I volunteer myself as such a machine, though I won't say I can perfectly predict the actions of another human and many others are better than I am. Nevertheless I find I can make pretty good estimates from time to time depending on information.

I'm curious about your "I like to believe" stance. Are you saying you'll ignore any evidence for any theory if you don't think the theory benefits you more than another theory, even if your theory has been shown to be wrong?