關於這一集
Modern AI can perform astonishing tasks, but does making a system work mean that we understand intelligence?
In this episode of Creative Difference, Maxime Gabella speaks with Aude Billard, professor of robotics at EPFL, about robot learning and the scientific mystery of how humans acquire physical skills so efficiently.
Starting from their shared background in physics, they explore why apparently simple actions such as grasping an object, manipulating watch components, or performing microsurgery remain extremely difficult for robots. Their conversation examines the tension between mathematical models and black-box neural networks, the limits of data-driven robotics, and the importance of closing the loop between perception and physical action.
They also discuss learning by demonstration, transferring skills between different robot bodies, active inference, humanoids, detachable crawling hands, shape-shifting machines, and the possibility that future robots may develop senses and world models very different from our own.
The episode concludes with a broader reflection on scientific understanding, publication overload, education, creativity, and why we may need to slow down to understand what we are creating.
