James_Miller comments on Notes on the Safety in Artificial Intelligence conference