My guides to AI coding agents

Where to run agentic AI — a quick tour. A short overview of the surfaces — VS Code, terminal, desktop, web, mobile — to help you pick where to start.

Claude Code in VS Code. A hands-on guide to setting up and using Claude Code inside VS Code for empirical research.

Codex in VS Code. The same workflow, but built around OpenAI's Codex CLI.

Codex vs Claude Code — a practical comparison. Side-by-side notes on how the two agents differ in everyday research use.

Linking VS Code and Overleaf. Sync your LaTeX papers through GitHub or Dropbox so you can edit locally with AI tools while coauthors stay on Overleaf.

My curated tips

I share a brief personal guide to doing research on Substack, and I keep a deck of presentation slides with tips for writing and doing research.

People whose resource pages I've learned from:

Doing research

Paul Niehaus on the transition from student to independent scholar — researchers often experience a dramatic loss of structure when the dissertation ends.

Writing skills

Editor Amitabh Chandra notes that clear and accessible writing, including an explanation of where the paper breaks down, is the strongest predictor of acceptance. Empirical work by Jan Feld, Corinna Lines, and Libby Ross shows writing quality significantly affects publication outcomes.

Concentration

Cal Newport's Deep Work and Study Hacks blog, on attention management in the age of email and social media.

Staying current

Subscribe to NBER, IZA, and RePEc working-paper alerts, plus journal notifications. Finance researchers: the Insights for Young Researchers in Finance newsletter.

Data visualization

Kieran Healy's Data Visualization, Andrew Heiss's course, and Paul Goldsmith-Pinkham's best-figures page.

Workflow

Organize data construction, graphs, and results so you can return to a project six months later without frustration. Asjad Naqvi's Stata workflow generalizes well. Jeppe Druedahl has good Python and coding notes.

Talk to junior faculty

Not just about topics. Ask about practical research matters — how they organize projects, handle referees, choose journals, and split time. Their perspective is often more actionable than senior faculty's.

AI, LLMs, Claude Code & Cursor

Claude Code skills

Writing

Books

Guides and articles

Publishing, refereeing, discussing

Workflow, graphs, tables

General workflow

Tables

Graphs

Tip. Install the lean scheme package for clean Stata graphs.

Presentations

Tip. Have fellow PhD students take notes during your own practice talks. Solicit feedback from colleagues beforehand.

Productivity & finding ideas

ChatGPT & large language models

Being a researcher

Getting an overview of the literature

Networking

#EconTwitter / #EconSky

Coding & website building

Computational resources

Stress & being a grad student

Job market

Teaching

Have a teaching resource that should be here? Email claes.backman@gmail.com.