mischlevi

Introduction Most discussions of AI tuning and the properties of consciousness revolve around a "bottom-up" approach: everyone expects the complexity of computation (mind) to ultimately give rise to a subject (consciousness). I propose considering a "top-down" hypothesis: A machine cannot possibly surpass the level of consciousness of its creator. 1....

Summary: I propose a top-down hypothesis: an AI system cannot transcend the level of awareness of its creator. Current architectures focus on discrete "attention," which is a property of the narrow mind. To achieve true understanding, we must shift toward an architecture of "periphery" — a state where the world...