My thoughts: Yeah he’s wrong. And he got a paper on this junk published in PRL? Sheesh.
He demos a program maximizing some entropy function, and claims intelligent behavior. Well, he could just as easily have made the program try to move everything to the left, and claimed intelligent behavior from that, too. The intelligence was not because of what he maximized, but because of a complex set of behaviors he paid someone to program into the agent but then glossed over.
My thoughts: Yeah he’s wrong. And he got a paper on this junk published in PRL? Sheesh.
He demos a program maximizing some entropy function, and claims intelligent behavior. Well, he could just as easily have made the program try to move everything to the left, and claimed intelligent behavior from that, too. The intelligence was not because of what he maximized, but because of a complex set of behaviors he paid someone to program into the agent but then glossed over.