Expert Asks AI About His Own Field, Loses Trust Over 'Confident Small Errors'
No_Note7752 · reddit · 2026-09-27
A Reddit user asked an AI to explain his narrow area of expertise and spent ten minutes cataloging subtle errors — nothing wildly wrong, just confident small mistakes and oversimplifications a beginner would never catch.
His takeaway isn't that the tool is bad, but something colder: in every field where he isn't the expert, the same confident small errors are sailing past him undetected. He suggests everyone try this with their own field to see if it changes how much they trust confident answers elsewhere.
More from AGI Musings
- Gary Marcus mocks AI safety camp that sees shutting down computers as 'killing' them — GaryMarcus · 2026-09-27
- Josh Achiam: 'agency' is courage with a neutral valence — jachiam0 · 2026-09-27
- Why the AI Scaling Hypothesis May Never Be Falsified: A Duhem–Quine Argument — burny_tech · 2026-09-27
- The Superintelligence Threat Is Terrifying Because It's Not Magical — AaronBergman18 · 2026-09-27
- Researcher laments the industry is abandoning structure: "just trust the agent, bro" — ivan_bezdomny · 2026-09-27
- "A purely digital job with zero offline component is dangerous today," argues AI commentator — itsOmSarraf_ · 2026-09-27