Sovereign Interoperability or Get Rugged | Karel Kubat - Union.fi

[Music] all right everyone thank you all for attending hi team um so today I'm going to talk about Union The Sovereign interoperability layer and to be honest about secure interoperability in general I'm Carell ZX Kaiser Carol on Twitter I'm a developer and systems engineer by background and founder of Union Labs uh where we do light CL based interrup uh and basically I think some of the best tech in the space our belief really is a massive horizontal scalability we're building towards a unified rup Centric future I kind of believe in the 10,000 EP specific l2s and everyone having full control of their entire stack a state machine for every single product out there all asynchronously composed but in the end achieving the same ux as monolithics Tech so similar to salana or ethereum that's really the goal what we're building towards what we're seeing right now happening with Celestia and such perhaps it's going to be kind of the next 30 years of crypto perhaps it's only this cycle but we're Believers in horizontal scalability first I want to dive into a little bit why we do kind of web three composability and what we're really building in the end of course the objective of crypto is to build the world's computer for those that read obviously the ethereum uh yellow paper that's what we're building initially uh single Global State machine where code is law all trustless verifiable without intermediaries uh this is obviously a great use case for defi because Finance in the end is kind of the Forefront of where we right now compete with each other and where we do want code is law U but in general this is what uh really can enable the future um of I think almost any application and kind of set us free from uh not necessarily regulations but free from unfair regulations or inequal standards what's really great about the world computer is composability which allows for a Cambrian explosion of products really with smart contracts we build tiny money Legos which we compose into F D5 products right you build a simple decks you build perhaps a reaking product and this allows a third party team to come in and just build an entirely new product which is why ethereum is so powerful right there's no need to WID list new products the ethereum Foundation does not need to interact with the smart contracts being deployed and you might build a product that in the end forms the backbone these individual protocal interactions are of course extremely secure when one smart contract calls another smart contract that's no issue whatsoever uh you remain within this secure World computer state where no one can mess with your API calls they don't disappear however for me and for Union we believe the one Global World computer in the end is dystopian uh single world computer isn't sensorship resistance it's basically a single sequencer queue and most of all it also cannot scale it cannot actually uh Bank the unbanked it cannot handle Global Finance uh in the end we're still stuck with at most like 400 TPS not something that can handle World Finance modular architecture solved this in a few cases and uh on the security perspective is extremely competitive we've got shared security which across different l2s and app chains leads to massive Economic Security for any app we obviously do execution proofs in ZK rollups which to be honest are even better than shared security because they provide mathematical indisputable correctness and finally right now with like Celestia and aill and such we finally have like data availability and with Austria and conduit uh we separate out sequencing into different stack which helps against uh sensor attacks uh in the end right now in the M Stack uh I think it's already better than kind of the monolithic L1 stack and modular is on the rise uh the amount of chains connected to each other uh through kind of standardized protocols mainly IBC has grown tremendously and is only more growing however as it stands right now kind of this whole modular thesis completely falls apart uh it just cannot succeed yet because what we gain with all this fancy technology in security we losing composability at the moment we like to refer to asynchronous composability which really is send a bridge message from one chain to the other between two smart contracts and handling that and right now it's simply not as good as synchronous composability we buildt all these fancy bridging standards right perhaps a single signature relayer perhaps a multi- bridge which I think everyone is aware it gets hacked all the time uh maybe NPC based like XLR um and finally you can go into consensus verification State verification but we really don't see many of those Bri yet out there we're really in the kind of leftmost quadrant at the moment um and without proper asynchronous composability really uh modular fails the current interrupt that we have can get completely attacked by many different parties out there um most common what we see right now is kind of hacker groups like Lazarus attacking Bridges and emptying them out but what we've really forgotten as well is that in the end we built crypto to create a system that's resilient against nation state attackers right and with current inro protocols we have oracles validators multi holders situated in the US so obviously that's a target for the SEC we have Bridges being completely in control of their team and to be honest like why do we do smart contracts because we don't want the team that has built the product to control the product we want something that's pure that's abstract that's mathematical so right now if I want to do composability between an app chain and some like app specific L2 for defi this composability can be completely attacked in the ethereum space an interaction between Unis Swap and a vault contract obviously completely secure and truly we cannot ditch uh interop security as I said right now we basically enshrined oracles as intermediaries uh deviating from codus law we completely lack Atomic execution uh right now for those building on top of bridges so many different broken states can happen inside of your protocol funds are kind of in limbo they might reappear on ethereum they might stand in uh intermediate chain really synchronous composability at the moment is web 3 and to be honest asynchronous composability and web 2 is is really web 2 uh bridging to be honest has been a bit of a disappointment really with Union what we ask ourselves is how do we make asynchronous composability as good as synchronous uh What uh we are truly obsessed with is bringing this trustless security level that you see in the evm to the full modular stack and this is probably some of one of the hardest problems to solve in crypto at the moment uh touching probably every single point in the stack really it comes down to a few key points and one of them is a very philosophical one what we need to do is truly invert control if you look at how interrup is being done right now basically your bridging Co protocol feds different chains and might choose to support them for those that have tried to get support from perhaps layer zero or Wormhole this is only really possible if your chain has a massive community raised 100 million from a16z and it can be next in line otherwise you might be waiting for 12 to 16 months to get support to me interoperability is not a privilege dydx was capable of building like a great app chain specific product because they had the community in the following to get bridging support where we have to ask ourselves how do we enable the next 10,000 startups to also build their app specific chain or the app specific L2 and G full control of their product if you think about it doing this like like app specific model it's far better for the product in the end who really wants a product where your users pay 300 million in gas a year to the underlying L2 right you want that value to flow back into your token or you simply want your user interactions to be cheaper and for them you use your product way more but we need on top of that is Unstoppable systems uh so I've already alluding to this but if we're going to get this 10,000 rollup Future these specific app chains not all of them will be aligned with certain jurisdictions there's differences in regulation between the US and Europe for example does that mean that if Europe Outlaws privacy chains but the US allows them intro providers need to completely outlaw them as well is that a fair solution I don't think so and finally we need to hypers scale what we see right now is Bridges kind of support 80 to 100 chains uh some of them struggle to scale up and really no protocol out there right now can handle this massively scaled uh future and in the end we need to to abide by a few things two of them obviously has to be permissionless ties into the inversion of control and it has to be fully trustless uh we cannot have a system with oracles we need to have the same quality as ethereum has to me security really is not pick and choose uh just like every website should have SSL TLS interoperability should be fully secure right now we seem some bridging protocols kind of go into the pick your security model choose what security you want for what message but really what this leads to is fragmentation again this makes it completely impossible for Builders to reason about their product what they're integrating with and just adds vulnerabilities uh tornado cach was a fundamentally sound protocol uh worked perfectly fine except people didn't have a proper look at the Dow itself and how that could be used to take control of the protocol um seeing that as kind of a lower security standard than you need for the actual mixer and that was used by an attacker to gain full control imagine perhaps a protocol saying that they want the higher security for asset transfers because those are most often stolen but then doing some of their governance messes with a lower security standard allowing for exactly the same attack to happen now really if you want to get this highest security level across the stack you need to do consensus verification consensus verification is what's used by suin Labs telepathy which is what all of this like ZK light client hype is about and inherence the underlying security of the chain instead of relying on oracles Rel layers or perhaps an igen layer AVS instead you inherit the security of the two chains connecting to each other and when you think of consensus verification you really have to go towards zkps zkps make sure that you can do consensus verification across this full stack because in the end we don't want to be spending about 50% of block space on interop right if we have this future while all of these different l2s are communicating with each other we can't be filling up the blocks and wasting millions of gats on just doing these bridging transfers so we need something that basically compresses all of this and that massively scals it if it's going to be as good as smart contracts communicating with each other and in the end this permissionless nature leads to hyperscaling uh interoperability is a right and like I alluded to if the ethereum foundation had needed to Whit list every single contract we'd have about 100 contracts on ethereum right now and single no traction at all on defi ethereum in the end would have failed in the end for mod to succeed if the intop player itself need to vet all of these different rollups appearing it will fail and won't be able to outscale monolithic Stacks to us we bet big on IBC the interal protocol that originates in the cosmo stack but right now is scaling outside of it all of this IBC traction that we seen right now supports most uh more chains than any other intro protocol is completely organic it's because chains can integrate themselves to the interop layer instead of the inop layer needing to integrate chains and that's what to me at least composability in crypto is all about you can choose to integrate with other parties without needing their permission to integrate with you so what are we doing uh with Union about this we're working on basically the first ZK IBC solution initially we focused a lot on the cosmos ecosystem uh we support any Cosmos chain that's out there and more recently we're diving into Celestia and aail specific rollups supporting any Sovereign rollup that's out there more recently we've started working together with scroll on their uh lightline technology to connect as well to IBC chains and so this is kind of the architecture that we're releasing this year that's going to enable us to scale to 500 to 1,000 chains this to me is still kind of the suboptimal architecture because in reality what we're doing is generating many different CK proofs and sending them to many different chains so in the end the more ecosystems we have the more proofs we are generating and the the more proofs need to be verified on chain the truly cool technology is the next step to this which relies on proof aggregation where we take all of these light client updates generated on all of these different chains and create one proof to rule them all this is where you can do the really cool stuff like coincident of once uh just in time liquidity and basically you end up in a state where every single blockchain just needs to allocate about 230k gas a block to have full interrup with whatever other chain they want be connected to 10,000 chains at the same cost as being connected to 100 chains this is kind of what we're working towards in our 5-year plan at the moment proof of proof technology does work it's simply too slow you cannot generate this for every block yet for this we need GPU acceleration and ZK S6 but this is coming and this is kind of the end state of interrupt the go protocol we can do better than this so my take-home message for you guys and what we believe in horizontal scalability is the future uh even though we scale ethereum with rollups we are sharding it we are creating more rollups more app chains are created every day uh different l2s with different architectures move l2s Eclipse l2s uh we're seeing this garine explosion of products happening right now to enable this future to actually work and compost and be attractive to end users we need live clients we need zkps um and finally the take home message the future is modular if you guys want to follow more on us uh check out our progress go to union. build check us out on Twitter at Union build and give me a follow at zerx ker Caroll I tweet about this regularly and I think do most of the Mind share on this thank you guys