No. 15Content & blogging10 min read

Editing what the model wrote

10 August 2026

Readers can now identify AI-written prose in about two sentences, and the recognition costs you their trust for the rest of the page. The tells are consistent and mostly structural. This is what they are, why the model produces them, and the editing passes that remove them.

The short version

→The tells are structural more than lexical. Uniform paragraph length, tricolons everywhere, every section the same size, and a conclusion that summarises the thing you just read.
→The deepest one: relentless hedging. Models are trained toward inoffensiveness, which produces prose with no position, and prose with no position is unmemorable.
→Delete the introduction. Almost always. It restates the title and delays the reader by four sentences.
→Replace generalities with specifics wherever you can. 'Many companies find' → 'we found, across 340 accounts'. This is the same edit as adding first-hand material and it does more than any other single pass.
→Read it aloud. Every unnatural sentence surfaces immediately and none of them survive being spoken.

Subscription

The Library

Everything running a software business asks of you and none of it is code: onboarding that activates, churn you can actually see coming, marketing with no budget, SEO that finds buyers, and the AI-plus-Ahrefs workflow for writing the blog that feeds it. One new playbook a week, plus the whole archive from day one.

$9/ month · cancel any time

✓A new playbook every week — onboarding, churn, growth, SEO, content
✓The full back catalogue the moment you join, not drip-fed
✓Vendor explainers: Stripe, Supabase, Resend, Vercel, the AI APIs
✓The research shelf — real papers, decoded, on learning and on LLMs
✓One email when a new one lands, off by default in one click
✓Cancel any time; you keep reading to the end of the month

first — it takes ten seconds.

No card is asked for yet — the subscription is recorded and really opens your account. See all plans →