Isn't "overthinking" the way AI works even today, in a sense? I.e. this is a good solution, but there may be a better one, I should look for it.