Maybe 10-20 years ago genetic algorithms were hyped. Mutation and mating were thought to be good methods for optimizing towards some goal. What happened to the field? Was it out-competed by fitter learning methods such as gradient descent?