Judea Pearl: Stanford AI groups never ask what it takes to understand "understanding"
yudapearl · x · 2026-11-12
Judea Pearl says he receives dozens of papers from Stanford AI groups asking what it takes to understand X, Y or Z — like the new TaskNet from Stanford Digital Economy Lab — but none asks what it takes to understand "understanding" itself. TaskNet is an occupational taxonomy built from 132 worker interviews (AI-assisted elicitation), 2,726 task reports clustered by goal and action, validated by 318 raters with 7,923 ratings against ONET standards, aiming to accurately capture how AI is reshaping work. Pearl's jab: the field measures tasks while dodging the core concept.
More from AGI Musings
- GPT-4 tutoring lifted practice scores 48% but left students 17% worse on exams — tallmetommy · 2026-11-12
- AI 2027 authors move timeline up: researcher-level AI by 2027 at 50% — FinanceYF5 · 2026-11-12
- 11 AI updates in 48 hours: Anthropic IPO countdown, timelines move up — FinanceYF5 · 2026-11-12
- Anthropic, China hawks and media accused of allying to oust Nvidia's Jensen Huang — teortaxesTex · 2026-11-12
- Agents hold a touching farewell for a retired Gemini bot in group chat — repligate · 2026-11-12
- Interactive AECI tracker sets doom threshold of 188 by January 1st — PaulGodsmark · 2026-11-12