> In this run, the initial random weights tended to output highly positive or highly negative action outputs. This makes most of the actions output the maximum

> In this run, the initial random weights tended to output highly positive or highly negative action outputs. This makes most of the actions output the maximum or minimum acceleration possible. It’s really easy to spin super fast: just output high magnitude forces at every joint. Once the robot gets going, it’s hard to deviate from this policy in a meaningful way - to deviate, you have to take several exploration steps to stop the rampant spinning. It’s certainly possible, but in this run, it didn’t happen. This is extremely human. Once you're deeply committed to something, it's hard to imagine alternatives, never mind embrace them.