Thoughtforms Life Conversation Between Richard Watson, Iain McGilchrist, and Michael Levin
4 snips
Feb 20, 2023 
Richard Watson, futurist studying evolution, learning and machine intelligence, and Iain McGilchrist, neuropsychiatrist probing brain hemisphere differences, join Michael Levin for a wide-ranging conversation. They explore whether evolution can be intelligent, how brains balance induction and deduction, where biological form and memory reside, and how simple feedback scales into goal-directed anatomy. Short, speculative, and provocative.
AI Snips
Chapters
Books
Transcript
Episode notes
Learning Beats Blind Selection
- Learning systems can perform smarter optimization than blind natural selection by implementing inductive biases.
- Richard Watson argues physical systems with adaptive connections can learn beyond Darwinian random variation and selection.
Induction And Deduction Work Together
- Induction forms general rules from examples and enables novel responses, while deduction applies those rules to new cases.
- Iain McGilchrist emphasizes both hemispheres use induction and deduction but play different roles in checking and asserting conclusions.
Generation Precedes Selection
- Models must be generated, not just selected; generation needs open-ended, biased schemes that produce useful candidates.
- Michael Levin links this to basal cognition: embryos build flexible models from prompts and solve problems from scratch each time.
