The pre-machine-learning formulations of AI focused on symbolic reasoning through the dual problems of search and logic. Many problems can be reduced to enumerating legal steps, and SAT/SMT/ASP and related systems can churn through those in a highly optimized and genetic manner.