Software, robotics, machine intelligence, and spatial computing have collided to present us with one of the greatest UX challenges of our time: How do you design an interface for a humanoid robot that can theoretically do anything? This talk will dive into the design challenges of teaching a robot about the world in three parts: Training robots to develop skills, building application interfaces to author task behaviors, and integrating with facilities to monitor robots’ effectiveness and realize the vision of the software-defined factory.