AprilSR answers Any real toeholds for making practical decisions regarding AI safety?