Rather than censorship, just give readers better filtering tools. This was largely solved on USENET years ago, has been largely solved in email (to a great degree) -- it's just Twitter, FB, etc. which seem to have a hard time solving it.
If you made certain sets of filters the default, or available in groups, and did filtering based on your best beliefs about the desires of readers (based on keywords, senders, and maybe other characteristics of the conversation), it would go a long way.
I think this is the real solution. Instead of giving people useable, client-side filters, we've let companies push us to using shitty, broken website filters so that they can control the flow of information. Client-side filters work great, and they are how we generally solve these types of problems with other communication methods, and do so relatively successfully.
How would you do that in such a way as to prevent a Twitter army from piling on someone? Or at least make it so they don't feel they're being bombarded?
At some point go whitelist or raise thresholds very high (like, on twitter, only people you follow, or blue checks, show up in your timeline if abuse is detected, as a very rough pass. Could easily do something far better.)
If you made certain sets of filters the default, or available in groups, and did filtering based on your best beliefs about the desires of readers (based on keywords, senders, and maybe other characteristics of the conversation), it would go a long way.