New Preprint Synthesizes Symbolic Models of Human and Machine Reasoning on ARC at Scale
burny_tech · x · 2026-11-10
A new preprint, 'Synthesizing symbolic models of human and machine reasoning at scale,' introduces Behavior-Conditioned Symbolic Synthesis (BCSS). LLM-based program synthesis has mostly been applied to classical psychology tasks (learning, memory, decision-making); the authors generalize it to open-ended reasoning using ARC, where humans excel and machines have been improving. By studying incorrect inferences from both humans and machines, BCSS builds scaled symbolic cognitive models offering process-level understanding of behavior, beyond task-level accuracy. Team includes Brenden Lake and Todd Gureckis.
More from AGI Musings
- Leetcode interviews will die as agentic coding rounds become the new hiring standard — rachittshah · 2026-11-10
- EEG vs implanted BCIs: researchers clash over physical limits and AI decoding potential — burny_tech · 2026-11-10
- Math notation is just poor communication — don't let gatekeeping cap your technical ceiling — TivadarDanka · 2026-11-10
- "Every Internet-Era Methodology Is Wrong Today": A Take on AI-Native Minds — oran_ge · 2026-11-10
- Mistral CEO warns two or three dominant AI providers would squeeze their own customers — ylecun · 2026-11-10
- Remote labour index on track to saturate by April 2027, says Ben Todd, based on real Upwork tasks — ben_j_todd · 2026-11-10