François Fleuret: Super-AI Shatters Cognitive Constraints, Then Energetic Ones

francoisfleuret · x · 2026-11-17

François Fleuret offers a terse take on super-AI risk: objects, people, and phenomena generally have a reasonably clear "perimeter of confinement," with constraints falling into two big families — cognitive and energetic. The problem with super-AIs is that they shatter the first, hence the second (e.g., nuclear fusion).

Original post →

More from AGI Musings

AGI Musings channel →