📈 TRENDING
Somebody dressed the Mona Lisa in a studded leather jacket and left everything else exactly where Leonardo put it; all three of today’s outliers borrow an era, and the era does the expensive part. Pick a century and it has already settled your light, your clothes and your streets before you make a single decision of your own.
A snowy nineteenth-century street seen from an upper window pulled 2.3 million views.
The Mona Lisa restyled as a modern fashion model took 83,600 views.
A traveller alone on a salt flat with a suitcase over his head pulled 952,000 views.
Three borrowed centuries, three sets of rules nobody had to invent. Here is what to take from each.
A whole town lit one window at a time
@pskjii_art opens on the street rather than the gown the title promises. It is night, and every window in the row is burning, so a wide shot with no one in it reads as a town full of people you cannot see. The snow softens every edge that would otherwise repeat.
📈 2,300,000 views — 25.5× the account’s median reel (@pskjii_art)
Why It Works:
Light the windows before you light the street. A row of small warm rectangles suggests a populated town far more cheaply than putting a crowd in the road.
Let the period buy your detail. Chimney pots, a carriage and street lamps arrive as a set, so one decision about the century furnishes the entire frame.
Let the snow carry the texture. Falling flakes and loaded roofs break up every hard line, which is why a frame packed with buildings never reads as flat or repetitive.
The Mona Lisa keeps the pose and changes everything else
@matsky1 leaves Leonardo’s composition untouched — the same three-quarter turn, the same folded hands, the same hazy blue-green valley behind her — and swaps only what she is wearing. Studs and tattoos read as a costume change on a set the viewer already knows, which is why it lands in a second.
📈 83,600 views — 5.2× the account’s median reel (@matsky1)
Why It Works:
Change one layer and leave the rest alone. The pose, the hands and the landscape stay exactly as painted, so every bit of attention falls on the wardrobe.
Work from a picture your viewer has already memorised. The comparison happens in their head, so you never have to put the original on screen beside yours.
Keep one thread of the original showing. The gold sleeve under the leather is the seam between the two centuries, and it stops the jacket looking dropped on top.
One empty plain and one impossible load
@kelly_boesch_ai_art gives you a single figure on a salt flat with almost nothing to read: a rust-orange suit, a case held overhead, and a cluster of balloons roped to a wooden cart. The emptiness is the move. Take the background away and one odd object has to carry the shot.
📈 952,000 views — 3.0× the account’s median reel (@kelly_boesch_ai_art)
Why It Works:
Empty the frame before you add the strange thing. A flat horizon and a plain grey sky give the balloons nothing to compete with, so they read as deliberate.
Give your subject something to do with its arms. A case held overhead turns a person standing in a field into a person mid-journey, and it costs you one prop.
Let the decade pick the props. A wooden cart, a hard-sided case and a wide-leg suit all come from the same years, so the scene holds together without a location.
Not one of them invented a world. They picked a century and let it make the small decisions, which is the cheapest move on this list and the one we pull apart post by post inside the community.
✨SPONSORED
Analytics on Live Data Without Leaving Postgres
When analytics on Postgres slows down, most teams add a second database. Then they manage pipelines, sync lag, and drift forever. TimescaleDB extends Postgres instead. Analytics run on live data, in the database you already have. No pipeline. No migration. No new query language.
🤖 NEWS & UPDATES
Tripo launched a plugin that runs inside ChatGPT, so a typed prompt comes back as an editable 3D asset without you leaving the chat window. Tripo points it at four end uses — gameplay, animation, commercials and 3D printing — with GPT-6 Astra handling the prompt side and Tripo doing the geometry. Worth a look if you have wanted 3D in your pipeline but never wanted to learn a modelling tool.
Runway calls its Fall 2026 collection the biggest model drop it has shipped, adding Fish Audio S2.1 Pro, MiniMax H3 Max, Cartesia Sonic 3.6 and a Flux video upscaler and edit pass, all sitting alongside Runway's own models. The pitch is that you stop hopping between vendors: generate, voice and upscale in one place. Useful if your current edit spans three tabs and two subscriptions.
Krea showed Krea Agent generating several camera angles from one source video, and published the prompt next to the examples so you can run it yourself. It is the video-side version of something that has mostly lived on stills — one take, then coverage you never actually shot. Handy if you have a single usable clip and need a second angle to cut against.
fal added xAI's Grok Voice to its platform, putting a speech model that answers in about 0.70 seconds inside whatever you are already building there. It covers text to speech in 30+ voices across 25+ languages, transcription with word-level timestamps, and voice cloning from roughly two minutes of audio. Useful if you are dubbing shorts or need one consistent narrator across a series.
Cline made Union Alpha — a stealth multimodal model built for agentic coding — free inside the editor, with a 256k context window and, by Cline's own numbers, performance near GPT-6 Astra and Opus 5 at roughly an eighteenth of the expected cost. The free access is a preview rather than a permanent tier. Worth grabbing if you build your own small tools between shoots.
🤫 THE DAILY SECRET
The accounts you’re copying have never been worth coming after.
Recently Scotty talked about a guy on one of his calls who wanted to build a whole channel around superhero characters — skits, in-character posts, the works. The angle was great. The characters weren’t his to use. Plenty of people are making Homelander videos right now, nothing has happened to a single one of them, and Scotty’s read was that those people are going to get hit eventually.
That’s a huge chunk of the AI category right now. Anime fusions, movie characters, faces you recognize — the stuff that pops on the first upload pops because a studio already spent decades making you care about it. Nobody’s said a word to you either, and you’ve been reading that silence as permission.
His advice wasn’t stop. It was make spin-offs of your own characters instead. Copying what works is the whole game — he says it on nearly every call and he’s right. Copy the edit, copy the pacing, copy the joke, copy the exact beat that made you stop scrolling. The one thing you can’t copy is the guy in the suit. It only gets taken once it’s worth taking, which is why nothing has happened to anybody you’re looking at. None of them has made anything worth the trouble yet.
“Everyone’s doing it.” — your entire safety check is a headcount, and every head in it is exactly as small as you are.
“They’d have to notice me first.” — correct, and that means the plan is to win at this and stay invisible at the same time.
“I’ll swap in my own stuff once this takes off.” — by then the thing that took off is the part that was never yours.
Put a character you made up in the next one. It’ll probably do worse. That’s what the one you get to keep costs.
Ask yourself
“If I had to hand back everything in my videos that I didn’t make up, what’s left — and what would it take to make the next one entirely mine?”
Here’s the thing. You can go as big as you want in this category — IF you’re around creators who make up their own characters instead of running the same ones everybody’s already running. If you’re ready to grow something that stays yours, 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.