To be honest I have showdead on and while I see the things you see, I also see a lot of blatant AI slop getting killed off too. I'm sure most people on here are at least somewhat attuned to Claude's 'voice' by now.
Point taken, I probably file a lot of that away under "generic HN spam" (or maybe I've gotten so good at skimming Claude output that my eyes glaze over automatically now in response :P)
If a LLM-voice comment has [dead] next to it, either the account was already banned in its entirety or HN's LLM text detector marked the comment auto-dead due to suspected LLM writing. If the comment has [flagged] [dead] next to it, the comment was killed by human flagging. You can see both kinds of comment if you browse https://news.ycombinator.com/noobcomments
If you have enough karma to flag comments you can also flag them directly from the noobcomments feed, which is useful to help manage the abusive users who keep re-registering new accounts.