Garrett Baker comments on A framing for interpretability