Timothy Lee: "most domains are shallow" captures what the superintelligence worldview gets wrong
binarybits · x · 2026-09-10
Timothy Lee quotes Ryan Greenblatt's remark that "most domains are fundamentally pretty shallow" as a concise statement of what the superintelligence worldview gets wrong. His own view is the opposite: many domains are surprisingly deep, and AI won't be able to "quickly pick up a bunch of domains." The tweet sparked follow-up debate over whether AI can rapidly generalize across fields.
More from AGI Musings
- Multi-Agent Systems Can't Be Backpropagated Into Reliability, Says Safety Bubble Is Collapsing — gerardsans · 2026-09-11
- "The safety and alignment bubble is collapsing under market reality" — NathanpmYoung · 2026-09-11
- Jensen Huang calls AI safety remarks "deeply untrue"; critics hit back — connoraxiotes · 2026-09-11
- "Pause AI to do what?" "Breed super babies to solve alignment" — jessi_cata · 2026-09-11
- Is Technology Positive-Sum, Zero-Sum, or Negative-Sum? Mapping the Three Takes — jessi_cata · 2026-09-11
- Grady Booch: Insiders Who Fear Doom but Only Quit and Talk to Media Are 'Feckless Cowards' — Grady_Booch · 2026-09-11