Stop writing prompts like this
@aiweekly.raj · 268K followers · 8d old · 77s · Sample snapshot, 2 Nov
1.2M views / 95.9K creator median (14 posts), computed 2 Nov
The numbers
- Views
- 1.2M
- Creator median
- 95.9K
- Views per hour
- 5.7K
- Likes
- 114K
- Saves
- 11.6K
- Shares
- 7K
No early snapshots yet: the first-72h curve appears once we have tracked this video.
Why it worked: the evidence
- Views 12x the creator's median of 95.9K (14 posts).
- No engagement rate stood out against this creator's usual: the reach likely came from distribution, not one standout reaction.
Other rates vs usual: Like 1.4x, Comment 1.0x, Share 1.0x, Save 1.0x
AI breakdown
The hook
0:00Stop writing prompts like this.
It is a direct command that tells the viewer they may be making a mistake. It names a specific task (writing prompts) and does not say what 'like this' means, so people stay to find out.
Why it worked
1. The video got far more reach than the creator's normal posts, and the caption and opening line use the same 'stop doing X' hook.
1.2M views vs a 95.9K median across 14 posts, which is 12.1x.
0:00“Stop writing prompts like this.”
2. Likes were above the creator's usual, so viewers who watched liked it more than they normally do. This is the only engagement rate that moved.
Likes 9.80% vs 6.96% usual (1.4x). Comments, shares and saves all sit at 1.0x of normal.
3. Comments show curiosity and a wish for more. Viewers ask how it was done, ask for a part 2, and say they will save it.
Top comments: 'wait, how did you do this?' (100 likes), 'part 2 please' (85), 'saving this for later' (70). 2 of 5 top comments ask a question. Saves were only at the usual 1.0%, so the 'saving' comment is not backed by a higher save rate.
Structure
- 0:00Hook: A one-line command: stop writing prompts 'like this'. The transcript gives no more content after this.
Keep (the mechanism)
- Open with a short 'Stop doing X' line that names one specific task your audience does often.
- Leave 'like this' open so the viewer has to keep watching to see the bad example.
- Keep the caption the same as the spoken hook so the message is the same before and after play.
Change (make it yours)
- Swap prompts for a task from your own niche that your audience does and might be doing wrong.
- Write your own bad example and better version from your own experience.
- Do not rely on claims tied to this creator or to AI tools, such as the '2026' question in the comments.
AI analysis by claude-sonnet-5-5, 8 Nov. Quotes checked against the transcript. This is sample data, not a real video. The transcript is two short lines and says it is not a real recording, so the body of the video is unknown. Only likes are above the creator's norm. The reasons for the 12.1x view lift are inferred from the hook and comments, not shown by the data.
Originality check
Signals we can see on the cover and caption. TikTok judges originality on the full video and account, so this is a starting point, not a verdict.
No originality signals on the cover and caption.
Make your version original: your own voice, your own research or examples, your own footage or edits.
The hook
Stop writing prompts like this
Template: numbers and amounts become blanks (rule-based, not AI).
Transcript
SampleStop writing prompts like this. Sample transcript, not a real recording.
Top comments
Sample- wait, how did you do this?100
- part 2 please85
- saving this for later70
- is this still working in 2026?55
- this changed my week40
2 of 5 comments ask a question: demand for a follow-up.
When this creator's hits land
Darker = higher multiplier at that hour (UTC). Creator level, from their recent posts.
Creator: @aiweekly.raj
- Followers
- 268K
- Creator median
- 95.9K
- Outlier rate
- 17% of 30 posts at 2x+