Doesn't Contradict the Bitter Lesson: Soft Inductive Biases Plus Scaling Win
irinarish · x · 2026-11-09
In a reply to Richard Sutton's discussion, irinarish argues the new viewpoint does NOT contradict the Bitter Lesson or the need to scale: the lesson is to scale search and learning without rigid hand-crafted constraints. The right inductive biases, imposed as soft rather than hard constraints and combined with scaling, can be even more powerful — exactly what evolution did, aligning with @andrewgwils's point.
More from AGI Musings
- Rich Sutton: Brute-force compute alone won't get you intelligence — AlexTensor · 2026-11-09
- Bitter lesson debate: Rich Sutton's two views aren't actually in conflict — allenainie · 2026-11-09
- First/last author incentives are the biggest blocker to interdisciplinary team science — RexDouglass · 2026-11-09
- AI commoditized hard skills — 'giving a shit' becomes the moat of the future — brandon_galang · 2026-11-09
- Essay challenges AI's computational functionalism: artificial consciousness may be impossible — sebkrier · 2026-11-09
- Dev reframes LLMs as "universal translators" translating between representations — brandon_xyzw · 2026-11-09