Hacker News new | ask | show | jobs
by sixo 135 days ago
Ideally it would learn a relationship between the sensor input and the correct actions, even if the sensor input is not realistic for the GG in reality.