Reinforcement Learning Supercharges Spot's Speed
Boston Dynamics' Spot just got way faster—thanks to reinforcement learning. Researchers at the Robotics and AI Institute used reality-grounded AI techniques to triple Spot’s running speed, pushing its performance to new levels. The secret? Direct control over its joints via a low-level API. And that’s not all—they’re also working on an autonomous bicycle that can jump. Cutting-edge robotics just took another leap forward.