Jim Buhler comments on Rethinking the Value of Working on AI Safety