From Imitation Learning to RL: How Your Data Strategy Changes
Imitation learning training data and reinforcement learning data are not the same thing. Most robotics teams discover this the hard way when they try to transition from one paradigm to the other using the same dataset. Imitation learning and RL: two robotics training data philosophies Imitation learning and reinforcement learning are not just different training […]
Humanoid Robot Training Data: How Much Do You Actually Need?
How much humanoid robot training data do you actually need? The honest answer depends on three things: your deployment tier, your model architecture, and what “enough” means for your specific use case. The humanoid robot training data question every team asks differently Ask five humanoid robotics teams how much training data they need and you […]
Lessons from 50,000 humanoid trajectories
Five things we learned getting from zero to 50K trajectories across Figure, Optimus, and Apollo. Operator pacing matters more than skill, failure-recovery data is what actually moves production performance, and annotation cost is the dominant cost — not capture.