Stop AI Slop

Your AI writes fluff, jargon and walls of text. You ask a simple question and get rambling paragraphs you read twice and still feel like you don't know English.

You want clear, useful writing instead: replies you understand the first time, explanations anyone can follow. Anthropic engineer Andrej Karpathy posted the fix, and it was seen 7.8 million times. Tell your AI to write in ASD-STE100, the simple English used in airplane repair manuals so a mechanic can't misread a step. He finds it "a lot more readable." Anthropic's Claude Code team uses the same rule in its own plugin.

This book gives you 100 before-and-after examples to use on your own writing. Each one shows what the AI wrote, the clear version, and the exact line to paste into your AI to get it. Use them on your replies, emails, explanations, reports and documents.

Examples 1 to 28: the writing rules, shown on real AI answers.
Examples 29 to 50: set it up once in Claude Code, Codex or ChatGPT, pick how strict it is, and check a draft before you send it.
Examples 51 to 90: plans, READMEs, emails, reports, proposals and SOPs.
Examples 91 to 100: when to turn it off, what it can't fix, and how to check it still works when a new AI model comes out.

More than 70 books promise to make AI "sound human." Most of them have you rewrite the text yourself. This book sets your AI up to write clearly the first time, and gives you 100 examples to copy.

Everything you set up comes in a free companion GitHub repo: the one line, the full writing kit and the check, ready to copy into ChatGPT, Claude or Codex.

9 new "sound human" books come out every month, and half the shelf runs under 90 pages. This one is 194 pages built on 100 before-and-after examples. Stop AI slop today.