Andrew Carr: strong priors and RL solve reasoning, math, games, 3D vision—everything but jokes
andrew_n_carr · x · 2026-10-31
Andrew Carr observes that the "strong priors + RL" paradigm now solves essentially every problem: reasoning, math, games, 3D vision, kinematics. The one exception, he notes, is writing good jokes—a brief but pointed take on where RL-based methods still fall short.
More from AGI Musings
- Getting Old Means Watching Norms Cycle for Decades — But Also Falling for Grifts That Feel New — RexDouglass · 2026-10-31
- Cybernetics shows leaps come from new representations, not new data — CatAstro_Piyush · 2026-10-31
- Researcher predicts AI disclosure norms will be dead in 5 years as compelled speech — RexDouglass · 2026-10-31
- 'Intelligence turned out to be mostly statistics': the one-liner reframing the LLM debate — khademinori · 2026-10-31
- Why AI can't build the next Linux: 41M-line kernel and the verification gap — AryHHAry · 2026-10-31
- When a BCI can change the person who would have objected to it — PierceLilholt · 2026-10-31