This is cool. It seems like the robot is responding to the mushroom's activity rather than the mushroom controlling the robot with intent.

I was hoping the mushroom learned to control the robot in pursuit of some goal. Maybe that's where this research leads.