Patrick Leask comments on Why I’m Working On Model Agnostic Interpretability