> Why do people always assume that if something is difficult to put in words it is impossible for a computer somehow? It seems like an invalid argument to me.

> Why do people always assume that if something is difficult to put in words it is impossible for a computer somehow? It seems like an invalid argument to me. Because if we don't really know how we are doing something, how can you tell the computer how to do it? Simply measuring a bunch of values doesn't mean it will be interpreted correctly. For eg, see the lie-detector.