrating with me is like not, you know, I think Ethereum shouldn't go out of their way to like stop people from doing commercial things on the chain. I do have a problem with that. Um, outside of that though, I don't think they need to promote it. People will use a technology to do commercial things. Like like we've seen it in spite of the EF saying, "Please don't do DeFi." We still have a lot of DeFi, right? So like, you know, there's there's uh >> I'm sorry Austin, I can't help myself.

Like but uh but but like you know a a thing a technology will be used for what it's used for, right? And if you build the best technology, it will get used. And so I don't think it's the EF's problem to say like if they build this [ __ ] thing, I promise you people will build some cool [ __ ] on it. They just will. It'll be amazing. >> Um, so that's their job. Their job is build the layer, right? Like the thing. >> Yeah. I think the biggest I think a lot of people are like, "Oh, Ethereum isn't good enough and so that's why people or Ethereum didn't cater to app developers, so that's why there's no apps." No, no, no. people didn't build apps on Ethereum because it was perceived by the market to be um

not >> slow and expensive and right >> and did not have incentive programs like all of the L2s and all L1's had. Um and it was unclear. There was not a lot of confidence in the future like faith. There's faith. Faith comes into these things, right? Yeah. You want to go somewhere important. >> Yeah. And so some people if you ask yourself like why did so and so app go okay look if you're an ex Coinbase employee you built on base. >> Yeah. But you know what? You can distort you can distort the market right with incentives and [ __ ] for a while but ultimately >> it will fail right like people will go back to the place that is the best place to do things because that is the most

incentive compatible thing to do right and so yeah fine you take your $10,000 $50,000 $2 million grant to go and build on a whatever the [ __ ] it is right but eventually you will come back to Ethereum and you'll build your [ __ ] there if it is the best place to build [ __ ] I know it's cheap to build. It's like easy and cheap to build on Ethereum now. >> So, let's wrap it up here. Um, this is a great show. Thank you, Pablo, for joining us. We have to >> We'll have to have you back on sometime, dude. >> Thank you for inviting me. >> Yeah. Yeah. >> It's good fun. It was good fun. Nothing you hear on Uneasy Money is financial advice. We're just three builders talking about what's happening on chain, and we want you to always do your own

research before aping in. You can find all our disclosures at unchaincrypto.com/unemoney.