I3L is an integrated hardware and learning system for whole-body mobile manipulation that pairs a low-cost, open-source correction interface with iterative policy refinement. For an operator, the change is to stop relying on offline teleoperated demonstrations alone and instead run repeated deployment-and-retraining rounds in which you take over the robot to correct the current policy's failures.