Humans choose more diverse goals than LLMs, potion-game study with 175 people finds
ValerioCapraro · x · 2026-11-06
A new paper by Valerio Capraro and colleagues compares 175 humans with five LLMs in a game where players mix potions and discover recipes, with a "goal" defined as making a specific potion by combining ingredients in the correct order. Results reveal substantial differences in goal selection: humans explore more broadly and pick more diverse goals, while models converge narrowly—empirical evidence for human-AI differences in goal diversity.
Related event: LLMs Pick Far Narrower Goals Than Humans, Study Finds(2 posts)→
More from Research
- Optimistic Hedge regret is O(n log n) for 2 players but exponential for 3, new paper shows — roydanroy · 2026-11-07
- Exo: Berkeley's agent runtime that edits its own harness, cutting model cost 35-61% — dair_ai · 2026-11-06
- Symbolic belief revision plus geometric projections lifts RAG reranking to 0.481 nDCG@10 — letandrewcook · 2026-11-06
- TMS-based architecture models how claims retract as evidence shifts — letandrewcook · 2026-11-06
- Charter: A Lean 4 DSL for Formally Verified Rules Governing AI Agents — QuixiAI · 2026-11-06
- ICLR 2027 sees 42,790 submissions; 57% of reviews are a 2, 2-2-3 most common — CSProfKGD · 2026-11-06