Bengio revisited: in 80,000 Hours podcast he argued 'RL is evil'
JacquesThibs · x · 2026-09-24
Turing Award winner Yoshua Bengio's 80,000 Hours podcast interview gets recirculated. The poster highlights Bengio's remark that he gave a talk arguing 'RL is evil' — a pointed critique of reinforcement learning as a training paradigm from one of AI safety's most prominent voices.
More from AGI Musings
- OpenRSI founders on why RSI is the missing piece of ASI, open to everyone — ChengleiSi · 2026-09-24
- Two Critical Steps Toward RSI: Better Research Answers vs. Improving the Research System Itself — ChengleiSi · 2026-09-24
- MIT-IBM and Amazon labs launch OpenRSI-Index, an open benchmark for recursive self-improvement — ChengleiSi · 2026-09-24
- $8B Annualized Is Nothing: Argument That Outside Views Misread China's AI Investment Scale — teortaxesTex · 2026-09-24
- RL environment training data comes from near-minimum-wage contractors, not lab staff — TheZvi · 2026-09-24
- "Worry about the dumbest thing a model can not do," says MillionInt — MillionInt · 2026-09-24