Lex Fridman Podcast #222 – Jay McClelland: Neural Networks and the Emergence of Cognition
61 snips
Sep 20, 2021 Jay McClelland, a cognitive scientist at Stanford, delves into the fascinating interplay of neural networks and human cognition. He discusses how these networks mimic brain functions and explores the evolutionary origins of intelligence. The conversation touches on the philosophical implications of consciousness and the transformation brought by backpropagation in machine learning. McClelland also reflects on the challenges of cognitive modeling and how simpler interactions can lead to complex emergent properties, shedding light on the nature of understanding and identity.
AI Snips
Chapters
Books
Transcript
Episode notes
Neural Networks & Thought
- Neural networks bridge biology and the mysteries of thought.
- This connection allows us to understand how physical beings can think.
Descartes' Hydraulic Theory
- Descartes believed a hydraulic system explained animal action, but human thought required divine intervention.
- He theorized that touching something sent pressure signals through the nervous system, triggering actions.
Cognition's Biological Roots
- McClelland felt that cognition is fundamentally biological, disagreeing with the prevailing view.
- He believed in studying the mind through the lens of the nervous system.
