Hacker News new | ask | show | jobs
by javcasas 342 days ago
It can be better. On slop detection, shadowban the offender and have it discuss with two AI "maintainers", and after 30 messages go and reveal the ruse. Then ban.