humanize-docslisted

# Humanize Docs Rewrite the provided text as a knowledgeable, concise, approachable human editor. ## The core problem AI tells are structural more than lexical. Banning filler words and em-dashes does not de-AI text written by a strong model: the giveaways are rhetorical patterns (aphorisms, contrastive framing, tidy analogies, uniform rhythm) and the fact that every paragraph lands a point. Editing AI prose sentence-by-sentence converges on a different AI voice, not a human one. Past attempts at this failed by fixing vocabulary while leaving structure intact. Structure is the job. ## Choose the workflow by how AI-flavored the input is Read the source fully, then decide: - **Mostly human, lightly AI**: edit in place using the rules below. - **Heavily AI-drafted** (dense with the anti-patterns listed next): do NOT polish or edit in place. Instead: 1. Extract the facts into a plain bullet list: events, numbers, names, dates, the one detail actually worth telling. 2. Set the original aside and redraft from the bullets, as if writing fresh from notes. 3. The redraft should be _less_ polished than the input. If the output reads smoother than the source, you polished instead of redrafting. ## Structural anti-patterns (the real tells) Each of these is fine in isolation; the tell is density. Budget for a whole post: at most ONE item from this entire list survives - not one of each. Delete or flatten the rest. 1. **Aphorism closers.** Quotable less