[trustless://interop.landscape] Steven Goldfeder - User-Centric Interoperability

Transcript

Now our next session is going to center on user centric interoperability offchain labs's work on interoperability for arbitum and ethereum ecosystems. Please welcome to the stage Steven Goldfeder. It's not me, but it's a nice picture. All right. Hey everyone. Uh, thanks for coming

today. Really, really excited to be here. And I'm going to talk today about user centric interoperability. And what I mean by that is what are the tools that we need to build and how do we actually go about building interoperability that works for users. There is an absolute ton of interoperability protocols. Every single L2 or us on the bottom here at arbitum zkync optimism has announced one or more interoperability protocols. There's an entire industry of interoperability frameworks and protocols and that's fantastic. But the problem is if you think about it for a second and you just like pop the stack for a second, if there are 100 different interoperability protocols, we didn't really solve

interoperability because you just have to have all these different pieces that themselves don't work together. If every two chains has a custom interoperability protocol, that is not interoperability. So the idea that the most trustless the ideal path and by ideal I mean the most trustless and cheapest path from chain A to chain B. The truth is it may not be the same path from chain B to chain C. And actually it may not be the same path from chain B back to chain A. Right? So if you go from Ethereum to arbitrum you might use one bridge. If you go from arbitum to Ethereum it might make sense to use a different bridge uh depending on what your speed or other uh price tolerance uh preferences are. So it's good that we have all of these different

interoperability protocols, but the problem is how do we actually make sense of it in a user centric way? How do we make these cottage industry of interoperability protocols where it seems like every pair of chains has their own way to go from one to the other actually work for users? So the first thing to note is users can't keep track of this many paths. It's impossible to expect users often who transact on tens or dozens of chains to be able to actually keep track of all the different ways they may go from one chain to another and know oh if I'm going from arbitrum to Ethereum I'll use this if I'm going from base to zkync I'll use that if I'm going from arbitrum to achain I'll use yet a different path it doesn't really make sense users can't

keep that state and even if they could users can't analyze the benefits the relative security benefits of different paths right we heard today a lot about EIL but your average user probably doesn't except for trusting the the folks who are saying these things doesn't isn't able to actually understand the difference between that and say an intentbased bridge or a native bridge. It's very uh opaque to many users. So how do we actually make sense of this? And what I propose is really it's a two-front approach to uh interop and it's really a wallet centric approach and the idea is rather than selling all of these uh interop protocols to users or even to intent protocols it's the wallets that need to inter inter integrate these protocols and the idea is if I go from one if I

want to go from arbitum and deposit something on on ethereum my wallet should figure out for me how do I go from step A to chain A to chain B how do I do the steps and I shouldn't have to do all these these manual flows and all the protocol work actually should be protocol researchers and developers presenting those to wallets because the wallets are the ones that are able to actually make these choices and make these decisions and the users then can basically make a single decision is which wallet is best for me and there might be one that prioritize cost one that prioritizes security but ultimately the users can make those decisions and it has to be a two-step program where the wallets really should completely uh ingest the interop protocols if you have interoperability imagine this circles interoperability. I think the the world

that most people think that this is is way we solve interoperability is like 90% protocol design and then 10% sort of everything else. But I think that's actually completely wrong. I think real interoperability today is 90% just wallet support. We actually have all the tools to do this and 10% everything else. And over time that 10% will get better and better and better. But ultimately we have enough today to actually have excellent interoperability. But the big big hurdle is getting wallets to actually support interoperability protocols. To take an example user journey today, imagine you're a user that has ARB on APC chain and you want to earn yield on USC on on arbitrum. Now note that APC chain gas is in Apcoin. So you have to

first get that you have to swap your USC uh and if there actually swap for USC if there's actually liquidity there bridge to arbitum deposit and at the end of the day actually get ETH on arbitum because it's a different gascoin at the end of the day you're also going to be left with some dust on ape which is not going to be consolidated in your wallet because you're not going to be able to like time to the or figure out the exact cost of bridging. It's a an awful and really really terrible terrible process. And this is actually pretty basic. Imagine that you wanted to deploy something on a on ethereum or arbitum. it would be even more steps and even more complex. What this should look like actually is just one step. The wallet has a button that says deposit USDC on a

swap your your your coins or whatever that one action you want to do and everything else basically the wallet has a path for you in the background. Then sure there can be an advanced tab where users can figure out that path and swap parts and say hey actually I want to pull this asset or I want to actually use this other bridge. But the idea is there's no reason today that this slide is exposed to users. Really the wallet should handle all of this and just say to users what is it that you want to do? And in the back we have uh the wallet can reach out to an intent protocol today to do that. And the good news is and this is why I say this is all in the wallet support is we know how to do this today. We should focus first on tokens

transfers and swaps because that's actually 90% of what people want today. And we actually have the tools to already do this today. We have intentbased protocols that wallets can directly incorporate today that would just solve interop and give users this one-step interoperability. This one-step UX that you just do what you want to do and the wallet figures out everything for you. So um across and relay other solvers we have excellent integr monitors the chain and receives a quote. The user deposits money into an escrow.

The user wants to go from chain A to chain B. So the user deposits money into an escro contract in chain A. The solver fronts the funds on chain B. The SO contract release those funds to the solver on ch on chain A once the solver actually did the task and then the solver takes care of actually bridging and rebalancing of its inventory. But the user just gets a really fast uh you know action, a really fast um really fast bridging or doing whatever action it wants to do. And we know how to do this today and it's just unfortunate today there's an extra step here where we have to go you know to a cross's website and actually figure out these different pieces. But there's really no reason why this is not directly integrated into every single wallet today. And the truth is um this is not

the endgame. This is not the best because there's a lot of trust involved here. And we heard about some good improvements today of how we can actually remove some of this trust. But the point is that can happen iteratively over time. That can happen in the background. Um we can replace the internals over time and in time actually we got get simpler and simpler. For example, as we have ZK protocols, we might not even need intense solvers anymore. We might be able to do this directly, but the idea is the wallet should basically take on this task today and use the tools that we have today to give you an excellent interop protocol and tomorrow that may actually may actually be more and more and more trustless because we'll take things like EIL and we'll swap that out for more trusted components and in the future

we'll actually be able to expand this with like not just token swaps and transfers but onchain messaging. But the core point is this these aren't things that should be sold directly to users. These should be users should actually this should be completely transparent to users. The user uses a wallet and in time over time as protocol researchers develop new things they sort of automatically get the benefits of what they're doing. But if you imagine your average user to be able to say hey uh I prefer EIL over uh this intent protocol and actually to understand that I think we're going to actually be in for a big disappointment at something that's not going to look very much like user centric interop. So core principle number four of mine is when we build these native interop systems or these cross-chain cross Ethereum interop

systems they should be built in the most generic way possible built in the smart contract layer. We don't want to see tight coupling between say a particular interop provider and a particular chain. It shouldn't be built into the protocol. The protocols might have custom rebalancing protocols. For example, you know arbitrum has something called fast withdrawals that allows you to move quickly between two arbitum antitrust chains. ZKSync has their protocol, Optack has their protocol, but generic interop should be able to ingest all of these and actually just be in a way that's actually not tied too much to a particular provider because the truth is in order to get cross Ethereum interop, we're going to need to weave together many, many, many pieces, many pieces in ways that users themselves won't be able to do, but wallets are in very, very good position to do. So, what is

offchain labs? You know, my my core premise until now actually is a lot of this is out of our hands and a lot of this actually is going to just require better and better wallets and better support for wallets. But obviously there is a lot that we're doing today. So what is offchain labs doing today? What are we doing today to actually bring forward this future of crosschain interop and how do we um how do we go about influencing that? So the first thing is um we funded something called the O of the open intense framework to actually build open solver uh software that any anyone can run on their chain to have open- source uh software that you can go ahead and opt into uh intent and solverbased systems. But the problem is um often you know uh large chains will be able to uh track the acrosses or the

relays of the world but smaller chains won't be able to and they'll often be left out of this interop and that's why we think it's so important to go ahead and actually build these pieces. So any chain doesn't have to rely on uh you know being big enough to get some solver to decide that it's you know monetarily or economically worthwhile for them to support your chain but as a chain operator you can go and do this yourself. So that's why we funded the O of for this particular uh component. So any chain can actually go and participate. And on the wallet level uh we got a little frustrated with the state of this. So we acquired a company called ZeroDev which does this publishes this API where any any wallet can go ahead and sign up to have uh to broadcast their intents to and the API

will actually go ahead broadcast those intents to solvers and you'll be able to participate basically uh in this uh today intentbased ecosystem. But the idea is tomorrow it might actually be a very different ecosystem. Uh but the wallet and the API are the ones that are going to hide some of those details u from the user. So why do we do it? Why do we believe uh why are we doing it this way? No complex protocol modifications. Again that's the previous principle. It's generic. It's smart contract based and you can plug in any chain and uh this will proliferate e ecosystem adoption of existing tools that we know work now. Again we have across we have relay today but how do we get the long tale of chains that those uh protocols aren't going to support?

And the answer is uh we create the OAF and open source software so any chain can themselves uh join this uh Ethereum wide interop ecosystem. But again it's only going to work if we have really good and embedded wallet support as part of this. Another thing that we did uh at at offchain at offchain labs as part of the is we created uh this uh broadcaster custom custom trustless broadcaster protocol uh based on ethereum storage proofs. The idea is if you have a solverbased system and the solver went ahead and did a task, it needs to go ahead and broadcast that message back to the uh escro chain, the chain where the funds are being escorted and say, "Hey, release those funds." Today, a lot of those protocols are trusted. Uh we built a completely trustless protocol and

contributed and actually the OIF uh has smart contracts to implement this. The truth is this is actually very slow today and no one's going to use it. But again, it shows the progress that we'll make over time. So today we'll have more trusted systems, but tomorrow we're already laying the groundwork to have more trustless systems where we'll be able to again transparently swap that away from users where users will just get a better and better and better interop system over time. Uh but they won't have to uh you know if you imagine that users will have to uh if you can take a look at the QR code, go ahead and read read the proposal here and understand it like again they won't be able to but that's fundamentally uh them going to be relying on their wallet service providers uh to go ahead and do

that for them. So we're also building something called the universal intent engine. The universal intent engine pairs very well with the O of the open intense framework. So over here on this side we have the O of what the O of does it standardizes around 7683 the messaging format for intents. Um and it defines how we uh lock those fonts in the escrow and it also defines how we go ahead and broadcast those messages back and you can combine different pieces. You can also use services like across and relay that go ahead and do that for you. But on the left side here um the open intents the universal intent engine uh is slightly more broad and actually provides that API where a wallet can go ahead uh this SDK where a wallet can go ahead and broadcast these intents um and actually have this API where they can

ingest these intents find a solver to go ahead and uh and participate and again with um the OIF anyone can join this themselves they can join this ecosystem whether or not the guys on the bottom decide to actually uh support that chain. So any wallet can opt in and any chain can opt in to what's going to be a broad Ethereum interop ecosystem. But of course uh one more time, this will only work well and the best UX will only come if it's transparent from users and it's actually just directly embedded uh inside the wallets. So we're not done yet. Again, we have a lot a lot that we have to do. Um everything we described so far really just works for token transfers and swaps. Uh we have something with zero

dev uh called the multi-chain validator that will actually be able to um do this in a way that works not only for swaps but for complex action across chains. This requires the use of smart accounts. So you can actually allow the solver to you know delegate authority to actually do specific actions in those chains like hey go ahead and don't just send my tokens but actually deposit these funds uh say in a so uh we are building that at zero dev and again today that's actually uh more trusted and requires a good deal of trust but we'll be able to uh make that more trustless over time. The EIL which was of course presented today leverages multi-chain validators just like we do in zerov and actually in the EIL uh zerov can actually be be both a bundler and a solver. So um the idea

is that we should just adopt smart account solutions that will enable us to go not only from token swapping and bridging but to generic smart contract support where you you have funds on one chain, you want to do some actions on another chain, you can go ahead and actually just make that happen literally in one button click and everything else uh will um sort of happen uh underneath the hood. So what is the future state? So, Ethereums and L2s as I said the wallets I believe firmly we should lean into the EIL and more trustless um uh interoperability protocols but again do that in a way uh that cross crosschain transactions feel like just single chain transactions right we'll have this UX

where we don't feel like we're a fragmented Ethereum but we feel like we're one ecosystem where everything just feels like one chain even though under the hood it's many different chains and that will come from uh direct wallet support and eventually direct app support and embedded wallets right eventually we may not even need these wallets. The wallet themselves may be embedded uh in most apps and users can rely on trustless protocols over time, not just more trusted APIs that we have today. So, we have a long way to go. Um but the idea is that over time we can improve on two different verticals. We can go ahead and we can improve the trustlessness of the protocols by saying okay today we have a little more trust but we have really good directions again uh EIL is a great example to make this more and more trustless over time. let's

swap out the internals and also the functionality of the protocol going from token swaps and transfers much more uh generically and broadly to uh just generic smart contract calling uh generic message passing from one chain to another. So I'll end with a few key takeaways that I think uh if you spaced out until now it's a good time to space back in. These are you know basically the key points that I have for you today. We have enough pieces to solve interop today. Interop will not be solved on a whiteboard. Interop will not be solved in a research paper. Interop's going to require us to take the pieces that we know how to use today and build those.

And by the way, that's how we built Arbitum. If you look at Arbitum in 2021 compared to it today, it was terrible. Terrible. It was slower. It was more expensive. The proofs were work were worse. But you have to meet the moment at the time. It was the best solutions for scaling at the time. And over time, we replaced it and we upgraded and we got to better and better and better. And we need to do the same with interop. We don't need to wait for new protocols. have all the pieces today. We must put them together and put them to work today. Key takeaway number two. We've built the components ready to allow any chain to participate. It's not going to be interop just between the most popular chains, but any chain even if they have very very little usage can opt into this

crosschain interop protocol and actually participate today in cross Ethereum interop with this onechain feeling. Number three is we'll replace the internals over time. As we get more trustless tech that becomes available, as we can get tech that is actually faster to do things like crosschain, generic message passing in a tr in a trustless way. We can make this better and better over time. But the time to start is not to wait until we have that future end state. The time to start is right now. Let's take the pieces that we have right now and you know just like that SpaceX image, we'll make it better and better and better over time, but don't wait until the end state. That's not going to serve users today. And that's when we need interrupt today. We don't want to lose the battle today for some future state for tomorrow.

And the final point, which was really my main point here today, is wallet support is the missing link. We need every single popular wallet to actually opt in and give users these flows that don't require them to keep state of all these complex complex steps that they want to do and figure out how do I go from A to B and then what do I do? And I have to go have to figure how to get gas. No, your wallet should give you a default path to do a basic user action. I have money. All you need to do is go to the app you want to do, press on the action you want to do, and your wallet should figure out where are your funds, how do I get them here to where you need to be, and you all you do is need to approve that. And of

course, you can go in and change that and not like that. But your wallet should do the heavy lifting for you of at least proposing a default path. Thank you so much. Thank you all for coming today. And uh really uh really great to be