Thread Author Proposes New Apprenticeship: Judge First, Then Compare With AI
krishnan · x · 2026-10-30
Part 7 of the same NBER paper thread, focused on practical takeaways.
- Proposed evaluation standard: after 90 days, test assisted performance and unassisted judgment separately on a new case — can junior staff spot and repair mistakes without the assistant, or only produce better work with it?
- The author suggests an alternative apprenticeship: make an initial judgment, compare it with the AI's answer, then explain where either failed. He stresses that whether this builds expertise is a hypothesis to test, not a result the study establishes.
More from AGI Musings
- ChatGPT lands on tech tree's central trunk with Google and iPhone; Facebook and TikTok called side branches — NathanpmYoung · 2026-10-30
- Reddit user maps how a productive AI chat can slide into "AI psychosis" — Individual_Visit_756 · 2026-10-30
- "Aligned with whom?": sharp blog essay argues big AI labs can't be trusted with AI morality — mimi10v3 · 2026-10-30
- Toward a cyborgism research agenda: finding human blueprints for the AGI race — repligate · 2026-10-30
- Silicon Valley buzz: security researchers will soon out-earn AI researchers — Yuchenj_UW · 2026-10-30
- "Lab employees have power only because you are useful — and that won't last," warns researcher — nabla_theta · 2026-10-30