Balaji Srinivasan - Towards A Pseudonymous Bridge

Balaji Srinivasan describes a system to move your reputation to a pseudonymous account without revealing your identity. Using zero-knowledge proofs, you can port verified attributes like follower counts to a new anonymous account, then bootstrap followers with an AutoFollow feature. He measures anonymity in bits and addresses concerns about cost, abuse, and the privacy-signal tradeoff.
All right, so I'd like to introduce our next speaker. Balaji is going to speak about. The talk is called Towards a Pseudonymous Bridge. I'm very curious to hear about this, so welcome to the stage.
Thank you. All right, let's see if this works. Should I use this one? All right, great. Let's move this one out of the way then. All right, so one thing I've been thinking about for some time is that pseudonymity is as important as decentralization. You know, we talk about decentralization a lot. You know, a network that is split across many different jurisdictions. It's hard for any, you know, one. One country to shut down, for anyone entity to shut down. But pseudonymity is also quite important because it means that even if the network is decentralized, you can't target the person who is creating it if they're pseudonymous. And there's other reasons for it as well. Like if you think about the founding of the United States of America, that was done by actually the Federalists or the folks who wrote the Federalist Papers, which are all pseudonymous. There's many important ways in which pseudonymity has allowed people to speak freely about things that otherwise would be attacked under real name. And so what I'm going to talk about today is sort of a high level construct for how we might be able to transfer some reputation to a pseudonym. And just to motivate this a little bit, with something like ZEC or more generally zero knowledge, we can now transfer wealth to a pseudonym. Okay. And the question is, can we transfer reputation to a pseudonym? And we need some definitions then of what reputation reputation is. So right now, one way of thinking about. Can we see the cursor up there? Okay, it doesn't. All right, fine. So nice. Okay, so right now we have basically a couple of choices. You can set up an account that has, you know, this is your main account and it has zero anonymity. And I'll define this in a little bit, but zero anonymity and all of your followers, all your reputation. Or you can set up a new pseudonymous account and that has 100% anonymity, but it has, you know, none of your followers. Right? And that's kind of the trade off. And one way of thinking about this is this is your main account, your Mark Anderson, who will be my running example here. And this is Mark's, you know, zero follower, you know, pseudonymous account where he has to, you know, start completely from scratch. The question is, is there something in the middle where we can, with a Construction reduce your level of, you know, have maybe fewer followers in your full thing, but have somewhat less anonymity, but yet enough that you still feel anonymous.
Right.
Okay, so first let's, you know, quantify anonymity via this concept of 33 bits. Right? The idea is that 2 to the 33rd power is about 8.6 billion, and 8.6 billion is bigger than 7.53 billion, which is the population of the world. So that means that with 33 bits of information you can uniquely, you know, identify any person.
Right.
And what that does, it turns your degree of anonymity into a, a measurable, a quantifiable thing rather than a zero run. Rather than just being, you know, anonymous or not anonymous, you can say, how anonymous are you? Well, how many bits of information do I still need to take you to just exactly one person? It's 10 bits of information, then you're one of a thousand people. If it's 20 bits of information, you're one of like a million people. Just to give some stats. The, you know, as I mentioned, 2 33rd is about 8.6 billion. The world population is 7.5 billion. Two numbers that will be useful to us. The Twitter maus monthly active users about 330 million. And the Twitter verifieds there's about 330,000. And so this is like the number of bits in that set. And it's a very simple calculation. Just log base 2 of N. We define it as the anonymity, the number of bits. So 28.3 bits is how much anonymity you have. If all I know about you is you're a Twitter MAU. Okay, now one way of thinking about. So this is quantifying the anonymity side, quantifying the reputation side. Well, we know that with buying Zack on Coinbase, Coinbase.com, you heard of it? Okay, if you buy Zec on Coinbase, you can send to a shielded address and, and you can get 50k zek with a pseudonym and 50k USD in your normal bank account. So you can use the wealth you've earned under your real name to bootstrap the wealth of a pseudonym. Can we come up with a construction that's sort of similar where we take, let's say some of the followers over here and bootstrap a pseudonym that has some of those followers. Is that possible? Because that's what I mean by transferring reputation to a pseudonymous. So another kind of data structure or concept which we'll need before we proceed is just this idea of an adjacency matrix. And you may all be familiar with this from graph theory, but if this is you and you're Mark and you've got followers over here, and here's another person who's following one of your followers, this adjacency matrix is basically defining in this case as Mark is followed by this person and this person, not this person, by this person and this person, and then she is actually following her. And that's that edge over here, right? So it's just a binary matrix of adjacencies, right? And so now what we want to do is we want to see, okay, given those definitions, can we trade off some anonymity for some followers? So the naive approach is you set up, you know, a new account, okay? Your little pseudonymous account over here, okay? And you try to preserve all of the same followers that you had over here, right? You have something where you can send over all of those followers, your entire account. And of course, the issue with this is anybody who's looking at the full adjacency matrix can see that. Oh, and by the way, these are not totally separate networks. The right way to represent it, or think about it, is really something that looks more like this, where you set up a pseudonym and then you had your entire neighborhood. All the people who are pointing into you are now pointing at that pseudonym. And the issue with this is anybody who is scanning the entire graph, the matrix of all connections, can immediately see that this vector of followers and this vector of followers are the same. And so therefore, you're basically just reduced to one person. It's unlikely anybody else has exactly the same pattern of followers. And therefore you have given up your anonymity, right? Okay, so naively, just transferring all your followers to the new pseudonymous account, that may not work. So what else can we do? What if we back up and we say, all right, not just think about your followers, but all the profile information more generally. There's really kind of three types of user profile data on Twitter. There's platform specified stuff like your verification check mark, which Twitter itself generates. There is user specified stuff which is like your name and your address, your bio, etc. Which you are typing in. And then there's stuff that is specified by others, which is, you know, your followers. And then also every, like every RT that comes from other people, right? It is the stuff that is specified by the platform and by others that's the most reliable to. To maybe to do something with, because it's harder to fake for an individual. And what you can do is you can take this idea of, here's your new account, your pseudonymous account. If, and this is one of the things I was mentioning towards the end. But if you assume a crypto Twitter, that is to say, if we have something that's like, if you've heard the saying, assume a spherical cow. It's from physics. Right? Okay, whatever. I thought it was funny. So if you assume a back end where you can do a construction like this, where you can do a zec like transfer, maybe that's like Twitch or something like that, you can move that verification over here to the pseudonym. And now this thing actually is a verified check mark over here. Okay, you've moved data in a similar way to how with ZK Snarks, you move money from one Zek address to another. And what does that give you? Well, what that gives you is some progress, because here, this was just under your full real name. So you were just one out of one people. If you just set up a pseudonym with no followers, you're one out of 330 million Twitter users. But you had no followers or anything. Now, when you're verified over here, you've given up some anonymity. You're one out of 330,000 verified. So you have 18.3 bits. We lost 10 bits of anonymity to put a verification flag on this profile. But what did that gain us? Well, let me introduce one more new concept, which I call Auto Follow. So Auto Follow would be a very useful feature for this, say, future crypto Twitter. And the idea behind Auto Follow is you'd set a flag in your profile which says, okay, I'm going to auto follow any of these new pseudonyms that have a verified flag, right? Or any new pseudonym that can prove it had more than 10,000 followers with some other pseudonym or any new pseudonym that was followed by ack, something like that, where you're auto following these notable accounts because that you think that they're going to say something interesting that they wouldn't otherwise say. So in this example, we have these three people all set up to auto follow someone who's verified. And notice that this lady over here is not actually following Marc Anderson. She's just set up to auto follow somebody who's verified. So if you have a normal pseudonym, nobody's following them. But. But if you have a verified pseudonym, you do that construction and you move the verification over. Well, then AutoFollow kicks in and boom, these three people are now following them. And so this is kind of interesting because now what we can do is we can say, okay, that gave up 10 bits of anonymity when we did this verification, but in return we gained three followers for that pseudonym. Now let's just plug in some random numbers here. I showed just five nodes over here, but Mark has 700,000 followers. So let's just assume that there were 150,000 accounts that would auto follow after this verification step. So in that example, we would go from a situation where either Mark had his full account with 700,000 followers and zero anonymity, or he had a pseudonym that was starting from scratch and it had no followers but total anonymity. Now he has the option of something in the middle where he uses zero knowledge proof to move over some information to the pseudonym, namely that he was verified. That drops the degree of anonymity by 10 bits. That's a reduction in anonymity from porting over the verification. But in return gains all of these people who are auto following any new verified account. Right. So that's actually pretty interesting. You know, this is a relatively small tweak to Twitter, the auto follow concept. Again, I'm not necessarily thinking Twitter itself would implement this, but if you could do a Twitter like thing on the blockchain, which a lot of people are working on, this would be a very useful feature. Now you can generalize this because while I talked about verification over here, there's lots of other digests that you can move over with zero knowledge. For example, you can have the, you know, the assertion that you're, you have more than 100,000 followers or you're followed by ack and other things like that. Anything that does not in and of itself reduce you to an anonymity set of just one you can potentially port over. And before you do that attestation before you, before you port this over, you can see how many bits of information you're giving up in terms of anonymity and how many new followers you would get with auto following. Okay, so this is kind of like a high level sketch of how we might be able to transfer reputation to a pseudonym, just like we've been transferring wealth to a pseudonym with Zec in terms of next steps, you know, as I said, you know, Sumospherical Cow, sumocrypry, Twitter. A lot of people are working on things like this. And this gives an interesting sketch of some of the API, some of the functions you'd want to call. You'd want to be able to make these kind of zero knowledge constructions. You want to have autofile, you Want to have a few things like that. But the broad idea is that if we set this as a goal of a pseudonymous bridge, to bridge between your real name's reputation and the pseudonym's reputation, I think this at least sketches some directions for that. And it can say what could go on chain for a crypto based social network that would support this kind of thing. So thank you very much. I guess I can take some questions here. Sure.
So thank you for the talk. Quick question. Are you assuming in this model that you somehow could limit the number of pseudo anonyms that the person could create for themselves? Or if it can be. If I can create 10,000 verified accounts for myself now I have an army of kind of reputable people on the platform that is actually represented by a single individual.
Yeah, great question. So I think a couple of thoughts relate to that. One is you could make it costly to create a pseudonym. It might cost you ZEC or money of some kind. Right. So that's one solution. The second is that it might actually be good for people to be able to have multiple pseudonyms so that for example, one of them you say hey, okay, I'm followed by ack. But the other one you say hey, I'm followed by this geophysicist. Right. And so then you can talk to the geophysics community over here without being character assassinated or whatever. Right. So I'm not against having multiple pseudonyms. I think that's reasonable. You may want to tax it to prevent abuse of the kind that you mentioned.
Yeah, it's a very cool construct. I have additional question. How do you handle modification on the attributes of the main ID like let's say mark goes below 100k followers. How does it go for the other?
That's a great question. And so what I've thought about is that you just have it be a one time transfer and then you know, auto follow kicks in and then if it's out of sync, it's out of sync. That's the simplest way of doing it. And I don't think you lose too much if you do that. Instead of, I guess an alternative to taxing pseudonyms, I guess you could also use linkable ring signatures to limit the ability to have multiple pseudonyms. Yeah, that's a good comment. Yep, that's true. Have you considered as another alternative, making the full graph of followers private so that you could just fork your following to a new account, but people couldn't see that map that you were showing earlier? That might be possible, but I think that if you Had a crypto Twitter, a big part of it would be allowing folks to be able to download and view all of that data because it might be possible to keep all of your followers private somehow. But it would seem to be different from how Twitter currently works and different enough that I'd have to think about it. Yeah, I'm really struggling with the auto follow part. If everyone's following everyone, doesn't that break down the signal versus noise? Ah, yeah. So it's not everybody following everybody. But you would set. So part of the attraction of a platform like this is basically Twitter, except folks would set up these like, you know, for example, a verified account that has at least one other account associated with that has 100,000 followers. And you would actually say you'd have a threshold for how significant those non accounts would have to be. Right. So you wouldn't just auto follow every new account that came out. You'd say, okay, I'm going to auto follow somebody who's got 100,000 followers and is verified, but somebody else might have a higher threshold and say I'm going to only auto follow a new pseudonymous account that has 500,000 followers is verified. It was also followed by acknowledge. But how's their ack if it's anonymous? Because ack follows like 4,000 people. No, but there is no ack in new Twitter where everyone's anonymous. There's no. What I mean by that is if you had a. Let's. I'm using ajack as like a, as a signal for a high profile account on Twitter. As you know, you've got both named accounts and you've got pseudonymous accounts. Right. So if a high profile named account follows your named account, that's actually something that you can porter to the pseudonym. My named account was followed by a high profile person. Therefore this pseudonym is actually valuable. It's one of 4,000 accounts that ack followed. Does that make sense? Well, so it's not new Twitter, it's hybrid Twitter with an identity. Yes, that's right. Basically you might have to actually like rebuild a whole piece of Twitter on there. But. Yeah, that's right. You know, I think I had a. There's one over, over on the other side.
Other side.
How do you. Oh yeah, go ahead. I can't see where you are. Wave or something. Ok, yeah, so if I followed just Mark and if I then got Auto followed, or rather if Mark got auto followed by me, I would know that it was Mark, right? Yes. So and now that's actually an important point. Basically before you actually put out the attestation. You don't want to put out an attestation that knocks down all 28 bits of information and something that, like you're describing. If Mark only followed one person and you put out my pseudonym is followed by Ark, then you would know before you put out that attestation that that would give up all your anonymity. So one of the features of this is with the auto follow thing, you know first how many followers you're going to get for a particular attestation. And second, with the 33 bits model, you know how much anonymity you're giving up for an attestation. You wouldn't want to give out an attestation that gives up your full identity. One or two more questions.
I'll grab the opportunity. So have you thought about refining this auto follow? Because you could potentially prove that you follow some of your original following.
And.
Make that proof, and then depending on how much of them you like forced to follow you, you'd reveal your identity more or less, because the auto following is just identical to being verified on this new crypto Twitter.
Right. So, yeah. So the way I was thinking about auto follow is at the time that you sign up for the platform and get set up, you have a bunch of checkboxes and it's opted into an aggressive auto follow where you're auto following these new accounts that maybe everything that's verified and has more than 10,000 followers, and then some people will mess with those defaults, but the defaults are for aggressive auto follow. And what that does is basically it says that pseudonyms that do these attestations will actually get a lot of followers. That's essentially what it's engineered to do. I'm open to other constructs that can get you there, but the basic idea is this way you can, before you do an attestation, quantify how much reputation you get, how many followers you get for a given level of decrement and anonymity. And there's possibly other ways to do it, but that's one thought.
I have a question.
Go ahead.
I'm over here.
Yep.
I was wondering, like, so the whole idea is based around the. The verified accounts, but is there an idea that the verified accounts would actually act better or more ethically or like, not lie?
Yeah, great question. So I just use verified accounts as one kind of piece of data that you could bring over, but you could have an account that wasn't verified that had 10,000 or 100,000 followers or those followed by a particular person. The key is just that it can't be data that that person profile itself is generating about itself. Verification is something imposed by the platform. Your followers are coming from something else. So it has to be some kind of data like that.
I think we got one more.
Okay.
My understanding this is like a theoretical construct form practicing the idea of Sudan and how you can transmit reputation. What's your incentive and what's your motivation to work on an idea?
Great question. So, like, you know, it is. It is sort of like the other part of what ZEC lets you do, right? ZEC lets you move money to a pseudonym in the medium to long term. I think that we want to move towards a pseudonymous economy where you have one name for earning, another name for speaking, and you use your real name, like your Social Security number, only in official context. And this is sort of like an HD wallet. You don't keep all of your balance necessarily in just one address. You split it into things and you can have a root address that generates things downstream the same way you have your real name and you generate all these other names for different purposes. We kind of already do this on Reddit. People will use this pseudonym or that pseudonym as it suits them because they don't necessarily want to give out their global identifier for all purposes. The idea here is one thing that deters people from using pseudonyms on social media is that they have to rebuild their audience completely from scratch. If that was not the case, then I think you'd see a lot more interesting content out there. People could speak more freely. So that's part of the motivation. Thanks.
Thank you so much for the talk.
More from Balaji

34 - Network State Conference 2025 - Fireside with Vitalik and Balaji on Zuzalu and d/acc

33 - Network State Conference 2025 - Fireside with Amjad Masad of Replit and Balaji

06 - Network State Conference 2024 - Fireside with Naval Ravikant and Balaji Srinivasan







