Matthew Barnett comments on One Way to Think About ML Transparency