I take a different view. I think moderation is hopefully not needed. What is needed is some way to differentiate topics within a community. As far as banning users, I would hope a group approach would work. Like maybe if enough community members mute a given user, your client can pick up on that signal and start muting that user also. Not ending up with a million communities for the same subject would be nice too, but not sure how you avoid that one without centralizing.  I think there is a lot of cross functionality between circles of trust and communities. Maybe circles of trust can have a community-specific circle. Should this thread have some community hashtag? I don’t know what that hashtag is.