Deconstructing Superintelligence: Identity, Self-Modification and Diff\'erance
A new paper explores the theoretical underpinnings of artificial superintelligence, focusing on the concept of self-modification. The authors propose a formal framework using an associative operator algebra to analyze how self-modification, particularly when it extends to its own supplementary mechanisms, can lead to the collapse of classical self-referential structures. This collapse is shown to mirror paradoxes like the liar paradox and concepts from philosophical theories such as Priest's inclosure schema and Derrida's diffèrance. AI
IMPACT Explores theoretical limits of AI self-modification, potentially informing future AI safety research.