Dr. Jeff Beck discussed the nature of intelligence and agency, proposing that complexity and sophisticated internal computations, rather than a simple input-output mapping, define intelligent systems. The conversation explored energy-based models and Yann LeCun's Joint Embedding Prediction Architecture (JEPA) as potential advancements in AI representation learning. Beck also offered a grounded perspective on AI safety, focusing on the risk of human AI
Summary written by None from 1 source. How we write summaries →
