

#115 – Dileep George: Brain-Inspired AI
Aug 14, 2020
Dileep George is a pioneering researcher in neuroscience and AI, co-founding Vicarious and previously Numenta. He shares fascinating insights on building brain-inspired AI, discussing the complexities of visual perception and the organization of the brain's structures. The conversation dives into Recursive Cortical Networks and how they differ from traditional neural networks, as well as the implications for memory and cognition. George addresses the hype around brain-inspired AI and the future of human-AI interaction, offering a thought-provoking look at consciousness and learning.
AI Snips
Chapters
Books
Transcript
Episode notes
Building a Brain
- Building a brain requires understanding its functional level.
- Simulating details without a functional theory is insufficient.
Blue Brain Project
- Lex Fridman mentions the Blue Brain Project, which attempts to simulate the brain by detailed biophysical neuron modeling.
- Dileep George notes this approach often yields oscillations but lacks debuggability without a theoretical framework.
Neuroscience and Computation
- Dileep George highlights the limitations of purely observational neuroscience.
- He emphasizes the need to integrate findings into a computational framework for understanding information processing.