Gradient Descent, Stochastic Optimization, and Other Tales