Viliam comments on Debunking the myth of safe AI