To continue with my previous post, I have thought of the practical upper limit for the strength of an AI. Mind you, this is not a practical definition in all its meanings and its definetely not without its sortiment of other problems, but I like it and at least its recursive.

The strongest possible AI is a one that can build another AI stronger than itself. How do we figure when this is the case? We ask them! Its non of our concern to wonder about the inner lives of immensely complex systems. Of course its probable that the AI that built the stronger AI uses this very definition to determine whether its AI is stronger…

This definition is practical in the meaning that a stronger AI that can build an even stronger AI is really not that important for us. The realization of an AI is fundamentally a programming problem, once this is worked out, AI’s take care of the Singularity. Easy! Simple!