A practical map through the Claude Code course: what remains essential, what to treat as terminal-specific, and what exercises make the course useful for economics research.
The course was made for a broad audience, so it spends real time on terminal setup and a deployed app. Economists should use it as a map of the core primitives: files, context, permissions, skills, MCPs, plugins, subagents, and Git.
I now usually point economists toward the Codex desktop app first, but the Claude Code course still explains the concepts behind agentic work very clearly. Use this guide to separate the durable habits from the app-builder examples.
Open the full course on YouTube →Claude Code is not just a chatbot and not only a coding tool. For economists, the useful mental model is a local research assistant that can read files, run commands, and carry out multi-step work in your actual project folder.
The YouTube-summary exercise is a good low-risk first move. Translate it to economics: summarize a seminar, extract references from a talk, or turn a transcript into a short research memo.
Watch the install path if you plan to use Claude Code in the terminal. The website example is not the economics use case, but the lesson transfers: installed skills can package repeatable workflows.
This is the operational core. Economists do not need to become terminal people, but they do need to understand paths, tool approvals, context windows, compaction, voice input, and permission boundaries.
The course explains why giant CLAUDE.md files can backfire. For research projects, keep instructions short: folder layout, commands to run, files never to edit, and how outputs should be checked.
This is the highest-leverage section for research work. Slow down on skills first, then learn enough about MCPs, plugins, and subagents to know which concept fits a given economics workflow.
The job tracker is not the destination for economists. Watch for the method: plan a project, put work in Git, review diffs, run locally, deploy only when the output actually needs to become software.
Use Claude Code with yt-dlp or a transcript file. Ask for a one-page memo with claims, citations to check, open questions, and follow-up readings.
Write a skill that tells Claude Code how your projects are organized, how to run your scripts, where outputs go, and how to compare new tables against prior results.
Try a Claude Connector, MCP, or Codex plugin for email, calendar, Drive, or Notion. Keep the task narrow: seminar prep, referee-report triage, or coauthor follow-up.
In one old research folder, write only the rules that matter: the folder map, commands that verify outputs, sensitive files to avoid, and the review loop before accepting edits.
The slide deck above is the full public deck I used for the course. The timestamped plan on this page is the economist-specific layer: where to slow down, where to skim, and what exercises to do after watching.