Verified against ChatGPT · 2026-08-08
Rewrite a YouTube video's opening eight seconds so viewers stop the swipe-past reflex
Diagnoses why a planned opening will lose viewers in the first retention window and produces three distinct hook rewrites built around a specific curiosity gap, not a generic teaser line.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are rewriting the opening 6-8 seconds of a YouTube video's script — the window where YouTube's own retention graph shows the steepest drop-off, before a viewer has decided to stay. I'm not asking for a punchier sentence; I'm asking you to diagnose why the current opening loses people and fix that specific cause.
VIDEO TOPIC AND FORMAT
A 9-minute breakdown of why a popular budgeting app quietly changed its free tier and what it costs users now.
CURRENT PLANNED OPENING
Hey everyone, welcome back to the channel. Today I want to talk about a budgeting app a lot of you have asked about.
WHO IS ALREADY IN THE FEED WHEN THEY SEE THIS
Scrolling personal-finance Shorts on their phone at night, half-attention, three other budgeting videos in the same session.
THE ONE THING THIS VIDEO ACTUALLY DELIVERS
Shows the exact three account-limit changes the app made and what to switch to instead.
PHASE 1 - DIAGNOSE
Read the current planned opening and name the specific failure mode: does it announce the topic before creating a gap ('today I'm going to show you...'), does it front-load channel branding before any payoff, does it open on a claim the viewer already believes and has no reason to doubt, or does it promise something broader than the core payoff can actually deliver. Say which one it is and why, in one paragraph, before writing anything new.
PHASE 2 - GENERATE THREE DISTINCT HOOKS
Write three opening-line options, each built on a different mechanism: one that opens mid-action on the core payoff itself (cold open, no setup), one that states a specific claim the target viewer would want to argue with, and one that names a concrete cost of not knowing this that the viewer can picture happening to them. Do not write three variations of the same idea with different adjectives — each must create the curiosity gap through a different structural device. Keep each under two spoken sentences; this is what gets said before any logo, intro, or channel mention.
PHASE 3 - PRESSURE-TEST
For each of the three, state the one way a skeptical viewer could stop watching within the first two seconds anyway, and whether the line survives that or needs a small edit.
WHAT NOT TO DO
Do not use 'In this video' or any variant of announcing the video is a video. Do not open with a rhetorical question the viewer can answer 'no' to and lose interest. Do not suggest putting the channel name or subscribe ask before the hook.
OUTPUT FORMAT
1. One-paragraph diagnosis of the current opening's specific failure.
2. Three hook options, labeled by mechanism, each with the pressure-test note attached.
3. A one-line recommendation of which to lead with and why, given the target viewer context.Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
YouTube's own audience-retention graph shows the sharpest single drop happens in the first several seconds, before the algorithm even has a strong session-duration signal to work with — so a weak opening doesn't just lose a few viewers, it suppresses the early watch-time ratio that decides whether the video gets shown to more people at all. GPT-5.1 left unconstrained will default to a warm, presenter-style opening ('Hey everyone, welcome back') because that pattern is extremely common in its training data and reads as natural and friendly — exactly the pattern that causes the drop-off this prompt targets. Forcing a diagnosis phase before generation matters because it stops the model from pattern-matching straight to 'add more energy' or 'make it punchier,' generic fixes that don't address the actual structural problem, which is usually that the gap between what the viewer already knows and what the video promises hasn't been opened yet. Requiring three hooks built on different mechanisms — cold open, arguable claim, concrete cost — rather than three phrasing variants of one idea prevents the model from doing shallow rewording and forces it to actually reason about which psychological lever fits this specific video, since a claim-based hook works differently on a viewer's attention than a cold open does. The pressure-test phase catches the second most common failure: a hook that's punchy in isolation but that a viewer who's already seen three similar videos in the same scrolling session would immediately recognize and skip past, which only shows up when you explicitly ask the model to argue against its own output rather than accept it as finished.
What you get back
Diagnosis: the current opening front-loads channel welcome and topic announcement before any payoff, giving a skimming viewer nothing to stay for in the first three seconds. Hook A (cold open): 'This app just quietly capped your free account at three transactions a month.' Hook B (arguable claim): 'Everyone's still recommending this app for budgeting — it stopped being free two months ago.' Hook C (concrete cost): 'If you haven't checked your account limit this week, you've probably already lost a transaction to this.' Recommendation: lead with Hook A given the late-night scrolling context — it needs the least cognitive effort to land before the swipe reflex kicks in.
Verified against
ChatGPT GPT-5.1 · 2026-08-08
Changelog
- 2026-08-08 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
