Rethinking AI Evaluation

Marcus challenges traditional AI evaluation methods, emphasizing the need for a deeper understanding of human cognition. Gary argues for the importance of symbolic manipulation in AI, suggesting current deep learning lacks critical capabilities.