The human-or-LLM authenticity fight is getting personal

A writer says they were banned from Lobsters for writing like an LLM (50365159). Commenters push back that the real reason was spam and ignoring site rules, but the thread's top worry is real: self-appointed LLM detectors accusing people with no evidence. A related Ask HN (48886741) asks for a flag for AI-generated articles, and the replies split on whether detection is even possible. Hinton's 'we might have been better off' (50363554) and 'Disaster is still better than AI' (50365009) show the same mood in the opinion pieces.

The pattern: communities want human-only spaces, but there is no reliable test. So moderation drifts toward vibes, and honest writers get caught in it. Even the Show HN game where you question lying LLM suspects (50358352) drew a 'I do this all day for work, no thanks' reply.

The counterweight is the 'Verification Is Not Knowledge' thread (50363019), where people argue that answers from AI are not the same as human understanding. That frames authenticity as a quality issue, not just a rules issue.