Thought Cloning Insights
Jeff discusses the limitations of traditional coding in replicating human-like intelligence, particularly in dynamic situations like conversations or sports. He emphasizes the importance of a dual approach to intelligence, where systems can switch between different types of reasoning. The concept of thought cloning is introduced, highlighting how large-scale data collection from human actions can train neural networks, yet points out the gap in planning and generalization that current models face.In this clip
From this podcast
Machine Learning Street Talk (MLST)
Jeff Clune - Agent AI Needs Darwin
Related Questions