Sutton restates the Bitter Lesson; Dileep George fires back: 'don't fall for it'
anshulkundaje · x · 2026-11-08
Richard Sutton distilled his famous Bitter Lesson into 26 words: don't be distracted by human knowledge — focus on methods that scale with computation, like search and learning. MIT professor Dileep George reshared it with a chart pushing back: a "better lesson" is to push the envelope in new directions rather than assume the scaling curve keeps paying off. The exchange captures growing skepticism about the marginal returns of pure compute-and-data scaling.
More from AGI Musings
- Albert Gu: Bitter Lesson isn't 'just scale Transformers' — algorithmic innovation matters — sewon__min · 2026-11-08
- The 'evolutionary mismatch' chart explaining why humanity will ignore AI warnings — sebpaquet · 2026-11-08
- The meta-question as agent swarms scale: can models beat gradient descent at its own game? — JensHonack · 2026-11-08
- "Don't pause AI for healthcare": Plinz argues AI can cut US healthcare costs — JHochderffer · 2026-11-08
- Richard Sutton to keynote RISEx 2026 on 'Robot Kindergarten': robots must learn online — RichardSSutton · 2026-11-08
- Critic says LeCun, Marcus, Sutton are active detractors of serious AGI discourse — danfaggella · 2026-11-08