davidhartsough comments on Counterarguments to the basic AI risk case