The problem of common-sense reasoning has plagued the field of artificial intelligence for over 50 years. Now a new approach, borrowing from two disparate lines of thinking, has made important progress.