Peter Todd @pete@petertodd.org

@bryancyan Tor has bridge nodes that specifically are _not_ part of the consensus. Those get blocked via what's essentially a white lost of allowed encrypted traffic.

@jgettbtc https://mastodon.petertodd.org/@pete/105530857173539502

@untappedgrowth @kexkey Implementing tribes well is a very tricky UI/UX problem!

@bryancyan The reason why China manages to block Tor has nothing to do with the consensus mechanism. Any consensus mechanism can be blocked, by simply looking at what nodes are in the consensus and blocking them.

The actual problem for Tor re: blocking is china has basically moved to a whitelist model.

@untappedgrowth @kexkey well, now you're getting into very tricky UI/UX problems. :)

@bryancyan Tor needs to come to consensus over bandwidth and other things that don't translate well into decentralized consensus schemes.

@kexkey sorry, specifically, sybil attacks.

@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.

https://bitcoinhackers.org/@emzy/105531731278274913

@bryancyan That's incorrect. Tor requires a consensus over all Tor nodes. That consensus is what's centralized. There's no good way around it - arguably the very fact the consensus is centralized part of the argument as to why Tor is secure in the first place.

@XBT He's welcome here. He can have all the blue ticks he wants: ☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️☑️

@benthecarman good point!

@alanturing Scuttlebutt is one. I honestly don't know much about matrix, so I can't really comment on it.

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.

@orionwl @harding @michaelfolkson @k3tan @stephanlivera I looked at that too and it wasn't clear to me how federation would even work. IIUC the identifiers rely on DNS names, so if a DNS name isn't reachable, things simply don't work. So if you had to move to Tor, after some censorship, you'd wind up having to restart the identity from scratch.

Scuttlebutt is better suited to that, being crypto based.

@alanturing That we need low bandwidth, latency tolerate, protocols that don't care if any particular computer is down. Unfortunately that's not easy to achieve.

@michaelfolkson @k3tan @stephanlivera Not as likely to do it to individuals, and not too hard to move somewhere else later. Most important thing is to setup your own server, even if the hosting is imperfect.

Mastodon itself is imperfect compared to the alternative of Scuttlebutt. But it's a practical choice.

@waxwing last I looked I got the impression that many though signing with addresses was a mistake. It is oddly low level to expose so widely.