The document describes experiences with an active learning course in artificial intelligence that used a semester-long group project approach. The project involved students developing an agent capable of playing chess, with the goal of winning an end-of-course tournament for extra credit. Feedback highlighted both benefits like improved programming skills, as well as challenges including low agent performance, unequal work distribution between group members, and stress from unclear project specifications. Lessons learned emphasized the need for group meeting time, backup plans, and relating theory to code.