AI coding agents

Coding agents are only as good as the setup around them. These essays cover choosing between tools, giving agents the right context, and supervising several at once without handing over your judgment.

Essays

Questions

What is an AI coding agent?

An AI coding agent is a model that can read files, run commands and make changes in a codebase on your behalf, such as Claude Code or Codex, rather than only answering questions in a chat.

How do you give a coding agent the right context?

Keep durable instructions in a project file such as CLAUDE.md, and load only the files, examples and state the current task needs. The agent acts on what is in its context window at inference time.

How many AI agents should one person run at once?

Only as many as you can still review with real attention. Each parallel agent adds cognitive load, and approvals given without attention quietly hand over your judgment.