Armchair Expert with Dax Shepard Max Bennett (on the history of intelligence)
94 snips
Oct 2, 2024 Max Bennett, a researcher and author renowned for his insights into intelligence evolution, discusses his unique path into neuroscience without traditional academia. He shares fascinating ideas about reinforcement learning and how curiosity may shape artificial intelligence. Bennett explains the paper clip conundrum, the significance of emotions in both humans and AI, and the evolutionary history of language. The conversation intertwines personal anecdotes with deep explorations into the cognitive similarities and differences between humans and other species.
AI Snips
Chapters
Books
Transcript
Episode notes
First Brain and Navigation
- The first brain emerged to facilitate locomotion and steering, using a taxis navigation algorithm.
- Nematodes demonstrate this by crossing a copper barrier based on food concentration and hunger, showcasing early decision-making.
Binary Origins
- Early organisms' simple detection of 'more' or 'less' food inadvertently established the concept of good and bad.
- This binary wiring, seen in nematodes and Roombas' dirt detection, underlies our attraction to binary thinking.
Reinforcement Learning
- Reinforcement learning, where actions are strengthened by rewards, was key to vertebrate development.
- Marvin Minsky's AI maze navigation highlighted the temporal credit assignment problem, later solved by Richard Sutton's actor-critic model.
