Episode 332 - Brussels catch up with Hart from Across

Transcript

[Music] welcome to zero knowledge I'm your host Anna Rose in this podcast we will be exploring the latest in zero knowledge research and the decentralized web as well as new paradigms that promise to change the way we interact and transact online this week to and I chat with Hart from Uma and across and just to note this episode basically has two parts for the first half of the episode we explore Hart's project across a cross-chain interoperability solution AKA a bridge

and a sister project to Uma we explore how a cross is constructed how the different actors function in the system the trade-off space the choices they make and how a cross compares to other interop Solutions then as we are in Brussels in person recording this we pop a bottle of champagne and very much shift gears so it is becoming a bit of an every six months or so tradition for tun and I to grab a bottle of bubbly and a friend for a candid drunk episode this time Hart is that friend and we begin with thoughtful observations about our week in Brussels during ECC events we've been at talks we've given research that's coming but very quickly we devolve into somewhat drunken ramblings

and interrupting one another a lot classic please forgive us all we were drinking on an empty stomach hope you enjoy it before we kick off I want to let you know about ZK hack Montreal an IRL hackathon happening on August 9th to 11th in you guess it Montreal this is the fourth ZK focused hackathon produced by ZK hack and it's a great place to learn about ZK tools meet fellow ZK Builders and build the next generation of ZK applications you can already apply over at ZK montreal. it's set to be the ZK event of the summer and it's heating up so get your application in soon now Tanya will share a little bit about this week's

sponsors namada is the shielded asset Hub rewarding you to protect the multi-chain built to give you full control over sharing your personal information namada brings data protection to existing assets applications and networks namada ends the era of transparency by default enabling shielded transfers and shielded cross-chain actions to protect your data even when interacting with transparent chains namada also introduces a novel shielding rewards system by holding your assets in a shielded set you help strengthen nada's data protection guarantees and collect n rewards in return namada will initially support IBC and ethereum based assets but will ultimately serve as a single shielded hub for all assets across the multi-chain learn more and

follow namada mainnet launch at n.net Alo is a new layer one blockchain that achieves the programmability of ethereum the privacy of Z cash and the scalability of a rollup driven by a mission for a truly secure internet Alo has interwoven zero knowledge proofs into every facet of their stack resulting in a vertically integrated layer one blockchain that's unparalleled in its approach Alo is ZK by Design dive into their programming language Leo and see what permissionless development looks like offering boundless opportunities for developers and innovators to build ZK apps this is an invitation to be part of a transformational ZK Journey d deeper and discover more about Alo at al.org and

now here's our [Music] episode so I'm here in Brussels our favorite city in the world with tun and Hart hi guys hey Anna hey excited to be here and uh for the record even though I'm going to get impaled by our pair of Canadians on the show Belgium really is Francis Canada I thought you were going to save this joke for like later in the pot it's okay we might as well pull the knives out get spicier faster so for for context Anna and I are both proud Canadians yes and I don't think we take kindly to Belgium or Canada being compared to what was what's the joke uh Belgium is France is Canada Belgium is

France is Canada yeah I don't like it yeah I'm not sure also Canada's so much bigger than America and Brussels is so much smaller than France it doesn't make any sense actually Chun I came here from a beach in Canada on the west coast where I saw humpback whales flowing through our gorgeous oceans yeah on my way here Belgium doesn't have that no all right well you know I I'm just the messenger but then I just heard this joke so many times and he thought it was really really funny this is the second time we're hearing it anyway for today's episode the plan is we're going to do I think we're going to try to do this in two parts the first part is actually getting to know heartt

and getting to know the project across and Hart I've tried to have you on the show I think now for a few months and it's just cool that we're in the same city and so we figured we'd actually sit down and do this interview now the second part maybe we can do a little review of what's been happening this week in Brussels because we're here for the the big ECC Extravaganza that happens every year but this year it's here in Brussels which is kind of unusual I ate zero Sprouts so far is Sprouts the thing I thought waffles Sprouts oh yeah is your dad jokes man what is going on no sugar-free Red Bull here it's been impossible oh no this is what happens he gets so sweet true there's an supply of sugar-free Red Bull

in my hotel oh I'll take you up on it cuz I kept finding Red Bull Zero which is a poor imitation of sugar-free Red Bull really yeah I think sugarfree has this particular ansy like aftertaste that I love you feel so unfocused because I think you have haven't had enough sugar-free Red Bull it's true it's true it's true it's it's truly a drug all right Hart I think we're going to start in on a little bit of an intro to you the audience may not know you um what have you been working on for the last few years yeah actually I know tun from before Gauntlet actually started oh wow yeah what year uh I guess this was 2018 ter 2018 yeah yeah we paid you as a contractor at one point do you remember

that I remember in two Sigma office yeah in the two Sigma office which is one of our investors anyways this was back when I found it so okay I'm like I joke that I'm the second oldest guy in defi after Chris Dixon uming yeah it's probably I know igov is kind of old huh igov is old I don't know him I think igov is actually probably might be older than both of you he finished his PhD in the early 2000s yeah okay well and I worked in finance uh as a a young Canadian I went to University in the US in New York then worked in finance for eight years uh as like a bond Trader um where I met my co-founder of Uma we went to MIT and new friends of

tun that's like kind of how we originally got connected I started a fintech business uh that that was from like 2013 to 2017 and then got sucked into crypto in 2018 okay before defi was a thing and made a protocol called Uma uh the goal there is to try we're trying to bring derivatives on a blockchains and our problem there was how do we get data and we me meant more than just price data but like arbitrary data on a blockchains and we built something called the Optimus gorol and that's been really cool and really fun um but the thing I think we're going to talk about today is the second protocol our our sister protocol we created called AC cross um which is really morphed into an intent based interop Tech mainly thought

of as a bridge and maybe we can talk about why bridging isn't a thing too but sounds good but before we do that like Uma that project where was it like what ecosystem is that in what where did it live all ethereum so oh it's all ethereum for some reason I thought it was Cosmos but it's not no I like Cosmos but we're not on Cosmos um it's actually quite possible that if tender existed Canadians are too nice to be Founders in the cosmos ecosystem just wait there are a ton of Canadians in the cosmos ecosystem actually well they're all do they as much no but I think everything was founded for Cosmos in Toronto J and Ethan and I think they're all true true Henry I just I'm just thinking about all

the people who fight with each other in Cosmos I'm like The Stereotype is you guys are really nice guysa Canada is a country it's a big country and people are people they're nice Canadians and there's a lot of also they just hide it better it's true and I've spent enough time in both the US and in crypto to become less nice that's been my goal all right you heard it here first okay so Uma was an ethereum project was it an application on ethereum uh essentially yes so Uma you can think of as an implementation of italic shelling coin blog post that he wrote in 2014 and this has other like intellectual predecessors in like truth

coin and augur actually um and the idea is how can you build a game theoretic machine that will uh tell you whether tell you a truth really oracles an oracle um but not like a price feed Oracle not like we're going to take the API data from binance and coinbase and put it on chain this is like I want to ask a question like who's the president of the United States or uh like an arbitrary question like is this set of data valid um and the origin story here for across actually was we were actually looking for use cases for Uma and this was sort of before any type of cross chain messaging existed and and we realized that um transacting across chains or sort of verifying things

across chains is a form of an oracle problem uh and we realized we could use the Uma Oracle to help with that um which we then do in a very specific way that we can maybe get into I want to just check this though because like you're saying sort of it was the first message passing but like had IBC been specked out or are you thinking more like in the ethereum in the ethereum world talking more like um arbitri and optimism sort of had just launched and they had a way to send messages uh they still do the canonical message Bridges go from ethereum to those chains in like 20ish minutes but take seven days to go the other direction even though everyone can kind of see that they're correct well before that 7-Day finality okay and so we were trying to use this Uma

chilling Point Oracle to do messaging faster so you build a cross for that particular use case and what is it exactly like you're talking about this message passing bridge but yeah maybe can you go in a little bit of like what that looks like maybe what it was and also how it's evolved yeah the origin story might actually confuse what it's become now so across is an intent based bridge now now and so the way I like to explain this is if you think that you have imagine you have two blockchains and we have no video so I can't use my hands but you have blockchain a and blockchain b and you want to send value between the two of them um what naively

the way you think that would work is okay I'm going to have blockchain a I'm going to send uh a message or send a deposit into blockchain a I have to wait for finality on blockchain a and then I send a message from A to B and then I release funds on B to the user oh wow okay okay so that's like typically how what I'll call like gen one Bridges work um was this the lock unlock yeah then you can get into there's sort of a different parallel conversation like okay am I locking and minting assets oh yes right um or you could imagine you have a pool of assets on chain a and chain B and you're going to use this message to release assets that are uh like canonical assets on chain B kind of ignore that for now that's sort of like

depending on your use case but this this sort of bridging framework is how I think a lot of people thought that value should be moved between blockchains MH for sure um and in something like IBC actually this kind of works pretty well because you have pretty fast finality and you can send the message pretty quickly and that all kind of works right uh but for a lot of evm uh rollups that's not the case and it becomes quite tricky got it so you go back and you think about this native framework or this this naive framework of trying to send a message from A to B and wait for finality on a to do so and the problem is it's slow uh like you have to wait for finality and then sending a message is expensive and slow and kind of scary

from a security perspective too and so our approach is actually to do something quite different and this is the nature of these cross-chain intents uh our approach is to not send a message but to rather have a network of third-party solvers uh or relays or fillers use the words interchangeably and those third-party uh solvers will observe blockchain a for when a user makes a deposit and that deposit goes basically into an escro contract they'll observe blockchain a and then they will front money to blockchain B uh front their own Capital to blockchain B and fill the user extremely quickly um the user happy they get their funds fast and then the protocol can

take its time in verifying quote unquote that that intent was filled before releasing the user's initial funds to the solver I want to really understand what these thirdparty solvers actually are like are they actors on one blockchain one of these two places or are they like really sitting outside of the entire system so they're they're kind of actors on every blockchain that they fill transactions on okay so you can think let's make tun be the third party solver here and so Anna you want to bridge from I'll make it arbitrum to base so it's a to B right so Anna's bridging from arbitrum to base Anna does a deposit transaction on arbitrum of say one eth tun is the offchain actor that

observes that so he can see you did that deposit on uh arbitrum um if he trusts the arbitrum sequencer which he does he can say Okay Anna's definitely done this deposit on the arbitrum sequencer I am going to go and fill Anna out of my own inventory on blockchain B on Bas yeah and he can do that extremely quickly like our median fill time is 4 seconds depending on the chain it's down to like less than a second on some of these fills and so he does that for a small fee and the fees here are really small which we can also get into too but he does that um and he makes this loan to you on blockchain B uh and then the protocol verifies that this F happened

and pays him back uh later so is it that 7-Day kind of like fraud proof or is this like a different kind of settlement yeah so it's not the seven so to the protocol like from the across perspective all across has to do is get comfortable enough or be sure that the fill the intent was filled um which we can talk about how we do that but then they can immediately release funds to tun On the Origin chain we can send a message to arbitrum in this case and say release funds to trun we do a bunch of like optimizations um to help trun manage its inventory he can actually request the protocol pays him back on base because that's where he lent his funds okay um but this sort of like an

optimization to help tun the third party actor with his inventory management um and yeah the so we're not waiting for this 7-Day uh finality period we're sort of this fast this fast fill layer uh on top of uh the the the L2 and like evm ecosystem the thing I still don't get though is how does a solver see both of these things like I've always had this impression that like if you're working in one blockchain it can't really see into another blockchain it doesn't know the state of the other space and I know you said so this solver will have like will be on both of these chains but still like

where's the information being connected and how is it being connected it's a really good question and insight so blockchain B base cannot see what's happening on blockchain a yeah but chun's not on a or b trun is his own independent special person um and so tun he's not making a decision on behalf of like a protocol or consensus mechanism he's making decision on behalf of own money okay so all he cares about is does tun think that deposit transaction on a happened if he believes it happened then he can go and make the fill on blockchain B with his own money and so basically by introducing these independent third parties that are

observing and making their own judgments on like what happened between blockchains we can shortcut all of these like messaging latency problems does that make sense I think so I just I still don't I think this makes me want to dig into who the third party solvers actually are well one important thing to think about is that um when you rely on like solvers you can lose liveness right like in theory all the solvers could not want to hold the risk because they don't want to hold a risky asset and then no one fulfills right that's a little different than a blockchain where like a layer one where someone can always step in and sort of Costless or low costly earn part of the network for participating here you're earning fees

but you you know you have to stay live and stay so it's a little more like Market making that's what it sounds like actually and that's why when you're talking about these solvers I'm really curious like what form they take are they a company are they like a market maker that's like just use like using software that you've provided but like has already checked like I trust all these things and it's sort of like happening in the background yeah it's so people actually think across is either us running our own solver okay or like winter mute running a solver for us okay um and neither of those are true like we do run a solver much in the same way if you think of uh like a or compound they at least early days they ran a

liquidation bot to execute liquidations on their protocol we kind of do the same thing but most of our transactions like 80% of them or more are filled by uh third parties okay and these third parties it it's actually kind of quite cool they're Grassroots folks that are like we know them by Discord handles and or we don't know who some of them are um and we've across has written a reference implementation of how to be we use the term relayer in our system of how to be a relayer um in across and other people have taken that and heavily optimized it so they like out compete our reference implementation by a lot um and we really don't really know who these folks are um it's

permissionless can they cheat no no how how do you make sure they don't cheat so they have to lock their funds or something is that why so let's we'll step through it so you doing your transaction on arbitrum you deposit one eth on arbitrum with instructions that say and this is your intent we'll go back to this intent concept so you deposit one eth on arbitrum with your instructions that you want to have one eth on base less a small fee let's say the fee is like I don't know a dollar um that's your instructions mhm any one of these relays um can observe that deposit transaction and those funds go into an escrow contract into like the across

protocol escrow contract which I I generally call like our settlement layer you use the word settlement earlier and like I think it's very appropriate here um okay so that happens and then we have um tun as one of the solvers he goes and he fills you on base his fill transaction goes through a contract that basically nonces it that makes sure you can't get double filled okay and he records that says okay Anna on arbitrum said she wanted one eth sent to her on base less than small fee and he fronts you money the protocol then like verifies that that happened but won't pay tun back if he screwed up okay so like tun

has a huge incentive to because he paid an eth out for this fee he he has a huge incentive to not get it wrong if he gets it wrong he doesn't get paid back your one eth on arbitrum and you might get like double filled by somebody that does it right and you'd be happy because you get twice the money but he doesn't get slashed or anything there's no like punishment it's just he he's already handed it over exactly he fronts money so there's like the slashing is sort of built in if you will where if he did it wrong he doesn't get paid back the money he fronted and it's your it's like the the overall protocol that's verifying it's not this like reference implement okay CU that the minute you said that this reference implementation that

people have optimized it I start to think like if they dramatically changed it couldn't they like well they still have to follow the rules of the smart contract right like they how they come to the inputs that they send onchain is up to them right and that they reference impementation basically computes inputs given some state of the world uh you could write something that maybe takes other inputs like you take in binance data and you compare prices and you you lower your fee because you're using some offchain it but that's that's all you doing it yourself but all you're doing is really constructing inputs that go to a public smart contract and you can only kind of take your money out if some covenants the contract holds are true and that's that's secured by the validators but there are validators he

means validators on validators of ethereum like the oh the ethereum validators okay okay talk a little bit about that protocol then the one that's doing the verifying what have you had to do to make that happen is that just like a series of smart contracts is that like yeah do it run from a central point is there a validator set there yeah okay so conceptually what we're trying to do is verify that uh tun's fill so we go back to our example right you were filled by tun you're happy you as the user move on with your day and like you got filled in a few seconds and you don't have to think about it again so that's that's great um now TR needs to get paid back so what we conceptually need to do is just verify that truns Phill was in fact

valid before we release the funds you deposited on arbitrum On the Origin chain back to him so um the the kind of naive way you would do this is you would like send a message from base back to arbitrum and say yes this fill did happen and release the funds that way but then we kind of go back and like well we're just sending messages again like it's not um didn't really it's faster but there's just as much overhead so what we do is we actually batch together uh thousands of fills uh every period periods right now are about an hour um so over course of an hour and this could be changed as a parameter in

the system but over the course of an hour we batch together all of the fills offchain so a bundle of fills is constructed and anyone can construct this bundle of fills because you can observe all of these blockchains right and um that bundle is then proposed optimistically this is now using the Uma Oracle back to what we talked about at the beginning it's proposed optimistically to ethereum main net and anyone can then dispute it and say actually I disagree with this construction of fills and is this like is there is there a fraud proof in here somehow or is this just like is there a time frame there's a there's an hourong uh roughly hourong challenge window okay so it's quite short um but again this is

deterministically verifiable um and we have people that run dispute Bots to verify this anyone can kind of run infrastructure to reconstruct these bundles and verify it and then uh provided that the bundle is this sort of bundle of repayment instructions is not challenged it then gets processed as kind of this batch repayment to all the solvers over that period and is this why you were saying before that you could actually repay it over on base or something like because it's batched because there's like you know action happening in both in all directions that like you could just say well somebody was sending it this way somebody was sending it that way instead of doing those two we'll just cross them out and

you can actually receive it on the one you sent it on yeah you're you're jumping ahead and this is like no it's it's really smart it's like this other optimization in the system where basically like you can kind of think left hand right hand like if you Anna were going from arbitrum to base and I was going from base to arbitrum right well if tun filled uh you on bass and then filled me on arbitrum and then kind of like just has funds all over like you can you can kind of net these off um and so the cross protocol has people that provide liquidity so we do have like LPS in the protocol and that liquidity is used so that if tun wanted to he could

ask for his repayment to happen on the destination chain and we kind of do some netting and some sort of financial optimizations to minimize the amount of money that actually has to get bridged between blockchains basically with the understanding that if you have offsetting flow like you don't need to send funds between arbitrum and base you just have offsetting flow and kind of net users out and the reason you would do that is because of I guess like when you're sending money back and forth if you actually do it if you actually have to do those settlements yeah are you saving money are you saving like what would you do otherwise you'd be like would you actually be moving the funds off the L2 onto the L1 and then back on the other L2 or is it even just like even just calculating the repayments on

each side would just be annoying like what what are you saving all the above right like the dumbest way you would do this is Tun has filled the user on base uh he gets paid back or the protocol releases the funds on arbitrum Bridges them over the 7-Day slow Bridge back to ethereum okay right and then uses the faster ethereum to base Bridge that's like 20 minutes and so tun gets repaid in s days and 20 minutes type thing yeah and obviously like that doesn't make sense so when you have this type of netting or these other advantages you could first of all repay trun on arbitrum itself and then trun has this like cross chain position where he's

short on base long on arbitrum or you could do other kind of inventory management type tools to say okay maybe we have LP funds in AC cross which we talked about cross does maybe the cross protocol has uh some some funds some sort of I I use the concept of like working capital there's like working capital in the protocol that could pay trun back on base or maybe trun is like a sophisticated solver that has his own Inventory management going on and he's just doing sort of his own netting left hand right hand where he's doing fills from arbitrum to base and getting uh paid back on arbitrum well making fills on Bas but then he's got other fills going the other direction and he just does his own netting um anyways all of

these systems sort of work and are Poss but those are the less optim like the optimized one is what you're actually providing yeah and I think it's one of the things I've actually thought is very cool and kind of been exciting about across is the architecture itself really hasn't changed in almost two years oh wow um it's just had all of these improvements and optimizations made to it really in terms of how we net flows and how solvers and these third parties operate and how they do their own inventory management you know I think for the user I think it's actually quite intuitive what across does but maybe given that across was created to in some ways dog food the Uma protocol or be you

know be your own customer of your own protocol maybe it would be good to talk about where you use you know what parts of work building one protocol first and building another one aside it how that influen your design choices and the you know the fact that you have had the architecture Frozen maybe is one of them there's actually probably a bunch to respond to uh there so first off our team does have like a financial engineering background and generally speaking I thought that was useful or has been useful for us to come up with this whole intent framework in the first place so it's almost like if you think from a very uh engineering first or kind of uh computer sciency first perspective

you think of the idea of oh I need to send a message to be able to bridge funds and and it's it's a little bit more of a foreign concept to say wait let's have a third party make this loan that was kind of useful for our own Financial engineering background then all the other things we've learned from building Uma is first of all build something that has actual demand and traction um and so in this case another learning is there is no demand to send messages between blockchains except to move value right and this is one of those this this I think is a spicy take like like all of these arbitrary messaging protocols they don't get used for anything other than moving value really yeah it it's like 95% of the use

case here of of moving between blockchains is moving value it's good we're going to the next question is going to be a comparing across other protocols although wait I had one question before that which is actually about fees but maybe that could tie into comparing them well we can get more spicy takes on fees there too because I actually think the economics of sending messages are AR particularly good the economics of moving value are much better which we can get to as well but there's another really important Insight here too that um I actually tried to hammer in my presentation at ECC uh as well which the the fact that 95% of the use cases are moving value between blockchains does let you take advantage

of the fungibility of money and so if you are moving value you don't care whose money it is that gets there this is where the third party solver works well so like tun if if I'm if Anna if you're trying to move money from arbitrum to base and you get filled by tun you don't care that it wasn't your money on arbitrum it's just money right and so that fungibility of value is what lets us like shortcut a lot of the latency issues around messaging um by again having these third parties do that and then tun the the solver gets paid back by other money somewhere and it's all fun ible so the point that I I tried to make at my presentation at the

conference it's uh people have traditionally thought about interop as equivalent to TCP IP we want to like connect servers with like Network pipes and send messages between them when I think in the blockchain concept that's not true the thing that we're moving is value and so you're moving money around and you have access to like different weight of doing that that actually shortcut many of the the problems with the hurdles and tradeoffs that you face in arbitary messaging would that also be like in even the software development itself and maintenance kind of like there's always sort of a way to fund this kind of work like that thing you

just described like with the relayer or the the solver like they're all of a sudden incentivized because they can get a fee whereas if it was just info you're sort of relying on these like altruistic Rel layers of info okay you want to be funny it's almost like arbitrary messaging is like communist and what we're building is like capitalist uh but or or the messages is like the roads it's like in it's like City infrastructure which doesn't make any money that's a better analogy it's it's public goodish right so we do have to have these roads between blockchains but they kind of get built and they have limitations and I actually like this analogy a lot um and

I'll extend it into like native interop and where this is going for a split second we'll get back on topic is the base layer the sewer system in theology like ouch But but so okay so we have this intent-based protocol that incentivizes offchain actors to compete to fill users as quickly or cheaply or both as possible that seems good that just seems strictly good and as they get more competitive users like the protocol gets better users get better fills which is really cool the messaging we have to do to verify those fills you can think of as your roads like analogy here and what's really cool is like to tun's

point we use the Uma system because that was one type of road we could build um to send messages between these blockchains but we're not at all wed to it um and we are I I think it's very possible that we'll move off it at least in some ways in the near future oh and where this gets really interesting is you start thinking about all the native interop systems that like L2 ecosystems are building so optimism has uh a design for Native interop for their super chain Y and what they're effectively doing with how this affects AC cross is it's like building a better Road between those chains so now we have like a highway that allows for faster messages

and more efficient messaging between let's say base and optimism here but the across protocol and the solvers they can do exactly the same thing they did before however now they get repaid faster since there is a better Road connecting those roads oh interesting and and I I guess you'd extend that past the opce with this beware like the agay exactly the hyper what is it CK elastic chain elastic chain yeah there there was so much beef on Twitter about that recently I know what the Z the a versus elastic chain yeah and and claims over TPS or whatever which are always gerrymandered anyway but I guess in your case where you're sitting you're happy if these things exist so it's worth zooming out and thinking about this from

like vitalik's perspective so vitalic wants a unified ethereum um and yet you have these ecosystems within ethereum that are very uh knife fitish right now so well arbitrum optimism ZK sync polygon cdk and others are all Fighting For dominance and none of them are going to dominate in my opinion there is going to be use case for all of them and like tribalism and token incentives and everything else right um and I think it's pretty cool From ethereum's perspective that we have these different ecosystems but they are all building their own form of like interop um and they're not compatible no right and if

you take this intense standard well they're not going to be compatible I'm I'm pretty confident about that um but if you take this intense standard and you think of it and delete a cross from this don't even think about AC cross think about this architecture of specifying an intent to go from one chain to another and Define it as a very lightweight standard uh that then allows these solvers to interpret these intent orders and fill users very quickly wherever they want to go what then happens with all this native interop is if the origin destination are both within native interop the silvers just get paid back faster which makes it cheaper um so this intent standard

becomes this simple unifying layer across these native interop systems and like turn to your point the like Uma way of sending messages maybe goes away um I was going to ask you if the Uma way of sending messages goes away because you want to support salana or other chains well we can we sorry was that spicy no that's not spicy because we really do want a sports salana and we actually probably are sooner than later um it doesn't it's actually weirdly compatible with that the way we are because we can use the Uma way of verifying all of these uh intent fills we can verify intent fills on salana too in exactly the same way the only thing we're

missing is a bridge to send occasional messages we're talking like one message an hour from uh ethereum to salana but there are people that have built those bridges so we could use them right I have I have two more points I definitely want to cover here one is I want to go back to fees because one thing I didn't even get from the original example was like where does the fee actually get paid like if if going really really back to like I'm depositing one e on arbitrum but I want to get it on base aren't I paying the fee on on arbitrum like I'm paying the fee so wouldn't the fee be when I'm like at the site where I pay or is it like on the other side so we're being precise if you do an onchain deposit transaction you pay that fee for

that onchain deposit transaction there there right um there's a gasless version of this too where you wouldn't pay that but let's IGN the fee is for the solver right like fee well you you pay you were saying I want to bridge one eth from arbitrum to base for uh as little as possible basically and so who do we have to pay along the way we have to pay the solver for the service their and the gas the solver then has to pay the gas on the destination chain to send you the message so they're going to bake that into what they charge and um in our version of some of this more um fancy like Inventory management if the solver borrows money from the across liquidity providers

that's kind of this extra convenience function um they pay that fee too um but basically from the user perspective they're just paying a small fee to get funds from A to B and I'll add two points so this is compatible with the user not just moving value but they could first of all also swap between assets we don't do this today but it's not a far stretch um so this could swap from eth to usdc in that in that bridging so like show up on another chain exactly and the solver can also execute code on behalf of the user so they could also say hey Bridge Anna your transaction could be Bridge one eth from arbit from to base

and execute this arbitrary code that for example buys an nft right so the thing I'm pushing for is for us to uh adopt a standard for what these intents look like we've published one that we built with Unis swap it's called ERC 7683 and it's an open standard that we want people's feedback on and you can go to ERC 7683 org and read about it and comment on it um but we want have an open standard for this intent structure that would unify ethereum and also allow bridging to salana and all these other fun things and just creates uh this it sort of emulates this seamless one chain

experience even though it's powered by solvers under the surface interesting the last thing about fees is I still don't understand where this the solver gets paid oh I'm sorry so the solver simply when they when you were Bridging the the one eth from arbitrum to base they don't pay you one eth they pay you 999 e or whatever is they take it on the way and then so they they pay you less than one eth and then they get paid back one eth so their fee kind of gets extracted that way okay so it is kind of on the destination chain because they're paying you less and they're sort of taking it there I guess I don't know why I'm stuck on I'm just really curious while I wrote it out I was like where's the fee so now I need to figure this out

you one unit on a and you get 99 units on B you the user and then the solver gets back depending on where they wanted to get repaid they get paid back one unit wherever right so yeah okay question answered thank you for that I'm sorry to for being so pedantic but uh the last one that I definitely want to cover I think you wanted to cover this to tun is the comparisons because there's a lot of bridges out there and actually Hart I don't know if you know this but the reason I reached out to you is I had a friend who was working he was using a lot of different Bridges and he cited across as like the one that he really liked to use and it was the one that was sort of gaining steam people were kind of noticing that this one

worked really well there have been a lot of bridges out there there are a lot of bridges out there I feel like I've had a bunch of I think I might have even done a series at some point about bridges like in 2022 or something well a lot of the people who are working on Bridges don't work on Bridges anymore had let be real that's be real yeah that's true but let's talk about the existing Bridges is like we don't have to go do the history of all of them I we'll keep this short to not like lose all your listeners right but um first off let's talk about how bridging is not a thing anymore or like shouldn't be a thing okay um and we'll we'll keep this short so in my example of like what I called gen one bridging at the beginning of the episode

that's like the I send a message and a lot of those bridges were also what I'll call like lock and mint or mint and burn Bridges where they create a synthetic represent a of the asset on some other chain I hate that structure because in that structure you have Perpetual forever risk to the underlying like Bridge protocol right you the user when you have a wrapped representation of that token you like at any point if that bridge gets hacked you lose your assets and the Wormhole hack in particular was about um a malicious operator figuring out how to control the mint function wow and then they just minted a ton themselves and then trade it against for

real assets right so and maybe this gets solved in the Far Far Future when we have like really good ZK proofs like this is the zero knowledge podcast maybe we can really securely prove these minting and bridging functions but it just scares the out of me right and so layer zero has their of standard which is effectively this uh Wormhole has their version uh axar has their version I'm not on Twitter I'm known for picking on layer zero but I'm I actually don't like any version of this standard because I think this like rapid Min thing is just scary but anyways that's like this kind of message based like lock in mint uh token standard delete those those are uh things that I I don't

think are ideal um the hyperlane hyperlane is a message Bridge yeah right that is intelligently designed that also has a a a rap and mint and like a of like standard they call them warp routes or warp tokens um too um and I think their code is pretty good but if you do have a wrapped version of the token you have Perpetual risk to the underlying security of that bridge which I don't like but hyperlane as a messaging protocol I do like a lot and I think they're quite great um yeah Axel are again they're not a token Bridge they're a messaging bridge but they have a token

like rapid Min standard yeah could could keep going there in terms of token Bridges Stargate is uh a a a bridge that actually blurs these Concepts too but where is that where layer zero layer zero has a yeah so they're they're big they use this send message to release tokens from a pool on the uh destination chain and I mean I wore with them a lot on Twitter so anyone listening to this will be like okay heart's Waring with Stargate on Twitter but okay I think their architecture is inferior in the sense that you have a minimum 40c you have to wait for a sense of finality On

the Origin chain and send a message every time to do a bridge transaction um which I think is inefficient and slow uh compared to this intent based framework um yeah what is your take on IBC IBC I think is great within like tender chains and Cosmos chains okay right IBC has always been explained to is it's not specific to Cosmos but I feel like the actual Integrations for everything outside of Cosmos just haven't worked very well you probably understand this better uh I guess like other people who've who've weirdly spent a lot of time implementing ibc's like Avalanche the subnets actually communication protocol is is is is IBC and there have

been all these people have tried to do it between rollups but it's just so expensive to have a like client I think like without Mak modifying and like having extra pre-c compiles and stuff to like make it gas efficient to have a light client in ethereum Chains It's kind of Dead on Arrival in ethereum you you other or you're just like stuffing into a client and making and so it it never becomes standardized so I think it's very hard to get adopted as a standard unless you're starting fresh that's like kind of that's kind of where I yeah so so I'm like maximally bullish on this intent-based framework right where well how do you compare to other intent based Frameworks so okay let's talk about this for a second too you also have like good views on this so intense can mean a bunch of things some

people like hyper generalize this and they think of intense as like anything I want and I get something exactly I'm going to describe in natural language that I want Uber Eats Sushi delivered at 500m at my house and this is like and please pay using my usdt on Tron and figure out how to convert it to whatever the pizza place takes exactly right so like okay I I actually think some of this is kind of cool but like not now type thing um look everyone you know you know people love making that uh we were promised flying cars but we got X meme but like honestly this is f flying cars though cuz it's like hey you're asking for like everything you want like an AI to interpret your natural language request and then find out a way to have

competition for it to yeah exactly so when I say intense I'm specifically referring to cross-chain intense in a much more constrained way it's really simple I like my mental model is it's an order ticket your order ticket says I have asset uh X on chain a and I want asset y on chain B I'll pay this much for it do you give it time uh yeah sure I want it by this time and I maybe want you to also execute this arbitrary code on the destination chain that's it so it's like heavily constraint from this General wait wait wait but execute arbitrary code does not exactly constraint bu but it's like buy an nft or like

deposit an A or something some some but like you have some computationally bounded code arbitary code it's bounded in more ways than just computation too like you there's another example like I can't send governance parameters that might be worth a billion dollars right I could like send a dollar Bridge transaction with like the code to execute the government that doesn't work right so there are constraints here on what you can do but I'm basically saying from the perspective of I'm sending money to like buy something or deposit it somewhere you can do that so anyways the intent thing or TR TR what other intent Frameworks do you want to compare

it to well we can start with aoma I guess they invented the word and then we can kind of go downwards through everything from cow swap through cow swap invented it didn't they they inv ANA used the word intent first though like Thea blog is yeah that that is it was the vi what is it called VI Vin vver viver yeah I can I know I see it spelled but yeah it's hard to say so maybe I should say cross chain intense every time I say intense but I don't know yeah I I think I think it is a kind of word that just got like in the same way that like rollups um you know there's always a new adjective in front of rollup there's

like X rollups and the lexical complexity is just growing over time there's like a million new rewards for things that may not be that different uh I kind of feel the same thing with intents except they're hugely different in some ways yeah yeah so I'm saying Crush chain intent from this like Bridge or or swap or call those are like what I'm kind of constraining it to and I I think there's other protocols in that sphere so like I don't think that's a Noma um what do you think about unop X because it's close to that right very close to that but like we worked with them to write this standard for 7683 I mean long story short I thought Unis swap X was going to like eat our lunch when they announced at ECC a year ago

they announced cross chain Unis swap X and the cross chain version hasn't shipped yet but they've actually figured out that there's useful ways to work with us um which we've been working really closely with them on developing a standard because we realize that the best thing for all of us is having a deep and liquid solver Network and our solver Network we want to be the same um and there's more I can talk about here too where yeah it's a great time to to move into talking about etcc because we just spoke on a panel about where the audience was all solvers so how deep is the solver Network in E well so before we do that why don't we we have a bottle here that we've been eyeing and waiting waiting to pop this this is a so why

don't we actually pop it and we can keep talking about intense actually cuz I think it was a topic of this week this let's do it while we drink this is now the third annual uh Anna andun drink and recorded podcast episode always recorded in Europe apparently wait there's I think there was another one wasn't there there was one with Zaki and Uma that was that was last year then there was Dev connect with Geo where that one was drank too drank way too too much yeah wait really you guys drank a bottle of tequila brought he brought BT of tequila and I was like Brendan was there too oh my God I was so my head hurts so much cuz I don't like that is one liquor I try to avoid but G was enthusiastic uh since then we've known

we know better and we stick to bubbly and it's fine and it doesn't hurt us too badly um so yeah I'm gonna you guys keep talking intense I'm going to pop the bottle the audience might hear it but do we record it I'm G to try I'll put the mic near it but you guys keep going um so where were we Unis swap standard yeah the standard so we worked with Unis swap on this I actually think uh cross and unap X are complimentary since Unis swap X is focused on swapping um so whoa I hope that got recorded that sounded great I see I see I see a pop it seems to have worked all right um so unit swap X is focused on swapping uh cross chain version is focused on swapping we're focused mainly on single same asset

bridging um and I think there's a sort of speed are you optimizing for Speed versus optimizing for price and sorry there's a small debate between little cups or big cups cups uh and the little cups won um so yeah I I think I think that the other intent-based uh bridging protocols like us uh I mean they're all pretty complimentary it's like one of those things where I want to push for a standard because the solver Network like okay andna what we didn't touch on is like what is the biggest critique of this architecture and the biggest critique maybe trun slightly mentioned

it earlier it's just what if no solvers show up oh that's the biggest critique actually I had I did have this question I forgot to ask which is like how many solvers are there yeah so how many solvers are there we have about 15 to 20 solvers in our Network and we have seven or eight of them that are like very meaningful like do and there's no one dominant one I uh that's good but it does mean that's a number of people that could maybe stop solving too do you remember do you remember when uh layer ones would always fight about like we we're a blockchain but our blockchain favors safety over liveness or or our blockchain favors liveness over safety and there's it's it's kind of they the the marketing speak was one or the other

but really was like some something in the middle you could think of intense as being favoring safety over liveness like it could lose liveness if all the solvers show up but you you're never losing funds right there's no there's nothing minted or burnt or you know you can remove from the escar contract after a certain number of blocks if it wasn't fulfilled right but in the other case the ones that try to give you safety and liveness they generally have to make some other trade-off either there's a an inter mediary who's running both blockchains or there's this kind of wrapped some control over over wrapped assets and uh I I I think like there it actually reminds me a lot of L1 design decisions it's just that in l1's you really can't

tolerate loss of liveness whereas here you kind of can cuz like what do you do you retry the transaction it's not it's not like the end of the world in the same way where like if you lose liveness in salana it's like oh everyone got liquidated because the offchain stuff moved so far yeah I think that's exactly right and the other thing I just push on is it's like the liveness properties are pretty good like competition is a wonderful thing and um it well empirical liveness proper not not like theoretical I think that's that's a key distinction right where whereas in in layer ones you actually need the theory and the empirical to be the same because your threat model relies on the theoretical

thing right like of like all the possible attacks I show that none of the attacks work but here I mean the attack is like everyone suddenly loses the internet at the same time and it's like unlikely yeah I think that's right even though it's possible it's very unlikely compared to like in cryptography or in dist systems where you need not only does it need to be basically impossible but like the likelihood needs to be effectively zero in some very controlled manner here it's it's not that but it's like sta super unlikely but what if it was slower that like you kind of threw that out there that solvers just stop solving like they they just take their money out it it but like this is the thing where it sort of like breaks the

entire concept of like a blockchain in the first place it's like what if Bitcoin miners just stop showing up right it's like why do they show up well they're economically motivated to and so I I get slightly frustrated when people push like in weird directions on this I'm like no man like blockchains work because they incentivize actors for economic reasons to show up I will take the I'll take one other side of this though which is like I wrote this paper on intense and and it's sort of was really trying to analyze the unisoft X style intent not necessarily this the single chain version yeah well actually I'm not sure actually if I think our result probably still applies to I'm not

sure that it does but that's another debate for another time but I think it doesn't affect your use case because your use case doesn't have this idea that like the action that's being taken implicitly changes the price of the fee so it can go from being uneconomic to economic depending on the the actual asset traded right if it's just the same asset Crossing you don't have this problem but the problem that we point out is like if the solvers are competing but they're sourcing inventory from somewhere and they all go to the same place then they all push the price up right if everyone went to binance at the same time to buy the asset to give fulfill to you then well their profit goes down in that case competition's

worse you may want oligopoly and fewer participants and so the the swapping example is much more complicated to reason about than the single same asset okay same assets is like basically hard to get around there's something else we're working on that I actually think might solve that that we're not ready to talk about but I'll tell you after the podcast okay who teaser yeah spicy damn do you spill champagne I picked I pushed for the small cup and I regret it the small cups are too small they're too small and I just spilled on myself um so the only last thing I'll say on liveness and then let's talk about ECC um it also gets better over time or like the pro the the the livess

guarantees get better over time like more volume more orderflow means more solvers in the network more liquidity in the solver Network and just better guarantees so I feel very good about the general direction of like that quality and you also look like our our our data and our stats and across right now our our 90th percentile fill times are under 10 seconds which we're pretty proud of and we're going to push those to under five seconds this quarter and we're doing that because there's enough solvers and we've made the the protocol easy enough for them that they they they show up and do their job because they're economically motivated too nice okay we are at the one hour mark whoa recording

I don't know if it'll be that on the final but I would love to take some time to do part two of this episode which is to talk a little bit about Brussels the city let not talk without Sprouts let's not talk about the city let's talk about the events and things that we've seen this week so far we're kind of halfway well yeah I don't know how far you are along are you almost done tun are you tomorrow I view as my last day and then I'm like I'm I'm in the middle still of everything that I'm doing but this is something we have done every time right so we get together we usually have a guest or two guests we haven't had on before heart you're the guest this time um and you're also a fellow podcaster we didn't say that

early on you're the star of a popular podcast called the star of a popular bell curve yeah no I did a a season with Mike my buddy Mike EPO although we did record Another bell curve episode yesterday I'm recording eight different podcasts this week wow yeah I'm doing nine but you but I'm not a professional podcaster but you are now heart yeah welcome to the club it's fun I like it tun is also a professional podcaster I compared I'm only on one podcast you're on many many podcasts yeah you know remember that time there was this one year where you

just gu you were guest on every podcast and I started to get jealous that's you know that's that's the sign of the B market we're back we're back in a bull if you see me on too many podcasts that that should be your like astrological sign of Hit the buy button sell button sell oh sorry yeah well you should start sell unless you want to be theity end your tour yeah okay so yes let's let's set up the scene though so and I said this earlier but we're here in Russel kind of you know partway through this week of events or for some people close to the end so let's just share with the audience like where we've been and maybe like actually let's start with that like what did you see what did

you actually go do this week because I feel like we've all been in slightly different scenes like I you know I was in a lot of ZK events I feel like maybe you were over in the reaking tun I actually don't know where you were heart defi stuff some defi stuff some yeah yeah okay why don't we start with you though where were you then well I showed up late which was a great um so I I basically got here Tuesday afternoon R rolled off the plane off off of a Redeye flight and did My ECC talk um which was great and uh apparently I was a clear enough speaker which I I'm personally I'm impressed with myself by um so yeah thanks for the support

um but was that like P you just did the Obama giving himself a metal meme buddy I was tired I got off the plane and like I did a really good job I did it in sarcasm guys that's sarcasm um true Canadians are too nice for real sarcasm note toelf we're not nice we're not nice I'm less and less Canadian these days but um love my country um we are not the Belgium of France um just love that joke so much okay so I rolled as a plane gave My ECC presentation had a bunch of meetings was at modular Summit uh today uh but I've been like I'm kind of annoyingly been

more focused on the multi-chain uh interop type how do we solve interop things um yeah that's where your head has been at that's where my head has been at I have been at I was at a risk zero event on the first day doing a panel was kind of state of ZK I've been I've been very state of ZK actually not even before this uh zkv put together we do this every quarter I I don't think I've ever mentioned it on the show but we do this like state of ZK report once a quarter and we've been doing it for about a year it's a great like review of everything that's happened and so the first panel that I did was about that sort of reviewing it and then I oh I was on an investor panel

me but you are an investor I know but it was like very I like I don't think of myself that way somehow and I was like I thought I found it very funny to be there to be like Oh I'm giving advice yeah it was strange um I was the only angel investor on the panel too though so I I got to be the one who's like the big guy don't trust PES they're evil but um yeah so I would say like but still very very ZK so yeah my world was very ZK what's going on in ZK I mean one of the big topics was like sort of this emerging I know it's a bit of a meme but it's a ZK supply chain I've had a few episodes kind of leading up to this too that touched on that this is like I think we've always thought of

the ZK roll up just like connecting to the main chain but like what is actually happening in there and can we break down some of those things and decentralize some of those things and I think it's still a question mark if we can or should but I know there's teams that are trying and so these are things like Pro networks and then the verification aggregators and maybe there's even going to be more things in there so that's where my head's been at uh where has my head been um you know the last 6 months have been I less than six like four months have been like me kind of being really running into brick walls many times on this like research problem I've been working on for reaking but I finally think I made solved 99% of it uh and so I gave I actually gave the same

talk twice and maybe I'm going to give it the same talk Thrice tomorrow with this tiny edit because I think it's one of my one of my favorite talks I've ever written and what is this is all on rest it's all you figured it out I think I figured out the thing that I think is the most important part of the security model for reaking okay so I I I don't want to take too much time but I can give a little detour no do it what's the title what's the title tun uh well the title of the talk is how much security does your AVS really need lessons from matching markets um but the we spice that up a little yeah probably um and I uh basically okay let's let me I'll just give you the

really simple version of the problem and then two different versions of solutions so the problem is imagine you wanted to attack two proof of stake Network simultaneously say salon and ethereum well to do that I let's say I was trying to double spend and let's Suppose there I would need of both right separately now in reaking when I reuse stake I actually don't have to buy 50% of both I can buy a smaller threshold and attack both simultaneously so imagine if you could attack salana ethereum at the same time but you only had to own 20% of ethereum and 5% of salana but do you get less like would it be like you can attack but it's somewhat less beneficial

like it's not as good a score uh or is as good a score profit same profit same attack it's just that because you're sharing the stake the this sort of attack you know you're able to get economies as scale because you're reusing the same sake right so then a natural question you might ask is like hey okay if I can do that then there should be some kind of like economos scaly looking thing that like looks like defi where where you might have some notion of a Cascade where you can attack a set of services by you attacking a set of services you now remove a bunch of stake that stake allows some other set of services to be attacked and then those get attacked and then that allows not so it's like the it's the sort of proof of stake analog of a lending

liquidation Cascade or perpetuals liquidation Cascade it's a kind of interesting security model because it's it's even though it like looks a lot like the def stuff it's still mathematically very combinatorial like the proofs of safety and liveness and proof of stake protocol so it's sort of like somewhere in the middle of the two and formulating that Cascade thing correctly is is actually non-trivial and there's a paper that's still not out uh I don't know why it's not out yet because of I mean I think they sent me the original version I heard I heard about this in like February uh and I read the initial version of the paper like three or four months ago which is showing that like hey you if you're you have to be really over collateralized in order meaning

that the amount of stake people have to stake has to be so much higher than the total sum of all the profits uh in order for you to be secure these Cascades so what that means is like imagine I could profit a billion dollars from attacking salana and they could profit $10 billion from attacking eth and they can profit $1 from attacking cardano if they were in this reaking world attacking cardano even though only get $1 to make cardano secure against reaking attacks it would have to have 11 billion1 of stake like something proportional to that so it have to have something like near the total sum yep right so it's like insanely over collateralized right very inefficient but uh this paper which sets

out really nice definitions allows you to kind of talk about this Cascade stuff it is it ignores one thing which is kind of similar to the type of thing Hart was talking about earlier which is it ignores the incentives completely it kind of assumes the adversary who's attacking is arbitrarily powerful they can get arbitrary amounts of stake across all the networks for free like cost free there's no cost of attacking multiple and the node operators the stakers they're sitting ducks they can't change their stake they can't like oh I should saw something at slash adjust my Ste so it's kind of this like you know you have someone who has like the world's most powerful rifle and the other hand other side is like handcuffed node operators right it's like it's like

a very know so wait but did you figure out so so what I did is I it took me a long time to figure this out with me and my co-author that you can weaken if you can weaken the adversary a little bit and by weaken I mean they face costs for attacking like they if they try to attack one well when they buy the stake to attack one they actually push up the cost for the next one right CU you know and if they face cost and the node operators are allowed to be smart they can rebalance in response to slashing events then you can control you can bound to this Cascade risk and it doesn't have to be as overized but you need one important thing which is you need each service to offer a sufficiently high amount of rewards so that the stakers re know where to

rebalance you can think of the rewards as a signal of where where you should rebalance like literal like yield chasing but that's all they need so there's kind of a weird sense in which the choice of security model for reaking is actually even though on the surface you might think you want the exact same thing as proof of stake if you want the exact same model as Pro stake you need to be so over cized that you might as well have just been an independent networks like what crazy and so anyway I've been thinking about this for amazing I'm glad you did that though I'm glad you put that together and that people can now see it yeah yeah well we did just by the way we did just talk about our own talks we should also talk about other

people's talks or other things only our own I like tun's Talk's talk I'm joking I'm joking I I didn't see I didn't see that you didn't see any other talk you may not have seen another talk but I bet you talked to people yeah yeah no no I actually went to this FH Summit that was pretty cool I saw some interesting like performance numbers and really see I have heard through the great fine that like the performance is still unusable like who's doing great numbers so I will say is it that one company that you invested in no I think what I'm saying is the following there is a a a a growth rate that people have predicted for performance yeah and then there's Real Performance and it has not matched the

growth prediction kept up somewhat but then in some tasks has so it's like it's kind of like you can always kind of point to like where it's failing but there have actually been some advances that are still cool ex ex so I I thought that was cool that's fair uh the other thing is I feel I've I've gone to a lot of things where I over time am becoming much more favorable towards te's um I know it's our audience like to hear this this is a theme I've heard too yeah yeah I know a bunch of people that are like te's are bad te's are bad they're like yeah maybe it kind of works it's okay for some things it seems to work especially if you combine them with other things like the example I gave on the panel that we were on of

like slashing rules where like they become a lot simpler when you could do them in FH or MPC or te in that threat model you don't have to write as much code because the adversary can't read the thing and and it's it's like there's actually a lot of Beauty in the Simplicity and I I think I've come to appreciate that maybe we should say the name of the panel we were just on we were all on it just now Midway through this it was information flow control panel panel but control panel control panel sort of stand like it it was on the second line it was kind of confusing um but that was kind of fun that was one of the things we did I want to know like I think it's hard to not mention us

being here like the vibe generally because it's different from last year and the year before last year I remember us saying things like it's sober but this year I think it's like more sober I don't know about you do you feel excitement no big announcements I will say I didn't at least I didn't observe any there were some ZK announcements but yeah they weren't like massive they were actually projects I didn't know but they just done something cool but nothing like nothing that would make kind of the mainstream news outside of the ZK space that's interesting in a way it's almost like uh the game theory of doing a big announcement at ECC is risky cuz you don't know who else is going to do a big announcement this time they all didn't they all didn't right or there were none

so we need a te so what happens when the solvers all disappear exact well you need a te where everybody submits their big announcement and decides who should announce the big announcement oh right an auction yeah who gets it who gets it CU yeah this is the beauty of crypto you crypto people can can be so good at making products only for themselves to pick the T the the auction to pick who gets to do the announcement at the next big thing do you think people are waiting for Devcon I mean honestly I'm excited no offense the Europe and America kind of suck in cryptand but every time I go to Asia it's like even though you're sometimes it's like crazyer everyone is like so much more

enthusiastic and excited and like I feel like that does not feel you know you're saying it's I definitely don't feel like that here yeah and I don't know if it's Brussels and I don't or if it's this year moment is like oh there's some new exciting even eth Denver felt more exciting because like people were going crazy about reaking eth Denver felt more eth Denver oh because of reaking because no no but there was like this huge because we're also living through and this is something like I I almost put into the state of ZK but realized it was had nothing to do with ZK but there is this you know the last six months there's been like and I it's not really I don't want to blame the teams doing it because it's kind of not their fault it's weird timing whatever but there's

been a very lackluster saying it nicely response to launches I mean you know this right I'm saying like it's actually worse than lackluster it's it's like it's sad when like the response to free money yeah and also like the mem coin stuff I actually feel like that's made people very cynical like this financial nihilism stuff seems to have bled Beyond its borders and right now we're in the like but isn't that didn't you once argue for this didn't I feel like last year you were like the nihilism is good or something and I was like no and now it's taken over I think I think it went too far like once we got to the celebrities thing it's like okay but we were there already but it's so much sadder this time no it's much wor yeah

yeah it's it's like I don't know I forgot Jason dero existed I really would like to go back to forgetting he existed you guys are getting kind of cynical right now yeah wait har do you are you feeling good I feel great all right give us the counter well the counter I mean I showed up here on Tuesday afternoon like halfway through which I think makes a big difference cuz these events are getting too long like oh another thing related to this there too many side events yeah too many but there were actually too many side events this time I I think it'll normalize we did two of them I am responsible for did two of them we never done an event two with two this our first ever event we did one and it was like oh my God this is when did you do one on Wednesday I

flew ining research we didn't know you didn't invite us you didn't invite me wait I did it public it was on Twitter it's on dude I want the personalized invite yeah I'm sorry I'm sorry first first of all I I like I like did I like came late and I almost missed we're just passive aggressive bullying you don't worry about it it's okay but I do agree there are too many side events the incentive structure there is also kind of messed up right where people want to have their own Side Event because it's their thing and I think it's the tradition too like you're here you're supposed to but there was definitely too many but like I I so so I feel like a lot of electronic music festivals have the same thing where like

one of them gets successful and then people start throwing all these side festivals and then it cannibalizes the main one and then they like cuz like how many talks did you go to at ECC zero I didn't even go to the you guys have like a ticket did you go to I didn't have a ticket yeah just I mean that's kind of like the problem right well are you supposed to go to the big event I think Devcon if I was I'm not actually going to be there but if I was I would also probably not spend much time in the main I don't know I feel like I feel like everybody should go to some talk sometimes yeah because it's like and I didn't I'm kind of full of because I didn't actually get to do that because I show up late but people

should people should and like the modular Summit I'm going to go to some of the talks tomorrow and on Saturday and I feel and I did go to a talk to you can see my talk tomorrow yeah I'm only slightly modifying maybe I should crash that talk too I should or if you're doing a panel I should just like I have a which is the also like this drinking episode this the third annual one I have a third annual panel of me Mustafa I can't crash that one you're doing that again tomorrow really yeah that's actually kind of cool that's at the end of the day yeah you did it last year I did it last year I did it the first year yeah ni I'm running out of questions to ask them do you guys have any ideas of like what's happened in the

last year I mean mem coins I mean that's the number one thing I'm ask about comes to ECC no he does it by video um that makes more sense but respect to all the salana people who who came cuz I feel like I feel like it's it's kind of interesting like they they they are clearly very curious people well on that like what do you guys feel about like the salana vibe here like how is the minimal uh other than Haven noticed Kyle and mert I didn't talk to anyone who was like Focus Kyle's here yeah he was here he left yesterday this morning sorry all right yeah I didn't see any salana personally I didn't see do you think it's getting more it is getting more heated right what salana eth well oh I guess we didn't talk about the the make

block times shorter in eth uh and like all the salana people were like wow ethereum's becoming slow Lana I thought I was like that's like a funny I like had not until I saw that part of Twitter I was like well slow Lana is pretty good I was like that's crazy I do think there'll actually be a move over though I feel like the people who are underneath the PE the real dedicated folks like I don't think they're going to leave and go there but it's fair weather friends that'll go there that's not what it's about it's about like the new people right and it is the the War I like the War I like the competition I think it's healthy right but the next billion people where do they build um or the next bilon users where do they go strange because like

where do they I mean there are also the sort of like Silicon Valley what do we call them like the those teams that are made of X meta X sort of Legacy easy to invest in projects which I wonder if those won't also potentially take I mean if you're a corporation do you want to trust I I think the most the most important thing that we've learned from the failures of a lot of layer ones is that a the people working on the protocol core protocol can't be dicks like they AR these aren't as they're just they're just more like I will I will say very few have been able to pull off what ethereum and salana have where

they have this community that even when there's a gtic Justin son he's a dick he's an exception okay we'll get we'll get to we'll get to that we'll get to that we'll get to that and and you have to have a lot of respect for for for Justin he's the largest liquidity provider in ethereum um is that true he is one of them yes uh he's but but you know a way to kind of conceptualize this fight is people in salana really care about a very different thing they like they care about the end user they don't necessarily care about all these particular theoretical properties that we've held near and

dear and they haven't nerd sniped the same way e they bond over that and and it's it's they they care about the performance stuff whereas like you look at the other high performance chains and even when they do amazing research or they have this like amazing consensus protocol or whatever right like they just don't seem to be able to communicate that to their users like their users don't get that and that Community aspect is actually one of the reasons I think salana was able to get a big developer Community get users who wanted to use it was they have a leader and developers who are kind of not dicks and not like oh we know everything right whereas like news flash I can't really think of that many l1s that don't have

Founders or CTO or devs who who fall into Noles or like all I mean he isn't but he's not that accessible actually he he's a good explainer to technical people wait who is this vitalic but when you look at like the snark blog and stuff like that like yes they but he is still a community inv person right in a way that like that's not you cannot say that for these other El ones they're like we are the dev team you're the user you know like it it has that and doesn't have like anat is like still literally constantly in all these GitHub threads like you and he's like he's like

you know he's he's obviously crazy and like Manic and like tweeting all the time and like you know but he's like he's still he's he's very focused on even the smallest user which is crazy right A lot of people they they they get rich and they tell their user to themselves like that's how he doesn't tell what was what was the founder who was it who said like do you know who I am when somebody said like directed him to customer and they said yeah hoskinson that's cardano that's why I bring like I think TR it's less about being a dick and more about humility right like aren't those the you can be a dick and not be humble but I think you could be you could be a humble dick but I think that humble dick is a good name for like a men's underwear brand humble dick you

could be a humble like not dude guys guys we got to start this men's underwear humble dick who's going to buy that uh I mean you buy it as a gift you buy it as a gag gift it's like wow it's a great G it's sorry any make it happen people I agree with you there's probably some difference but like yeah the humility humility and accessibility are vitalic and Anatoli both are humble people um and that's really very very it's also genuine they're not like there's a lot of people who like fake it and then they get caught and then not humility that's not real that's not humility that's but it has to be like sort of authentic

humility yeah it can't be um Obama giving himself a medal oh that didn't really happen tun I'm kid that's a meme poor Obama I was just referring to his no I know yes know trolling me um that's what I'm here for what else do we want to talk about from this last week oh cuz actually I did want to say you know I I mentioned I'm I was all in the ZK world and then I was at modular Summit I've also I'm on I'm going to be doing nine interviews this week we're actually recording this is a little little hint I I mean I hope this comes through but we we got a grant to do the second season of The ZK whiteboard sessions which is a series that we

created from like the ZK hack side and we just recorded the first module of the second season today so I'm very excited that's going to be coming back sorry didn't mean to be talking about myself I actually meant to to get outside of my own but what I wanted to say was um the things that I've been hearing about outside of the ZK World um and especially through these interviews I've been doing it's actually about interoperability about bridges and across kept coming up that's why it's also very cool and fitting that we just had you on to talk about it such a nice Canadian it kept coming up and I'm not kidding it just kept like there's new bridging there's new bridge ideas or people who work with Bridges and it's

just like mentioned over and over again so that kind of came up for me Simplicity is key in some some some ways you know I feel like that's another thing about bridges is that there was this race at some point where everyone everyone wanted to make their Bridge the most complicated right they wanted like every Bell and whistle possible wasn't it always kind of to make it the safest it was like safety but they they conflated complexity with safety which you know sometimes I think that that happens in this space this happens in Ai and crypto more than any other space yeah man it's so interesting now having been in this space for I I you said 2017 was when you got into it yeah 2018 2018 yeah for me it was 2017 I remember having you on the show beginning of 2019

tun that's when I met you but like you see through this you see these cycles and there's this moment in a cycle where you don't know what the outcome will be but there's like a couple projects that are like pitching you their vision and you're looking at them and you're like I don't know how this plays out but then you get to see it play out and I think what's a bit fortunate I think people forget to look back and think about like what was being pitched then what were the like four Bridges the four Bridge ideas and what actually played through because in the moment that they're pitching it to you you're just like this looks amazing this seems really cool you have to see it analyze it and then see what actually happens yeah do you feel like we just Chase narratives a lot I

feel like are you joking in crypto are we are we just like Anna can I tell you a little bit about an asset called a mem coin but like this is the the perverse incentives right where it's like you chase a narrative because you want your token price to go up yeah right and and you make the narrative sometimes you like craft a narrative which is hey all Marketing in the world did that I mean like big Concepts in the world have done that so like take celesia for example like they crafted the modular narrative very very well it's so it's so nicely done it's so well done it's like a work of art in my opinion but spread they backed it up with a great product I

think or like they built real Tech too and so like to me that's the that's the dream yeah I don't know I mean way too long a pause right there we might cut that I think I think like there too much gravitz I just sort of think the the narratives often times just run at a faster speed than reality right and that's in and of itself that's not the worst thing but I do feel like in crypto it's like we're still a space that needs to prove itself to like 90% of the populace right like it's not it's not like AI in that sense where like you

don't need to prove to the average person that that there could be something there I still feel like we still have that that issue outside of you know using stable coins and turkey that's the only thing I could probably con conceivably tell my grandmother I just I do think I mean last year I think it started Shon on me the the AI you know we we got really excited T you and I on the show and I mean you were into it before but like what was it 2023 was when Kobe modded my voice and created zk. and I got like really freaked out and last year I was really like I think what we're building I don't know if it was crypto but it was something like maybe immutable Ledger plus ZK or something like it feels like the human it's like

this creation out of the human mind and the human populace just the fact that there's so many people that have to be involved this like this is the crowd in a way that it would somehow stand as like a guard against the AI AGI kind of world but I don't know like lately I feel especially with the mood and the focus of a lot of what people are yeah the focus of the industry that I'm like are we are we going to be able to be like the human counterpart or the human I I say count like counter actually like thing that keeps it kind of in check or like you can actually control it somehow through this like Mass Hive like what do we call or like just

participation yeah I think an example of where crypto like the incentives seem to to be kind of poorly made in in regards to this is like the 5 million AI tokens that are about decentralized training and like stuff that's computationally basically infeasible not very useful to most people most people are going to take a me a mega model and maybe find tune it a little bit but like it's not not a real thing but yet there's like a million of these protocols for doing this and they all have tokens with market caps yet you know what would actually be a much more useful Ai and crypto thing is just making it so I can do use intense while with natural

language like I don't want to that's that's incorporating AI into it for sure but that's it that's more honest the other thing is like computationally useless like anyone anytime I talk to my friends who work in AI they just like laugh at me when I explain to them these these crypto things they're like why are you people such clowns and I'm like yeah I agree it's true but I do agree with tun it's like what is the crypto xai kind of yeah what is the actual point of it making it easier for users that's what it should be right like to me that's what it seems like it but I feel like they were trying to use AI to make crypto easy but shouldn't that seems like a way higher value than the other thing so I was

talking about like using like all of this stuff as like a counter to you mean world coin style I think the way you're supposed to don't say that but go back so okay so we want to make um use AI to make crypto easier that was jun's Point yeah okay fine that actually does make sense sure it's just not a business the problem is that's like so no one still a thing but I was actually saying use everything everything we're building to keep AI in check I'm like completely on aerous maybe maybe don't make it keep AI in check although that's let's like say make AI better yeah like nobody's come up with anything that uses crypto to make AI better and I think there could be there are I mean in the ZK land there's a few ideas but yeah their ideas

they seem untractable particularly consumer friendly and I'm not convinced yet of like the use case I think the idea of like hey could we build stuff that I've thought about a little bit could we build a Marketplace for AIS using crypto like technologies that reduces hallucinations type thing okay like there's stuff like that that might be kind of interesting of like how could you take what AI does and make it more accurate maybe crypto crypto could help there but nobody's really going down that vein yet fair I feel like we're about to get kicked out of the room we're recording in yes I keep feel kicked out half an hour ago we did get warned yeah that was we it was actually

half an hour ago I think I think I think this is actually I think it was like 40 also we're all drinking this bubbly have drank this bubbly on an empty stomach clearly because it wasn't that much but I do feel like we're being kind of silly so maybe time to wrap up it's time to wrap up post production will be great post intruction will be minimal thank you Henrik thank you Henrik I apologize for brain cells all right thanks everybody thank you guys for doing this this episode thank you both so much it was so fun to get to know across too first time on but I hope not last time on I hope and to our one listener GMO we're never doing tequila [Music] again unless you're here then maybe all

right uh I think I'm supposed to say thank you to the podcast team right Rachel Henrik Tanya sometimes yonas and to our listeners thanks for listening