My reaction to the Black Friday deals: OMG.

My next move: open Dreamina, then CapCut PC. 🎬

I’m creating fresh visuals in Dreamina and turning them into a trend-inspired edit in CapCut PC. Dreamina Basic costs $0.29 for the first month, and CapCut Desktop’s Black Friday sale % off Pro and 40% off Ultra.

A little surprise, a little inspiration, and one more video ready to post.

Dreamina + CapCut offer: tinyurl.com/5xbtv4x6
CapCut Desktop sale: tinyurl.com/2v23rn22

#CapCutDidThat #Dreamina #Smallbusinessmarketing #CapCutSale

made a fight scene on a misty mountain and honestly the pole-vault double kick got me 😭

Created with Seedance 2.5 on @FishCreativeHQ

Prompt:
16:9 widescreen, cinematic live-action, 15 seconds, high-momentum parkour fight in a continuous-feeling sequence. Hard cuts only lves, morphs or transitions; motion and direction match across every cut. No dialogue; only natural footsteps, fabric, splashes and impact sounds.

PROTAGONIST: @image 1 (facial identity, hair, rose-pink streaks, long side braid, plum shirt and cream cargo trousers locked to the reference throughout). Her movement is sharp, decisive and fluid. She dominates the entire fight and takes down ten masked attackers in a rapid chain.

OPPONENTS: Ten masked "Ashen Guard" fighters in charcoal hooded rain coats and slate face wraps, armed with bamboo staffs and short batons. They close in from above, from the sides and from behind. Every one is hit cleanly and reacts visibly: knocked back, off-balance, or thrown into a wall, rack or railing before leaving frame. The next attacker steps in immediately.

SETTING: An abandoned terraced tea plantation on a steep misty mountainside just after rain. Mossy stone terrace walls, narrow wet stone staircases, a shallow water-filled irrigation channel running down the slope, a collapsed wooden footbridge with a gap, rusted tea-drying racks, leaning bamboo scaffolding and a thin waterfall at the cliff edge. Low golden sun cuts in from the side through light haze. Palette: moss green, wet slate grey, warm amber, deep plum. Kicked-up water spray, mud flecks, torn tea leaves and mist react to every impact. Realistic natural lighting only.

SHOT 1 (0-2.5s) Low-angle, ground level, FOV 35 degrees. She bursts up from the bottom of the stone terrace, launches off a step and drives a flying front kick into attacker 1's chest, throwing him back into a stone wall. She lands and slides sideways under attacker 2's staff swing, then snaps a rising elbow to his chin; he drops flat into the wet tea bushes.

SHOT 2 (2.5-5s) Tracking from the side, FOV 40 degrees. She plants her foot on the mossy terrace wall, runs three steps up it and delivers a diagonal kick to attacker 3, who loses balance and tumbles down the staircase. Without pause she uses a leaning bamboo pole to spring into the air and drives a knee into attacker 4, who collapses onto the stones.

SHOT 3 (5-7.5s) Camera glides close to the ground and sweeps upward, FOV 30 degrees. She sprints downhill in a low crouch under hanging bamboo and ropes, spins into a sweeping kick that sends attacker 5 crashing into a rusted tea rack. Hand-plant roll, then a flying kick into attacker 6, who skids backward into the edge of the terrace wall.

SHOT 4 (7.5-10s) Camera orbits her, FOV 32 degrees. She slides on her knees through the shallow irrigation channel, throwing up a wide sheet of water, and pops up at the collapsed footbridge. She vaults the gap and twists mid-air to ram her shoulder into attacker 7, sending him to the terrace below. She lands, catches a bamboo pole thrown by attacker 8, plants it and pole-vaults, driving both feet into attackers 8 and 9, who fly in opposite directions.

SHOT 5 (10-12.5s) Camera pushes forward as if chasing, then pulls back fast, FOV 38 degrees. She runs along the leaning bamboo scaffolding, then back-flips over attacker 10's sneak attack from behind. Mid-air she twists and lands a back kick between his shoulder blades, sending him rolling forward down the slope.

SHOT 6 (12.5-15s) Camera drops to a low wide angle, FOV 30 degrees, then locks front-on at eye level. She lands and slides through a puddle in a low crouch, water spraying around her, then rises into a sharp ready stance. Her braid and the hem of her shirt are still swinging from the speed. The sequence ends sharply on this strong, still pose, final frame in crisp focus.

CAMERA: Momentum-driven tracking that follows her center of gravity (jumps, wall-runs, pole vault, kicks) rather than sitting static. Reactive impact cinematography: a sharp small shake on every hit, the camera briefly following a thrown opponent, then snapping back to her. Natural motion blur on fast moves.

PACING: Speed builds through the explosive moments. Her actions are clear and decisive; enemy reactions are fast and heavy. No slow motion, no pauses, no posed moments, no wasted movement.

VISUAL QUALITY: Ultra-high definition, cinema grade, realistic skin, fabric and water physics, live-action photographic look.

POSITIVE LOCKS: Pink streaks stay a soft rose tone, braid stays on her right shoulder, plum shirt and cream cargo trousers unchanged, boots and gloves unchanged, exactly ten attackers in charcoal coats and slate face wraps.

CONSTRAINTS: At least ten enemies, all hit. No empty kicks; every strike connects. No repeated moves. No fantasy light effects, no neon colors, no neon glow, no cartoon or CGI-style look, no distorted limbs, no clipping or objects passing through each other, no stuttering motion, no extra people, no text or watermarks.
#AIvideo #AIart #Seedance #AIFilmmaking #Parkour

How did I only just find this? You can save up to 60% on CapCut through the desktop app. Desktop only, and it wraps up Oct 30 — just putting that on your radar.
Also, Halloween marketing doesn't have to mean filming an elaborate haunted house. A product shot, some eerie effects, and a few well-timed cuts can do a lot. CapCut PC lets you shape the mood in post instead of reshooting everything.

tinyurl.com/2v23rn22

#capcutpc #capcutblackfriday #capcutsale #capcutmarketing

Same hallway. Same two people. One of them just got a glow-up. 👀

I ran the same scene twice with @Flovaai : first with a default AI face, then with the High-End Character Face Design Skill. The second one actually feels like a lead you’d root for, and the character sheets keep m shot to shot.

Watch the side-by-side, then steal the workflow. Quick tutorial’s in the video 🎬

@Flovaai If your characters keep coming out with “generic AI face,” this is the fix. 🙌 Describe your lead, pick a look, and try the Skill yourself:
flova.ai/en/skill/?id=6…

You don’t talk to your investor the way you talk to your group chat.

So why does your AI?

We spent weeks reading Reddit threads from people who live in their inbox. Here’s what they figured out ↓

and most importantly, try tonebird.ai :)

Okay i’m officially calling myself an AI creator on X now. nobody appointed me but here we are.

my latest film: a founder at 11:48 pm, an unreasonable number of notifications, and a small green bird who somehow knows what’s going on.

been making little films for Tonebird.ai about the situations we’re building it for. this one’s for everyone whose “one last reply” somehow became another hour at their desk.

the bird is getting suspiciously comfortable being the lead.

The best part of a celebrity face-swap video? The reveal. The challenge? Making every transition feel intentional ✨

You can streamline the process with CapCut Web × Codex. Install the CapCut plugin through the Video Editor prompt and ChatGPT, then let Codex help execute the ructions. Describe where you want the reveal, how fast the cuts should feel, and what kind of transitions fit the mood. CapCut's editing tools can help shape the sequence with transitions, filters, and text effects, while the editable draft lets you make the final creative calls.

AI doesn't replace your taste. It helps you get from a rough idea to a polished first cut with less repetitive work.
#capcutweb #capcutchatgptplugin #capcutcodex #capcutmarketing

Try now
tinyurl.com/2xeju4vv

Seedance 2.5 on Flova keeps surprising me. The motion holds together shot after shot, and the whole thing came out of one sitting. 🎬

Made with @Flovaai

@Flovaai If you’ve been waiting for a good time to try Seedance 2.5, this week is it. Flova’s savings event runs through Oct 27, with up to 80% off and the discounted rate locked for 365 days. Grab the Skill and make your own 👇
flova.ai/?refCode=9JVWH…

EAGLE-4 is here.

From EAGLE-1 to EAGLE-3, we’ve kept making inference faster.

EAGLE-4 is our fastest yet.

961.6 tokens/s on @Alibaba_Qwen’s Qwen3.8-27B with TokenSpeed (BF16, 4× @nvidia B300, TP4, one request, T=0).

Github: github.com/eagle-hq/EAGLE…

POV: Your LLM inference just got a serious upgrade. 🚀

Vanilla decoding: one token at a time.
EAGLE-4: let's pick up the pace.
📈 961.6 tokens/s and 4× the speed of vanilla decoding, as reported in the EAGLE-4 paper for Qwen 3.8 27B on 4× NVIDIA B300 GPUs.

What's powering the e drafting that responds to hardware and workload
🧠 40%+ lower draft-model overhead with the Efficient LM Head
🛠️ Open-source code for developers to explore and benchmark

Inference acceleration is heating up. EAGLE-4 is worth a closer look.

Explore the repo:
github.com/eagle-hq/EAGLE…

Reported benchmark results; performance varies by setup and workload.

Not gonna lie, I didn’t expect this one. Higgsfield went free, and it’s packed: one-click local generation, a real 4K/8K-ready timeline, and a full script-to-storyboard pipeline.

If you’ve been stressing over generation costs, go try it today.

Most “spooky idea → finished video” workflows fall apart the second the creature shows up in a new shot. The face changes, the proportions shift, and the location stops matching. 🕷️

This time I let @Flovaai build the character, the creature and the location references first, ot prompt around them. The design stays locked in every cut, and the whole thing went from one creepy idea to a video I’d actually post.

Made with @Flovaai.

@Flovaai Want to try this workflow yourself? Grab the Skill here and take your own spooky idea from references to a full video 👇
flova.ai/?refCode=BEL73…

I launched the first image and video generation models at Google, and kept asking why all AI designs look the same.

AI has brought us the WORST era of design.

It made content creation effortless, but it also made 90% of it look painfully generic.

My co-founder and I spent 2 th to change that.

Meet Vyb, the first AI design model that understands your taste like a designer would.

RT + comment your website URL and we’ll make your first design.

Silver blade or ember blade? 🤍🔥

Two girls, one rain-soaked shrine, and a fight that ends with a truce nobody saw coming. Which side are you on? Comment below 👇

Created with Seedance 2.5 on @TapNow_AI

Prompt:
30-second cinematic live-action video, 16:9 aspect ratio, 24 fps, lenses with fine film grain. An abandoned mountain shrine courtyard at blue hour, just after sunset. A thin layer of rainwater covers the worn grey flagstones, mirroring the sky and the blades. Moss-covered stone lanterns line the sides, a broken red-lacquer torii gate with a cracked crossbeam stands at the far end, and a flight of mossy stone steps leads up to a dark shrine hall with a hanging rope and a rusted bronze bell. Fine mist drifts between the pillars. A few dim, warm amber paper lanterns glow softly in the background. Pale petals and wet leaves drift down in slow, thin rain. Color grade: cool slate-blue shadows, soft grey-green mist, muted earth tones, gentle warm highlights from the lanterns. Natural and filmic, nothing oversaturated.

CAMERA RULE: dynamic shoulder-mounted handheld with real weight and micro-shake, mixed with occasional low floating glides on the wet floor and a few whip pans. The operator stays close to the action and is never perfectly stable.

CHARACTER A (use Image A as reference): a real young woman in her early 20s with fair skin, a short wavy dark-brown bob with wispy bangs, and a small black bow clip on the right side of her head. She wears her oversized lavender varsity jacket with cream striped trims and a cream “B” patch over a black tank, light-wash wide-leg jeans (hems soaked and darker), and chunky black sneakers. Her sword blade is a solid beam of cold silver-white light with a soft pale sapphire edge, reflected in the water under her feet.

CHARACTER B (use Image B as reference): a real young woman in her early 20s with fair skin, a beauty mark on her cheek, very long straight jet-black hair with wispy bangs, silver hoop earrings and a thin silver necklace. She wears her black satin bomber jacket over a white crop tank, washed-black cargo pants and white chunky sneakers that get muddier through the fight. Her sword blade is a solid beam of deep burnt-copper and ember-amber light, casting warm reflections on the wet stone.

PHYSICS RULE: every step, slide and landing throws sheets of water. Droplets catch the blade light. Where a blade touches the surface it leaves a brief glowing wake and hissing steam.

CHOREOGRAPHY RULE: no long pauses. Each exchange ends in a 2-3 frame freeze pose (back-turned stance with the blade pointing behind, low crouch with the blade skimming the water, one-handed overhead guard, mid-air spin).

TIMELINE

0.0-3.0s | LOW GLIDE THROUGH THE MIST. 35mm, camera 40 cm above the water, gliding slowly forward past a stone lantern. Rain taps the puddles. Through the mist, A stands on the steps in front of the shrine hall, B at the broken torii gate, 8 m apart. B drags her unlit sword hilt along the torii pillar with a metallic scrape. A rings the rusted bell once, and the low tone ripples the water. At 2.2s both blades ignite, silver-sapphire and ember-copper, lighting the mist around them. Sound: rain, distant bell tone, rising electric hum.

3.0-6.0s | THE CHARGE. 24mm handheld, tracking sideways at running speed. Both sprint across the flooded flagstones, sending sheets of water outward. They meet at the center with a huge flash. Three fast strikes and parries. At 4.9s A slides on her knees under B’s horizontal slash, drenching herself, and swings up. Sound: splashes, crackling impacts, sneakers slapping on wet stone.

6.0-9.0s | CLOSE-UP, MIRRORED FACES. 65mm, shallow depth of field. Tight on A, breathing hard, wet strands of bob stuck to her cheek, silver light on her face. A fast whip pan at 7.2s lands on B, who smirks, shakes the water from her hair and flips the sword to a reverse grip. Behind her, the reflection of her blade ripples on the floor. Sound: whoosh, breaths, hum.

9.0-12.0s | THE LANTERN BREAK. 28mm, handheld, orbiting 120 degrees. B pushes A back toward a moss-covered stone lantern. A runs up the lantern’s side, flips over B and lands behind her. B’s slash misses and slices the lantern’s cap clean off. The stone top slides slowly and drops into the water with a heavy splash, sending a ring wave across the courtyard that both girls jump over. Sound: stone grind, huge splash, echo.

12.0-15.0s | LOW ANGLE, THE STEP LEAP. 24mm, camera 30 cm off the ground, tilted up, dragged beside them. B bounds up the stone steps and leaps off the top step with an overhead arc at 13.2s. A meets it from below with a rising slash. The blades lock above the lens in a burst of sparks and a spray of droplets lit silver and amber. They land on opposite sides at 14.6s in mirrored back-turned poses, water rippling outward between them. Sound: heavy impact crack, low boom.

15.0-18.0s | OVER-THE-SHOULDER, RAPID EXCHANGE. 35mm, circling from over B’s shoulder to over A’s within 1.5s. Six rapid strikes: horizontal, diagonal, a low sweep A jumps over, a thrust B sidesteps. At 16.8s B skims her blade along the water and flicks it up, throwing a curtain of glowing droplets into A’s face. A answers by sliding under B’s next slash. Light trails of 2-3 frames follow the blades. Sound: rapid overlapping strikes, splashing steps.

18.0-21.0s | THE TORII CUT. 32mm. A leaps, runs three steps up the red torii pillar, flips off it and slashes down. B blocks and sidesteps, and A’s blade slices through the cracked crossbeam. The beam falls slowly and lands across the water, sending a wave outward. Both girls jump back from the splash. Sound: wood crack, heavy splash, ringing echo.

21.0-24.0s | THE BOW, SLOW MOTION. 85mm, 40% speed, shallow depth of field. B’s blade slices past A’s head and cuts the black bow clip from her hair. A few strands float with it. Time slows. The bow turns lazily in the air, a single silver droplet rolling off it, and lands on the water, floating on its own ripple. The blades hum on either side. A’s eyes flick to it, then back to B with a small, challenging smile. Sound: sustained low hum, muffled heartbeat, a tiny splash.

24.0-27.0s | EXTREME CLOSE-UP, BLADE LOCK. 85mm, focus at the crossing point. The blades cross in the center of the frame, sparks hissing in the rain. Faces in profile on either side, A lit silver-blue, B lit amber. Raindrops sizzle and burst into steam on the blades. At 26.3s they push apart in a flare and each steps back. Sound: sustained electric grind with sizzling, then a sharp release.

27.0-30.0s | FINAL POSE AND TRUCE. 32mm, pull-back 4 m with a slow 8-degree rotation. They lunge at 27.4s, pass each other and freeze back-to-back with blades pointing away, water settling around their feet. The blades dim together. Without turning, B scoops the floating bow from the water with her blade tip and flicks it back over her shoulder. A catches it, wrings it out and pins it back in her hair. The bronze bell rings once more and they walk off in opposite directions across the mirrored floor, rain falling over the empty courtyard. Sound: hum fades into rain, one last bell tone, no music.

NOT IN FRAME: no cartoon or cel-shaded rendering, no glowing auras on the bodies, no magic effects other than the two blades, no neon or fluorescent colors, no text or logos, no other people.
#AIvideo #AIart #Cinematic

Flova × Seedance 2.5 is sticking around a little longer 🎬

Thanks for all the support! The offer is now extended through October 27.

Save up to 80% and lock in your price for 365 days. 480p starts at $0.022/sec, and 720p at $0.044/sec.

That scene in your head? Consider this your cue.

CapCut’s Black Friday deal gives you 60% off Pro or 40% off Ultra, October 20–30, available on CapCut Desktop only.
The ad needs to be ready soon, and a small visual flaw is holding up the final shot.

CapCut PC’s Local Inpainting gives the selected area a focused revision e updated shot in the campaign sequence and keep developing the material that already fits the creative direction. A more specific next step helps the finishing stage move forward when launch is close.

Useful when the remaining visual work needs to fit around everything else on the campaign checklist.
#capcutpc #capcutmarketing #capcutsale #capcutblackfriday