Wallet Infra 101, & where innovation needs to happen
Devcon 7 SEA·Thu, Nov 14, 2024, 04:42 AM · 08:59
Speaker
In this talk I hope to go over the infrastructure stack of a standard wallet, and then also go into where in this stack I think innovation should happen/is already happening and why that's exciting. This will also broadly cover other related topics such as: - The future state of crypto UI & dapp interactions - What crypto wallets can learn from web2 apps - My framework around "What users can do with their balance?", and how wallets can use that to build new features
Transcript
Cool. Hey guys. So as you can see, I actually changed the title of my talk really last minute. But my name is Meida Kothari and I lead product for all of the wallet related things at Uniswap Labs. So in this talk, I want to cover some of the incredible infrastructure innovations that have happened over recent years, and why I think they're going to bring in the next wave of users in crypto.
Forgot my little thing. So when we talk about onboarding the next billion users, often the narrative online is focused on blaming crypto UX for why we haven't had mass adoption yet. But what does better Crypto UX really mean? Like if you look at a lot of our apps, they're already beautiful, they're clear, they're comparable to Web2. And based on recent breakout successes for dApps like Polymarket or TradingBots or whatever you use, I feel like it's safe to say that users are willing to overcome a lot of friction to do what they want.
In my opinion, what crypto often lacks is more opinionated products. And what a lot of these infrastructure innovations that I'm about to highlight do is give developers optionality to build more opinionated products. Historically, I think we've put a lot of burden of choice on users for every micro decision, whether it's how they store their seed phrase, whether it's how they pay gas, whether it's what chain they're on, everything else. So I find a more useful framing for the crypto UX problem is how do we make it easier for users to do more with their balance? Before I get into these innovations, I kind of wanted to set the framing and identify core problems to a really quick case study.
So for this, I'm going to use my dad as an example. He's a software engineer. He's technologically savvy. He trades a lot on apps like Robinhood and E-Trade for stocks. And he knows what ETH and Bitcoin are and holds some but doesn't really know much about it.
He doesn't really know about ERC-20s, L2s, the underlying technology, anything else. So he somehow finds out about a token. This actually happened a few months ago. Let's call it Blah Token. And he asked me for help.
And so I helped walk him through how to buy it. So I walk him through the process. He only has crypto in his Robinhood centralized exchange account right now. And so I had to guide him through downloading a non-custodial wallet, explain what that non-custodial word meant and then what a seed phrase was and why it was so important. And then I had to walk him through transferring his ETH from the centralized exchange into this new wallet.
But oh, it turns out the asset was on base and he didn't have ETH on base to pay the gas and so I had to explain what bridges were, what L2s were, what gas was, and then we had to bridge his ETH over to base. And then he was finally able to swap the asset using Uniswap. But of course, he had to reserve some of that ETH for gas. That took so many steps, so much education, so much time. And one of the biggest issues I think all of us see in the space right now is onboarding.
And my dad specifically in this case was a very high intent technical user. And it took him several minutes to create a wallet and was thrown so many new concepts right up front. Currently, I feel like we're taking high intent users who probably come in wanting to do something, wanting to try out a wallet or buy an asset, whatever. And we end up surfacing so many friction points for them that they end up churning before even completing the action that they initially set out to do. Luckily though, there are so many infrastructure advancements that have been made already that make this whole process easier.
There are quite a few, but for the sake of time, I want to focus on four. So the first one is account abstraction, specifically through EIP-7702. So account abstraction and smart contract wallets have been in the zeitgeist for a while now, but one of the biggest issues with most of these implementations is that they're not compatible with EOAs, which are the most popular form of existing wallets. 7702 basically allows EOAs to get some, not all, of the benefits of account abstraction without forcing users to fully migrate into a smart contract wallet. So while this isn't fully comprehensive account abstraction, it does two things that I think are really exciting.
One is gas abstraction. So in my dad's case, he wouldn't have to worry about the concept of gas or gas tokens, etc. And another thing it does well is batch transactions. Second is this growth of multi-chain standards. So it's very clear we're going through a multi-chain world.
Block spaces are getting more and more specialized. And I feel like we don't give enough credit for how much Ethereum has evolved here. Like I remember just even a few years ago, going from one chain to another was so, so cumbersome. But now with intense space standards like ERC7683, which a couple other speakers before me talked about, and with block-space standards like the super chain that Optimism came up with, and just generally more options for different types of bridging, we have way more of a unified experience for cross-chain interactions that are happening on-chain. So now it's up to dApps to figure out how they want to represent these multi-chain interactions, like how much they want to abstract and how much they don't.
But it is so, so much easier. The third thing I want to mention is embedded wallets. So as I mentioned before, onboarding is a huge, huge issue in the space right now. And all the innovation happening around embedded wallets allows users to create wallets in seconds. And also users can use familiar concepts like email or social auth or even pass keys, which are becoming increasingly popular outside of crypto.
And sophisticated traders are already taking advantage of this tech when they're using certain telegram bots and things like that, or dApps that are using things like Privy or Capsule under the hood. The last one that I want to talk about is AI. So can't forget that as much as our industry is innovating, so are others. Everyone I think is aware of the massive impact AI is having on consumer apps right now. And I'm pumped about the confluence of that and crypto.
And I think there's a ton of stuff here that I could talk about with AI, obviously, but the ones that I want to focus on are context and discovery. So with AI, we can rapidly give users more context and information on the actions that they're trying to do on chain, as well as the assets that they're trying to buy. A big issue, another big issue I see in crypto right now is discovery. And I think fine-tuned AI can really help with this. So clearly right now is the best time to build consumer apps in crypto because of all this infrastructure innovation that's been happening.
So I want to end with a call to action to builders, especially new builders. All of these new consumer app devs, I hope you guys build more opinionated apps utilizing this tool chest of infrastructure to give users the most smooth experience possible. Thank you. Thank you, Medha. And we have time for a question.
Let's see. Okay. And there you go. You got it. Thanks for the talk.
It's very inspiring. I would like to ask maybe a Medha question. If there's a top three app, because we're on a consumer app, or you want to build, what would be the top three you can recommend for us to build on that? Did you say top three dApps I would recommend building? Did I get that right?
Oh, yeah, yes. It's more like you want to get your thoughts on that. For sure. That's a great question. Let me think about this.
There are so many exciting things right now. I would say one is just like a better information aggregator, so something that's like more personalized based on your trading history and also like assets that you own in a fun way I think that would be really exciting another one is like a multi-chain dApp that's like super completely abstracted so I saw some demos for this that some optimism devs did and I think stuff like that would be really cool where the end user doesn't have to care about the chain that they're on, but they're able to interact with people who are using multiple chains. Yeah. Thanks for the question. Lovely.
We have time for another question. Yes. Just speak up. In your opinion, what's the biggest drawback to 7702? That's a great question.
I think there's two. One is smart contract risk. So you're introducing just more smart contract risk since you're delegating to a contract. And then the second one is just lock-in. So if we don't standardize the 7702 implementations that wallets do, it could lead to a case where, like, let's say I have a seed phrase on, like, Uniswap wallet, and then it implements 7702, and I get all these, like, cool features like gas abstraction, but then I export that private key and put it into, like, let's say Metamask or Coinbase wallet or whatever, and then suddenly those guarantees are gone if they're not using the same 7702 standard.
So that's something I do worry about. Thank you. Thank you for your questions. That's all the time we have for them. Thanks again, Mera.
Give it up for Mera.
Automatic transcript — names and jargon may be misspelled.