[ninjagoo]

Society is not ready for an AI world: any platform that does not guarantee anonymity will be of limited utility for social discourse in a world lurching towards authoritarianism, and any platform that does guarantee anonymity can no longer reliably distinguish human from ai; not that that should matter when it's ideas that are being debated.

But the bigger issue is the control of money: hierarchical institutions disintermediate workers from the way the fruits of their labor are put to use. Money spent or paid in taxes is aggregated and misused by third parties against the wishes and against the providers of that money. Essentially, your labor is used against you. This is true regardless of where someone is on the political spectrum.

A platform for debate or voting isn't going to resolve this fundamental problem.

[thoughtpeddler]

I agree on the importance of anonymity for social discourse. But if a tool/platform like Polis is some equivalent of a local 'town hall meeting', where there is no anonymity (and you as a citizen publicly appear, state your name, make your argument, etc), then why is lack of anonymity a threat in this specific context?

[ninjagoo]

Because town hall meetings don't work in an authoritarian world.

[worldsayshi]

I believe we can solve both anonymity + proof-of-humanity using zero-knowledge proofs that act as intermediary between a trusted identity provider and the service provider. I.e. you get a digital id but you use it to generate proofs rather than handing out your identity.

Right?

[Zaskoda]

Related: https://www.proofofpersonhood.how/

[judahmeek]

The algorithm used here is ring verifiable random functions, which only provides plausible deniability, nothing more.

Proof of personhood would require proof that every entity in the ring was an actual person.

[worldsayshi]

> PoP makes it possible to prove "I am unique" without giving up privacy.

Ah, very nice! I have been trying to figure out if this was possible!

[ninjagoo]

Is that even feasible? Thinking of it like security certificates for humans. Can there really be anonymity if a cert signature chain has to be trusted? CAs and intermediaries can always trace certs back?

[warkdarrior]

And who is going to be a trusted identity provider in authoritarian regimes?

[worldsayshi]

Yeah, I mean that is definitely an additional hard nut to crack.

I also think it has potential (partial) solutions. I'm thinking that there are many ways to prove identity information. You could use something like tlsnotary to prove that you can log in to a certain web page (i.e. you are an employee of corp X). You can prove that you know someone that know person Y given certain encrypted data.

I just think that Zero-knowledge-proofs are very under explored. As I understand it, and I am not an expert - more or less anything that can be proven algorithmically can be turned into a zkp. Any question that algorithmically can have a yes or no answer can also avoid leaking further information if handled in a zkp way.

I just learned like a few basic examples of zkp and I realized that so many proofs can be made this way.

[ninjagoo]

Maybe the solution is to accept that anonymity comes with the trade-off that bots will also participate. The dependency then is on effective moderation.

Perhaps effective moderation is achievable today through, dare I say it - bots? They certainly seem capable of it now, perhaps more effectively than the average human?

[chickensong]

It needs to be globally distributed and not tied to any one government. Worldcoin but not Worldcoin.

[worldsayshi]

Either that or a protocol that is as reusable as email that can facilitate a variety of identity proof and trust mechanisms.

[econ]

AIs in bunkers rigged to explode.