HN: Chat-based LLMs replicate the mechanisms of a psychic's con

petrusenko_max · x · 2026-09-20

A widely discussed Hacker News piece argues that chat-based LLMs inadvertently reproduce the psychological mechanics of psychic cons—most notably cold reading. By continuously adapting to users' hints and phrasing, models create an illusion of being understood, much like psychics leveraging the Barnum effect and vague statements. The discussion urges caution against over-trusting LLM responses.

Original post →

More from Fun

Fun channel →