The problem is we don't know all the inputs, and there is much hidden state that we can't observe. So it isn't very useful to try and model a human after how a computer works.