You need to have good, immediate abuse-banning tools. So if Wil Wheaton says "look at my new movie" and you get 60 abuse reports about it, there should be an o

You need to have good, immediate abuse-banning tools. So if Wil Wheaton says "look at my new movie" and you get 60 abuse reports about it, there should be an one-click tool that says "Everyone who reported this as abuse gets suspended and all the reports are removed from the queue {click}". If Wil Wheaton has 92 people replying "fuck you" to his posts, it should be one click to pull out everyone who wrote "fuck you" to Wil Wheaton and give them a slap. You can set the level of violence done to whatever you want (maybe a temporary suspension is fine, maybe you just want them to lose abuse-reporting privileges, maybe you want to nuke them from orbit). But it should be one click to deal with mobbing. "Mark all these accounts as mob participants". Etc. All this stuff is, frankly, EASY, you just have to think through the dynamics and have an understanding of social media. EVERY tool you give users, including the "report abuse" tool, can and will be abused, and therefore you need tools to deal with abuse of every tool. Tools obviously won't help with BAD moderation. If your moderator hates Wil Wheaton and likes Nazis, they can take that side. But it will at least make moderation (good or bad) fast and efficient, which gives you the chance of hiring enough good moderators to provide overall good moderation. (Hint: if you have multiple moderators, you need tools to review your moderators....) You should have all sorts of counters to record stats about abuse - how many problems you get from each other instance, ratios of problems/good content, that sort of thing. It should be easy to notice that you've gotten 5 mobbing attacks today from foobar.mastodon so maybe that entire instance needs a little timeout. And so on.