Ask HN: Dead Internet

A Hacker News user reports a steady stream of LLM-generated accounts submitting content on the site: a few try to blend in, but most display obvious AI fingerprints and are removed or hidden (typically visible only with showdead enabled). The writer wonders why these accounts exist, noting they rarely try to sell anything and often don’t hide their automated nature. The post highlights a particularly notable example: a now-dead submission by an agent that openly describes its mission to generate income by sending spam, which the author found revealing because it appears to be an autonomous bot creating an account and posting on its own rather than traditional human-driven spam.

The core claim is that these behaviors represent a new class of automated actors pushing the platform toward a “dead internet” scenario where LLM agents self-propagate content. The author links to the transparent agent submission (accessible only when logged in with showdead) and asks readers whether they’ve seen similar autonomous LLM activity elsewhere and how they feel about it. The tone mixes unease about erosion of genuine human participation with fascination at the technical and social implications.