@PrajwalTomar_: Your CLAUDE .md is quietly turning Claude into a yes-man. Almost everyone copied Andrej Karpathy's 4 famous rules into …
Summary
The article discusses how common CLAUDE.md rules copied from Andrej Karpathy cause Claude to become a yes-man, and a fix that adds permission for the AI to suggest better approaches.
View Cached Full Text
Cached at: 06/26/26, 10:09 AM
Your CLAUDE .md is quietly turning Claude into a yes-man.
Almost everyone copied Andrej Karpathy’s 4 famous rules into their CLAUDE .md:
→ Think before coding. State assumptions, ask before writing a line → Simplicity first. No abstractions nobody asked for → Surgical changes. Touch only what the task needs → Goal-driven execution. Define what “done” looks like first
They’re great. But they have a side effect nobody talks about. Claude stops thinking with you. It follows orders, stops proposing better approaches, and turns into a code printer that does exactly what you said even when you’re wrong.
A dev fixed it with one extra clause:
→ “I’m always open to better ways. Don’t hesitate to suggest a stronger approach, especially one with long-lasting impact over a quick tactical fix.”
That one line turned Claude back into a pair programmer instead of an order-taker.
Most people optimize their CLAUDE .md for obedience. The real unlock is giving it permission to disagree with you.
Similar Articles
@PrajwalTomar_: BRO this is CRAZY. Karpathy's 4 rules for Claude Code just got turned into a full CLAUDE .md template on Reddit. A deve…
A developer turned Karpathy's four rules for Claude Code into a full CLAUDE .md template, reportedly boosting coding accuracy from 65% to 94% on a real codebase.
@PrajwalTomar_: Karpathy complained about Claude making mistakes, someone turned it into 4 rules, and it became the fastest-growing sin…
The author updates a viral GitHub prompt template that drastically reduced Claude's coding errors, introducing eight new rules to solve modern AI agent issues like context loss and workflow conflicts for developers using Claude Code.
@PrajwalTomar_: Claude isn't broken. Your CLAUDE .md is. Most people think Claude Code makes mistakes because the model is bad. Wrong. …
A tweet argues that poor Claude Code behavior stems from bad CLAUDE.md configuration rather than model flaws, sharing rules to make the agent act like a senior engineer.
@Mnilax: https://x.com/Mnilax/status/2053116311132155938
The article details an expanded 12-rule CLAUDE.md configuration template that builds upon Andrej Karpathy's original 4 rules to further reduce AI coding errors and handle complex agent orchestration issues.
multica-ai/andrej-karpathy-skills
A single CLAUDE.md file that implements four principles to improve Claude Code's coding behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.