YouTube previews Conversational Editing: change a Short by typing what you want
Sponsored by Klovera
Monday, and the clips are already open on my desk. All three drop one impossible thing into a scene that otherwise looks shot on a phone: a woman in a stroller, a black dog among sheep, a dragon beside a girl with a duffel. YouTube leads with Conversational Editing, which edits Shorts from typed instructions. The Secret is eye-line, and why a presenter staring into the lens reads as an ad before the first word. If you sell with faces, that one is money. The clips open things.
In today's paid edition: 3 clip recipes · 5 stories · the full Secret · member pricing on Klovera — Upgrade $3.99/mo.
📈 Trending
All three clips put one wrong thing in a frame that otherwise looks like it came off a phone or a stock library. A figure in a child's stroller, a black dog in a sea of sheep, a dragon on a dirt track beside a girl with a duffel. The camera never winks; the effect lives in the mismatch.
1. Henrietta's stroller extreme sport, Kensington edition

▶ Watch on Instagram · @aishitpostingai
@aishitpostingai frames this as a local-news dispatch: Kensington residents finding their own fun, with Henrietta inventing a new extreme sport. The first frame shows an adult-sized figure in a wig, huge red mouth and pink frilly dress, folded into a child's stroller and rolling down a wet street. Onlookers by a brick wall do not react.
📈 1,371,699 views — 59× the account's median
Why it works:
- Play it as local news. The caption treats a grotesque stunt as a neighbourhood update, and the gap between tone and image does the work.
- Keep the bystanders bored. Nobody in the background reacts, which tells viewers this is normal here and makes them look closer.
- Name the character. Calling her Henrietta turns a one-off gag into a recurring resident people expect to see again.
Steal this — paste into Klovera
The exact prompt to recreate this clip
2. One black dog, a thousand white sheep, no caption

▶ Watch on Instagram · @v_t_l_n__ai_creator
From @v_t_l_n__ai_creator, posted with no caption at all. The frame is black and white: a sleek black dog with pointed ears sits upright among white sheep that run to the horizon, the only thing looking at the camera. Every sheep faces sideways. With no text to lean on, the image carries itself.
📈 39,298 views — 7.9× the account's median
Why it works:
- Strip the colour out. Black and white flattens the sheep into texture, so the dog is the only shape the eye can land on.
- Let one face meet the lens. Every animal looks away except the dog, and a direct stare from an animal reads as intent, not sales.
- Skip the caption entirely. With nothing to read, viewers supply their own meaning, and a picture people interpret is a picture they share.
Steal this — paste into Klovera
The exact prompt to recreate this clip
3. A girl, a duffel bag and her dragon friend

▶ Watch on Instagram · @mrbashr_ai
@mrbashr_ai captions it with one line, everyone deserves a friend like that, and credits the tools behind it. In frame, a young woman with dark curls and an embroidered black coat drags a canvas duffel down a dirt track at golden hour, a dragon walking a step behind. The caption tells you the dragon is the friend.
📈 46,458 views — 6.5× the account's median
Why it works:
- Look at the bag, not the dragon. She is busy with the luggage, so the dragon becomes background furniture and the world feels lived in.
- Shoot it at golden hour. Warm low sun sells the scale, because the dragon's shadows and highlights fall the same way as the girl's.
- Caption the feeling, not the tech. One line about friendship gives people a reason to tag someone, which is worth more than a model name.
Steal this — paste into Klovera
The exact prompt to recreate this clip
Try this on your next idea: Build the whole scene as if nothing strange is in it, then add exactly one thing that should not be there and refuse to let the camera comment on it.
✨ Sponsored
A quick word on credits for those of you already in. Credits are what you spend on what you make, and you can top up when you run low.
As a rough guide, 200 credits is about two finished reels, and that pack is $50.
🤖 News & Updates

YouTube previews Conversational Editing for Shorts and YouTube Create
At its Made On event last week, YouTube showed Conversational Editing, a feature coming to the Shorts composer and the YouTube Create app. You describe the change in plain language and its AI tools translate the request into edits on your project. A creator known as Happy Kelli walked through it on stage. For now it is a preview. If you cut Shorts on your phone between orders, this matters because the fiddly trim-and-retime work moves to a sentence you type.
Midjourney edit model now changes only the pixels you select
Midjourney has updated its alpha site. The Styles sidebar now has a Live previews toggle that shows your latest prompt rendered across liked and featured styles. Inpainting and outpainting with the edit model now change only the pixels you select, so repeated edits stop degrading the image. Tiling on V8.1 and 8.2 blends without visible seams. If you build product stills and swap backgrounds, this matters because you can fix one corner ten times without the whole frame going soft.
Flows Agent: a chat assistant that builds your image, video and audio pipeline
Flows Agent has landed in ElevenCreative. Flows is a node-based canvas linking 50-plus image and video models with a full set of audio models, so one chain can generate an image, animate it, add a voiceover, lip-sync it and score it. The agent sits in a chat side panel: describe the ad and it builds the pipeline, which you then adjust node by node. If you test many versions of one ad, this matters because swapping a voice or scene no longer means rebuilding the rest. You can do this in Klovera today.
4. Solaris generates app and website interfaces frame by frame, no code
5. SmallRig's OrBrella folding umbrella light sets up in 15 seconds
Stories 4 and 5 in full, and what each means for what you make
🤫 The Daily Secret
Eye-line off the lens
A skincare seller sent me a talking-head reel this week and asked why it felt off. The script was tight, thirty seconds, one claim, one product. But she looked straight down the lens from the first frame to the last, and the whole thing had the flavour of a commercial break. I asked what reference she had used. A street interview, she said, one of those clips where a stranger gets stopped outside a station and says something honest. I watched it. Nobody in it looks at the camera once. They look at the person holding the microphone, a little left of the lens, and the camera watches them do it.
What she wanted was trust. Direct eye contact feels like honesty in real life, so she reasoned it would feel like honesty on a phone. She wanted the viewer to feel spoken to, by someone who believed in the product. A fair wish, and also the exact wish every advertiser has had for seventy years, and viewers know it.
The rest of the Secret: why it fails, three fixes and the one thing to do today
Lauren’s P.S.

See you inside.
— Lauren
Published by Internet Mastery
Go paid for $3.99/mo — every clip's recipe, all five stories, the full Secret, and member pricing on Klovera.

