Nam Chu Hoai I Open Standards' Critical Role in Protecting Ethereum Interoperability I Pragma Taipei

[Music] We are ready for our next talk. And uh speaking of L2s, I want to bring up Namchu to tell you about how we can actually reduce fragmentation and and do better interoperability from hyperlane. So please give him a big round of applause. Thank you, Cartik. Uh, hope everyone's having a good day.

It's nice sunny day out in Taipei. Um, I have 20 minutes, so let's get started. Um, the title of the talk today is open standards critical role in protecting Ethereum interperability. A lot of big words. I'll swear I'll try to make it justice.

Um, the agenda for today is in three parts. One is Ethereum fragmentation, right? not just with respect to interoperability but even historically speaking. Um two I want to use account abstraction as a case study of like how standards can help with uh solving this fragmentation. Uh and then three kind of talk a bit about how that like applies to like interop interop so to speak um that I think a lot of people have been talking about.

So unifying Ethereum very brief history. So I only joined the space in like 2018. So like I can't pre 2018 is mostly from like secondhand uh reports but I think for folks who were there probably remember that when Ethereum first launched there were actually no standards at all right it was just like soly uh smart contracts and people could just build all kinds of applications but quickly people realized oh like tokens are super interesting and so people built their own tok like versions of tokens um and that kind of created some kind of fragmentation because right like every token was slightly different and so like wallets had to like custom integrate each token, right? Like apps had to custom integrate each token and so ERC20 uh that came around really helped Ethereum become a lot more interoperable because uh you had these kind of like default integration points and that's really what ultimately allowed DeFi to um kind of get created on Ethereum. Um but quickly thereafter Ethereum ran into kind of like scaling issues, right?

Um and so first we're looking at sharding and like how can we uh yeah scale Ethereum for sharding but really in 2020 we got to the rollupcentric road map right like we just said like hey let's do execution off the L1 uh we call them roll-ups and kind of like scale Ethereum that way. Um that means that today Ethereum is quite fragmented, right? Like we have a lot of like now we no longer have this problem of like oh man like Ethereum is super expensive, right? Like I think I'm always in awe when I see like sub 10 gas prices on Ethereum. Um and so you have great rollups, but like as a user you have like this experience, right?

that you have now instead of a single chain of Ethereum you have like at least like 10 20 like rollups and other chains in your metam mask and it becomes quite difficult as a user like defy sum was great because you just like open metam mask interact with all these apps now you have to think about like how do you go to these chains how do you go off of these chains so that's what we call I'd say the ethereum fragmentation problem of our times um and there's been a lot of like kind of like angry tweets about this right especially last year when um you know all the Salana meme tokens came out everybody was like hey love Ethereum but like the experience is just like much worse compared to Salana um and this is going to be something that we'll be talking today about like why this is actually fine and what we can do about it so the goal that um we kind of stated uh and by we I mean like uh in defcon for example was like Vitalik on stage of all the L2 uh founders was like we want to make Ethereum feel like a single chain again right and don't want people to have to like go through all of this friction um to compose um and so um a similar goal that we've had earlier was we want to have Ethereum have consistent account experiences like before before we had standardized account obstruction um right like most users had like kind of like the different wallets and they were not interoperable as well right like whenever you had like they were interoperable in the sense that you could use your mimmonic in other places. But if you wanted to use right like smart contract wallets like at the time I think argan was very popular or safe right it became like you were effectively vendor locked to that account. Um and so when you wanted to build an application you had to kind of like to use the state-of-the-art like account abstraction you needed to effectively put your users in that world garden. Um and so that that's not great right like it forced a lot of applications to basically have to build things from scratch and so the litmus I usually use for like oh why standardization is whether somebody can build a business on that standard and you know like I think it's not a black and white thing right like you can build a business on let's say like no safes but we can also see from right like the bbit hack that um uh I would say like there's like limits to like just how much you can think of safe as a completely vendored independent standard versus just like an ecosystem just for Nosis. Um, and so that's kind of the last point I want to say about like you know like do the standard authors or like the core developers so to speak support cannibalization of their own product or service.

Um, and there were standards prior to 4337, right? I think there was um I forgot the number but 2982 or something. Um but it was really like after the uh AA team at um the EF came up with 4337 I think spearheaded by YOF that kind of created a shelling point for the community to come together and do the similar thing that ERC20 did for tokens but just for account abstraction right and so now we have like this very rich kind of like ecosystem of different players contributing to native account abstraction in Ethereum where like you as an app developer if you wanted like state-of-the-art art account abstraction. You no longer had to build everything from scratch and we have like these auxiliary standards who kind of like help with um uh help us with kind of like yeah like again bringing this uh continued innovation on with respect to account abstraction uh to the market. And so this is kind of like the like lens I want people to kind of like take away with is how stand how does standardization actually happen and like what is important for it to happen.

Um and in this case I think there was again like this like the important part here is the social shelling point that I think the account abstraction team at the EF was able to create by facilitating a lot of the like um interfaces between right like the bundlers and the pay masters and all that kind of stuff. Um and so the coming back to the goal from uh interoperability right like we want to make Ethereum feel like a single chain again like if you think of this timeline of like again 2017 2020 bespoke smart contract wallets then trying standardization 2983 and now again mostly solved account abstraction like what would it look like for interup um and so I'm going to go into a bit of like the history of that but I would say right now where we're at is in this middle point like we have some standards but they're not quite there. So we're in like I call pre4337 days of interop and I say this because like clearly interrupt and in like in the definition of like hey I am a smart contract on a destination chain and I want to know what happened on an origin chain like this is something that we've been working on for ages right like all roller bridges are basically that we have a lot of like we have I would say like three or four generations at this point of bridges um that are like um externally validated that call themselves standards as well. But again, similar to like the pre 4337 days, um they're mostly like uh I would say standards pushed by those authors, right? And like you can't really build at least currently today, there's very few businesses that are built their kind of like differentiation on top of these standards.

Um and I I think actually Cello who I think Merrick was on stage earlier today was a great example but generally speaking migration of verification or execution in respect to interrupt is actually like a substantial effort like right like sell was probably like the most successful recent experience there but you don't really ever see that like once people choose like a like interop stack they're basically stuck in it um fact effectively forever um and so there has been attempts prior to like standardize interoperability and I think probably the most prominent example is ERC782 also known as XGC20 um and uh they were originally authored by the connect team which is now goes by Everclear and I actually remember like when it came out like a year and a half ago and I was one of those people who had a perceived bias right like I was like oh man like looks great it's basically the same thing as everybody else but it like looks like connects thing Right. Um, and I think this is like a really important point to kind of like remind yourself on when you work with open standards is that uh there's part of it which is the technical merits, right? And like again in account obstruction there are like obviously technical merance to be had about 437 or 2982. But um even besides the kind of like technical matters there's like these social matters of like hey do I believe this is actually a neutral thing or not. Um and that's been I think the most difficult thing about interop standards is that pretty much whenever um yeah somebody basically like trying to push a neutral standard it just like looks like oh you're just trying to like pump your bags and look I'm on stage here like you can say the same thing about me but that's just like something that I want everybody to recognize.

And then I think the second point um speaking as somebody who is probably has the best XVC20 support out there is that there's a lot of lack of tooling like right like there's a lot of people who want to use open standards and want to be part of the solution but the problem comes when you kind of doing the right thing actually is like more expensive for you right like because there was no tooling you had to build it yourself versus if you went with what I call one of the vendor standards like there's a lot more tooling because you have this primary like commercial party motivated to give you a good experience. And so for those reasons, I would say XC20 has had limited distribution. And so the next effort I want to go to, this is the most recent one. Um is what um you might have heard of as the open intense framework. Um and this was basically like our effort to try to learn our lessons from what happened with yeah XSG20 and what we can do to kind of like give it greater distribution.

Um so first why um intents are great and I'm not going to go too much into like why they're great but um more and more bridging goes instead of like kind of traditional base bridging over to intense a lot of people wanted it but to actually get it today you needed to talk to basically existing intent protocols and convince them to deploy it on their chain right and so as hyperlane our customer base often is like new rollups new chains and so it wasn't great that like we didn't have an answer for them of like how they could get intense. And that was kind of like why the idea of the open tense framework came about. But yeah, lesson number one was basically like let's not kind of repeat the mistake of like not having tooling. Like the experience of using or having the open intense framework should be the same as if you convinced somebody to do all the work for you. And so out of the box, the OIF comes with kind of like an open source solver implementation.

The goal is to have some rebalancing logic kind of ready to go smart contracts and a UI template, right? So like the goal is basically like hey you can have an intent experience ready to go within a week similar on like similar time scale basically as if you were like you know a big enough chain and you can convince like hard from across to deploy uh across onto your chain. Um and so what what we found as we were building the framework was we wanted to find ways to get winwin situations where like let's say right like if we're working with another protocol um like in 10 protocol and that wanted to use like that needed to have a solver and I think eco is one of was one example of that like it's a win-win because they get a reference solver with less work and the framework is a win because they have more people kind of building on these shared abstractions and so I think not just thinking about this from a perspective of like hey the idea of the standard is good for you but actually like here is how the framework can help you succeed in your business and doing so actually improves the framework I think was a very important thing for us but arguably the most important thing uh was like we originally funded this work but we recognized for this to be successful we needed to like not be in the middle of it um and so we've been talking a lot with like uh DF and Josh from DF who was kind of responsible for um for interop now at DF and yeah I think you know like most people who probably heard of the OF have never heard of hyperlane or like its role there but I think it ultimately uh has received I think the attention it has received came only from the fact that like we didn't put like hyperlane front and center of the so now there's like yeah many great parties like open zeppelin or lei who are getting involved and kind of can push right like the standard for uh as the O of much further than we would be have been able to do if we just kind of kept it as our thing. So yeah, combination of basically tooling and social uh is what hopefully I mean it's still early days so I'm not trying to jinx it but like hopefully we'll get uh to a better spot with the um but this is just for intents and so there's a bunch of other interoperability kind of like product service area I would say that needs standardization as well right whether that's kind of like storage proofs whether that's like kind of interop clusters that are like very invoked now with like the super chain and elastic chains and act layer there's a lot of like similar work that people are doing across the stacks um that um is fundamentally just like hey how can we share trust assumptions to basically have better interop between our chains and I think what we should recognize is that um yeah again there's no single right answer and like we can't afford to all copy um not copy but like uh we can't afford to like all build the same thing. Um, Vitalic last week made this post on the magicians forum about pretty much that same point of right like we used to always argue about like oh optimistic rollups or ZK rollups and like why do work optimistic roll-ups will eat everything but in reality like that world is not going to exist for quite a while and multi-prover architectures um are pretty much like the way to go but even with like things like like aggregation here right we I don't want to end up in a world where like oh to have multiprover with op stack chains.

It's basically like the same concept but like totally different code as if you did multiprover on like orbit stacks or if you wanted to let a chain right even migrate from opt stack to orbit or if you want like if you had like a different TE like there's just like a lot of like um kind of duplication that we're doing right now because we don't have interop standards. Um so you know how I earlier said be wary of people who are trying to push their own standard. Um here's me. Um hyperlane is an open framework for interoperability because we like we started 3 years ago. We like knew this is going to happen right?

Um, and again, I think this is going to be a bit harder to say like, oh, how can we like not have hyperlane in the middle, but I think the best kind of like I guess answer to the like, hey, how do you like look at the incentives and I tell you what's going to happen, right? Is that like Helen is fully open source, fully permissionless, which means that like if you want interoperability and you don't want to talk to us at all, right? Like you can do this right now. like our docs tell you exactly how to deploy hyperlane to your chain and you never have to talk to us. Uh so my my personal mental model of this is usually Kubernetes, right?

Like Google created Kubernetes and obviously Kubernetes is kind of like a Google thing or like most people think of it as a Google thing, but right like they plausibly did enough work that made like even their competitors view like Kubernetes as neutral enough for them to build a business around it. And so that's kind of like our goal as well here is that like while we have a hosted service that gives you hyperlane on your chain, right? Like in a super fast time horizon. It's free. We run all the infrastructure so that you all you have to do is like click a button effectively and get like hyperlane as interoperability.

Like we also believe it's like super important to make you believe that like you know that like if our company goes away that you can just yeah operate this entirely yourself. And so what I usually call this is like friends don't let friends build their own bridges. Like it's super hard. We can't afford to continue to duplicate everything. We can't keep being in the like 3 4 37 days um with respect to interrupt.

Um, and yeah, there's here's some examples of folks who kind of like use this kind of interoperability framework to basically build their business and because they know that like hey like I don't have to use hyperlane right like eco is using hyperlane to secure the intent protocol but they're also running their own relayer right there's basically no dependency on our relayer same with like uh roll-ups as a service right like they their customers are people who deploy chains and they recognize that they can't have a business where their customers uh need to convince another business to be worth their while to have a like core product experience. And so this is kind of like all to say that um yeah like we're in the pre3 4337 days of interop and like once we kind of get out of those days we will have kind of the same renaissance of like rich ecosystems um with kind of these standards then um we had once 37 came about or ERC20 with um D5. Um I believe that's it. Um, I guess I have like minute and a half left. So, if there's any questions, I'll take them.

But if not, I hope you have a good day in Taipei. [Music]