Markus J. Buehler’s talk unveils emerging approaches toward superintelligent discovery engines: systems that blend reinforcement learning, graph reasoning, physics-informed models, and generative AI to accelerate breakthroughs in materials, engineering, and biology.