How do you make "solving cancer" a verifiable task for RL?
iScienceLuvr · x · 2026-11-22
An open question making the rounds: if reinforcement learning thrives on verifiable rewards, how would you define "solving cancer" as a machine-checkable task? It probes the limits of the RLVR paradigm — math and code have crisp judges, while biomedical validation is slow, costly, and hard to automate, making closed-loop wet-lab feedback a core AI-for-science challenge.
More from AGI Musings
- Polymarket traders put 82% odds AI solves Riemann Hypothesis in 2026 — Polymarket · 2026-11-22
- Robotics veterans clash: general models are the key to the last 6% problem — BradPorter_ · 2026-11-22
- Adam Dorr: AI went from failing grade-school math to solving the Riemann Hypothesis in two years — jamesdouma · 2026-11-22
- Pausing AI Won't Work: Cost Per Solved Task Will Plummet for Years — teortaxesTex · 2026-11-22
- 1998 Red Herring piece on VC bubble resurfaces as AI-era investors say 'history rhymes' — MartinGTobias · 2026-11-22
- Pause AI won't work: cost per solved task will keep plummeting for years, argues teortaxesTex — teortaxesTex · 2026-11-22