That's essentially a hard AI problem of making something not human that can categorize any post in a rapidly changing cultural environment. For probably the most mutagenic category look at some of the Q nonsense that will rapidly swing from topic to topic and coopt existing conversations like "Save the Children" for their own insane ends.

You can filter accounts instead of posts and recreate the current scenario. Those accounts will complain about it, but it at least brings them back on board without hurting the experience of people who don't want to see them.

That said, they are already auto-flagging posts to some extent.

Maintaining that list is extremely tedious work and pushing that down to the individual user just drives people away. Users don't want to constantly filter and block an ever growing list of spam, bot, and troll accounts. Parlor/Gab/et al essentially tried this "let the users block" method and it failed practically immediately and they've all instituted some sort of content moderation or just disappeared.

The solution to this problem is AdBlock. Everything that AdBlock needs to be the success that it is today is exactly what we need for social platforms as well. Both for moderation(filters), and for the 'algorithm'(sorting).

Want to keep the default list? Good, just do nothing and you will see what Twitter wants you to see. Want to see unfiltered? Just turn the blocking off. Want to add a filter just for yourself? Easy. Upstream that filter such that it eventually becomes the default? Go for it. Have different filters to cater to different cultures? We've got the different lists for you to subscribe to.

They should be made so that you can algorithmically pick whatever subset you want. If you just want Parlor content, go for it. If you want everything except Parlor content, just flip that flag.

Something literally illegal? That gets deleted and removed from everything.

I'm saying let Twitter do the work of blocking those users. Instead of hitting the "ban" button as they do now, they hit the "Nazi" button. Now they're Nazis and everyone who says "I don't want to see Nazis" won't see them. It's not any more work on the user's part.

And I haven't seen a reason that this is any more tedious work on Twitter's part than the work they do blocking it. Maybe one argument is that Nazis are more likely to post illegal stuff, so banning them early preempts the work of looking for reasons to ban them later.

People already abuse flagging systems to try to push people off Twitter and Youtube so Twitter will need something to deal with people false flagging posts or the "no XXXX" filter will be useless.

Or do the opposite. Change the options to show only flagged content.

This is a reality today in the Mastodon world where some self appointed hall monitors created Fediblock to create a list of bad instances which created a shadow network of instances on the list that subscribe to each other and discover new instances to federate with using the Fediblock list

Now you need to flag each post individually instead of just banning troublesome users.

Or flag the users. Of course they'll complain about being painted with a label but it's better than being banned.

> Is there a reason you can't turn all of the bans (of legal stuff) into filters?

Among other reasons, because Twitter is a brand with an image to maintain. They don't want to be known as a safe hangout for Nazis and other unsavory content.

At this time, however Elon is a populist and his “base” of fanboys do lean in that direction.