The Jim Rutt Show cover image

EP 192 David Krakauer on Science, Complexity and AI

The Jim Rutt Show

NOTE

The Uncanny Valley of Complexity

Understanding complex phenomenology requires a first principles deductive framework to establish certainty. In high dimensions, models reveal regularities that shed light on the nature of complexity. Technologies like backprop, developed in the late 80s, work efficiently on very large models due to the miracle of ultra high dimensionality allowing gradient descent to navigate. The issue of the uncanny valley may be connected to the challenge of simplifying complex models, as seen in the shift from multi-parameter models to concise approaches in neural networks. The mystery of how deep learning models achieve their outcomes parallels the enigmatic nature of mathematics' invention, with even leading experts admitting a lack of full understanding.

00:00
Transcript
Play full episode

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
App store bannerPlay store banner