# Hook Anti-Patterns

Patterns to **never propose** as candidate hooks. If any of these match a candidate you have drafted, rewrite or replace before presenting to the user.

The anti-patterns evolve. AI tells in particular shift roughly yearly as model defaults change.

---

## 1. "In today's fast-paced world..."

Family includes "In an era of...", "Now more than ever", "À l'heure du tout-numérique", "À l'ère de l'IA", "Dans un monde où...". Zero information, fully generic, immediately disqualifies the writer. Cut entirely.

## 2. "Have you ever wondered...?" / "Vous êtes-vous déjà demandé...?"

The reader's honest answer is "no, why should I?". A question hook works only when the reader was already implicitly carrying the question. Bare "have you wondered" formulations assume curiosity that has not been activated.

## 3. The dictionary-definition opener (unsubverted)

Works only when immediately subverted (Tim Urban) or ironic. Played straight equals school essay. If your candidate starts "Productivity, defined as..." or "Selon le Larousse...", rewrite.

## 4. Throat-clearing meta openers

"In this article, I'll discuss...", "Today we'll look at...", "Dans cet article, nous allons voir...", "Avant de commencer, il est important de préciser que...". This is process documentation, not writing. Substitute the actual promise or scene for the agenda.

## 5. Generic statistics without source

"90% of people...", "Studies show...", "Les études montrent...", "8 personnes sur 10...". Either cite the specific study (year, sample size, source) or cut the claim. Vague stats signal made-up content.

## 6. The platitude quote

"As Einstein said...", "Comme disait Sénèque...". Almost always misattributed. The quoted maxims are already cognitively cached, so they trigger fluency without surprise. If you must open with a quote, use a specific one from a verifiable source plus your own disagreement or extension.

## 7. The apologetic preamble

"I'm not an expert, but...", "Je ne suis pas spécialiste mais...". Destroys authority pre-emptively. If the reader has to discount you in sentence one, why finish the article?

## 8. The rhetorical-question chain

Three questions in a row signals "I had no idea how to start." One question can work if it is the right question; three is a confession of confusion.

## 9. The "imagine" cliché

"Imagine waking up...", "Imaginez que...". Works only when the scene is vivid and specific (Tim Urban's "Imagine taking a time machine back to 1750"). The bare verb "imagine" without concrete sensory detail is now a near-universal AI tell.

## 10. LinkedIn-bro patterns dropped into articles

"Hot take:", "Unpopular opinion:", "Most people won't admit this, but...", "Voici la vérité que personne ne veut entendre...". These are social-media throat-clearing. In article position they read as desperate. They are increasingly anti-signals.

## 11. The corporate self-introduction

"At [Company], we believe...", "Chez [Entreprise], nous pensons...". Marketing-page voice in article position. Move company self-positioning to a footer or sidebar; do not let it eat the hook.

## 12. The vague time anchor

"Recently,...", "Récemment,...", "These past few years...". Either name the specific date, week, or event, or skip the time framing entirely. Vagueness in the opening signals vagueness in the thinking.

## 13. "You're not alone"

Empathy theater. The phrase is now associated with corporate-wellness content. Replace with the specific shared scene or specific shared frustration.

## 14. The double-meta

"Writing about writing is hard. But here's why I'm doing it anyway...". Self-referential Substack opening. Once charming, now exhausted. Just start.

---

## Current AI tells (rolling list, update yearly)

These phrases or patterns currently signal "this was written by or with an LLM with no humanizing pass". They are especially deadly in opening sentences because they prime the reader to read everything else as machine output.

- "In today's fast-paced world" (and any variation)
- "It's not just X, it's Y" (the formula construction, not the meaning)
- "Picture this:"
- "Imagine a world where..."
- "What if I told you..."
- "Whether you're a seasoned X or a curious newcomer..."
- "In the realm of..."
- "Navigating the landscape of..."
- "Unlock the power of..."
- "Dive into..."
- "Buckle up,"
- "Let's dive in"
- "Here's the kicker:"
- "At its core,"
- "It's worth noting that"
- "That said,"
- Em dash overuse in body prose (one per article is fine; one per paragraph is a tell)
- Tricolons ("X, Y, and Z") where the three items are near-synonyms
- "Crucially," "Notably," "Importantly," "Essentially," used as sentence openers
- French equivalents: "Dans un monde en constante évolution", "À l'ère du numérique", "Plongez dans...", "Découvrez comment...", "Il est crucial de...", "Notamment,...", "Par ailleurs,..." (→ See `samber/cc-skills@humaniseur-fr` skill for the canonical French AI-tell catalog)

Run candidate hooks through this list before presenting. If a candidate contains any of these, rewrite it.

If the user has a `humanizer` skill installed, defer to it for a final pass after the user picks a hook — for French copy, use `samber/cc-skills@humaniseur-fr`.