Comment by froh

uh --- which implications does this have for driving AI? Possibly none, if the acceptance criterion for AI is: "it kills fewer people than humans would. possibly others, but fewer overall."

?