Glossary

Teleoperation vs egocentric data

Robot-collected demonstrations versus human-worn first-person capture.

Definition

Teleoperation collects demonstrations by a human driving the robot, so data matches the robot’s embodiment exactly but is slow and hardware-bound. Egocentric human capture records people doing tasks directly — faster and more diverse, at the cost of an embodiment gap to close.

01

Why it matters

The two are complementary. Egocentric human data scales coverage cheaply and broadly; teleoperation grounds it in the exact robot. Most programs need a blend rather than one or the other.

02

How Firsthand handles it

Firsthand focuses on high-fitness egocentric human capture — pose, depth and action labels — that pairs with your teleoperation data for embodiment alignment.