In this interview, Woj (woj.eth), founder of Mazury, explains the idea of social recovery for crypto wallets. Instead of a single seed phrase, a user designates a set of trusted contacts—friends, family, colleagues, or community institutions—as guardians. If the user loses access to their wallet, a threshold (e.g., 5‑of‑7) of these guardians can attest to the user’s identity and reconstruct the seed phrase. The discussion covers the intended user‑experience (selecting guardians from phone contacts, keeping guardians unaware of each other), the game‑theoretic rationale that most guardians would not collude to steal identity, and the risks of coercion or collusion. It also contrasts the trustless nature of blockchain transactions with the need to re‑introduce social trust for identity recovery, arguing that decentralized communities could replace centralized identity providers. Finally, the speakers acknowledge that social recovery is still in a research phase, with many open questions about security, usability, and real‑world adoption.
Chapters
Social recovery is basically relying on this, that instead of you storing one seed phrase and your seed phrase being the entry point to all your accounts, you delegate 5 or 10 of your friends that, let's say 5, you delegate your 5 friends that are able to recreate the seed phrase to your system. So you delegate it to me, you delegate to your boss, to your mom, to your cousin, and to somebody totally different.
How is that a better experience? Yeah.
So right now it's a hard experience. Right now it's very hard.
And also, man, that's prime territory for coercion, them working together. If I just—
yeah. I'll let you finish. There's many problems again, but If done right, it should be as easy as you create a wallet, you have your contacts on the phone, you decide who are your guardians. They never know about each other.
Okay.
And if you choose a set of 7 people and there is your family and there is your friends and there is your colleagues there, or whatever type of relations, you need to be a huge asshole for all of them to collaborate against you.
Yeah, but huge assholes exist, man.
Yeah, but in theory, game theory behind it is pretty, it goes in your way. People will not probably collaborate to take over your identity because why would they? They appreciate you in your life. And at least in this positive scenario, it should work like that. But in general, imagine this UX, you create a wallet, you set your guardians or whatever. And you use your identity. If you lose your phone or you lose your access to wallet, you can ask 5 out of these 7 guardians or identity providers to just confirm that this is you. And if 5 of 7 confirms, then you get your wallet back. And this was also mentioned in this decentralized society paper I mentioned before. They used a method for attestations instead of these guardians. It's like, I can ask my university to confirm that this is me. I'm a college dropout. I don't have a uni, but if I had a university, they could attest that this is me. I love that smile that you're like, I'm a college dropout. I don't do this shit. But you have your university, your football club, Your family, or all the different communities that you hang out with. And maybe again, 5 out of 9, or whatever is the count of your communities, can attest that, oh, this is actually you, and now you deserve to recreate your wallet.
Okay. One thing, taking it back to first principle thinking, big word, but it doesn't mean anything. Isn't the whole concept of, or isn't part of the sort of Blockchain, doesn't it enable a sort of outsourced trust in the sense that you— we don't have to rely on each other, we don't have to rely on me knowing you, uh, for me to be able to interact with the internet, um, or, or sort of access the web? And so it's like, if we're trying to outsource trust, why all of a sudden we're like, oh, let me just put the trust back into people?
We go, we go the full circle, right?
Yeah.
So there are applications that again don't require any trust in this paradigm. And for example, sending money doesn't require trusting anybody. You can always set up your node, you can submit the transaction, it's going to get processed. And this is very much trustless. But at the same time, when we try to recreate some setups, some other setups that actually require trust, like this social aspects. And just if I want you to trust me, I need to earn this trust somehow. And the key here is not to rely on one identity or on your bank or on Google or on other private corporation to store this identity for you. But maybe your friends should be the source of the identity. Communities that you hang out with should be should be these guardians of your identity. And here comes one caveat that I wanted to mention. You mentioned thinking from first principles and doing some obviously good solutions. I wanna be clear, right now we are in very much a researchy phase of how can we improve this with this new tech, et cetera. It's not really like, thinking what's going to be immediately the best for the people. Right now we are in the phase of like, oh, this could work like that. Maybe it's going to be better. I don't think anybody knows the answer yet if this is actually the smashing success. Okay.