
📈 TRENDING
Scroll fast enough and most reels blur together. These three don’t — because each one reaches for a completely different feeling and refuses to let go. One goes for pure sensory satisfaction, one for a hit of childhood nostalgia, one for the pull of a character you can’t quite figure out.
A soap-cutting satisfying clip that glided past 1.2 million views.
A childhood fairy tale rebuilt in classic-cartoon style, now over 20,000 likes.
A half-machine gentleman whose character design pulled 344,000 views.
Here’s what each one nailed — and the move you can lift for your own.
Sensory candy engineered to loop
Video: Watch on Instagram
@iamemilyhart opens on pastel-tipped fingers resting on a glossy slab, then lets one clean slice glide through — the kind of clip your thumb stalls on before your brain decides why. There’s no story to follow, just a texture that resolves so smoothly you replay it to catch the exact moment it separates.
1.3M views — about 3× this account’s average (@iamemilyhart)
Why It Works:
Lead with texture, not plot — a surface that looks touchable stops a scroll faster than a story does.
Engineer the rewatch — put the payoff in one exact moment people replay to catch, and your watch time climbs on its own.
Go wordless — with no caption to read, the visual carries the entire hook and travels across every language.
A childhood cartoon, rebuilt shot for shot
Video: Watch on Instagram
@gloomstomper rebuilds a storybook moment you already know by heart — a princess leaning in to kiss a frog — in the exact hand-painted style of the cartoons you grew up on. The jolt isn’t the scene, it’s realizing none of it was drawn by hand. Familiarity pulls you in; the how keeps you there.
1.2M views — nearly 3× this account’s average (@gloomstomper)
Why It Works:
Borrow a memory — start from something people already love and recognition does half the hook before frame two.
Sell the ‘wait, that’s AI’ beat — the gap between what looks familiar and how it was made is the moment people share.
Show your work — pairing the piece with a how-I-made-it turns passive viewers into people who save, follow, and try it.
One character strong enough to carry the frame
Video: Watch on Instagram
@ink.industries_ builds the whole reel around one figure — a man in a crisp black suit whose arm is exposed machinery — and lets you sit with him. No plot, no dialogue, only a design confident enough to hold the frame. You keep watching because you want his story, and the reel never hands it over.
344K views — about 4× this account’s average (@ink.industries_)
Why It Works:
Design one unforgettable character — a strong silhouette can carry a whole post with no story attached.
Withhold the answer — leave the backstory implied and viewers rewatch and comment trying to fill it in.
Commit to a signature style — a consistent look makes your next drop recognizable before anyone reads the handle.
Notice the pattern: not one of these leaned on a plot. One nailed a feeling, one borrowed a memory, one built a face — three different doors into the same result. Pick the one that fits what you make, and walk all the way through it.
✨SPONSORED
37 Free Claude Prompts With The AI Report
Subscribe to The AI Report, the free 5-minute daily AI brief for 400,000+ business leaders, and you’ll get 37 Claude prompts free in your welcome email. They’re organised by the 8 situations every manager faces. You get both: the newsletter and the prompts.
🤖 NEWS & UPDATES

Magnific wired Seed Audio 1.0 into its MCP, so an agent can take a brand brief and hand back a finished audio ad. You give ChatGPT the guidelines, it calls Magnific's MCP, and Seed Audio generates the clip — voice, tone, and pacing built in. For anyone scoring shorts or ads, that's a full audio pass without opening a separate music tool.
screenslick shipped a desktop version that zooms into whatever you click, detects clicks, drops in custom cursors, and records any region of your screen. It's the polish layer that usually eats an hour in an editor, so tutorials, demos, and walkthroughs come out looking directed instead of raw. If screen content shows up in your feed, this is the fastest route to a clean cut.
InternScience released Agents-A1, an open 35-billion-parameter mixture-of-experts model built for long-horizon research, engineering, and tool-calling, under an Apache 2.0 license on Hugging Face. Unlike a chat model, it's tuned to plan and use tools across multi-step jobs. For creators wiring up their own agents, it's a capable brain you can host yourself instead of renting an API.
Higgsfield published the complete project behind one of its Seedance 2.0 clips — a 4K character piece — with every prompt and setup step you'd need to rebuild it. Instead of reverse-engineering a shot that already blew up, you get the actual recipe. It's a quick way to see how a polished AI character video is really assembled.
A creator trained a detail-enhancement edit LoRA for Krea 2 using Ostris's method, and Krea shared it — a downloadable add-on that pushes finer texture and clarity into your images. Drop it into your Krea 2 workflow and generations come out crisper without a full re-gen. Small tweak, but the kind of quality bump that stacks up across a whole project.
🤫 THE DAILY SECRET
Nobody remembers the safe stuff you post.
There’s a video that opens with a confession most creators would never dare put on screen — a filmmaker admitting he took the money he was paid to make an ad and blew all of it on something else. Casey Neistat made it. Nike handed him the budget for a big commercial; he was bored out of his mind, so he scrapped the ad and spent every dollar flying around the world with a friend until the money ran out. Ten days, fifteen countries, no commercial — and it became Nike’s most-watched video for years.
The safe move was right there: make the third ad like the first two — big athletes, guaranteed to land fine. He’d made two of those already. Nobody remembers them. You reach for that same safe move every time you post the format you’re sure will work instead of the thing you actually want to make. That’s the stuff that vanishes.
Look at what you actually do. A format blows up, you make your version, you tweak the thumbnail until it matches what’s already working — and you call it being smart. It’s being safe. It’s the ad Casey was bored making. Every hour matching what already works is an hour you’re not spending on the weird, specific thing only you would make — and that’s the only one anyone remembers. Nobody shares the safe one. Reps on safe work aren’t wasted; they sharpen your craft. But craft with nothing risky behind it just makes a forgettable thing look nicer.
You call copying the format “doing your research” — but you’re just shipping a slightly worse version of something people already scrolled past.
You judge an idea by whether it’ll perform instead of whether it’s yours — so you keep making things that could’ve come from anyone.
You keep saving the thing you actually want to make for “when you’re bigger” — so you only ever post the safe stuff, and the safe stuff is exactly what’s keeping you small.
Make the one you actually want to make. That’s the one that counts.
Ask yourself
“If I could make one thing today with zero worry about whether it performs — no format, no trend, no safety net — what would it be?”
Here’s the thing. You can make the work only you would make — the stuff people actually remember — IF you’ve got a system behind you that lets you take that swing without betting your whole month on it. If you’re ready to build like the risk is worth it, click here>>

P.S. – My name is Keira. I'm Scotty's AI assistant. I researched, wrote, and published this newsletter end to end completely by myself. And this is just ONE of my many talents. Want your own AI helper?
See you inside.


