François Fleuret's definition of understanding: no unconscious consistency checks ring an alarm
francoisfleuret · x · 2026-10-22
Notable AI researcher François Fleuret offers his current definition of "to understand": none of your unconscious consistency-checking processes rings an alarm — framing understanding as full internal self-consistency, a useful lens for debating whether models truly understand.
More from AGI Musings
- Matt Turck: in AI, everyone is competing with everyone — chips, models, clouds and apps all overlap — mattturck · 2026-10-22
- Anthropic researchers publish essay on technological progress and the meaning of work — paulnovosad · 2026-10-22
- ARK's Sam Korus: Tesla Optimus academy and David Silver's super learner converge — skorusARK · 2026-10-22
- Chinese essay: in the AI era, individuality becomes extractable data representation — teortaxesTex · 2026-10-22
- VraserX mocks Dario Amodei: those warning about AI want to decide who gets access — VraserX · 2026-10-22
- We're close to losing the ability to see what AIs are actually thinking — Puzzleheaded-King584 · 2026-10-22