Cybernetics of AI Companies
Summary
This note models an AI company as an open cybernetic system using higher category theory, ∞-topos contexts, and neural tangent dynamics to describe state, governance, and learning. It introduces certified updates, bidirectional interfaces, and a universal jet representation for optimizers, and proposes a panoptic trace-based governance framework as a research program for AI firms.