https://consensus-health.torproject.org/
Looks like the Tor consensus is broken at this moment. Unclear why. It'll take awhile for the network to fail. But if it stays broken it eventually will as the consensus becomes invalid.
Reminder: Tor is centralized. Knocking out a half dozen directory authorities is enough to shut it down.
@pete I2P might gain popularity from this, like mastodon did with the Twitter censorship events... I’m not even sure I2P is still being developed actually, I looked at it a while back. I remember it seemed to have a more distributed architecture than Tor.
@kexkey ...which is kinda scary: what stops MITM attacks? Tor isn't great against that. But at least humans run it, and node operators are a good mix of anonymous and not.
@kexkey sorry, specifically, sybil attacks.
@pete @kexkey we need to bridge meat space and cyber space with distributed trust networks somehow.
Like if we had communities that could determine who was a trusted actor couldn't each community maintain consensus for that community, then those communities form branches to each other based on a different trust model?
That way if one community goes down you can likely still have connections to route around it through trusted parties?
@untappedgrowth @kexkey well, now you're getting into very tricky UI/UX problems. :)
@untappedgrowth @kexkey Implementing tribes well is a very tricky UI/UX problem!
@untappedgrowth @kexkey Not clear really. Hardly anyone has tried lately. Keybase is probably your best example.
@pete @kexkey yeah, that nowhere near resembles a mimicry of the organic into the digital space.
We as humans are used to delegating trust, even in groups. This is doable. It just gets into authority structure & culture creation through incentives... which get screwed up massively by the majority of organizations and businesses once they pass a certain size, but it can be done well.
This is kind of like the digital version of the city state model we all foresee in Bitcoin future. @robingrant
@untappedgrowth @kexkey @robingrant one of the problems of delegating trust is moderation is a lot of work...
@pete @untappedgrowth @kexkey
could that be reduced via barriers to entry and/or smart contracts? for example, rules which take away your deposit if you create spam (with some formal machine definition of spam).
@robingrant @untappedgrowth @kexkey if you could define spam formally in a smart contract algorithm why not use just that algorithm to block the spam to begin with?
@pete @untappedgrowth @kexkey
true. I guess that is already done in some networks. enough down votes and you're deleted, then enough over time and you're out?
@robingrant @untappedgrowth @kexkey Voting algorithms like reddit's always have to put significant effort into weeding out bots.
@untappedgrowth @robingrant @kexkey Well, like I said, try doing something relatively simple like PGP web-of-trust first and see how far we get.
Satoshi model-
Game theory the system -> THEN build it 😛
Stepping all the way back to defining the problems clearly is what enables you to have the intuitional type flashes of brilliance on novel ways to address massive issues with simplicity
Thank you for your time Peter 🙏 this was very helpful
@untappedgrowth @pete @kexkey
human but emergent, bottom up?
Currently I'm thinking delegated but consensual. It needs to be effortless in implementation by the trusted authority though, otherwise it doesn't scale
@pete @robingrant @kexkey yeah, it has to be human. Just like we do socially where we "ignore" the village idiot who is just clamoring loudly in the street.
The breakthrough is we need the ostracisms to be effortless, just like how bitcoin ignores bad blocks, but prohibitively expensive for those trying to disrupt