This podcast interviews Jeff Hawkins and Viviane Clay about Hawkins' "Thousand Brains Theory of neocortex," a neuroscience theory informing a new AI technology. The discussion covers the theory's foundations in Vernon Mountcastle's work on cortical columns, the role of reference frames in representing knowledge, and the differences between this approach and current deep learning AI. Hawkins and Clay explain how their "Thousand Brains Project" uses sensory-motor learning and a modular architecture to create AI that learns efficiently from limited data, unlike current data-hungry deep learning models. A key takeaway is that this new AI approach aims to build systems capable of independent problem-solving and adaptation in the real world, potentially leading to more robust and versatile AI applications.
