3 ms·
It's actually a simple solution. Have a flamewar detector similar to HN, and then for that thread and those users, have it look like the other people in the th
by jennyyang 6y ago
It's actually a simple solution.
Have a flamewar detector similar to HN, and then for that thread and those users, have it look like the other people in the thread respond "Ok, I'm sorry." Obviously randomized using machine learning so that people will think that it's a real apology. Everyone on the thread will think the other person apologized and will simply move on. Or, if they respond, no one except the writer will see the responses.
- gallegojaime 6y agoI might see this working only if it doesn't curtail freedom of speech that much. But it's too eerie. Without UX cues I don't know if Twitter wrote something or you did. Train it slightly wrong and Twitter takes control of some important figure's public viewpoint, perhaps even replacing something worth being angry about in the first place.