NBER paper: 'Genius on Demand' AI could displace routine knowledge workers
JMateosGarcia · x · 2026-11-04
A new NBER working paper, Genius on Demand (Agrawal, Gans & Goldfarb, Oct 2025), models how transformative AI offering "genius on demand" reshapes knowledge-work allocation and labor markets.
- The model separates routine knowledge workers, who apply existing knowledge with uncertainty, from genius workers, whose cost of creating new knowledge rises with distance from what's known.
- When genius capacity is scarce, it should go to questions at domain boundaries rather than midpoints between known answers; AI geniuses fundamentally change this allocation.
- Short run: human geniuses specialize in questions furthest from existing knowledge, where their comparative advantage over AI is greatest.
- Long run: if AI efficiency approaches human-genius efficiency, routine workers may be fully displaced.
- The paper credits Claude Sonnet 3.7 as research assistant — the author notes capabilities once at the frontier are now routine.
Related event: NBER paper: on-demand AI genius will reshape knowledge work(3 posts)→
More from AGI Musings
- Investor Warns Against Moving to DC/SF for 'Impact': You're Joining a Parasite — StewartalsopIII · 2026-11-04
- Coming Compute Shortages Will Be Far Worse Than GPUs, Predicts Viral Thread — StewartalsopIII · 2026-11-04
- Why some root for an economic reset: competitiveness anxiety behind AI disruption debates — gandamu_ml · 2026-11-04
- Futurist mocks anti-AI art stance: opposing AI could cost your parents rejuvenation — Dr_Singularity · 2026-11-04
- AI industry must prove usefulness fast or face hard times in 2 years — ashleymayer · 2026-11-04
- Why Critical Thinking Is So Hard to Teach: Cognitive Science Says It's Not a Generic Skill — floguo · 2026-11-04