RoboHub🤖(@XRoboHub):More robotics teams are getting behind the same idea: let robots learn straight from people. Sunday’s ACT-1 learned long-horizon household tasks from human demonstrations. X Square’s TwinDEX used a few hundred robot-free episodes to train a policy that powered a 24-step chemistry experiment. Now @RewardAI_’s OM-1 learns manipulation from human demonstrations alone. The same policy runs across industrial arms and humanoids, handling contact-rich tasks that require coordination, force control and recovery when things go wrong. Reward says OM-1 can pick up a new task from less than 30 minutes of human demonstration data.

2026.09.15 11:49

More robotics teams are getting behind the same idea: let robots learn straight from people. Sunday’s ACT-1 learned long-horizon household tasks from human demonstrations. X Square’s TwinDEX used a few hundred robot-free episodes to train a policy that powered a 24-step chemistry experiment. Now @RewardAI_’s OM-1 learns manipulation from human demonstrations alone. The same policy runs across industrial arms and humanoids, handling contact-rich tasks that require coordination, force control and recovery when things go wrong. Reward says OM-1 can pick up a new task from less than 30 minutes of human demonstration data.

显示更多

Reward AI@RewardAI_

2026.09.14 17:40

Introducing OM-1, our first robot foundation model, zero-shot generalizing to any robot: table-top arms, industrial arms and humanoids. - learned directly from human manipulation data - no teleop/robot data - close to human-level dexterity and efficiency - multi-robot collab

显示更多