I actually don't mind a little censorship of illegal-ish stuff (or at least warning, that hey, this person is trying to metaphorically sell you something illegal !), but what really bugs me is the subtle (and not so subtle), targeted manipulation I sometimes feel I experience, where there might be like many different usernames all targeting you but somehow all controlled by the same person / interest, and you just think it's normal conversation... (Apparently Reddit is mostly bots?) I think we saw some (heck, a lot of this) in the 2016 election, and personally I try to keep myself to a human-usable number of usernames (I think I have like 5 to hackernews as I forget passwords, but tend to use at most one, or two obviously linked ones for an extended period of time, and then switch if I say something too stupid and want to get a clean start) I don't think something like necessitating a real id is a good idea, since then everything you say is tied to you forever (hail satan), but I'm sure some kind of manipulation detection would be possible... would be an interesting project to work on actually.