[joe_the_user]

Any of these neural networks is going to have the same basic qualities - lack of explainability and unreliability (even if they have many good qualities, are "fairly" reliable etc, they do fail and we don't exactly when/why they fail).