Scaling Laws
Scaling human experience improves robot manipulation, with action-grounded glove data delivering especially strong gains.
60 Tasks It Was Never Trained On.
Liber-0 reached 36.0% zero-shot success across 60 manipulation tasks with no task-specific robot demonstrations.
OUR TECHNOLOGY
Technology Overview
We ground human experience in action and align it to the robot.
Zero-Shot Generalization
Transfer skills from human experience to new tasks, without task-specific robot demonstrations.
Application Workflows
Adapt pretrained manipulation skills to real-world workflows, demonstrated through autonomous boba tea preparation.
