Stepping into AI || Let’s collaborate & create || DM for promos || CPP @RoboNeo_ai, @Flovaai , @Capcutapp [email protected]

Joined November 2023
Pinned Tweet
I asked Seedance 2.5 for an Edwardian action thriller with a bride, a gunfight and a very British one-liner. It delivered. Made on @FishCreativeHQ using Seedance 2.5. Would you watch the full film? 👀
60
48
195
1,159
CapCut's Black Friday deals are live: Pro is 60% off and Ultra is 40% off on Desktop through Oct 30. And if you've been curious about Dreamina, its first month is just $0.29. Making ads shouldn't mean spending hours hunting for the right visuals. Start by generating assets in Dreamina, then edit your video in CapCut PC. Product showcases, social promos, campaign teasers… there's plenty you can do without overcomplicating the process. Get the CapCut deal: tinyurl.com/2v23rn22 Explore the Dreamina offer: tinyurl.com/5xbtv4x6 #CapCutDidThat #Dreamina #Smallbusinessmarketing #CapCutSale
🤝 Paid partnership
34
2
148
17,130
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.
35
15
35
2,496
A dance video just hits differently when the audio fits. 🎵 Start with your character-rendering idea, use Codex to control the CapCut Web workflow, and explore the built-in music and sound effects library to find audio that matches the mood. Whether your AI character is doing a smooth routine or a hilariously awkward dance, the right sound can make the whole scene work. Find the sound, feel the beat, make it yours. #capcutweb #capcutchatgptplugin #capcutcodex #capcutmarketing
🤝 Paid partnership
49
2
510
59,495
One more week to stretch your AI video budget further. The @Flovaai × Seedance 2.5 deal is extended until October 27, with savings of up to 80% and discounted rates locked in for 365 days. ✨ Seedance 2.5 from $0.022/sec (480p) ✨ Image 2.5 from $0.011/image ✨ Wan 3.0 Prime from $0.022/sec More time to create, less cost per generation. 🔗 flova.ai/?refCode=9JVWHWYP
🤝 Paid partnership
31
2
422
46,950
A supercar that defies gravity while keeping every design detail intact. With Seedance 2.5, one carefully crafted continuity prompt turns a single car reference into wall climbs, midair flips, hovering shots, and full-on flight. Created with @Flovaai
🤝 Paid partnership
50
6
262
36,680
EAGLE-4 is out. On Qwen 3.8 27B with 4× NVIDIA B300 (SPEED-Bench qualitative split, T=0), the paper reports 863.8 tokens/s. That is 1.28× DFlash 2 and 1.40× DSpark, and 3.59× vanilla decoding. At 512 concurrent requests the same paper shows 1.70× the throughput of DFlash 2 and 1.99× DSpark. The code is open source. You can read the implementation, reproduce the setup, and test it on your own workload. github.com/eagle-hq/EAGLE-4 #EAGLE4 #OpenSource
🤝 Paid partnership
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-4
35
1
136
34,624
Would you jump off a Meteora pillar after finishing the sausages? 🤔 This guy did. Nobody was hurt (in the video, anyway). Made on @FishCreativeHQ using Seedance 2.5
🤝 Paid partnership
71
31
252
63,407
AI creators just got a major upgrade. Higgsfield going free brings local generation and professional video editing into one workflow, making high-quality content creation more accessible than ever.
🤝 Paid partnership
This Post is from a suspended account. Learn more
26
1
128
44,538
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 years building in stealth 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.
271
238
196
2,032
3,651,840
Wait — Krea 3 Ultra at #4? 1151 Elo, basically level with the models right around it on a pure blind preference board. Didn’t expect an independent lab to land this high. @krea_ai
🤝 Paid partnership
today, Krea 3 makes it to the top of image arenas. - "Fast" pushes the pareto curve of speed and quality. - "Ultra" ranks top-4, only surpassed by OpenAI models. the only company valued under $1T in the top-10! more below 👇
30
1
129
63,588
Shami retweeted
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.
122
182
104
300
1,611,053
Video Prompt: MASTER PROMPT (paste first) Format: 16:9, 30 seconds, 24fps, 1080p. Cinematic 90s anime film look, hand-painted backgrounds, warm golden sunlight, saturated blue sky with cumulus clouds, soft film grain, gentle lens flare. Setting: A seaside sidewalk in the afternoon. A tall chain-link fence runs along the right side, with the ocean behind it, palm trees, and an orange-lined road on the left. Characters (use the attached sheets as strict references): Mei Aihara (@Image1): wavy purple-black hair, round glasses, pink hoop earrings, pink and blue oversized sweater, green high-waist pants, white-pink-blue sneakers, black spiral notebook with stickers. Shy and bubbly. Yuki Kageyama (@Image2): long white double braids, black blazer, turtleneck, slim pants, boots. Calm, observant, protective. Luca Moretti (@Image3): messy blonde hair with a faded undercut, star face tattoos, neck and sleeve tattoos, ear piercings, white tank top, black trousers, leather shoes, cigarette in hand. Sound style: Anime audio. Exaggerated sparkle chimes, soft “doki-doki” heartbeats, cute kyahaha laughs, wind and wave ambience, light lo-fi guitar under the first half, and a low tense string drone from 19s. TIMELINE 0:00–3:00 | Luca, profile close-up Camera: tight side-profile close-up on Luca, fence in sharp focus beside him, deep-focus background. Action: He leans on the fence with the cigarette in his lips. Smoke drifts up toward the sky. In the soft-focus background, Mei and Yuki walk small and distant down the sidewalk. At 2:00 his eyes slide toward them and the frame blurs as the focus pulls away. Sound: waves, seagulls, soft inhale and exhale, faint lo-fi guitar. 3:00–7:00 | The girls chatting (medium two-shot, eye level) Camera: Front-facing two-shot at the same angle as the reference, with the fence and ocean at right, as they walk toward the camera. Action: Mei gestures excitedly with one hand, clutching her notebook to her chest. Yuki listens with a small smile. Mei (excited): “Gojo Satoru is so handsome, I could kiss and marry him!” At 5:00 both burst out laughing. Mei closes her eyes and laughs wide-mouthed (her “cheerful laugh” expression). Yuki covers her mouth and laughs with a quiet “pfft”, shoulders shaking. Sound: kyahaha! giggles, a cute sparkle “ting” on the punchline, and bouncy footsteps. 7:00–10:00 | The collision Camera: Low ground-level shot of Luca’s black leather shoe stepping into frame while Mei’s sneakers drift past, slightly out of focus. Cut to a wide shot at the fence with the ocean behind them: Luca walks left to right, Mei walks right to left, looking at her phone-free hands, and they bump. Mei (voice-over, still rambling, cut off): “I have his wallpapers plastered all over my room—” Her notebook and books fly out of her arms. Sound: shoe clack, quick “whoosh”, soft thump, “Eh?!” gasp, books slapping the pavement. 10:00–12:00 | Kneeling to pick up Camera: Wide high-angle shot looking down at both of them crouching, books scattered on the sidewalk. Luca bends low and Mei crouches beside him, flustered. Mei (flustered, soft): “A-ah, sorry!” Sound: fabric rustle, shy breathing. 12:00–16:00 | Hands and eyes (the key shot) 12:00: Extreme close-up on the blue book. Her hand and his tattooed hand land on it at the same time. Sound: all ambient sound drops out, one soft heartbeat “doki.” 13:00: Low-angle close two-shot over the book. Mei on the left with sunlit hair and glasses, Luca on the right looking down at her. Their eyes meet slowly. Her cheeks flush. 14:00: Extreme close-up on Mei’s eye. The pupil dilates and the glasses reflect sky and sparkle highlights. 14:45: Extreme close-up on Luca’s eye with the pierced eyebrow, face tattoos and a slight softening in his stare. 15:30: Return to the two-shot, slow push-in. Her lips part slightly. Luca (low, calm): “...Here.” Sound: double heartbeat “doki-doki”, a sparkle chime, soft strings. (Luca’s line is my addition, since the reference has no dialogue here. Remove it if you want it silent.) 16:00–19:00 | Yuki calls, Mei snubs him Camera: Wide shot from behind Yuki’s position. Yuki stands on the sidewalk on the left, waving. Mei and Luca are kneeling at center-right with the fence behind them. Yuki (calling out): “Meimei!” Mei jolts, grabs her notebook and stands up fast. She does not look at Luca. She turns her head away, bows slightly without a word, and hurries toward Yuki with her head down. Luca stays crouched and watches her go. Sound: gasp, quick footsteps, a “pip-pip” panicked blip. 19:00–22:30 | Yuki’s warning Camera: Medium shot, low angle, over Yuki’s shoulder. Yuki grabs Mei by the arm, and both glance back toward the fence. Sun flare glints off Mei’s glasses. Yuki (hushed, serious): “Meimei, that guy is with the Mafia. Be careful!” Sound: the tense string drone rises, and the ambience gets muffled. 22:30–26:00 | Luca stares Camera: Over-the-shoulder shot from behind Mei and Yuki (their heads blurred in the foreground). Rack focus onto Luca, who is now standing, hands at his sides, staring after them. A new cigarette smoulders and smoke drifts across the sky. Sound: wind, a single low taiko or bass hit. 26:00–30:00 | Final beat 26:00: Close-up on Mei’s profile as she looks back over her shoulder. Their eyes meet one more time across the distance. She blushes and snaps her head away (the snub). 28:00: Cut to Luca. A slow, small smirk. He looks down at his tattooed hand that touched hers, then back at her fading figure. 29:00: Final wide shot. The fence, the ocean, smoke curling up, the two girls walking away small in the distance. Fade out. Sound: a last “doki” heartbeat, a sparkle chime, then the music fades into waves. Negative prompt Realistic or 3D look, character design drift (wrong hair colour, missing glasses, missing face tattoos), extra fingers, text watermarks, camera shake, fast cuts during the eye-contact shots, western cartoon style.
2
344
Reference Character Sheets
1
4
289
Made this fully with Kling AI 4.0’s Native 30-second Generation. One continuous anime shot of an otaku girl gushing about Gojo, accidentally bumping into the mafia guy. No cuts, no stitching. Just smooth motion, consistent characters, and natural anime-style timing. Prompt is in the comment 👇 @kling_ai #KlingAI #Kling4.0
116
124
2
1,299
522,979
What if AI could understand a world instead of just rendering it? JST-2 takes a different approach, modeling environments as persistent 3D atoms with learned dynamics rather than simply generating pixels. This could open exciting possibilities for game worlds, robotics training, and truly interactive experiences.
🤝 Paid partnership
Today we are releasing JST-2, the world's first full-stack physics-native world foundation model. Many world models learn in pixel/video space. We went back to first principles and built the entire model stack in 3D. Generate a 3D scene in seconds and expand it in real time. Create characters with motion built in, all through one general architecture.
34
1
309
30,394