[YeGoblynQueenne]

>> This seems to be a redux of the 1x play (which was panned across the internet after the WSJ review).

For a moment I thought it was the same company.

>> This 'environment reset' problem is at the core of RL - there are no solutions for this yet, only workarounds.

That, and generalisation. A year or so back Aloha made huge waves but teleoperation doesn't solve the fundamental limitation of RL, that you have to train anew for every new environment or task.