ETHIstanbul Conference 2025 | The Next Generation of Ethereum User Experience | Kassandra

Transcript

[Music] [Applause] Okay. Hello everybody. Um I am these are the things you can call me. Cassandra kzg cass k anything that starts with a k and I'll respond to that. My favorite letter. Um this talk is about Ethereum wallets and the next generation of them and uh it's going to be quite simple. It's about going back to basics in my opinion.

So Ethereum wallets 2.0 I believe we are at a critical juncture now in the Ethereum space where uh I believe Vitalic at ETHCC this year said something really interesting. It's time to focus on concrete properties, not ideas of things we might do, not Uber on the blockchain in 2017, but real concrete properties that we can ship to real users. And I believe one of the most important things to update to make this possible is uh being bold about uh new wallet user experiences. uh not getting stuck in the same wallet uh

paradigms that we've had for the since literally the inception of the Ethereum mainet chain. So what are the main goals for uh for updating wallet user experience? In my opinion, they're super simple and there are two really big really important goals. The first one is a new and improved user experience. And to me, we're going to talk specifically about uh what I mean by this is about a different paradigm for how the user should think about their accounts and their holdings and the way that they interact with the chain through the wallet client.

And the other imperative goal in my opinion uh at this at this juncture is privacy by default in the wallet. So right now everybody who transacts on Ethereum almost more than 99% of users have zero privacy when they transact on Ethereum mainet or other EVM chains. And uh I believe privacy is an achievable goal that could really change the shape of uh the utility concrete you useful properties for users uh in interacting with the Ethereum blockchain. So those are the goals and as I said it's about moving from

theory from ideas from uh from uh like proofs of concept to practice. Now especially with the most recent hard fork that that uh that came online, we have all the technology, all the tools we need to actually ship these new user experiences. It's not something we can it has to be in abstract in theory. Wouldn't it be cool if we can actually build and deliver these things to users and give them these new properties. So let's move from theory to practice. I think this is especially true in privacy where we've had just an incredible amount of work in done in the theory realm. I think beyond almost every

researcher's wildest dreams, the uh way that ZK technology on the theory aspect has accelerated, um I think we're lacking tremendously on putting these things into practice in a way that real users can uh actually benefit from them. And so this is uh the big thing to focus on. This talk will have nothing to do with new fancy ZK schemes or uh you know more compressed proofs or faster proving times or anything like that. Those are amazing things that we should keep doing but we're already above and beyond. We've done amazing there. Uh where we're just like far behind is on the on the practice part.

So if you remember anything from this talk my slides are extremely simple by the way. So uh the the hope is maybe they'll inspire you uh be a bit provocative somehow. I think if you remember anything remember this slide. We must rethink Ethereum wallet UX from the ground up. I think we are stuck in old models that are hindering us tremendously. And uh I uh we're going to talk about the most important thing uh there in just a second. But yeah, I think it's time to be bold. It's time to believe that we can actually give a different user experience. Up until now, almost every major wallet, for instance,

let's say browser wallets have copied the model of the existing wallets with distribution in order to maybe get distribution themselves or to be palatable to users or to be a user experience that people already know and love. But this is holding us back from uh being able to like envision much more upgraded experiences and ways of interacting with the chain. Um so this is uh my uh biggest uh uh demon that I would like to slay and that I think is holding us back which is we have been relying on address by address um a way of think

oh excuse me hello our wallets and um our accounts and So we think about a user as a specific address or an identity as specific address or a wallet and uh sorry sound here. Can you guys hear me anyway? I think you can. Um and I think this is like the that is holding us probably my beard by the way. Um the >> which is uh yeah switch.

>> Okay sure. that seems to be having a few issues. No worries whatsoever. >> Hey, how's this? Hey everybody. Um, so uh yeah, I think this in the Ethereum uh wallet land that is a big problem. And if you think about it, even Bitcoin wallets don't have this issue. So Bitcoin much older than Ethereum. Uh a Bitcoin wallet is an abstraction over address space. Every time you uh receive or or uh create a new UTXO, that's usually in a different Bitcoin address.

And your wallet is an aggregated balance across these addresses. And when you do an operation of sending some Bitcoin, well, actually, you might be signing with a bunch of different uh ECDSA keys at the same time. And all of this happens under the hood. and a wallet is an abstraction over this kind of raw address space and ECDSA signatures. However, in Ethereum and there are reasons for this um there's this accountsbased model, there's the idea of gas, there's smart contracts and things that made uh the easiest and simplest first experience to ship uh more compatible with like address reuse and thinking about an address as your kind

of account or identity. But I think to get beyond this model, uh the uh the analogy that I like to give is it's kind of like if we had the Ethereum we have today, but we didn't have the solidity language. So you could you could write smart contracts in pure op codes if you'd like and it's totally feasible and solidity and other uh higher level languages are just an abstraction over this uh raw EVM bite code. But I think it would be very very hard to progress in Ethereum applications if uh we never shipped anything beyond just raw Ethereum bite code. I don't think we would have ever seen innovation like uh smart contracts like unis swap and and uh lending

markets and all other types of things we've seen. So I believe we need a new approach in Ethereum and now we finally have the tools to do this because you needed an EOA funded with gas in order to initiate an Ethereum transaction in the past. There was no way to avoid the I like it would be extremely cumbersome to fund a new EOA with gas every single time you wanted to do any interaction. Um but uh things on the chain uh have been changing and so uh yeah I believe there are a tremendous amount of tools which have already been

shipped to production that really changed the game about the way that uh we can uh build wallets and create a account user experience for users. Um what we have are chain specific addresses and crosschain intents. We have these ERC's that are currently being worked on and uh uh tremendous work going on there. I think this is another like another issue around the cardinal sin of reusing addresses and thinking of addresses as our accounts is that we also have this ambiguity around EOAs are often used across different chains and so you think of addresses on different chains as being

somewhat equivalent but smart contracts do not have this property. So a smart wallet might not be the same on multiple chains or different smart contracts might not be the same. And so this ambiguity is literally a security issue in my opinion and quite unsafe, but it's rampantly used and understandably so because if you're used to your, you know, a certain key in your MetaMask wallet um well why not just, you know, use that key for everything, use that on different chains. But I believe again if we had these higher level abstractions in the wallet client and the way you thought about your account uh besides using this super lowlevel address space and key bykey way of thinking about accounts um then this issue never would

have even arised. Um I think there's a lot of great work going on in terms of trying to unify the cross EVM chain experience user experience in the wallet as well. um great initiatives going on there with these ERC's. We also have um we have these privacy protocols that have gotten way more mature on Ethereum. Things like privacy pools, rail gun, um uh you even have tornado cache as another, you know, important uh like uh precursor to these. Um, these ones have built in maybe more more tools for uh for like a selective revelation of uh of like which part of the anonymity set

you're from. Uh so these are really powerful primitives which once upon a time ZK was moonmath but now these are things that have long time been deployed on the chain and could be utilized way more and baked right into your user experience into your wallet and we'll talk more about that later. We also very importantly have a new EIP that was just shipped in the recent hardfork that gives EOA superpowers. And this is a key unlock for this issue of address reuse I was talking about because it's no longer true anymore that you need an EOA funded with ETH in order to operate on that EOA and to initiate a a transaction from that address space. You can now uh

uh have your everybody. >> Yeah. >> Yeah. Yeah. Everybody I'm I'm a yeller. >> Ah for the Yeah. For the video. What's your definition of privacy? What do you think is missing? >> I think privacy is your ability to have agency over selectively revealing things about yourselves to yourself to other

parties. So, uh it it's not secrecy. It's it's about autonomy. It's about you having the choice to reveal. Um and I think the difference wh why uh blockchains it's not that everything should be secretive and there should be no ability for any kind of compliance let's say it's that uh we need to get back to a place for instance in financial privacy where um where you actually have to be asked to reveal not reveal by def revealed by default and uh today the way everyone uses the chain they are constantly being uh like uh revealed doxed by default. Um and this

is worsens every time you use the same address over and over. Like how many people in this room use like a few addresses for almost all the uh uh interactions they do on chain? I do. I know I'm extremely lazy and I do this all the time. And this is uh every single transaction that you do from the same accounts creates links. Um, and like all of these links together make it incredibly trivial for the entire chain to be deanonymized by, you know, sophisticated actors like chain analysis. Um, I'm I believe we can take steps, simple steps to create a mist of privacy that uh gives them a much harder time. And that's what they should have.

That's what they they're uh the it's not that these people's jobs to go find bad actors that they shouldn't be able to do them. it's that they, you know, they should have a job to do. Um, like, uh, it's doesn't necessarily have to be like everything revealed to them ahead of time already. So, I would like to make their jobs a lot harder if I c if I could. Um, we also have uh a really cool relayer infrastructure, account abstraction, and gas abstraction with uh 4337. We've got a stealth addresses protocol, which is super useful and something I'm very passionate about. um very basic could be built on top of any other blockchain.

Bitcoin it works on top of as well or it's a it's a primitive that's exogenous to the way that the chain works. Um we also need to update the way apps and wallets interact so that wallets can do these sophisticated operations seamlessly while interacting with app frontends and we have EIPs for this as well like 5792. Um, finally, privacy super complex. It's uh maybe you can do a lot of fancy things to be private, but then you leak your IP when you uh connect to a uh an application or you know talk to over HTTP to um someone else. So, uh we need anonymous HTTP coms. Things like tour and VPN and

Hopper and and Wayu messaging are all really cool initiatives. Um things that could be coming soon is a full account abstraction with EIP7701, less heard of often than 7702, privacy preserving RPCs either with a TE solution or maybe in the future with a private information retrieval and cool fancy cryptography, let's say, and uh eventually having really light local nodes. So light nodes that make it possible that people don't have to interact with third party RPCs at all.

So all of these tools are they're already available or available soon. Maybe even a last one I would mention is there's an EIP uh for the long term uh which uh could eventually make ETH's transfers uh be private at the actual L1 level, but this would require hard fork changes as well. So uh tons of tools either already available or coming soon to be able to do some of these things. Okay, so now my slides get incredibly simple. Uh let's see how much time I have to see how long. Okay. Got a couple minutes. Yeah. Yeah. Yeah. Yeah. No worries. I um uh for me this talk really is an offering to to uh see the tools that are available and see uh the things

that we can do the types of things that we can go home and build today to actually ship things that people could try to use that give a different experience. And uh I think there are some incredible initiatives going on around interop. I don't know if anybody saw the post update from uh Barnaby at the EF, but he mentions the uh open intents framework and the Ethereum interop layer as things that are being worked on all the way from the top at the Ethereum Foundation and a bunch of other uh like uh stakeholders and groups and bridges and uh and relayers who are all coordinating around uh creating better interop standards. So that the goal here, what's the goal, right? Of

course, there are lots of little steps to take in uh in between to actually get there, but the goal is simply to make it feel like using a bunch all the different EVM chains we have out there or a subset of them that maybe have the properties that the a user deems um you know, secure enough uh has this feeling of using a single chain, seamless interaction across chains. Um, so I believe there's a lot going on there which could be really cool. I I would love to see wallets for instance, a browser wallet much like the ones you might use today, Arabi, MetaMask, etc., etc. But, uh, when you interact with a DAP, if you have funds on a different

chain, you don't need to use some third-party popin application or whatever. Your wallet itself understands how to relay those assets from the other chain to the source chain that you need in order to do the act interaction that you like. Simple but extremely valuable things that could completely change the way users think about using these chains. Again, I think we're using them at way too low level, way too difficult uh user experience that people have to know about manually switching networks, manually switching across addresses. Is it funded with the gas token from this chain? what is the gas token on this chain? These are like uh these make these create a barrier for us to

actually reach what we might be calling real world adoption or let's just say again user experiences that are actually useful to people. Um so this is very important and uh the last thing I want to tell you about which is what I am working on. So I am a grantee by the uh Ethereum Foundation uh working with a number of different uh groups, people um on this Kohaku privacy wallet project. And our goal here is to ship a browser wallet similar in uh feeling to the types of wallets you might use today, but which gives strong privacy by default. And uh the way that

we're thinking about approaching this is um to seamlessly and embeddedly uh to basically embed the uh different privacy protocols that already exist on Ethereum mainet into the wallet in a seamless way. And so you can think about uh instantiating private accounts like an account on uh with uh with rail gun or a uh privacy pools seed which allows you to make a number of notes um uh for the privacy pools protocol. And think of these accounts as part of your wallet as

well. And every time you interact with a DAP, you would actually instantiate a new fresh the the default experience would be use a fresh account every single time you interact with a DAP. And maybe if you if you've already interacted with that DAP before and it's something that has to be longived for uh in terms of address reuse in Ethereum, I don't think we can ever get to the place where addresses are never reused more than once. For instance, let's say you make a uh a CDP with a maker DAO. Uh this might be a position you might want to adjust in the future. So you may still have to have that key uh live longer. Um however, this is something

you can do, but by default, every time you sign into a new application and connect your wallet, you're going to connect a new fresh public address. And like in the longest term with this project, what I would love to see, by the way, I'm this is all personal opinions of mine, uh uh my like desired UX is that uh if you have funds already sitting inside of one of these privacy protocols, you could atomically interact with uh some uh decentralized application in such a way that the uh the UI, the DAP front end doesn't have to change at all. But when it sends payloads to your wallet, your wallet prepends that with a bunch of other operations that basically pull the funds

that you need from the privacy protocol into a fresh public account, upgrade that fresh public account to uh with 7702 so that it acts like a smart account and then passes the interactions that you want to do, approve and interact with the DAP, let's say, uh through that public account and pays the gas to some relayer who's doing that with funds sitting in a privacy protocol. And if you do all of this, you could seamlessly be interacting with a DAP uh entirely privately but have the exact same experience you have today.

And why am I talking about all this and then I'm finally done. Uh I'm I think what's really important is many of these things could be done today. However, you have to be extremely technologically savvy and go to like six different frontends in order to actually achieve something like this, which means zero people do it except maybe, you know, a few niche researchers who are really excited about these ideas. And so the only way that we are going to get privacy uh like distributed to many users and thus actually have real privacy because if the anonymity sets are not big there's not really any privacy for the five users who actually have it. Um we need to embed these

things into this authentication layer into the wallet layer and that's what we're trying to do with Kohaku. So uh hopefully we can show you something exciting soon and I hope you enjoyed the talk. Thank you guys.