In this talk, Chelsea Finn explores the cutting-edge developments in physical intelligence and robotics, detailing how her team is moving beyond rigid, pre-programmed systems toward adaptable, learning-driven robots. The discussion highlights the shift from controlled lab settings to real-world environments, emphasizing the importance of scalable models and diverse data sources. Through trial, error, and innovation, the team is teaching robots to understand and interact with the physical world in increasingly autonomous and intelligent ways.