Did anyone else notice the similarity of Mitchell’s arguments in this post, and the one in his comment to one of my posts? Here he says that there is no color in a purely physical description of a mind, and in his comment to my post he said that there is no utility function in a purely physical description of a mind.
I think this argument actually works better here (with color), because my counter-argument to his comment doesn’t work. What I said was that in principle we know how to go from a utility function to a physical description of an object (by creating an AI) and so in principle we also know how to go from a physical description to a utility function.
But here, we don’t know how to go from a color to a physical description of a mind that can experience that color, nor can we tell what color a mind is experiencing or capable of experiencing, given a physical description of it, even in principle.
Did anyone else notice the similarity of Mitchell’s arguments in this post, and the one in his comment to one of my posts? Here he says that there is no color in a purely physical description of a mind, and in his comment to my post he said that there is no utility function in a purely physical description of a mind.
I think this argument actually works better here (with color), because my counter-argument to his comment doesn’t work. What I said was that in principle we know how to go from a utility function to a physical description of an object (by creating an AI) and so in principle we also know how to go from a physical description to a utility function.
But here, we don’t know how to go from a color to a physical description of a mind that can experience that color, nor can we tell what color a mind is experiencing or capable of experiencing, given a physical description of it, even in principle.