Guillaume Corlouer comments on Rethinking the Value of Working on AI Safety