Robot Arms Learn to Play Chess Blindfolded

Lucas and Luna explore how robot arms are learning to play chess blindfolded — using only touch sensors and spatial memory to move pieces on a physical board. They dive into a 2025 project at MIT's CSAIL where a KUKA arm with force-feedback fingers played a full game against a human without cameras, relying on a 3D mental model updated by tactile data. The episode covers the sensor fusion stack, the challenge of 'muscle memory' for robots, and why this skill matters beyond games: from assembly lines where fog or dust blocks vision, to surgical robots that must operate by feel when line-of-sight is compromised. Specific numbers include the 0.2-millimeter precision required for piece recognition and the 400-gram grip force threshold for lifting a king piece. #RobotArms #BlindfoldChess #TactileSensors #ForceFeedback #MITCSAIL #KUKA #Robotics #Automation #BusinessPodcast #TechPodcast #FexingoBusiness #MachineLearning #SensorFusion #SurgicalRobots #IndustrialAutomation #HumanRobotInteraction #TouchSensing #SpatialMemory Keep every episode free: buymeacoffee.com/fexingo