Comment by yters

If it is so easy to fool deep learning, why is it so hyped? Seems a great security risk.