Recognizing a plastic bag is tech on par with two of the techs you list (and arguably beyond sending a man to the moon.) Making a robot that can recognize plas

Recognizing a plastic bag is tech on par with two of the techs you list (and arguably beyond sending a man to the moon.) Making a robot that can recognize plastic bags and manipulate them is harder. And applying that level of expertise to a public works problem in a capitalist society is even harder. (In response to the obvious counterargument; the configuration space of "a plastic bag mixed in with a bunch of random junk" is actually probably larger than the configuration space for a road with a dozen other cars on it, and so is the action space of a reasonably dextrous claw/arm greater than that of a car. AI is full of easy-sounding things that are really difficult.)