Most of the post (~80%) is actually about decision trees: principles, how to train, choosing splits etc.