The Myth of Model Interpretability

Deep networks are widely regarded as black boxes. But are they truly uninterpretable in any way that logistic regression is not?
https://datafloq.com/read/the-myth-of-model-interpretability/1077