In this episode of Hidden Layers: Decoded, Ron Green, Dr. ZZ Si, and Michael Wharton explore the latest AI breakthroughs, including DeepSeek’s R1 model, Meta’s work on intuitive physics, and Stanford’s S1 model. They discuss the rise of cost-effective reinforcement learning, diffusion-based language models, and DeepMind’s advances in geometry-solving AI. The team also dives into AI-driven biology with Evo2 and the emergence of civilizations in a Minecraft simulation. Throughout, they reflect on the future of AI, from domain-specific models to the impact of world models on business and science.
Hidden Layers: AI and the People Behind It