5 ms·

Just to challenge the easy part (or is this a challenge to peasy), can you please expand on "somehow"? What is your strategy here to ensure that only "real" pe

Just to challenge the easy part (or is this a challenge to peasy), can you please expand on "somehow"? What is your strategy here to ensure that only "real" people can get access to the private keys (and not real people proxying for an automaton)? How do you maintain this assurance for a long running period of time?

Reelin 6y ago
Web of trust (or something substantially similar) solves this on an algorithmic level. The problem is that, to the best of my knowledge, we lack friendly and accessible UI tooling to support such a scheme. Naive web of trust algorithms are likely to threaten anonymity in their own way though. Zero knowledge proofs might be able to facilitate a solution to that. > and not real people proxying for an automaton You can't prevent this under any scheme, only detect it after the fact. An authenticated individual will always be able to deliberately take actions on behalf of someone else.
- amelius 6y ago
  > You can't prevent this under any scheme, only detect it after the fact. So do I understand correctly that if someone posts 100 messages "anonymously", and in the last one they accidentally reveal themselves, all the other 99 messages are traceable to that same person?
  - Reelin 6y ago
    The part of my comment you quoted was referring to preventing people from disingenuously proxying for another entity, not to maintaining their own personal anonymity. Regarding accidentally revealing oneself, that's always a risk. If you create an anonymous HN identity but then inadvertently link it back to yourself then all previous posts under that account are immediately traceable as well. Web of trust itself is easy and already widely used by many free software communities as a decentralized means of securing reputation based identities (anonymous or otherwise). The challenge is in allowing an established identity (established either via reputation, government ID, in person interaction, whatever) to perform an anonymous but trusted (ie authenticated) interaction with the network. In other words, to implement a web of trust scheme that supports anonymized but otherwise verified interactions via zero knowledge proofs.
cosmojg 6y ago
Heh, that's definitely the messier part of this. It all depends on your threat model and the scale of the community you're running. For a small group of trusted friends, one could distribute keys in person and enforce time-based key expiry. For larger groups of strangers... I'm not sure. Reelin's web of trust idea seems like a good start. The only ideas coming to my mind right now require burning down the community and starting over whenever compromised. This is definitely an interesting and difficult challenge! What ideas do you have in mind?

Reelin 6y ago

Web of trust (or something substantially similar) solves this on an algorithmic level. The problem is that, to the best of my knowledge, we lack friendly and accessible UI tooling to support such a scheme. Naive web of trust algorithms are likely to threaten anonymity in their own way though. Zero knowledge proofs might be able to facilitate a solution to that. > and not real people proxying for an automaton You can't prevent this under any scheme, only detect it after the fact. An authenticated individual will always be able to deliberately take actions on behalf of someone else.

amelius 6y ago
> You can't prevent this under any scheme, only detect it after the fact. So do I understand correctly that if someone posts 100 messages "anonymously", and in the last one they accidentally reveal themselves, all the other 99 messages are traceable to that same person?
- Reelin 6y ago
  The part of my comment you quoted was referring to preventing people from disingenuously proxying for another entity, not to maintaining their own personal anonymity. Regarding accidentally revealing oneself, that's always a risk. If you create an anonymous HN identity but then inadvertently link it back to yourself then all previous posts under that account are immediately traceable as well. Web of trust itself is easy and already widely used by many free software communities as a decentralized means of securing reputation based identities (anonymous or otherwise). The challenge is in allowing an established identity (established either via reputation, government ID, in person interaction, whatever) to perform an anonymous but trusted (ie authenticated) interaction with the network. In other words, to implement a web of trust scheme that supports anonymized but otherwise verified interactions via zero knowledge proofs.

cosmojg 6y ago

Heh, that's definitely the messier part of this. It all depends on your threat model and the scale of the community you're running. For a small group of trusted friends, one could distribute keys in person and enforce time-based key expiry. For larger groups of strangers... I'm not sure. Reelin's web of trust idea seems like a good start. The only ideas coming to my mind right now require burning down the community and starting over whenever compromised. This is definitely an interesting and difficult challenge! What ideas do you have in mind?