Prospects for Hybrid AI

Classical AI research is oriented towards the performance capabilities of a program-controlled computer, which, according to Church’s thesis, is in principle equivalent to a Turing machine. According to Moore’s Law, gigantic computing and storage capacities have been achieved, which made AI performance possible in the first place. But the performance of supercomputers have a price that can be equivalent to the energy of a small town. Human brains are all the more impressive, that can compare the performance of a computer (e.g. speaking and understanding a natural language) with the energy consumption of a light bulb. At the latest, one is impressed by the efficiency of neuromorphic systems, that have emerged in evolution. Is there a common principle underlying these evolutionary systems that we can make use of in AI.