Is agentic AI governance even a computationally bounded process?
Wrt to context drifting, goal misalignment, etc. Is it possible that a Turing machine could, in theory, handle all of the known issues wrt governance? Or is it a case where (say) 90% of the issues could be handled by a strict governance process, but th…