Is AI really a substitute for humans? What better way to answer that question than having AIs play a Dungeons & Dragons dungeon master? D&D follows an intricate, yet ambiguous set of rules, requires constant creativity and unexpected imagination, empathy, and playfulness. Coté has developed a sort of Turing test that uses a handful of D&D scenarios to test how human AIs can act: things like how it DMs a goblin ambush, role plays generic tavern encounters, and creates interesting open-ended adventures and world-building. More than benchmarking, this is an excellent way to learn how to create prompts, to understand what AIs are, and pretend like you’re working which you’re having fun playing games. It's also a good way to explore what the basics and best practices of creating and maintaining AI-based apps. To go beyond the limits of public generative AI models, this talk will build the needed generative AI apps on Cloud Foundry.