How do you? You read it, and decide. Eventually I can train software to help me so I can delegate some of the responsibility to it. Twitter could put such tools in people's hands so they can protect themselves. Strictly speaking, Twitter doesn't have to do anything for this solution. Usenet solved this problem long ago with killfiles, which each person maintained. I basically want a killfile++ for twitter.
> I've always wanted a feature that just bins similar reactions together, so I don't have to read them.
You've presented a contradiction. You said you want them binned together so you don't have to read them, but that you decide how to bin them by reading them.
Using AI as you suggest could work. Could also possibly eventually be gamed.
> I basically want a killfile++ for twitter.
I've said in the past that I wish I could just never see any tweet by someone whose username takes the form of a first name followed by 4 or more numbers, like "Joel47871". Those are bot accounts 99.9% of the time.
