I’m going to be doing a deep dive on a single paper in econometrics related to a technique called difference in differences with continuous treatments or “dosage”. Think of the minimum wage which isn’t just “do you raise it” but also “what is it”. Sometimes it’s one number in one place but a different one in a different place, which means it’s a dosage — a community takes a different minimum wage dosage. But I’m using Claude code to help break this paper down, and I start with just the two way fixed effects decomposition. I think we start with a standard regression model, understand it’s bias, then if you’re still interested you stop by for lunch to learn the new estimator. Today we focus on four skills I have in my mixtapetools repo: beautiful_deck (to make beautiful presentations), split-pdf, Tikz and referee2. You can read them in action. And the deck is at the end!