I think this also suggests interesting directions for studying the limits of open-ended cognition in humans & AI -- ad-hoc model synthesis is a hard problem, so how is it possible at all? (cont.) x.com/T_BrookeWilson…

One of the most exciting papers I've read in a while - highly recommended!🧠🤖 It gives a compelling account of how the human mind reasons so flexibly, and of what's missing from LLM reasoning

How do people reason so flexibly about new problems, bringing to bear globally-relevant knowledge while staying locally-consistent? Can we engineer a system that can synthesize bespoke world models (expressed as probabilistic programs) on-the-fly?

Modeling Open-World Cognition as On-Demand Synthesis of Probabilistic Models

Excited to read this.

Huge congrats to Raphael and Oxford — what a place to do philosophy of AI!

Come join us at Yale!

we think humanity’s biggest challenges won’t be solved by ai thinking for 1000 hours coming back with an answer they’ll be solved by many collaborating humans, and ai that understands them and their different skills, goals, values, etc to empower them to do more together

Seems very cool

How do people flexibly integrate visual & textual information to draw mental inferences about agents we've never met? In a new paper led by Lance Ying, we introduce a cognitive model that achieves this by synthesizing rational agent models on-the-fly--presented at EMNLP 2025!