We don't need moderators.
All the moderation/trust/ignore stuff should integrate in a single community-base thing.
How exactly?
Any flagging system can be abused, and if there's some kind of majority to vote for stuff, majority can still abuse over minority.
If instead you have someone preventing majorities from mistrusting minorities, basically you'd have moderators again.
So, either you build a system as complex as a State (which might easily fail anyway, as anyone can see), I don't see how your suggestion might meet the real word.
(and that's sad, because it would be nice. Only, it doesn't work)
