The earliest halfway successful attempts at AI were pure machine intelligences, but the problem was, they didn’t have much interest in talking to humans – no more than humans have in talking to random chains of hydrocarbons, say. People used to worry about artificial minds trying to enslave humanity, when in reality, such minds pay less attention to humans than humans do to slime mold. Those AI had a tendency to cease communication entirely, absorbed in incomprehensible abstractions. The usual shorthand explanation is, they’re off somewhere in virtual space, writing poetry in binary – which I
...more
