Niklas Todenhöfer comments on A Longlist of Theories of Impact for Interpretability