AI slop is turning into a trust tax on content and communities

A recurring request: add a flag for AI-generated articles on HN. Replies point out the guidelines already say not to post generated or AI-edited text in comments, but nothing covers submissions. Others argue detection is unreliable, voting can be gamed, and that a heavy anti-LLM stance made a commenter unwelcome on Lobsters.

The same worry shows up elsewhere. A "rumors of my death are slightly exaggerated" post, from the author of The Cuckoo's Egg, is about an AI-generated obituary. A commenter says AI slop is the easy way to grow social accounts. Readers also now spot the tics: the SlopCodeBench report drew complaints about Claude-style writing. The Decoherence Trap thread argues models with context limits create decoherent systems over long outputs.

The pattern: cheap generation has pushed the cost onto readers, who must work out whether a human wrote something. Nobody has a reliable fix. Moderation by vibes is the default.