tl;dr: Uber have not the slightest idea what they're doing, and should not be permitted to develop anything safety-critical until they can properly explain 'obj

tl;dr: Uber have not the slightest idea what they're doing, and should not be permitted to develop anything safety-critical until they can properly explain 'object permanence'. Also, would probably be a good idea to hold the creators, trainers and marketers of anything called 'AI' legally responsible (jointly and severally) for all of such a system's decisions for the next few decades. Might encourage some much-needed caution.