I don’t know if this would have any effect, but splitting content distribution and content moderation, in the sense that all providers are free-for-all, but then users would subscribe to block-lists depending on their preferences (just like one does with an ad-blocker), is a notion that interests me. How resilient it is against bad-faith-actors depends on the implementation, but I believe that if you try to do more than just a naive blacklisting, and instead analyse networks, you could manage to get relatively far.

A funny notion I know from some image boards is having a list of substitutions, that replace frequent buzzwords with silly phrases. After a while, you get what someone wants to say, but they frequently get annoyed their messages don’t get posted the way they want it to, which is amusing.