Why do we laugh, and can we teach this to computers? In this talk, we dive into the humor theory and the history of computational humor. We take a look at how computers wrote jokes 30 years ago, and why even large language models like ChatGPT are still struggling to come up with good jokes. We then explain how to optimally use prompt engineering to make language models like GPT more helpful in a wide range of tasks, and showcase these techniques for generating higher quality jokes in any humoristic style.