“OpenAI's "Weak to Strong" generalization concept is a method for controlling future superintelligent AI that is smarter than its human creators.”