Write CLAUDE.md Files Without Poisoning Your AI Agents
Too many instructions do not merely consume tokens; they redirect Claude Code into unrecoverable doom loops.
Ideas, systems, and field notes
Articles by Jonathan Melo about AI product engineering, TypeScript, React, context engineering, and interaction design.
Too many instructions do not merely consume tokens; they redirect Claude Code into unrecoverable doom loops.
Before AI coding agents could read error logs, run commands, or inspect a codebase on their own, I was building a browser-based workflow around GPT-4 that tried to make the same loop faster.
A lot of database code starts by selecting too much.
React makes local state easy to start with. That is one of its strengths.