We Finally Have the Elements of an Ethereum Identity Stack

Transcript

Hey everybody. Uh, thanks for being here. We're going to be talking about we finally have the elements of an Ethereum identity stack. I am Brantley. Uh, this is my onchain profile. Uh, you know, I'm building Ethereum follow protocol and Ethereum identity kit which I'll be talking about a little bit today. Also, I've been very involved in ENS for many years. Uh I'm the former core team lead. I'm on the security council. Uh I w I am a DAO delegate. I'm the creator of Stein with Ethereum. Uh so this is stuff I've been thinking about a long time.

So I want to go back to the very beginning of uh kind of where digital identity came from and it actually goes back to the 1970s. So public key cryptography was invented or some might say discovered in the 1970s and it actually was uh uh invented by two different groups of people independently. Some people secretly in the US government, some people uh in academia at MIT and you have to understand this um this is one of the most important inventions or or discoveries in my opin opinion of like the 20th century. I mean certainly in computer science extremely powerful.

Okay. Okay, so prior to this, if you wanted to encrypt a message and send it to the other person, right, you encrypt it with a private key, you had to send that person the private key so they could decrypt it. That was the only way for thousands of years, for all of human history up until the 1970s where we figured out a way that you could uh encrypt something with a private key um using that person's public key and then they were able to decrypt it with their private key and you didn't have to share any private information. extremely powerful. This was so powerful that the US government said this is like classified as like a defense weapon and it's illegal to share with people outside of the United States. And the US government was not wrong in uh

determining that this was extremely powerful. Uh but this sort of um shut down its use in technology though for a few decades until the 1990s. This guy named Phil Zimmerman. Who here has heard of Phil Zimmerman? Okay, great. He's a hero. We would not be here without Phil Zimmerman. So, thank you Phil Zimmerman. So, Phil Zimmerman, yes, he uh created this thing called PGP uh pre- good privacy, an implementation of uh public key cryptography that was public and he fought the US government to make this legal so we could use it for uh you know for uh general tech and he won in the 90s. So, by the way, most of the internet protocols were invented in like the 70s and 80s even though they became uh open to the public in the 90s. none

of them had public key cryptography built into them because it was illegal to do so. So since the 1990s, people have been attempting to put it back in, but it's kind of tricky to do because it wasn't built into it. In the 90s, uh there was a group called the Cipher Punks, and they realized that public key cryptography, among many things, could be used as the root of a digital identity. And they promoted PGP, and they had like the MIT PGP key store and all these things like this, and they um really pushed it, but they just never really got that much adoption. How many people here have have ever used PGP? How many people have stored it like on the MIT key store? Okay, so no one's gonna be able to find it anyway. How often who used PGP this week? Okay. Wow.

Amazing. Okay. Um it never really got huge amounts of adoption. Okay. Until 2008, Bitcoin is created and it uses public key cryptography not to like encrypt like hidden information but as like uh your wallet. And of course, Ethereum did the same thing in 2015. All all blockchains do this. Okay, this is why this so important. Crypto has accomplished with economic incentives what cipher punks couldn't with appeals to privacy and freedom. And namely this, the widespread adoption of private keys and the flourishing of a private key management industry, which is what we call wallets. Okay, really guys, amazing. uh prior to crypto public uh private key management software

wasn't very good and there wasn't any in uh economic incentive to make it or to use it or anything like this and although I believe in privacy and freedom it doesn't motivate most people uh many people in this room have experienced that but money does and if you can make money you know potentially make money by having a private key uh then a lot of people get private keys this is really important because once you have a private key for crypto you can use it for all sorts of other things all these other things that cipher punks wanted people to use it for. Okay, another key thing in the history here, 2011, Namecoin was the first serious altcoin and the key insight there was that you could use blockchains for things other than money and in this case internet names. So for those of you who are into

ENS, the very first non-monetary use of blockchains was naming. It's interesting goes back a while here, but it didn't really get adoption. It was a good try. Then in 2014 in the Ethereum white paper, Vitalik listed five areas that Ethereum could be used for and one of them was identity. And he mentions namecoin as an attempt at internet naming uh with blockchains and he says maybe we could uh try this again with Ethereum. When Ethereum launched right away, you know, inspired in part by his white paper, there were many attempts at uh Ethereumbased identity and they pretty much all failed. and you haven't heard of them uh because they failed they were

good tries and I think the mistake that many of them were making is that they thought of identity or social as a single app. I think the problem with this is that identity is not a single hard problem. It's like 20 hard problems. So you might solve a few of them but you're unlikely to solve all of them. You'll probably have to cut corners if you try. Um here's a key thing. Building things in a decentralized way is way harder than building things in a centralized web 2-way. This is a big mistake that many developers make. They're used to building like iPhone apps or or websites and things like this and they said, "Hey, let me try to make a web 3 app.

It's it's going to be the same thing. It's not. It's way harder to do." And in my experience, I think there are many web 3 devs who wish they were building web two because they're like, "Why can't I just centralize everything and it just works?" Like, and they say, "Users don't care about decentralization." Okay, great. We have an entire part of the tech industry where that doesn't matter. It's called web 2. You can go do it over there. That's great. I'm not anti-web 2, but over in web 3, we do care about decentralization. So then we had the Ethereum name service and it launched in 2017. Um, it originally was created for decentralized websites and as a result it was modeled after DNS. So you have like DNS and you had ENS, right?

And then uh early on also they said hey we could also use this for uh uh to name Ethereum addresses for crypto payments but it was not thought of as an identity primitive at all. Those words were never said for years was um was used for these other things. However then in 2021 four years after launch I realized that ENS was accidentally actually your crypto username and profile. people were already using it as like their crypto username and profile. We just hadn't uh crystallized that in our mind as like what we were really doing. And really what I realized was that the username is the meta use case for all these things, right? So it's like you want to send me

money, you send it to me. You want to find out more information about me, you go to my page, you go to my name on the internet. All these things ultimately are are sub use cases of the bigger use case which is that the you it's a username for me. So then also in 2021 I realized uh what I said before that uh with crypto wallets we'd given everybody a private key and there's all these other things we could use that for and um including for authentication. So people were already using uh uh crypto wallets um mostly Ethereum wallets for authentication but everybody was was making their own message and spinning up

their own uh message design. It wasn't very secure. It wasn't standardized and so we uh standardized it with EIP 4361. So if you have a crypto app and uh you need people to sign in. Now this is not just connecting your wallet. connecting your wallet doesn't require any authentication, but if the user has to access private information like on your server, they have to like sign in to your your servers. Um, you should use the sign with Ethereum standard. Why? Number one, we've thought of everything and it's secure and if you're rolling your own message, it's probably not secure. Don't do it. Okay? Friends, don't let friends roll their own uh message. Number two, it has support in wallets. So, for example, in MetaMask, if you ask a user to just sign a random

message, it just says, "Do you want to sign this message?" If it if it's a sign with Ethereum um formatted message, it recognizes it as that and says a sign-in request. And so, that helps the user know I'm not just signing a random message and I think it's signing me in or is it doing something else? They know it's doing authentication because should users be asked on regular basis to sign random messages? No, they should not. So, don't do that. that you sign with Ethereum. So around this time this sort of cris crystallized my mind that Ethereum identity right is not a single app.

Instead it is the sum of the information relationships held in various decentralized protocols and open standards which I call the Ethereum identity stack. What this means is that the identity stack itself can be decentralized rather than vertically integrated. uh and I think this is a good thing. So uh I think the path is that we unbundle the elements of identity to their smallest viable parts and have separate teams solve them individually with decentralized protocols and open standards. The key thing here is as a result of that each part can stand or fall on its own merits rather than being an all or nothing proposition. So for example, if you don't like ENS, you can

launch an ENS competitor, but you don't have to at the same time launch a sign with Ethereum competitor or vice versa. And so actually I think it makes the identity stack much more robust and uh much more open for improvement uh and competi competition on its individual parts. As a result of this, that has led me to create Ethereum follow protocol. So I was first talking about this in 2021. I tweeted a write up in 2023 and we launched finally in September 2024. So funny thing about that during that year I had a small team. We were trying to build this and get this out and my wife was uh expecting our eighth child and I told the team I said hey we need to launch before my wife has this baby and

we did just barely. We launched at the end of se uh end of September and we had our child um in early October. Um so uh EFP uh it is an onchain social graph protocol for Ethereum accounts. Notice there so it's onchain it's fully onchain. There's not any of this offchain like Okay it's a social graph. So what it means is that you can uh Ethereum accounts can follow other Ethereum accounts. They can follow you. You have this kind of web of relations. Um and it's for Ethereum accounts. So you don't have to have an EFP account. you just use your Ethereum account. This also means that even if you've never used EFP, I can still follow you if you have an Ethereum

account. It's designed of course to complement ENS and signed with Ethereum. It's been funded mostly by the ENS DAO because it's a complimentary protocol. Um some information about it. So we have 58 integrations so far. Um we've been growing that a lot. We're in things like Superfluid, Snapshot, the POAP app, Yodel, uh many others. Yep. Yodel is great. We have uh you know about 27,000 uh unique users. It's growing all the time. About 77,000 onchain actions. It's like follow unfollow tag untag. We let people uh tag people as well. And here are just some examples of how to use it. So let's you know you have someone's uh ENS profile.

All this information except what's in the red box is from this person's ENS profile. You have their avatar. The background is their header image. Uh they have like a bio. They have their socials. They have their name. All these things. Okay. Okay, so what e uh EFP does is it adds on all this other information. So it shows my relationship to them. So it says following. If I was not following them, that would just be a normal follow button. It shows how many people uh they are following, how many followers they have. That gives me context. And this is really important on the right. It shows people who I follow who also follow them. This is one of the most important use cases for this um because maybe I don't know them, but if people I trust say they know them, then it's like they're in my circle of trust,

right? So there it's this old school concept of web of trust and uh this is I think actually one of the most valuable use cases uh for EFP and why it can be put in all crypto apps. It can be in wallets, it can be in DeFi, all these things because if I'm interacting uh with any address or sending crypto to any address, I want to know is this really who I intended to uh intend to be sending this to? And if I have no connection to them on EFP, that's you know kind of like a warning of like are you really sure? Whereas, if I already follow them, it's like, "Oh, yeah, no, that is my friend." Or maybe they're a friend of a friend. Okay, this is probably the right person. That's very, very valuable.

Here's another example. Super fluid uh which is a great uh uh streaming uh crypto streaming app. And uh when you're sending uh crypto to somebody, when you start to type in, they grab the people you follow on EFP and they start to do autocomplete with the people you follow. So it's like uh it's like your address book that's cross uh cross app uh saved on chain very very valuable. Here's another example. Of course you know if you have a social graph you can have an activity feed. Here's an activity feed that we made partnered with interface which is a great app as well.

Also just so everybody here knows if you integrate EFP into your app you can get this shiny POA app uh integration first 100 if you're one of the first 100. Like I said, we're at 58. Uh we're going fast, so these will run out soon, but if you're in the first 100, you can get one of those if you let us know. So, are there other elements of the Ethereum identity stack? Yes, absolutely. Um you know, and of course, there's there's degrees here. So, like Ethereum attestation service, uh you can attest to anything. Very very valuable XMTP. You know, it's got a lot of adoption. There's a new thing called Ethereum Common Protocol and many others. I'm not trying to do an exhaustive list here.

So you might say, okay, I want to integrate the uh Ethereum identity stack, but there's all these different parts. Well, we've created a component library that puts it all together. So Ethereum identity kit ethdentkit.com. It's component library and API that makes it super easy to integrate the Ethereum identity stack into your crypto app. So for example, you're like, you want a profile, you don't have to design it. We've already designed it. Here's like a profile card. You can use a follow button. You can drop into your app. Um we have a public API where you can get all, you know, all this identity information. and you just grab it from our API. Of course, all this is onchain if you want to, but if if it's easier, you can for you, you can grab it from our API. We have a sign with Ethereum button that's welldesigned. We also have

LLM context files. So, everybody here's uh uh programming with LLMs. Uh they don't always have the right context for these niche things. You can download our LLM.txt files, plug it in, it'll have all the information it needs. Here are some links. And uh yeah, does anybody have any questions? We have a few minutes for questions. Yes. What's next? Uh I hear somebody's working on putting credentials on chain, connecting it to their ENS profile uh as a central place. I think that'd be really useful.

Does anybody else have any other questions? Yes. Yeah, absolutely. So, a nice thing about this being entirely on chain is that not only can humans have it, but uh machines can have these things as well. So, absolutely that includes AI agents. I expect that AI agents will have ENS usernames, profiles, they'll have, you know, you can they can follow each other on EFP that you can follow them.

Absolutely. All these things are open to machines as well. Uh yeah the first part of the question was front uh it was essentially about the AI agents also owning the identity. The second one is about the proof of humanity. Are there any plans to integrate it? Yeah. So uh absolutely. Uh so I'm not the gatekeeper of what is exactly the Ethereum identity stack. These are my opinions about it. There are many proof of humanity you know uh uh projects out there and those absolutely are a part of this as well. uh in Ethereum identity kit specifically. We just haven't done that yet. Uh you know, but um definitely could be something in the future. I mean, proof of humanity is a tricky

thing because anything you do ultimately probably can be gamed in some way or has some sort of results you don't like. I tend to think of proof of humanity as more of a probabilistic thing or like a web of trust thing. So actually I think EFP is actually a good proof of humanity because uh if somebody who I trust follows another person on chain then that person's probably a real person and as you go further and further away from other people you know you know degrees of separation it gets less and less likely. Um, so I actually think that the old cippher punk web of trust idea is I think that's actually the most

workable, but you know there's experimentation all the time. Yeah. Right here that day in Yeah. Yeah. for my AI agents to be able to um follow anyone if you desire. Absolutely. Yeah. So, it's funny you say that. We're actually working on it like right now in MCP server for all these things. Yep. And uh I mean look, AI is the future.

Everybody here knows that. It's it's massively sped up our development process and we want to uh make it even easier. So, absolutely. I look forward for it. Any other questions? Okay. Well, uh we uh we will be here to discuss afterwards if anybody else has any other questions. Thank you very much.