Michaels comments on Rethinking the Value of Working on AI Safety