AI slop anxiety, and where AI actually helps

Ask HN: Add flag for AI-generated articles (48886741) shows the community split. Some want AI text banned outright, and others note that detectors are unreliable and voters often can't tell. The guidelines now say not to post generated or AI-edited text. In 'Rumors of my death are slightly exaggerated' (48037336), Cliff Stoll posted that a rumor of his death was wrong, and commenters joked about an AI simulation of him and warned about slop growing accounts.

The constructive examples are narrower. The Philadelphia Inquirer built Scrape (49930872), an AI tool to surface hyperlocal news, and learned that prompting can control search depth, though cost is a constraint. The Jev thread (49933476) argued that decision models don't beat LLM-as-a-judge or traditional classifiers on simple tasks. The Jev side replied that the benchmark prompts were too trivial and its model wins on multi-step reasoning.

The pattern: trust is the scarce thing. People will accept AI that finds or classifies, and resent AI that pretends to be a person or an author.