After a year of pairing with LLMs on infrastructure code, I’ve landed on a setup that mostly stops the AI from over-engineering everything. Two pieces make it work: Anthropic’s Claude Code CLI and a tiny set of behavioral rules from Andrej Karpathy that I now keep in my global config. Here’s how I install both, and the philosophy behind why this combination beats Copilot-style autocomplete for anything bigger than a one-liner.<\/p>\n\n\n\n
Claude Code is Anthropic’s official terminal CLI for their Claude models. Unlike a browser chat, it reads and edits files in your project, runs shell commands, watches test output, and stays grounded in the actual codebase. The plugin system layers skills<\/em> on top \u2014 bite-sized behavioral rules that activate when they match the task at hand. The right plugin set turns Claude from a chatty autocomplete into something closer to a junior engineer who has read the handbook.<\/p>\n\n\n\n Karpathy posted a short set of rules for collaborating with LLMs on code. They’re four sentences each, and they shut down almost every common failure mode (over-engineering, scope creep, premature abstractions, confidently-wrong refactors). Here they are verbatim:<\/p>\n\n\n\n\ud83d\udccb Prerequisites<\/h2>\n\n\n\n
\n
cd<\/code> in and run it<\/li>\n<\/ul>\n\n\n\n
\n\n\n\n\ud83e\udded The Karpathy Guidelines \u2014 Expanded<\/h2>\n\n\n\n