OLogic CEO to present on teaching robots via first-person human demonstrations
OLogic CEO Ted Larson will lead a session at RoboBusiness on egocentric robot learning, where robots acquire skills from human demonstrations captured via wearable cameras and teleoperation. The talk will cover technical challenges and the role of embodied AI and large-scale learning in commercial robotics. The event takes place October 20-21, 2026, in Santa Clara, California.
Egocentric learning relies on first-person video and teleoperation data to train robots, an approach championed by major AI labs and robotics startups. OLogic, an embedded systems firm with over 25 years of robotics experience, has worked on projects for clients including Google, Hasbro, and Panasonic, building platforms like AMRs and social robots.
Larson's session at RoboBusiness, scheduled for the second day at 10:45 a.m., will address technical hurdles in multimodal data collection and large-scale learning. The conference, held October 20-21 in Santa Clara, targets commercial robotics developers across manufacturing, logistics, and healthcare sectors, offering networking and discounted registration for academic and corporate groups.
This approach could lower deployment barriers for small and mid-sized businesses, which often lack robotics expertise. If robots learn from ordinary human demonstrations rather than specialized programming, workforce training may shift toward demonstrating tasks rather than coding them. However, this could also raise questions about data privacy from wearable cameras and the reliability of learned behaviors in unpredictable environments. Companies adopting these systems may gain competitive advantages, while workers might need new skills to supervise and verify robot performance.