@servasyy_ai: Karpathy's CLAUDE.md topped GitHub Trending. 220k stars, but most developers haven't read it. It's only 65 lines. Yet it boosts AI coding accuracy from 65% to 94%. The 4 rules inside: → Think before you…

X AI KOLs Timeline Tools
claude coding developer-productivity ai-coding github-trending karpathy

Summary

Karpathy's CLAUDE.md file tops GitHub Trending with 220k stars. Just 65 lines of rules boost AI coding accuracy from 65% to 94%. Four core rules: Think before coding, Simplicity first, Surgical modifications, Goal-driven execution.

Karpathy's CLAUDE.md topped GitHub Trending. 220k stars, but most developers haven't read it. It's only 65 lines. Yet it boosts AI coding accuracy from 65% to 94%. The 4 rules inside: → Think before coding: State your assumptions. Ask if uncertain. Never guess. → Simplicity first: Write only the minimum code to solve the problem. Don't add abstractions no one asked for. → Surgical modifications: Don't touch code unrelated to the requirements. Every line changed must be traceable back to a requirement. → Goal-driven execution: Before writing any code, turn vague instructions into verifiable success criteria. That's it. 65 lines. 4 rules. 94% accuracy. Bookmark it before others see it.
Original Article
View Cached Full Text

Cached at: 05/22/26, 06:06 AM

Karpathy’s CLAUDE.md hit #1 on GitHub Trending. 220k stars, but most developers haven’t read it yet. It’s just 65 lines. Yet it improved AI coding accuracy from 65% to 94%. The 4 rules inside: → Think before coding. State your assumptions. Ask if unsure. Never guess. → Simplicity first. Write only the minimum code to solve the problem. Don’t add abstractions no one asked for. → Surgical modifications. Don’t touch code unrelated to the requirements. Every line changed must trace back to a requirement. → Goal-driven execution. Before writing any code, turn vague instructions into verifiable success criteria. That’s it. 65 lines. 4 rules. 94% accuracy. Bookmark it before others see it.

self.dll (@seelffff): karpathy’s CLAUDE.md hit #1 on github trending. 220,000 stars. most devs still haven’t read it.

it’s 65 lines. it took AI coding accuracy from 65% to 94%.

the 4 rules inside:

→ think before coding state your assumptions. ask when unsure. never guess.

→ simplicity first

Similar Articles

RTK reports token savings, but our cost benchmarks disagree

Hacker News Top

The article evaluates RTK, a popular tool for compressing terminal output to reduce token usage in AI coding, and presents cost benchmarks that challenge its savings claims, showing mixed results with slight cost changes and lower pass rates in tests.

Claude is no longer available for minors

Hacker News Top

Claude has implemented age assurance measures, restricting access to users over 18 and using Yoti for verification to detect and block minors.

Accordio

Product Hunt

Accordio AI launches admin tools for Claude, including time tracking, invoicing, and contract signing via a free MCP connector.