Dec 31, 2025

The Algorithm That IS The Scientific Method [Dr. Jeff Beck]

Dr. Jeff Beck argues that the future of AI lies not in scaling up large language models, but in building systems that mirror the brain's approach to understanding the world. He posits that true intelligence is grounded in physics and object-centered models, not language. Beck's vision involves creating AI composed of numerous small, modular models—much like a video game engine—that can be dynamically combined, updated through continual learning, and understand the world through causal relationships and forces. This approach, he claims, will solve key challenges in generalization, robotics, and alignment by enabling machines to 'know what they don't know' and reason about the physical world in a way that is fundamentally similar to humans.

Sep 10, 2025

Karl Friston - Why Intelligence Can't Get Too Large (Goldilocks principle)

Professor Karl Friston provides a 20-year retrospective on the Free Energy Principle, exploring its implications for life, intelligence, and consciousness. The discussion delves into the nature of agency in "strange things" that can model their own future, the crucial distinction between intelligence and consciousness, the potential for a "Goldilocks zone" for intelligent systems, and the profound challenges of building conscious AI, which may require a move beyond current computer architectures toward "mortal computation".