It's straightforward to create a list of restrictions based on types of information to prevent doxxing, at least from my understanding. Avoiding postings of addresses, employers, etc without the consent of the original person can be made an unambiguous rule. I don't believe such a restriction could be used as a weapon to silence legitimate speech.
My ultimate concern however, is the poor application of harder to define restrictions like "glorifying violence" being used maliciously. I could see pressure from various people that recent protest organizing would fall under the same category (not that I agree with such assertion). As Twitter has decided to filter some speech, bad actors are going to try to use that filtering to get opposing speech filtered.
Additionally, that's all assuming Twitter, Facebook, etc are acting in good faith. There are numerous ways the good intentions here result in dystopian outcomes.
[1] https://www.cbsnews.com/news/washington-dc-protests-unidenti...
I think this highlights my original concern even more. While an ideal world would have some board at Twitter able to make these decisions in an ideal way, the implementation you desire requires subjective judgments. Based on the recent stories on HN about difficulties with app store filtering that should have an objective standard, I am still highly concerned about giving the power to make subjective speech filtering decisions to SV social media companies.
Would you essentially let users choose to filter it out, but show them a garbage homepage before they do, but then also let companies pay to bypass the filter (advertisements), but block politicians from paying to bypass the filter for their advertisements?