HN has mechanisms to detect upvotes and comments that seem to be promoting a product or coordinated in some other way. I'm not sure what they do behind the scenes or how effective it is but it's something. Also other readers downvote bot spam. Obvious bot/LLM-generated comments seem to be "dead" quite often, as are posts that are clearly just content/ad farm links or product promotions or way off-topic.