And 1967 was 58 years ago, which was when the first deep neural network was trained with stochastic gradient descent. Yet, DNNs didn't take off until the 2010s when the hardware became powerful enough and data became plenty enough to successfully train and utilize them such that they were practical.