My Code Agent Experience
My Code Agent Experience
This page is a personal guide to using AI code agents as a junior economist — not from the perspective of a seasoned researcher with deep domain expertise, but from someone learning both economics and AI tooling at the same time.
My goal is to share what actually works (and what doesn’t) when integrating code agents into an economics research workflow, based on my own hands-on experience.
Two resources that have helped me a great deal are Claude Blattman and the claude-howto guide on GitHub. Both are written primarily with Claude in mind, but in practice I have found they translate very naturally to Codex as well.
In fact, I switched between Claude and Codex often. Currently, I am on Codex. I find it to be more efficient and understanding of my current workflow.
A note upfront: all of the content here is produced from my own input with the help of code agents.
My setup when I share this:
- Primarily Mac user, but sometimes use Windows
- ChatGPT Pro plan · GitHub Copilot · VS Code
- I use code agents for a variety of tasks by breaking it down as described in Sample Workflow and Learning New Knowledge.
- My work towards “allocative efficiency” of tokens are summarized in Saving Tokens in My Process and What Works vs. What Doesn’t Work.