Across + Intents: Solving Interop with Cross-Chain Intents - TOKEN2049 Singapore 2024

hey all um happy to warm up and kick off uh token 2049 here so uh I'm Hart Lambert uh I'm the founder of a protocol called AC cross I previously founded another crypto protocol called Uma um which is used to power prediction markets like uh like poly Market um and I'm here to talk about interoperability uh and how we unify all the blockchains so they feel like one user experience um all right so what do we want we want fast cheap and secure interoperability that is so seamless when I'm talking between blockchains it feels like I'm not talking to blockchains uh the mental model I like to use with my team is users just want to click buttons you want to click buttons and you want them to do things and you don't really want to care what blockchain you're on or what blockchain you're talking to and so um when people think about interoperability between blockchains they often think about tcpip they go and they analogize it to like okay it's like we have different servers all over the world and we're going to connect them with a cable and we're going to send messages between these different servers and I think that's a very like useful analogy but I don't think it actually applies to how we want to connect blockchains um and uh like to go a bit deeper on this my slides moved a little bit here but um bridging when people think about it in this kind of like tcip type structure bridging means I have blockchain a I have blockchain B what do I do well I send a message I if I'm going to move value between these two chains deposit on blockchain a wait for finality on blockchain a send a message to blockchain b release funds on blockchain B that's how I'm going to move value between these two chains and um that makes sense except there's problems for this and the problems the problems are that sending that message quickly cheaply and securely is basically impossible um and I'm going to walk through each of those aspects and why this intent-based architecture is is the right one um but you can really start focusing on speed waiting for finality on blockchain a and sending a message that's going to be slow and we'll get into that okay so I'm here supposed to tell you what the answer is what the the future how we solve this the uh the future is you just stop sending messages blockchain a blockchain b just don't send a message um and then you guys should be like wait how's that possible how am I going to move value between these two blockchains without sending messages and um there's an observation here that's important so in tcip when I have two servers and I'm connecting them with a pipe and trying to move information between them that information is bespoke it's custom It's like got some SSL signatures whatever else but in blockchains what I'm mostly doing is moving value I might be moving value to then execute a transaction like buy an nft or something like that but mostly I'm moving value and value value is fungible so I don't care if I'm trying to move ,000 from A to B I don't care which $1,000 it is I just want $1,000 on chain B and because value is fungible you don't care about how the money got there you just want the money there and so if you start thinking about intents which I'll introduce in a second this is why they work is we're leaning into the concept of F ability to be able to uh not send a message at least not initially all right so we'll go back to my diagram so bridging with intents what we've labeled intents and what across has really been focused on blockchain a blockchain B I deposit a message on blockchain a a and I don't send a message instead I've broadcast my quote unquote intent I've broadcast my intent on blockchain a and a network of of third parties we call them relays we can call them solvers you can call them market makers this network of third parties is competing to fulfill my intent on blockchain B and if my intent here was I want to move $1,000 between these two chains we've got a network of third third party actors that are competing to send me $1,000 on blockchain B first person uh to fill me wins they get paid a small fee and as you can see there's no message that gets sent here uh I get the users fill me on or the the solvers fill me on blockchain B um without ever having there to be a message sent between these two blockchains they are outside observing blockchain a and seeing that this is a valid fill and in doing so we can make this be extremely fast we'll get into this as well as being cheap and um and secure um so that's the core model for why why this intent architecture is so different from traditional Bridges is that these third party actors use their own funds to fill users on the destination chain very very fast and um you know this isn't set up for a demo but if I was here you guys could go to a cross. two and try to bridge few bucks between your favorite l2s uh it'll take two seconds the median Bridge time for a cross right now is I think 3 seconds uh we are extremely extremely fast simply because of this architecture and it's also really cheap so okay so our definition here with cross chain intents thirdparty actors solvers or relays are going to fill use their own Capital to fill users quickly and cheaply that happens and then the protocol will later verify that that intent was fulfilled and repay the users all right so um when you start thinking about the structure of this technology um our view of cross un intense looks like this cake structure and we're stealing this from our friends at Frontier research that that like this cake analogy so you have the the intent layer where the user says what they want to do and my mental model is I think of this like an order ticket the user says I want to sell a th000 usdc on chain a to get a th000 usdc or like one eth on chain B that's their intent and they sign that message and that's what what creates the initial order and once that order is created um there is then a competition at the solver layer at Layer Two for who's going to go and fill that um and our job as a protocol is to design the market structure that's going to efficiently create competition around that solver layer to have these users correctly compete and the settlement layer so when the user user funds on the origin chain they don't go directly to the solver they go into escrow um on the settlement layer which is part of the across protocol system and that escro contract holds those assets and uh and only releases them to the solver after we verify that the intent was fulfilled all right so data um intense this architecture user experience we want to make it seem like users are clicking buttons and moving assets between chains um this architecture can be really really fast uh this intent architecture which we have in production doing you know $20 million of volume 20 to $50 million of volume a day um 50,000 messages it's one to two orders of magnitude faster than uh traditional bridging systems um and the thing that I'd want to point to let's see if the laser pointer works I don't know if you guys can see this but the L2 to L2 uh median fill time is 3 seconds for us and there's room for that to come even faster and you compare that to traditional ing architectures like Circle cctp it works really well for moving USC but it's slow um Stargate which uses a traditional pool-based architecture they are 52 seconds for the fast version 195 seconds for the slow version uh it's just an order of magnitude difference and in my opinion if we want the user experience of united blockchains we got to get down to this like 2C like experience okay so intense can be really really fast I'll go over this quickly but I also think intents can be very very secure and if we go back to the traditional architecture of sending a message there's a lot of security assumptions into wait who's sending the message between blockchains how are you doing that and how are you doing that quickly and sending a a message between blockchains very quickly is a very hard problem to do it secure what the intent architecture does that I I think is is a useful observation is we're separating the Urgent part which is the user experience users want a very fast fill we're separating the Urgent part from the complex part which is sending a message between blockchains is hard to do and in doing so it buys ourselves time to do the complex part the sending messaging um slowly and potentially more expensively we can use the most expensive slowest bridging technology possible and it's okay because the user still has a fast fill and in doing so this intent architecture supports verifying the intent with any type of proving system out there you could use the canonical bridges that l2s have use ZK proof systems use optimistic systems really the the architecture can choose the interoperability messaging technique that best fits uh uh the the need for verifying these intents and in doing so it's kind of an interesting sidebar it commoditized generic messaging generic messaging anding the intent protocol can pick whatever whatever architecture works best there intents can be really cheap and this one I'll I'll do one more slide to explain this this one's a little bit harder to wrap your head around um why are intents so cheap well it's because we batch verify the proofs that the intents happened so if we think about a traditional Bridge brid architecture I've got to send a message for each and every bridge transaction and there is a cost to that there's gas costs and there's security costs and you probably have to pay the messaging layer or something so every time I do a a cross chain interaction I'm paying a messaging cost with the intent architecture we can have these thirdparty solvers filling thousands or tens of thousands of intents and then batch verifying that these fills did happen with one message and that batching means we're able to compress our messaging costs and effectively commoditize and minimize our messaging use um so what's this look like at kind of a theoretical level say I want to move $50 million of volume over 50,000 transactions and let's assume that I can borrow money at a 10% interest rate annualized and let's assume that each message cost me 5 cents which is probably like you know border magnitude right my loan for borrowing $50 million for an hour if it takes me an hour to pay back these solvers is 500 bucks so every hour I'm going to create a batch and batch verify all these intents and then I send 10 messages to verify them this is roughly how the cross Works cost me just over 500 bucks in in in in costs here whereas in a traditional bridge if I'm sending 50,000 messages and I'm paying 5 cents a message that's $2,500 of messaging costs and you can kind of look at this that at um what the intent architecture by batching the verification can get really really cheap um you know I'm short on time so I'm going to skip through some slides the tldr here um I'm a I mean I am biased it's what we're building but I'm a firm believer that uh our our cross-chain interoperability defragmenting ethereum having that seamless user experience of Click buttons and actions happen on other chains um it's going to be producing billions of cross-chain messages as we scale up crypto and web 3 um and if you're going to have to send billions of messages between blockchains that's going to be expensive and it's also going to be slow so uh our pitch here is this intent architecture will compress our messaging usage and allow you to have that really fast experience which is going to just keep things cheap um okay so across and I'm going to run through some slides here um our vision intents we're going to help uh applications uh create intents to abstract chains help solvers unify intents settlement layer to uh to settle these cross-chain intents and the last thing I want to touch on is uh what we've announced with Unis Swap and what we've been working on with Unis swap around a a standard for cross chain intense so we want to define a common interface for what this order structure looks like so that we can have solvers easily support cross chain intents of all flavors and styles and um we've defined this uh in an ERC and a a common um a proposal for a common standard in the ethereum ecosystem um encourage you to go check out ERC 7683 org and we've also been working on this with optimism optimism is doing a lot to make super chain interoperability be a thing um but they also care about interoperability outside of the super chain and the optimism team is using this 7683 uh cross chain intense standard as the basis for how they're going to have seamless and super fast interoperability um between the super chain itself so with that got to flip through the last slides and just say thanks uh I'm Hal 2001 on Twitter uh I love talking about intense uh and uh appreciate you guys all taking the time