Field Notes
Field notes.
Working notes from the classroom, the field, and the research pipeline. Most of what shows up on LinkedIn lives here first.
-
Teaming with an AI Agent. We both were necessary, but not sufficient.
I built an interactive map of a century of Senate voting with an AI agent. It wrote the code. I caught the numbers that were plausible and wrong. Neither of us saw that it still needed explaining until a network economist told us so.
-
Learning in a different order: the judgment work no longer has to wait.
For generations, learning ran one way: master the basics, prove you can execute, and only much later earn the judgment work of deciding what to make and whether it is any good. Generative AI does the execution now, so the judgment no longer has to wait.
-
The hammer didn't build the shed.
Why we forget our own grammar when the tool is AI, what we lose by refusing to say what we actually did, and why doing AI-assisted work honestly is its own skill.
-
Orchestrating work: when AI agents become the players, students must become conductors (and composers).
Every orchestra has three roles: players, composer, conductor. Until recently we trained students to be players. Generative AI is changing which role they need to inhabit.