Sometimes there's an obvious AI comment, how is it auto-hidden/downvoted?

There is some LLM detection for comments that auto-flags them.

It has some unfortunate side effects, though. Namely on a thread where people are discussing a text-heavy post that is LLM-authored, quoting a large enough block of that content appears to trigger it sometimes. Makes sense that it does, but..

Contrary to popular belief, it's quite easy to spot something written with AI. You just need the skill/gift. I usually spot them rather quickly and will flag on sight. Other users will do the same.