Maze robots show why big goals fail when you aim at them, per Stanley & Lehman
ankurhandos · x · 2026-11-16
A shared excerpt from Kenneth Stanley & Joel Lehman's Why Greatness Cannot Be Planned:
- Big goals are rarely reached by aiming at them: maze robots rewarded for getting closer to the exit kept walking into dead ends next to it — what the authors call deception.
- When stuck, chase novelty: robots rewarded only for reaching somewhere no robot had been (novelty search) found the exit far more often than those chasing it.
- The takeaway for personal plans: write down the number you track and ask whether hitting it moves you toward the thing you want, or only toward the number.
More from AGI Musings
- Rutger Bregman: AI denial is the left's new climate denial — and more dangerous — marigo · 2026-11-16
- Healthcare AI's bleeding edge is 'on par with dinosaur SaaS like Workday' — RishabJainK · 2026-11-16
- Will human intellectual life survive when AI does the frontier thinking? — zetalyrae · 2026-11-16
- Dev: more worried about AI builders justifying harm 'for the greater good' than AI doom — Dan_Jeffries1 · 2026-11-16
- Old world: write code, then test. New world: build the verifier, then let agents grind — Yuchenj_UW · 2026-11-16
- Physical AI hits the factory floor: reliability, safety standards and energy per task decide scale — ingliguori · 2026-11-16