Sebastien Jefferies

Hot take: Kling 3.0 quietly killed the entire ‘talking head’ AI workflow.🔥 We just stress-tested Kling 3.0 and the lip sync is ridiculous. Ran a Deadpool scene + a few celebrity-style tests (parody) and it kept character identity locked across cuts! Check this out👇👇👇

Google Veo 3.1 ultra realistic prompt 👇👇👇

This is the best workflow for creating multiple angles with AI while maintaining character consistency.👇👇

Here’s how to do it: → Go to Higgsfield AI → Select Nano Banana 2 → Type in your base prompt to create your model, steal mine below . → Use my exact prompts to generate different angles → Keep the same character and build a full campaign feel

No subtitles. No music. SHOT — Extreme close-up shot focusing on the grandmother's face, framed tightly to capture every wrinkle, tear, and expression; 100mm macro lens perspective emphasizing texture and detail. SUBJECT — Elderly grandmother approximately 75 years old with soft aged skin showing deeply detailed wrinkles and natural skin imperfections; grey hair slightly messy, framing her face; emotional expression full of sadness and nostalgia with tears welling up and slowly falling down her cheeks; eyes slightly cloudy but full of life; trembling lips; subtle shaking breath. SCENE — Indoors, sitting near a window during late afternoon; warm golden light from the window softly illuminating her face and casting gentle shadows that enhance the depth of wrinkles and facial features; a blurred, softly textured background including softly out-of-focus elements such as framed family photos and cozy home decor. VISUAL DETAILS — Slow motion capture showcasing tear reflections and skin texture with extreme clarity; film grain overlay for natural, cinematic texture; lighting highlights moisture on skin and tears, creating complex highlights and glistening effects; soft, diffuse shadows enhance emotional depth. ACTION AND CAMERA MOTION — The grandmother blinks slowly as tears well up and roll down her cheeks; her lips tremble as she breathes with subtle shaking; gentle camera push-in and slight lateral tracking to accentuate intimacy and emotional storytelling. CINEMATOGRAPHY — Warm, golden window light creating contrast between illuminated and shadowed parts of the face; emotional tone is intimate, heartbreaking, and reflective; color palette includes warm earth tones with creamy skin hues and natural greys from hair; soft focus background to isolate the subject. AUDIO — Voice: Soft, shaky female voice aged 70+; Language: English; Tone: Emotional and nostalgic, subtle breaths and slight sobs; Dialogue: "I remember when life was so different..."; Primary sounds: gentle breathing, tear drops softly falling; Ambient: quiet room with muffled outdoor sounds; Environmental details: faint creak of chair. STYLE — Ultra-realistic, naturalistic visual aesthetic emphasizing human imperfections and emotional authenticity; 4K ultra-detailed resolution with macro-level detail on skin and eyes; film grain applied lightly for texture; prohibited elements removed including any on-screen text or subtitles.

🚨 Most AI events talk about the future. This one shows it live. March 19, London (The Drum Labs) A room full of people actually building with AI: → Real workflows → AI-powered production → Scaling creativity without losing it With leaders from @freepik & more. I’ll be there, come say hi.

Five Guys burger ad… created with AI. 🚨 Here’s the exact workflow we used 👇

@maxescu So many friendly facesssss 🔥

No matter who you are, we can agree on one thing: We all love Nano Banana! A few days ago, I was singing to myself, "I love Nano Banana..." so I made a song. Naturally, I fired up Higgsfield and made a video. See how many familiar faces from the X AI community you can spot:

@Endlessedits Background

Stop wasting time making backgrounds just drag & drop them into your editing software.🥳❗️ I’ve collected 50+ Animated backgrounds🎉. Want them?🤑 Comment “Background” Follow + Repost ♻️ I’ll send it instantly 📥

@lorenzo_pravata Research

Winning ads need a winning research process. After generating $100M, I’m giving ours away: 👉 Step-by-step instructions for research 👉 The frameworks we use to break down competitor's strats 👉 Proven prompts to speed the whole process up by a ton(Must be following) 👉 A plug-and-play Notion template you can duplicate for your own brand Want it? Like + comment “RESEARCH” and I’ll send it over. (Must be following)

@zeely_ai @sachinmaya1980 Avatar

🚨BREAKING: Influencers aren’t getting replaced. They’re getting AI-powered. With new avatars, your product can speak for itself — in ads that actually feel human. RT + like + follow + comment “Avatar” to get 500 credits in 10 hours.

→ Head over to Higgsfield AI → Create the VFX reference image →Start with Nano Banana 2 to generate the visual style of your scene. Think of this as your VFX concept frame the image that defines what the final shot should look like. The goal is to create a strong visual reference that captures the exact look you want. → Capture a driving video Next, record or choose a video. This can be: – a simple phone clip – a walking shot What matters most is the camera movement and framing, because this motion will drive the final shot. → Import everything into Kling →Open Kling and use the Motion Control feature. Upload: →the Nano Banana 2 reference image →the video Set the reference image as the visual guide. → Restyle the motion Kling analyzes the motion from the driving video and restyles the entire scene based on the reference image. So the camera movement stays natural and consistent.

AI VFX is entering a new era. 🚨 FULL WORKFLOW 🧵👇

Give this a try👀

No editing. No manual upscaling. →Just Nano Banana + one prompt. Here’s the exact workflow: →Step 1: Take any blurry or low-resolution image (movie scenes, old photos, memes, screenshots) →Step 2: Upload it into Nano Banana →Step 3: Paste the prompt below →Step 4: Generate →For best results, copy-paste this exact prompt: Ultra-premium professional image enhancement. Preserve 100% original identity, face structure, expression, pose, clothing, accessories, background, framing, and composition. Do NOT alter, redesign, replace, or add anything. Transform the uploaded low-quality, blurry image into extreme high-detail cinematic quality. Recover micro-details: sharp facial features crisp eyes natural skin texture visible pores realistic hair strands clean refined edges. High-contrast clarity, deep depth, and balanced cinematic lighting. Poster-grade realism with dramatic but accurate detail. Photorealistic textures only. True-to-source enhancement only. Output in 8K resolution, ProRes quality, studio-level sharpness. Keep everything exactly the same only enhance quality.

Wake up! Low image quality is optional now. 🚨 The only workflow you need👇👇👇