LR would be rendered untrustworthy in an instant upon such a feature launch and all indiscretions leading up to such a situation could be scrutinised so it's not unfair or even censorship
I suppose you envision distributed effort of classification of people/posts/topics. It runs into the problem of different individuals using different classification methods
I am not saying this is an unsolvable problem, but a problem it is, and I can understand the wish to do such things "correct the first time", which makes it a hard one.
