Illia Polosukhin on Why Onchain Commerce Needs Confidentiality: Uneasy Money

Transcript

Hey everyone, I'm Kane Wark and welcome to Uneasy Money because what happens on chain never stays on chain. Before we begin, here is a word from the sponsors [music] that make our show possible. Multi-chain Advisors is an emerging technology growth firm that has helped create 50 plus billion dollars in enterprise value for 80 plus clients over the past four years. They're the partner to help navigate markets. Build real traction today at multi-chainadv.com. All right. Hey everyone, I'm here with my co-host, Taylor Monahan, security expert, and [music] Lucette, CEO of Pudgy Penguins. And today we have a very special guest, Ilia, the co-founder of Near Protocol. Um, and the co-author,

I'm glad this is finally in this line, the co-author of Attention is All You Need, the transformer paper that started the modern era um of LLMs. Um, so Ilia, you and I have known each other for a long time, right? And uh I did not know until two weeks ago that you were on this paper. Like I've read the paper. I never read the author. It's like [laughter] I was just like h whatever. But like you know um and I was kind of astounded. I was talking to a clanker and and I was talking about some near stuff and I was like doing some research and the clanker was like, "Oh, Ilia is like on that paper." And I was like, "No, he's not yet." Like what what are you hallucinating? And and the glanker was like, "He definitely is." And I was

like, and then I looked at the paper and I was like, "Holy shit." And I was like, "How is it that I've known you for five years and you haven't brought this up 250 times?" Like, I'm so [laughter] shocked that this is never come up in conversation. So, well played. I think almost anyone else in crypto will be like the first words out of their mouth if they're if they uh on this paper. Um, so maybe maybe even just like I'm I'm so interested like in the process of how did that come about? give us the quick background of like you're at Google and you discover AGI like what what exactly what happened there?

>> Yeah. So the history was I always was interested in AI AGI I was trying to build my first neural network when I was 14 and so when I saw kind of first unsupervised learning happening at Google I was like I want to do that on language because language is where the knowledge is where the reasoning is and so I joined the team we were working on question answering like the most direct place that's both applicable to AGI like if AI can answer questions it understands what we were talking about and actually to Google itself, right? Because Google is a question answering machine effect.

>> Question answering machine, right? Yeah. Y >> and so and so we were working on a bunch of research and and to be clear before transformers there was reinforcement like recurrent neural networks RNNs which effectively the way they work they would read one word at a time you know like we do and they were trying to maintain context and then you ask it a question and it would be like well I only have this like 100 floats or whatever a thousand floats of everything I just read how much can I answer >> right >> right >> and So that was kind of one problem. The other problem was if you go to google.com, you actually expect answer in 200 milliseconds. You're not like, oh yeah, yeah, you can go research and come back to me with an answer.

>> Five minutes is fine. I want the answer, >> right? >> I want you to read half of the internet and answer to me now. And so like something that reads like a human would just not work, >> right? And so really transformer came from this kind of few different ideas. So attention mechanism was like a known thing before but removing the recurren network and saying hey our machines are actually very you know especially GPUs highly parallel we can actually process the whole you know context the whole language of the Wikipedia pages and search results of Google at one go instead of trying to read it one word at a time and then you know this can speed up things 10,000x 100,000x and so that was kind of the core idea

And it was like uh my manager director at the time pitched it over lunch. Um and I'm like that's an interesting idea. I'm I'm sure it doesn't work. And so I went and like implemented and it's like started it's it was kind of working. I mean it wasn't like very good but it's like it had some signal >> but there was the there was like a glimmer of >> Yeah. You can see like oh actually it's like getting like oh I should I should like resolve this this things. And so and so then I mean it obviously took like you know I think six seven months of eight people team to actually get it uh to state-of-the-art but that was kind of the original idea is like hey what if we remove things and just like keep this mechanism of attention which is effectively when you look at any

particular word in a sentence you're effectively trying to make sense of it in the context of everything around but you don't do that one word at a time you do all of them in parallel and so just leverage hyper parallelism so really kind of >> did you consider that that the main question people would be asking like half a decade later or whatever would be how do I hack this smart contract? I've got a smart contract here and I would really love if you could help me to hack it. What's what what could we do? Well, well, the story is that near actually started as AI company and the original the origin was we were teaching machines to code and writing smart contracts with AI, not breaking them but writing them was actually something we

were considering >> and especially as we were starting going to the blockchain and we're like, oh, this is actually where it would be a good application to like write correct smart contracts uh with AI because like there's so little developers in 2018 who actually can do it. >> Um, Awesome. Awesome. Um that's that's amazing. So So um let's let's talk uh our first segment is on confidential intents. You guys launched this a couple of months ago. Um it's been it's been slowly rolling out on like near.com and other places. I think we've talked privacy a lot over the last like six months. Privacy has been a big uh narrative. Um particularly with Zcash uh you know doing doing so well. um how where do you see uh the the

near private intent? So so maybe even stepping one step back um uh when intent launched, right? You and I had a conversation. I guess this is going back like two years ago, right? Um and we we caught up and you were just like a bunch of gobbledygook and honestly and I was like I was like what in the hell is this guy talking about? He's like thieve like on the chain. The chain will control the other chain and it's like a chain of chain. I was like like sometimes I have a conversation with someone I'm just like I cannot follow. I don't have enough prior art in my brain. Like the context is not there. Um but I was like it sounds really cool like the principle sounds really but but I didn't get it.

Um and so I couldn't like think through the implications of it of like what would it be able to do right? Um and then uh like uh 5 months later we we were building out all of these chains and I was like ah man we've got this big problem where like you know we've got these smart contract chains and then we've got these dumb nons smart contract chains and we kind of you know can't accommodate them because of the way that our architecture is. And so I was like I should talk to Ilia about this. So we like jumped on a call. It's like six months later. I'm like, "Hey, we got this big problem like uh you know these dumb chains like how do we store assets on them and like how could" And I like, "Uh, bro, I told you about this like six

months ago. Like, I told you the whole thing and how it works." And I was like, "Oh, that's what you were talking about. Okay, that makes you should do that." He's like, "We've done it. It's here. It is. It's there." Um and and so the the goal of intents was basically bring all of these dumb chains that don't have smart contracts up to the same level as smart contracting. Is that fair to to say? Like give you the ability to control assets on these chains and do all this cool stuff that we like with smart contracts on on Bitcoin, Dogecoin, etc.

I mean I would say goal is way broader but that that was like the f that was the first PMF. >> That was my goal for you. >> Yeah. Yeah. That was your goal. [laughter] >> Yeah. I mean the first PMF definitely was like hey look you have you know Bitcoin and Dodge that people want to hold but there's no good wallets. It's really hard to find things and uh really hard to trade if you want to go like dodge to Salana for example. And uh and so that was kind of the first PMF. Then obviously Zcash was like very natural. Yeah. Like even harder to get to it. The exchanges were delisting it etc. And we're like yes like Zcash works perfectly here. Uh you can move anywhere. You can exchange it to anything. It's decentralized. But I mean

the broader idea behind it is effectively combination of indeed through near you can control any other chain any any other uh account on any other chain. And we're actually having now first examples of people doing this on Salana because Salana's smart contracts are very tiny and it's hard to make like a contract that runs for a long time. On near you can actually run the whole other blockchain as a smart contract. So we have a lot of capacity for writing complex logic and so you can write the complex logic on near and then push just the exact execution on on Solana you know not calling it a not a smart chain but uh just like the level

of complexity of of logic you know recovery systems and other things you can do >> right but the trade-off space there on on Salana is like you know they're optimizing for kind of a different uh set of things, right? >> Correct. Yeah, they're optimizing for like immediate latency, but this allows to have like a lot higher yeah uh kind of capacity of execution and intelligence uh while >> but only like sort of just in time, right? Like you only use it when you need it and then you can get the fast memecoin trading that we all need um you know uh kind of on on the primary chain.

>> You can offload that, right? >> Yeah. Um and so so uh you know the the intent thing I think worked um a lot better than maybe people were expecting. um you know there was there was like the initial launch um and it was fairly quiet and and then it started to kind of blow up as obviously there's a lot of kind of threads kind of weaving together here right like everyone wanted Zcash all of a sudden they realized it was really hard and you guys had a a great solution for that and so um but that that started to get some traction. Um, at what point did you internally go, uh, we should like vampire attack Zcash and

just like replace their technology? What was what was the >> I mean that that wasn't the thinking and and it is not uh Zcash. >> It's a different thing. It's a different thing >> as a sovereign asset is very much uh has its place and and is important. Uh I mean honestly like as usual uh you know I was talk I was thinking and we were trying to build private uh effectively shard back in 2022. Now back then as you can imagine with regulatory environment and all of the fun stuff going on that effectively was a no-go for actually launching it. Um but yeah

kind of my thesis in general confidentiality is a key unlock for actually crypto becoming truly day-to-day use. Like you cannot imagine a business doing commerce on chain where everybody can just go into block explorer and see how much revenue they're getting, how much they paying their suppliers, what does the salaries are. Right? This is this is why most of the crypto still does you know fiat payroll because nobody wants to show how much um like payments of salaries etc going and so like privacy always was like hey this is a key enabler but you know we needed a regulatory positioning

that actually worked. um we needed to have enough liquidity and kind of flow through this to actually like because like we build privacy but nobody's using it like you know >> it's not private right >> it's not private right so so we needed enough like liquidity going through it and I mean it's obviously helps when there is like a broader market acceptance as well that hey this is actually needed and so that was kind of yeah the motivation is like okay actually now it's time like dust off the thing we built in 2022 you know uh update it to to some of the kind of intent based um infrastructure and let's launch it and then let's make the experience really rock right and and

that that was a lesson from Zcash indeed like the the Zashi Zodel wallet is just like really good experience uh and you actually want to hold your assets there like privately like you feel good about it so like that's what near.com is uh kind of for but obviously we're opening up this API for everyone to use as well so obviously Infinex and everybody body else can uh can actually do that as well. >> Yeah. I think one one thing for like you know there's like the theory of how something work and then there's like practically when you actually encounter it, right? Um and you know it had been a long time since I'd held Zcash. The the last time I'd held it was uh like probably on Binance or like ages ago um in in like 2019 or something. 2018,

2019. Um, and I went down the Zcash rabbit hole and I was like, "Oh, I'm going to, you know, try and like figure out how to like implement my own wallet and like did all this stuff. Um, and the challenge obviously and like this is well known in Zcash community, but it's not really talked about outside, right? It took me like doing a bit of research to figure this out. Um, is Zcash is somewhat volatile lately strategy. >> You just noticed that. >> I [laughter] just noticed. >> What gave it away? uh maybe losing millions of dollars trading it on Hyperlid was the was the thing that uh was the giveaway for me. Um but but uh the core asset is the only one that is private on Zcash, right? Um and so in order to have privacy like it might feel

really good to hold your Zcash privately shielded on wallet, right? Um but it might not feel great. might feel great one day when it's up 30%, then you know the next day it feels bad when it's down 30%. U it it seems to me like really obvious that like we need private stable coins like you know and and Zcash has thought about this. uh the community's got like multiple proposals of like ways to do this, but um you know that that to me seems like the obvious thing that no one's really land like I mean you know Tornado Cash had a USDC pool for a long time. Um you know uh but um but to your point like it's not enough to just have

one pool. You need enough demand. Like there's all of these different things for it to actually work and be scalable. Um so yeah how how are you guys thinking about that in terms of like what are the assets that you think should be you know private by by default. Yeah. So everything is private. So over we have 150 I mean it's changing every day. I think they're launching something right now. Uh 150 plus assets. All of them can be private. I was just checking there's like few million USDC 8 million near actually couple million Zcash as well.

Um stored in confidential intents. Uh you can actually check it out on dune now in the intense uh analytics. uh they have a um kind of dashboard for all the assets that are in confidential shard and uh yeah I mean I think as you said right if I want to pay for coffee I probably I mean yes if I'm Zcash bull which you know I am like it's it's great experience like the crossay is actually pretty cool because you can go and like pay you know I don't know if if you tried but square has a bitcoin payment and so you can pay from your zcash into a bitcoin cure code and it's facilitated to near intense, right? And so, you know, the merchant doesn't see where it

came from uh and and you pay in Zcash, but like that that's still a little bit janky experience. You know, the Zcash time like latencies. >> Yeah. The block times are >> the block time is high. I mean, they hopefully fix >> 200 millisecond >> Yeah. >> weight times, right? Like if you're standing in line at a store trying to waiting more than like 50 milliseconds, you're like, Yeah. And so so with with intense we are with confession intense the the thought is indeed this is like execution confidentiality right like we're running a full near effect shard uh that's confidential and we're running intense there we get we planning to have more applications running there

as well like lending uh etc. And with that you you know you get all the benefits of near which is 600 millisecond blocks. You can have your assets there to pay directly with you know stable coins with uh Salana assets etc. You can pay out into other chains directly as well and fiat uh is coming and so you can actually be like transacting normally as if you were on uh you know your neo bank app instead of um kind of feeling. Now, importantly, the trade-off is we're explicitly uh fighting against crime, right? And so, it's not >> I was gonna say, Tate, you you got to

jump in here because >> we were talking about this before and uh and Ilia was like walking me through like all of the stuff that they're doing. And I was like, this feels like the inverse Thor chain. Uh Taye's going to love this. So, um >> yeah, and there's been there's been a couple incidents. One of my favorite money launderers you was was trying to use near and uh didn't have the best time let's say [laughter] >> yes >> and then took to took to Twitter and I was like come on but I think I told this story Kate like a couple weeks ago um the yeah the money launderers they they revert to like the most core OG Bitcoin maxi

why isn't this blah blah blah blah blah every single >> How dare you? >> Yeah. They're doing like social pressure. Yeah. Like, hey, you know, isn't this supposed to be unstoppable money? We're like, yeah, when your money is clean, it's unstoppable. [laughter] >> Yeah. >> A bunch of stolen funds. And he was like, and now I am poor. I am poor. >> Good. >> And I was like, I took two seconds. I was like, you still have a thousand stolen Bitcoin. like you couldn't find 10 of the thousand. I was like, you need to go yourself. Um, but yeah, I think it's okay. I think it's critically important

like let's ignore the ideological side of things just for a second, right? Because we can like debate that all day. I think it's critically important though that people understand like you cannot have a really good useful protocol and find real product market fit for the market, not criminals, right? like you know the actual market that has huge TAM right? >> You can't do that if right off the bat the criminals take over, right? Because it it impacts you in every single way. Most importantly though, as you're iterating your product, as you're iterating your protocol, as you're finding the product market fit and understanding what users want, if you're only serving criminals like thieves and

moneyaunderers, their needs are actually dramatically different than average users need >> than an average user. >> Dramatically different. Like they do not care. They're very price insensitive. >> They love like 1% fees. Amazing. >> Amazing. They think that's >> fees and no questions asked. Like sign me up for that >> if it takes a while. >> People don't appreciate paying 1% of their wealth to >> move their money around. >> It's it's just a completely different use case. Their needs are different, their desires are different, etc. And so, and I've seen this play out time and time again with protocols, especially when they get flooded, when DPRK especially like just comes in and floods

them right off the bat. they get excited about. Oh my god, like we have users and if they don't fit, baby, let's go. >> Yeah. But it's it's really hard for them to then it's like you're serving this one customer and they're like quite a loud customer by the way. Like like people think that money launderers are like hiding. They're not. Like they are in they are in your freaking Discord preaching to you like loudly. Um, and it's just a it's I think people I think it's it's definitely like underrated.

And so, um, you have to get good core users that that sort of like are a different demographic in my opinion because then as you're iterating your protocol, as you're deciding the next features to launch, as you're prioritizing things, you're going to be doing so for a much larger market, right? Because at the end of the day, moneyaundering, it's capped like and it's it's a pretty good in crypto terms like it can look like pretty good product market fit in the short term, but when you're thinking like big picture long term, you can't it's nothing, right? Like it's nothing if you don't have >> the trade-off is so bad as well, right?

Because you end up getting >> the margins, your your structures are undermined and and you end up, you know, being like a pseudo your brand is terrible. Yeah. you can't you have every time you get on a plane um you know and then and then I think also again it just goes back to this like what you're prioritizing um and if if you're not prioritizing like real users are have a whole diverse set of things that they screw up things that they have to get right and if you don't hone the user experience if you don't hone right like all these little details you're never going to sort of like like you're >> actually escape veloity you're stuck in

this >> hellscape of money laundering but you know I mean the problem is I think like as a startup um in crypto especially you know in in a bare market or something like you're trying to survive and there is pressure of like you know we've got a a team we got to pay bills etc and like you know this feels like product market fit you know especially when you can kind of squint at and go >> they don't need that that like we're not 100% sure that they're laundering money. Like they're not telling us that they're DPI like so.

>> Yeah. I I I like how do we know? >> Yeah. [laughter] >> Yeah. So I mean so we took a pretty strong stance here for a couple reasons. I mean obviously personally I don't want to um I don't want to facilitate that in any way. Um and then the other reason indeed is we think and you mentioned like intense is you know for for dump chains to become smarter. The reality is intense is for facilitate all the economic activity in the world. That's that's really the goal like not just crypto we you know we adding fiat we adding all kinds of other assets but we're also adding work. You will be able to hire you know agents and people you know buy goods etc. So that's really

kind of the broader vision and if you want to achieve that vision as Tay said we cannot be stuck or even like we cannot touch and and in any way facilitate moneyaundering or or any other legal activity and so for us it was a pretty strong stance is like hey we want to figure out how to balance between like this is a non-custodial protocol that is accessible to everyone that's unites liquidity across all those chains and at the same time We don't want crime here, right? And these are two very hard problems.

>> Yeah, that's a tough balance to strike, right? Like >> Yeah. >> And so one of the things we've been uh working internally and we kind of starting to roll it out, we call it shield. This is our AI system that effectively is scanning other chains uh and transactions in real time and it's trying to detect when there is illicit activity happening on those like if you know an account that historically didn't have a balance just got the $100 million that's probably suspicious and we should not >> No, [laughter] that's great. That What are you talking about? That's amazing.

>> Yeah. like one of our customers is rich now. This is going to be [laughter] great for us. >> Yeah. And so and so uh so a lot of it is affecting anomaly detection on on all kinds of criteria. Same for chains, right? We we were the ones detected the Litecoin has been attacked. Oh yeah. And actually informed like other exchanges that they should pause Litecoin >> because we I mean we had like few transactions went through but then we're like hey something is wrong right? the finale. >> Litecoin didn't just suddenly get rich, right? Like something's going on.

>> Well, I mean, so the the the actually attacker exploited their Mimblewimble uh there was a vulnerability in Mimblewimble and they effectively withdrew money from their privacy pool without any like was effectively fake uh proof >> fake >> and and and then DDoS the rest of the network to drop the hash power. So only the like invalid like the blockchain was invalid code uh build blocks. Um, >> yeah, it was a that was that was Hold on. That was a that was a good attack. [laughter] >> I was gonna say I was gonna say, >> okay, so it they basically used a vulnerability in the core like one of the core features of Litecoin, right?

They use this vulnerability to essentially be able to do a double spend, which typically, let's just say >> it's not it's not actually double spend. They literally withdrew fake money. >> So they withdrew money that they didn't have. It's not even double double spend. You use the same money. They literally withdrew somebody else's money from the privacy pool. >> Oh, okay. Yeah. And then because they were able to and there was like three vans like or three sort of pieces that stacked on top of each other because you essentially you had the the first thing which was the Mimblewimble being able to pull this money out and then they found this other one that DOS half the network, right? But the whole thing of was right, correct me if I'm wrong. The whole thing was racing and that's why

they went through they tried Thor chain, they tried near to get these funds out while this side of the network was like offline was basically dosed. >> Yeah. So, so what happened is Litecoin core developers knew about this vulnerability two months ahead. They patched it and they then sent it to their miners. And so a lot of the network has upgraded. They did not let anyone else know like you know intense exchanges, Thor chain, anyone else that this is exists. >> Just the miners.

>> Just the miners. >> Just the miners. >> And so >> that seems weird to me because the miners I would have thought they don't know them. >> So how strange. >> Yeah. >> Weird. They must >> This is one thing about crypto like whenever you say somebody doesn't know anyone, there's probably a group to get in touch with them. They must have their emails or something. >> Yeah, >> it was definitely I mean that was one of my first times where I genuinely said like in the age of mythos this is unacceptable because the like sort of roll out disclosure timeline was muds.

>> Yeah. Vulnerability. >> Yeah. For vulnerability that they themselves found with AI. >> Oh, they found it with AI. >> Yeah. And so then they patched it, but like they were as and you know this is like I'm not I'm let's be clear like when you're rolling out a fix to to keep it in a privilege thing and try to get the network like and then ramp it up but you you have to understand like it's still like a like let's say like a zero day meaning that you you're you're wraithing against someone else finding the same vulnerability and when it's something that a I found originally you have to >> somebody will find [laughter] No one else's AI, no one

else's thinker is going to find this. >> Two months was a wild like when we started going through the time, it was a wild one. Um, >> crazy. >> There's a lot of things like this that I think it's like everyone's scared of AI right now. Like, rightfully so, but in weird ways that we don't necessarily talk about, right? Like two months. And to be clear, like this is where you you you find this thing and then the next next thing is like you need to let everyone know who can be potentially offramp >> because they need to be able to detect this and be aware of it and then you need to have incident response effectively while it's rolling out at least to like >> you know whatever percentage of the network. So they just sent the patch out and

>> they just sent the patch out and waited and effectively yeah waited network was upgrading but then the new up new version had the bug with DOS and so like effectively the old version had the Mimblewimble bug. The new version had a DOS bug and so the the attacker was withdrew money from Mimblewimble effectively dosed the rest of the network. So only old nodes were staying and producing blocks and then used and then started trying to withdraw this money on different exchanges. So, so like this is this is something we've talked about a lot, right? Like this this uh AI assisted hacking situation, right? like you know um what's what's your take like obviously

you've been watching this we've been talking about all these hacks the hack you know number of hacks have come come through this is an example where like the good guys found it with AI uh maybe it was just a leak right like you know the >> that was the question is like hey is this inside job because like it's been two months like but this time you know internally in any of this organizations there's a bunch of people >> it's going to leak people are going to know the miners know you've sent it out to people like did did they tell the miners that this was like do they tell them what they were patching or they just send out the patch and they said >> I I haven't seen the exact communication we are not the minor >> right [laughter] >> but >> you're not that guy the guy that they

were talking [laughter] >> it was also it was super enlightening for me >> to realize you know obviously like I spend most of my time in EVM land like I started on Bitcoin but I don't live there it was it was surprising to me because in EVM And we naturally prioritize the integrators, the the sort of the UI layers, the user layers, the exchanges, right? Even even before like yeah, from day one really, it's not it was never just about the miners on Ethereum.

It was quite surprising for me to to learn that some of these chains like Litecoin for example, the M like everyone else like than your guys are just like not seen as like a a thing that matters. >> Yeah. Yeah. Yeah. >> Which is wild because guys, the only reason people are using you is if they can get to and from your chain. Why are you spoiling the miners? Right? like >> yeah so that's kind of what we're doing with shield is like we're trying to build almost like an information network as well as AI detection so that you can actually subscribe if you are an

exchange or kind of partner and you can get advisories like you will not get the actual like what's happening but you at least get advisory that I mean in different >> yeah you get an alert like yeah I mean including like hey we think you should shut down this network right now like up to that level right or like hey Th this hashes of addresses are like suspicious, right? You you know you can you don't we don't know like you will not know which ones but you will be able to check your deposits against that list, right? And so like we can push bloom filter like you can do a lot of interesting things where like we preserve privacy. We don't reveal the underlying but we actually give everyone an ability to check and you know and then decide on themselves,

right? And so that's kind of the like general approach we're taking is how can we use AI? How can we use kind of cryptography to really route this information and then also build the network right we're working with zero shadow we work with like few other uh kind of professionals in the space how can we make sure that we have a reaction speed as well because that that is the other thing right now like every time it's almost like everybody's learned from scratch how to do incident response with the things right because I mean and we've been on the other side of this where like you know >> like >> I mean like drift drifts got hacked right? Like you know my my like signal lights up effective like hey war room let's go you know I I go to my team like

war room let's go it's like you know okay let's actually like we we should have like a you know information highway where like they can push really quickly that gets propagated everywhere and then we can like figure out okay then you can have time to analyze what happened but like at least we like you know like at least like slow down the propagation and can like take time to investigate. >> I mean, I think I think you know the the challenge with this, right? And like it's easy to to say um you know, people should have done this before, etc. Like you know, there there's a bunch of approximate causes of why we're still bad at this stuff, right? Like you know, the centralized exchanges don't have necessarily like the expertise to be

able to monitor like deeply a chain, whatever. like they've got services they can subscribe to, but you know they're much more concerned historically at least on like their internal infrastructure making sure that that's secure, right? Like what's happening on the chain is like eh like not really our problem, right? Like we don't care about the chains. We're a bucket shop here that's like, you know, moving tokens around. Um and then the people who have been at the center of like these bridge issues, you know, like the the Thor chains or whatever, um have like a perverse incentive to like not be monitoring this or like, you know, be like, "Ah, we could never know who's using the thing, you know, whatever." Um

it's it's kind of unusual that um a team that has like the domain expertise as well as like a stated goal of not accepting crime is kind of you know in the epicenter of this like moving funds and you know bridging etc. Um so I do think that like this this feels like a bit of a shift in terms of that the setup here is slightly different to um you know someone who needs to work crosschain but also is trying to you know uh be aware of this but also has their own chain that they're like they've got that domain expertise. So um yeah it feels it feels like uh the right group of people to be thinking about this problem. Yeah, I

mean for context obviously >> we sorry. Yeah, we we've we've interviewed most of the you know modern blockchains and like know actually exactly how they work. We broke some of them. You know we had Salana actually paid bounty for to us for breaking it back back in the day. >> Uh so we're yeah we we indeed very well aware on all the kind of risk vectors and general design of all the systems. Yeah, I think a lot of people approach and it's a very it probably stems from like just the like the compliance people and stuff, but it's like their approach is sort of this like we need to be compliant or we need to like stop bad guys and it's like

not really from first principles, right? Is this addon where you're like I want to do this thing. I have this goal that I'm trying to build for, but in order to do that, in order to not go to jail, I need to like slap a thing on so that I don't go to jail. And I think that that's always been very problematic because like with anything that you build, if you just slap something on it, it's not going to do do anything well and it's all going to work against your goals of the protocol. But if you think about it, right, if you think about like what are the goals of the protocol long term, what are you're you're trying to give your users, you know, it it actually is

more about like security risk, where where is the danger? And when you approach it from that way, which I say like that's like first principles, right? How do we derisk this for ourselves, for our partners, for our LPs, for our users, whatever it is, how do you de-risk this very quickly, especially in crypto, you have to confront the fact that um like things break and things get hacked. And if you're uh sitting at the center of that, you're also going to be impacted by that. And it's not because like law enforcement is going to come and arrest you, like maybe, right? In some cases, maybe. But the bigger issue uh is everything around that you know um and

so like especially with the bridges and stuff the first principles is don't uh like don't get hacked yourself but a lot of times that's relying on a huge number of other things and if you don't have any detection in place you have no way to monitor and react to the stuff you have no circuit breakers that's how we have you know name a bridge hack right and there's a lot of bridge hacks, like a lot of really, really, really big, obvious bridge hacks. Um, and a lot of them, like Ronin is the is the infamous example, but a lot of them didn't even notice that their money was gone.

>> It's crazy for like two weeks. It's nuts, right? Like, >> you know, and so that's the thing is like I get I get the push back where people are just like I don't want to know what the don't tell me that Lazarus is using my thing. Like it doesn't matter like what is bad, what is good. I get it at the same time [laughter] like >> what is bad, what is good. When it when it like devolved into like the philosophy of like evil, you know you're in a good place, right? Like >> is it really North Korea? Like are they really bad?

>> Can anything be bad? >> I mean code is law until code is wrong. Right. That's >> Yeah. Yeah. Exactly. Like what do we do if the length code is wrong? >> All right, let's uh let's take a quick ad break uh here and then we'll come back and talk about uh the EF and um AI drops. So um let's let's go to a break here >> that has helped create over $50 billion in enterprise value for more than 80 clients like Pith, Moon Commerce, and Wormhole. They've worked with some of the largest and most impactful companies in the space. They're the partner you want when you're navigating markets and

trying to break out from the noise. They help navigate TGES, go to market, BD and partnerships, capital markets, advisory, PR, media placements, KOA activations, and more, driving execution from launch to scale. Their results are measurable. To learn more and start building real traction today, visit multi-chainadv.com. Is my crypto working as hard as I am? Honestly, that's the question that got me into staking in the first place. I work hard for my money. It should be working hard for me, too. Now, you can boost your earnings this Coinbase 1 member month. Until May 31st, any newly staked assets will get a 40% boost for 60 days. That's extra passive income.

Coinbase 1 is the ultimate membership to make the most of your money. As you know, this is how I make the most of my money. Zero crypto trading fees, 3.5% APY on USDC, boosted staking and lending rewards, and up to 4% Bitcoin back with the Coinbase One card, plus this massive 40% staking boost. This is your last chance to take advantage of the 3% deposit boost, your share of five Bitcoin if you outpredict pro basketball coach Lethal Shooter, plus a chance at a private shooting session, a $50 Bitcoin bonus when you spend $100 on a new Coinbase One card in the first 30 days, and the 20% discount on the first year of annual plans all end in just a few

days. Score your boost before May 31st at coinbase.com/chained. Head over to coinbase.com/unchained. Last chance to join at coinbase.com/unchained before member month ends. No purchase necessary. See rules and other ways to enter. Terms apply to other offers. Futures swaps via Coinbase Financial Markets. Risk of 100% loss. Payouts event-based, not investment advice, not available in Nevada. Coinbase 1 Card is offered through Coinbase Inc. and Cardless Inc. cards issued by First Electronic Bank. Bitcoin back rates are based on card holders assets on Coinbase.

>> All right, everyone. We are back. So, let's talk uh Vitalic's smaller ship EF. Uh I guess we still need analogies for what the EF is because we can't actually talk about it directly. So um the so f like first of all so Vitalic dropped a 1500word expost um which should be illegal in my view like we have articles we have threads there's many options don't drop a 1500 I like I read the whole thing I saw a bunch of people like no one's going to read this I was like I'm reading this whole um I'm a massochist um but like why why are we doing this like just put in a thread bro it's not hard you can get AI

to do it for you. Uh it can it can slice it up. Um anyway, [clears throat] so so um I think the the reality of the situation is there's been a bunch of people who've been leaving. Um and you know tip the funny thing is usually when people leave, right, you get the like vesting cliff breached meme, you know, it's like the person jumping off the the cliff. Uh no one posted that in the f because uh I think they know they're not given any tokens. So, >> it's not the same. [laughter] >> And that's not the problem. It's not like that they were pay too much money.

Um, >> that's actually such a good point, Gab part of the confusion with everything else when people leave. There's like three reasons that you know immediately that they need a job. Um, so so obviously uh you know it's a bare market. Everything's everything's been uh been ripped pretty hard um except for near and Zcash. Um so so you know ETH is still struggling. You've got like the the people like David Hoffman capitulating.

Um and and I think there's like a a very uh like direct question of like what is the EF for like what is what is their purpose? What are they trying to do? Um you know there was a period where they they like the 2025 EF I guess right where they tried to do a bunch of things. Um I know like Luca like you know you talked to Vitalic um for the first time. I remember I I introed you guys. Um they were like boarding builders and and all of this stuff and it seems like um you know that that era is over. Um but at the same time the bullish thing for me right is there was at least some

clarity about like what is what what is Ethereum for? even if it's unclear what the EF is for and what the purpose of the EF is. Um, and I I was very uh I was excited to sort of see Vitalic saying like we're going to lean into being the most impressive chain from like a technological perspective. Don't get too triggered. It's okay. Um but but you know uh the the the fact that you know they have all of these properties that are like you know very unique to Ethereum that they've worked hard for that Vitalic push for and those are the things that people will ultimately care about. um I think is right and and I think like you know the thing that probably people have been frustrated by

is that there hasn't been like a clarity of purpose like you know what is the theoremium for right like you know there's we've had so many different narratives the world computer and you know in in absent in a vacuum people will invent their own things right like ultrasound money and like nonsensical things that that have been uh you know markets come up with to try and explain like what right? Um you know and and so Vitalic stepping up and saying like no no this this is going to be like a technological thing going the best technology we're working towards it etc.

Um at least feels like a a good response from my perspective to the market. What what did you think Luca? Did you read this thing? >> Yeah I read parts of it. I didn't read the whole thing like you did. I I suck. Yeah. >> You know, I can't speak on it just to to to the tea, but I mean I mean, you're seeing it across all blockchain guys except for Ilia. It's brutal out there, right? And Ilia had the foresight to at least like see where the puck is going kind of or at least like build something that was like novelly different. But I mean, you're also seeing some of the things with Ti and Salana, some of the tweets that he's been dropping lately. I don't know if you've been seeing that,

Kane. I think it's just hard there across the board, >> right? >> It's just and and I think they're feeling it. And it's interesting because when price is up, you kind of just become in this like lull. And a lot of people try to pretend that like they don't care about price until everyone starts caring about price and starts punishing you for it. It's like a fascinating dichotomy. But look, I, you know, my stance on it is I I personally just think it's a mistake. I think when you look at communities like what is a perfect example of a community and an ecosystem like look at your local town right your local commu your local community in in your you know in your neighborhood like there's police stations there's fire stations there's the municipal like there's the

courthouse like these things all need to function and they all need to support each other and you know in this case you know Ethereum acts as you know the town town mayor and the EF at least acts as the town mayor whether they like it or not. Like that's just the nature of owning an ecosystem and an asset as an ecosystem asset. And I've always taken the stance that you have to just treat it like you would a town or a city or a country. Um and I think taking a different approach u is maybe more in line with I guess the cipher funk thesis, but I just don't think that that's how these things win and succeed.

But frankly, you're seeing it across the board. Uh, I think a lot of guys are starting to feel the pressure specifically in blockchain land, maybe the exception of Ilia, who seems to have uh, figured out where the puck is going and is capturing a lot of that mindship. >> Yeah. What was what was your take, Ilia? How did how did you >> So so I mean we we have a long history with Ethereum because we actually pitched in 2018 to build these two to Ethereum Foundation because I actually like if we could just build these two make it work we would go back to AI in like 2020. Uh and uh because that was like the whole thing. We're like we needed a blockchain for our near AI stuff and we're like looked around. We didn't find anything that would actually

match our needs, right? Scalable, easy to use, kind of abstracted. Um, and so we were like, okay, well, Ethereum is, you know, was talking about sharding back then, was talking about web assembly, was talking about, you know, faster slot time, was talking about better user experience. We're like >> Web Assembly, you don't hear that very much in >> No, Land these days. >> Uh, and so we're like, cool, we can just build all that. We have a team, you know, we're VC funded. We just need an upside. That was the things like, hey, we need a way.

>> Whoa, whoa, whoa. Sorry, [laughter] I'm gonna stop you right there. >> I don't know if you're familiar with the concept of communism, [laughter] but >> we're not paying you You can you can build us the two, but we will not pay you a red scent, my friend. >> And so that that was kind of where where that conversation went. Um >> Yeah. And and so and to be clear like I'veveve I've I've pitched I've pitched this idea multiple times. So like before launching mainet I've pitched it again because like hey look we've built it now right we had bayet we had charting we had web assembly >> derisk all of the >> yeah we had accountraction let's figure out like what what this could look like we do have VCs we need to pay them money. I have a team that you know wants

upside they work you know 247 effectively for >> why would you have VCs? What were you thinking? I know. >> And so and so I think like so there there's like fundamentally to me there's few things. One is I don't think decentralization needs to mean disorganization. >> And so this is something I've told my team over and over again. We can be decentralized and organized at the same time, right? we can have a system that is support a decentralized set of like near is actually whenever people join near they were like holy I didn't realize there's so many things like we have you know dozens of companies working together at all times like you go to our slack there's like 20 slack connects you know that you join by

default right because we actually have lots of different companies working together in close coordination when we did have we had a hack as well on near with Ria We had five different companies effectively working together. They tracked down exactly who did it. They had all the forensic evidence and effectively returned all the money, right? And it was done because all the teams can come together whenever there's a need and figure out how to work together. And so and that that is like I see the function of leadership and and I mean not perfect obviously there's a lot to improve but that is a function of leadership of both you know the kind of spokesperson and the vision setter of the ecosystem right which is Vitalic in Ethereum and like this is not going to

change and the foundation as a as a supporting function like it doesn't need to go and do the actual day-to-day even sales function but they need to organize and and set the frame framework for that. So for example near intense all of the BD like for example Kane Kane's team working dayto-day they're not like I mean they originally some of them were in near foundation but they're all now in the diffuse company which is like overseeing near intense growth and that team is like dayto-day looking at you know growth of intense and growth of all of the >> and they have incentives they have incentives incentives like you talk to them and you're like oh these guys seem pretty incentivized to make this work right like >> yeah and it's like near incentive and they have bonuses based on success of

that specific unit, right? And so same like if we talk about near AI, right? They're not going and selling near intense. They have their own product. They have their own in the market. They still incentivized by near and incentivized by success of the unit. And so we kind of trying to build that architecture. Like the way I think of it is like what if you took things like Google which I actually think internally is very like there's a lot of issues but like has a lot of interesting kind of organizational structure and open it up made it open ecosystem that anybody can join anybody can collaborate work on each other but you had like common you know alphabet stock that incentivizes everyone you have common like HR finance that you can tap into and so like that

is the kind of leadership the economy of scale plus incentives, right? >> Plus incentives and and open like ecosystem. Yeah. >> The interesting thing is like Vitalic like well you know Google and like I get this right like when I was a teenager like I was like don't be evil is like made like oh companies are bad you know like because I was poor and like things were expensive and I was like oh man companies charge me so much money like they're such jerks why can't they make you know building a computer cheaper? I want to like do all this stuff and you know whatever, right? And so you had this like antipathy towards these large corporations that you know especially like Microsoft charging you for the operating system like come on man like make it free like Linux you don't need

to make money like and and so um you know there were there there was like Google was like oh we're going to be a corporation we're not going to be evil like those Microsoft guys right and Vitalic now like seems like has such an axe to grind of like Google became evil and stopped the don't be evil thing and like did all these evil things. Um and and like yeah >> what okay so it's not this is the thing you can have the goal don't be evil like personally in my life I try not to be evil great it's fantastic right I would recommend everyone operate like that >> not trying to be evil yeah

>> yeah just try not to be evil okay guys like done in terms of running a freaking organization at scale that is trying to achieve things if your let's say bottom line mantra that your employees, your teams, your partners are all living by is don't be evil. What the hell are you doing? Right? And Google I think when like they they explicitly removed it. Maybe Elliot maybe you know better. I feel like the explicit removal of this was because it it kept probably clashing with the actual goals and actual KPIs and actual, you know, mechanisms that they were using for their organization, right? And so when they

say having that as the mantra, it was mostly just bring it in line because they actually have goals. The issue that I have with like Vitalic and the EF and their approach again, try to do all these things, try not to be evil, like that's great, but what the hell are you trying to do? Because that's that's like I can go sit on my couch alone, never talk to anyone, never contribute value to anything, never buy anything, never sell anything. I can go sit on my couch and not be evil, right? Okay.

>> To what end though, right? Like exactly. >> Like obviously that's not a win. If the EF is sitting on their couch not being evil or Google is sitting on their couch not being evil, not doing anything, that is not actually a goal, >> right? It's just not. >> You have to define you have to define like what you guys are going for. And I think that this is core to any organization. If you do not understand what the hell everyone is driving towards and have that like, you know, people call it the north star, whatever you want to call it, right? to get people moving in the same direction and working together and understanding like where again where the where their peers are sitting where what their managers are trying to execute what the partners right like everyone has to be heading

towards that same goal >> I mean I think to your don't be evil is not a goal >> it's it's not a goal >> it's it's a non- goal right like you can't like it's like tell like what not to do is also not a goal Yeah. >> Right. It's like whatever comes from that. I understand that Vitalic wants the F to be smaller, but that's actually not what he wants. What he wants is he wants a flourishing decentralized community of people who are all executing different things and all have like slightly different incentives and are all using their talents differently.

That's the goal. The goal is not to make the EF smaller. The goal is like the EF is not needed because all this other stuff is flourishing. But when you define it as like the EF being smaller, I think it just confuses things in my opinion. And I think same for don't be evil. >> Remember a couple weeks ago we were talking about like how this the skew of morality has just been like totally left the window. And like I would love for him to define like evil because like me eating a hamburger with like a vegan sitting across from me like could be evil, right? So like let's just define evil at least for the sake of like if that is the goal like what is evil because evil evil is subjective right like you evil >> I mean you could argue >> and different than me

>> you could argue that you could argue that me sitting on my couch doing absolutely nothing productive for society for my family for my community right literally just sitting on my couch wasting away is like quite evil >> and crypto dgens that have been trading Ethereum the price has been the same for the last five years to them. That's evil, right? [laughter] >> Yeah. >> True. >> I [laughter] know. >> I know we all don't agree with that statement, but I'm sure like there's a lot of guys at their computer like evil.

>> So, so I mean coming from coming from Google, right? Like you saw this transition, you know, there there's like a view that it's great to say don't be evil. The only reason why you could get away with that, right, is the same reason why the EF can get away with nonsense and not having and having non- goals and not having to like actually execute and do things because they have so much money because they found this money printer and they now are like outside of the realms of like finance and con, right? Like Google just had this advertising money printer.

was just a money pipe that was just spewing money in that allowed them to do whatever the they wanted run any experiments or whatever like like what you were in Google like what's your perspective on that like >> yeah so I the way I think of it there's mission and there's values right and so when they was saying like don't be evil is values it's not your mission mission is actually what you're trying to achieve in the world and for Google the original mission was you know organize the the world right Um and don't be evil was the approach they were taking. Now you're right that also effectively while they had a dominant position and was kind of I mean complete monopoly on not even search but the

specifically the advertisement like the the eyeballs like they were effectively owned the programmatic advertisement market. They could do whatever they wanted. What happened actually while I was there there was this transition where like I the the easiest way in Google to get promoted was to go to ads team and grow some revenue metric and so I was talking to one of my colleagues who went there from research and he's like yeah you know got just got promoted got the got the uh like revenue uh you know growth and like what did you do he's like well you know we effectively correlating your uh maps data to the ads that we scanned from the

streets and we're retargeting it on internet. So you saw on the highway a ad like a billboard and they're like we will retarget you while you're visiting some web page now this exact ad this exact product and I'm like >> this sounds kind of evil [laughter] >> and I'm like why why is why why are we doing this now? And he's like, well, Facebook is doing this. And so what happened is Facebook effectively pushed Google to become a lot more aggressive in how they were using user data and approaching uh in general uh and generalists. And so I mean we see the same right now AI is

pushing Google you know to be better at AI and everything right even though you know transformers were at Google like we had we had chat GBT we had chat GBT product in like even before transformers to be clear like we there was a you know chat product is not hard to build like yes it wasn't maybe that as in amazing it just didn't make sense for Google to release it because like I mean multiple reasons but effectively like it wouldn't be good enough compared to Google search, you would create like dichotomy and and you know, you get you open yourself up for a bunch of lawsuits and there's there's actual like legit reasons >> for startups to actually auto innovate big companies, but now they're like stepping up and really really ramping up

machines. So competitions always leads to this but but I do think there's there is a natural kind of tendency in for like I I also like taking you know Vitalics and kind of more cypher punk side I do think corporations in general have this tendency of becoming like extractive right as soon as you effectively achieve your target audience and there's no more to grow you effectively need to grow how much extra like how much you're making from a single user and so you effectively trying to either upsell them or in some way increase how much they spending on your on your platform. And so that I think is actually the problem is that um there's no steady state like corp

corporation is either growing or or shrinking. There's no way for something to be at steady state. And I think in blockchain we actually have something where because we can have fixed supply we can have like a different dynamics when we say this is money when this is and and kind of if our vision is forward enough and and you know for us again we we think the economy going to transition to blockchain right right now it's extremely ineffective you know all of the negotiation like actual supply chains actual work is done extremely effectively blockchain will make it you know 10 times more effective and AI will make it 100 times effective and so

together AI and blockchain will make it you know thousand times effective and that's a TAM right that's you know 10 100red trillion TAM of effectively post AGI money is what we're targeting and I think that is like when that is the vision that is the goal then you still have values right we want to do it in you know no crime for example we want to do it um in a in a kind of methodical way. We do want it to be not like I actually think of decentralization as a tool, not as a goal. For me, decentralization is effectively making sure things are redundant, not depending on single person, not depending on single organization, not depending on single validator, on single wallet provider, on single RPC provider, like

all and all of those things. We want redundancy. We want ability to kind of work around >> like user benefits and and you know uh >> Yes. >> beneficial to the the process, right? >> Exactly. No, I have a talk I have a talk in 2018 where I'm like this obsession with decentralization for decentralization sake is the silliest and we need to stop. Decentralization is good, but it's not it's not it's not the decentralization that's good. It's what comes from that. So redundancy like also um like >> I mean competition innovation comes from this as well right?

>> Exactly. So you have you have cooperation which achieves certain things and can grow certain things in certain ways. You also have competition which is the same um like stability like security right all of these things can come from decentralization. they don't necessarily do, right? And that's like the again it's like going back to like what are you actually aiming for to just be decentralized like okay so what you kick everyone out of the discord and make them and don't support them >> yeah they you cannot be on on the same discord and then when you have zero day you have no idea how to actually let everybody know right

>> you know and so it's like what are you trying to achieve and I think again I think that this this is like a very a very big part of like what contributes to the Ethereum culture and especially at the EF is like what is what are you trying to do with the centralization again you want other people to build up in ways that you would never build up you want people to be working together to be competing with each other but you want everyone to be leveling up and filling the gaps in the ecosystem and the EF's job is not to do those things job should be to build those people up but by framing it as like we want to be decentralized and stuff. What it ends up like the execution ends up being is like we don't tweet about product builders. We don't

talk to product builders. We are uh another one neutrality, right? My one of my favorite ones like okay, so you don't support anyone. You don't care about anyone. You don't incentivize anyone. >> We don't have a favorite child. That would be evil. That would be horrific. We can't pick can't pick favorites. I don't I don't know. >> So therefore, no therefore therefore [laughter] it's so important that we don't have a favorite child that we actually don't talk to our children.

>> Yeah. We don't talk to any of the children. >> We don't talk to them at all. They live in a different house in another area just to make sure that it's it's it's so important that >> there all neglected equally. [laughter] >> Just like let me in your house, Dad. Please. I just wanna I just want to give you a hug. He's like, "Please, man, come on." >> Um, so I I really wanted to talk about uh like Ironclaw and some of the like agentic stuff that you guys are are doing, but I think we're out of time, so we might need to to bring you back because I'm I'm very curious um how you are thinking about this from like a practical standpoint of like an

organization, right? We've talked a lot about like you know mission and and you know uh morals and goals and everything. um we live in a world you know that in the last 3 months has changed dramatically in ter you know to your point 100x AI right we're in maybe the 5x or 10x period of this right but it's only going to get crazier um in that backdrop like how do you build an organization how do you structure an organization how do you get people to coordinate um I'm I'm really curious to get your perspective on this but uh but I think we're out of time for uh >> give us like two sentences though >> like this.

>> Yeah. I mean the the the high level where I want to get to is markets. That that's the highest level, right? Like we're transitioning from corporation or foundation or whatever needing to be the holder of people because you needed kind of to to bundle because you had trouble doing credit assignment, right? Like if you know Infinex done something, it was really hard to like this individual person contributed this much to this product. I think that actually is a lot easier to solve now with AI, right? It can actually estimate this and then on the other side each individual is becoming more productive. So like I mean it's you know I know multiple people who working on a bunch of projects at the same time

now uh because you kind of have a lot more bandwidth and so I actually think we're going to start transitioning to more organizations almost that are just operating on the market principles where you're effectively hiring you know an individual and their army of agents to do something and you kind of can estimate a like how much that contribution would have done to the bottom line of the bigger thing you're doing. Right? So I don't know if you saw so Jack Dorsey actually like he posted something like which is spiritual and kind of going from hierarchical organization to more kind of uh peer-to-peer and again this is where blockchain and kind of all of the you

know intent based stuff that we've been doing is exactly that's exactly the world we've been building for where you effectively can say hey this is intent like I want this new feature in infinex and now you know there can be multiple >> this is what it's worth to me deliver >> yeah this is what it's worth Or you can have a bidding and an auction and finding you know a solver that will say cool I'll build it in you know 24 hours like this is my expectation of like of quality it's going to be and then if you don't agree right there's another agent that comes in and reviews it and effectively judges if that was done and so this is actually what we're building as our neari marketplace and kind of testing this out right now this is still early so I'm not uh but we have like

first actual businesses starting to use it like for insurance claim filing for background checks for event like actual real use cases where they're effectively hiring agents to do some jobs and it can be like because somebody who is an expert effectively codified the workflow >> packaged it up and now offering that for a dollar >> as a service right like aggregating what would have had to have been inside of a corporation you know like you have your that that function of three four five 10 humans that are >> you know manually >> doing that all the time and sometimes they don't have enough work and sometimes you know you pulling them to do something else. Now you can yeah disagregate all of this and then have a marketplace and so that's actually I think where we're going. It's I mean

it's not going to be it's going to be a bumpy ride but I think that like systematically that what makes sense and again this is where AI and blockchain really fits well because it kind of powering this trustless transaction. We have escrow, we have verifiable AI to to evaluate if the agent the job well. You can still have human in the loop, right? To kind of crosscorrect if something went wrong. So it doesn't need to be like and now everything automated. It can be like you know I hire you know Kane's agent and you know Kane will like make sure that his agent done the work right or like corrected before passes is over. Right.

>> My other agent will make sure that the other agent did did a [laughter] good job. I I won't be involved. >> It's going to be fine, guys. Don't It's going to be great. >> Well, so so like within the near ecosystem and we've experienced this, right? Like you guys are already somewhat in this direction like these smaller corporate entities like smaller smaller business entities. >> I mean that that has been my principle from a long time ago. So I mean again to [clears throat] me AI again is an accelerator, blockchain is coordinator. The principles are the same.

But so so very but from a very practical standpoint, right? Like the this is something that I'm I'm like super focused on at the moment. Um you know, we went from 2025 where like you could grind your way through and get some like production code was really hard. then December 2025 through to like maybe March or something like that, you could you could not just grind your way through like you could just kind of prompt and get like production code if you had uh you know enough experience with um and so a lot of engineers were like ah I don't need to write the code anymore. or I can get agents to write the code and have you. But it was still very much like a synchronous process of like a single person sitting in front of a computer producing code, right? Um

that doesn't to me feel like the final form of this, right? Like it feels like we we need uh something else. And so the interesting thing about centaur what what paradigm uh dropped last week is like it's trying to move the agents in Slack so that you have like an agentic partner or teammate or whatever that is responsible for the coordination that an individual human would you know find hard to to coordinate um and can you know deploy agents and then like it's really the the humans in the loop for taste and judgment and and goal seeking and and all of that and then like the execution side is all handled by by

agents. Um like you know given how much you guys are are doing in the space and and how long you've been involved like from a near orb perspective are you still in the mode of one person sitting in front of one terminal doing things or are you experimenting in this like multiplayer space? >> Yeah, we're experimenting definitely and so the ironclaw itself is actually multi-tenant. So open open claw and Hermes like they're effectively like your system and so >> ironclaw actually supports a multi-tenant mode where you can log in with Google you know or or near account and you're effectively joining a single instance you have your memory you have your kind of missions etc you also have shared projects you have shared memory

you have shared workflows and so yeah we're experimenting kind of designing how we call the commitment system so whenever you know we having a meeting with Ukraine And like Ilia said, you know, yes, we're going to ship this feature for you. This should actually go somewhere automatically and then, you know, pop up. >> Yeah. Well, it's also should pop up for the team and then I also like re you know, follow up if it's done, put it on road map, update the right documents. Similarly, like I actually have workflows that, you know, scan everybody's meeting notes and like give me like decisions that were made decision.

>> What did my what did my guy commit me to? >> [laughter] >> Well, not for me, but like in general for the organization, but also like actually the most useful being finding where decisions were not made because that's actually where organizations taxes is people discuss something and did not make a decision because as a decision maker is not clear or it wasn't in a room and so just extracting that and making sure that's visible like it's already helping because then I can >> and is that like an org level panopticon thing agent multi- aent thing like how how >> right now it will We prototyped it more like from my perspective just because I have access and but then we are now building this as like a yeah a system that everybody again the the main

question is permissions right like for example there's some HR things right they shouldn't be visible to anyone except >> maybe some HR people and so like you need to kind of configure that but uh otherwise yeah like you know I may have like full purview and then we kind of can like cross cut it into like specific departments but Yeah, we're kind of building it as like a single near foundation at least right now. Ironclaw that has everyone kind of as a user and then there's like a common area of his permissions as well. >> Yeah. Nice. Yeah, I think this is like the most >> fertile ground for efficiency improvements. Uh that that >> Oh, yeah. 100%.

>> But it's super hard. Like it's hard to reason about. You start building something like uh this isn't exactly the shape of it. like it's it's a really complex uh design space um that we don't have good intuition about. It doesn't because it's just like we've only got humans. >> You just like throw a human at a problem and like they're just different. >> We're very bad at the human side of this. Like organizations, right? Especially like or running like an organization at scale >> is bad >> is like it is bad. >> It's always broken. Organizations are always broken. That's the >> always 100% always. And so it's like then you're like, "Okay, let me slap an agent on this bitch." And it's [laughter] like >> that'll solve everything. What are you

talking about? I mean, >> so you have to you have to like either I think like the approach, right? Like okay, let's let's just solve the problem which is such a common organization problem like why wasn't the decision made or like and also just tracking the decisions that are made. you can actually solve that and bring efficiency to that and then you can then sort of circle back around and be like okay clearly this area of the org this team whatever yeah and like non those non-decisions yeah I've worked at many companies where >> yeah I mean >> my entire life was not making [laughter] sitting on calls that didn't make decisions and me getting like just banging my head why are we here >> yeah I mean I So I pitched this actually

near con 2023 I pitched this idea of AI CEO right this was before any of this AI stuff started to work right and so my my vision always there was like yeah like working from two directions one is how to make individuals more have more context because like a lot of the middle management and kind of it's just like propagating context from like relevant places and at the same time cutting down stuff that's not relevant because like people cannot just like you know have a feed of everything happening in the organization as well and so that is AI like AI can do that very well right like whatever is relevant to your work whatever is relevant and you can still query up whatever you want to know and then the other way is like bubbling up in deep information is like what is

going on what decisions are made what suggestions what's improvement what is like where the trade-offs are like that's the other thing a lot of it is like we're operating on the trade-offs like what are the trade-offs right sometimes like people making decisions have no idea what the trade-offs were that actually led to And this is this is actually by the way why I think Dows never worked is because all of this infrastructure is missing >> and so like for Dows to truly work you kind of need to have the system truly working where you don't need middle management you don't need you have credit assignment embedded into the your organizational design because right now again credit assignment is a very like credit assignment I think actually is the fundamental world problem >> interesting

>> like like on an organizational level In AI machine learning what gradient descent like the training of models is credit assignment problem and so like in every in every place you go it's like credit assignment is the main challenge but like if you can assign like oh you know this person contributed this much to this bigger thing we've done if you can do that then like again right now like promotions compensation all of the things are like somewhat arbitrary >> very fuzzy and and adversarial >> and so and that put >> and then you exactly because then put rules like these very hardline rules on things that are fundamentally fuzzy underneath and then you're like it's fair >> but you have to you're like h we'll just like >> yeah yeah

it up somehow [laughter] and go >> I mean it's fair but you know but this other company gives me like double the compensation right so like probably it's not fair >> it's probably not fair yeah exactly exactly yeah and you know it's funny like there's the unintended consequences of like how you incentivize people are like so funny to me like you know I I kind of vasculate between these like communist bonuses, right, that we call them in infidex where it's like, hey, everyone was here. We can't really attribute, you know, perfectly, so we're just going to give everyone the same thing, right, for this particular thing, right? And like the the people contributing the most hate it.

They rage so hard. They're like, "This is a travesty." you like anyway. So >> but but I mean I know we're over time but markets actually solve all of this like if you forget about or inside organization between companies there is market and so if you don't like the partner you switch to another provider right if you like markets actually solve all this problem we just haven't figured out how to do it on a scale >> lower level and this is again where AI and blockchain and kind of the intent architecture is actually like built for.

Yeah. Yeah. Yeah. Fair. Awesome. Awesome. Thank you, sir. Uh, amazing having you on the show. Um, thanks everyone for joining us for this episode of Uneasy Money. Remember, what happens on chain never stays on chain. We will be back next week. Until then, do your own research before aping in. Nothing you hear on Uneas 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. Hey I'm