François Fleuret: AI will force us to decide who can understand what
francoisfleuret · x · 2026-10-02
Meta AI researcher François Fleuret offers a hot take: interacting with AI will force us to clarify who can understand what, pushing either for "intelligible paths" into model reasoning — or an admission that "trust the AI bro" is the only option for keeping the new tokamak stable. The tension between explainability and blind trust is becoming unavoidable.
More from AGI Musings
- Verifiable tasks are solved: Opus 5.5 hits 100% on accounting work — CurieuxExplorer · 2026-10-02
- A plane flies without flapping wings — so does AI — RileyRalmuto · 2026-10-02
- Two questions decide if an AI deployment is smart: cost of error vs cost of verification — YvesMulkers · 2026-10-02
- CC Tokyo hosts Oct 2 screening of 'AM I?' and panel on whether AI can be conscious — kaixhin · 2026-10-02
- Bindu Reddy: Industry Overinvests in Cheap Inference, Hopes SSI Pushes the Frontier — bindureddy · 2026-10-02
- DeepMind vet Graepel: ten years after AlphaGo, LLMs still don't reason — nordicinst · 2026-10-02