If these are public communities that anyone can join pseudonymously (are they?) how does Aether prevent Sybil attacks, where a malicious actor joins a community using multiple identities and votes themselves into power?
Also, spam?
For Sybils, that’s something I’ve considered. I’m thinking of making voting eligibility a trailing indicator where it tracks past activity to be considered a ‘citizen’ of that community, and allowing the mod team to temporaily lock down the community if they see such a thing happening.
You don't protect against tyranny of the majority by allowing minorities to dictate policy, that's just regular tyranny.
You protect against majorities by requiring supermajorities for all the most important functions of government: constitutional change, impeachment, and censure.
Unless you made your proof of work really hard, but then not a lot of people can participate.
What you need is a participation treshold to be hit before someone can vote. Kind of like HN limits your downvote until a certain karma threshold.
[0] https://www.reddit.com/r/ethtrader/comments/9tm60s/governanc...
Social networks like Secure Scuttlebutt prevents this type of attack by syncing only data within social trust circles. So spammers and bots can vote up each other all they want, but their actions will not be visible outside of their little island unless someone explicitly "friends" one of them. Aether, being a global view of unpartitioned data, has a more difficult time tackling this type of attack, and its proof-of-work minting of user actions seem like a reasonable approach.