Super Data Science: ML & AI Podcast with Jon Krohn 969: The Laws of Thought: The Math of Minds and Machines, with Prof. Tom Griffiths
81 snips
Feb 24, 2026 
Tom Griffiths, Princeton professor bridging psychology and computer science, explores mathematical models of thought and AI. He discusses probabilistic versus symbolic approaches, how autoregressive training shapes LLM behavior, engineering inductive biases and meta-learning, and modeling curiosity as information-seeking. Short, clear dives into building and evaluating minds and machines.
AI Snips
Chapters
Books
Transcript
Episode notes
Mathematical Laws For The Mind
- The Laws of Thought propose mathematical principles for describing internal mental processes analogous to laws of nature for the external world.
- Tom Griffiths traces that idea from 19th-century pioneers to modern cognitive science and AI frameworks.
Mind As An Intuitive Data Scientist
- Griffiths frames the mind as an 'intuitive data scientist' that forms hypotheses from presented data.
- Lab experiments present controlled inputs (like gamble choices) to infer the mind's inductive rules and heuristics.
10,000 Decision Problems Revealed Human Patterns
- Griffiths' lab collected over 10,000 decision problems online to study human choice, far beyond traditional undergrad samples.
- They then used neural networks and modern ML to analyze patterns and derive new behavioral theories.
