Made a Stranger Things x Adidas Samba spec ad using Nano Banana Pro, Kling 2.5 Turbo, and Kling O1 on @openart_ai , @Kling_ai and @Weavy_ai .
Still figuring out the best way to mix AI tools with my usual workflow.
Working on a BTS breakdown now. If you have questions about the process, drop them below and I'll cover them in the next video.
#StrangerThings5 #NanoBananaPro #klingai
#adidas
Introducing ChatGPT Images, powered by our flagship new image generation model.
- Stronger instruction following
- Precise editing
- Detail preservation
- 4x faster than before
Rolling out today in ChatGPT for all users, and in the API as GPT Image 1.5.
One technical challenge: getting the backpack to fly into frame from off-screen.
AI video models struggle with this because they can only animate what's already in the image. Can't create something from nothing.
Kling O1's new multi-element/reference feature solved it. Took a few tries, but I got the shot I wanted.
What surprised me: the video rendering quality from this feature is almost as good as traditional image-to-video. That's a big deal for complex shots.
Made a Stranger Things x Adidas Samba spec ad using Nano Banana Pro, Kling 2.5 Turbo, and Kling O1 on @openart_ai , @Kling_ai and @Weavy_ai .
Still figuring out the best way to mix AI tools with my usual workflow.
Working on a BTS breakdown now. If you have questions about the process, drop them below and I'll cover them in the next video.
#StrangerThings5 #NanoBananaPro #klingai
#adidas
Interesting discovery: Chinese models like Seedance Pro had no idea what a Demogorgon looks like. It created a completely different creature with a separate mouth.
Cultural training data gaps are real.
Nano Banana Pro just changed how we create stories.
I've been developing a Stranger Things-inspired Adidas Samba spec ad, and the workflow is completely different now.
The old way:
Hunt for visual references
Use images as style guides
Generate, adjust, repeat
The new way:
Just describe what you need
Nano Banana Pro pulls references from its training
Get it right in one or two tries
The breakthrough? You don't need reference images anymore. I just typed "demogorgon or Eleven from Stranger Things" or "Upside Down atmosphere" or "1980s gym lighting" and it understood exactly what I meant.
Season 5 drops in a few hours. Here's a sneak peek of the project. Full ad coming in a few days.
Made with Nano Banana Pro on @openart_ai
🍌 Nano Banana Pro is now UNLIMITED on OpenArt.
* 2 weeks of unlimited 1K/2K generations
* Available with the Wonder Plan
* Now 60% OFF for Black Friday
Retweet + comment for a chance to win 1 year of Wonder Plan.
I started the project with Flux before Nano Banana was released. I designed the main character in Flux 1, while the rest were created with Nano Banana 1.
The motion was made in Kling 2.1 and 2.5 Turbo. AI moves so fast that I can’t keep up with just one tool or maybe I’m just a bit slow 😅.
Introducing SAM 3D, the newest addition to the SAM collection, bringing common sense 3D understanding of everyday images. SAM 3D includes two models:
🛋️ SAM 3D Objects for object and scene reconstruction
🧑🤝🧑 SAM 3D Body for human pose and shape estimation
Both models achieve state-of-the-art performance transforming static 2D images into vivid, accurate reconstructions.
🔗 Learn more: https://t.co/yXcvts8Ogc
She collects men by turning them into ponies 🖤🐴
30-sec teaser from my AI music video "Jenna Ortega" by @rnbstellar
2000+ images generated
200+ videos created
98 final shots
10 days production
Full video: https://t.co/6584qZ9kjD
Full breakdown dropping tomorrow.
Created for @openart_ai #OMVA #OpenartMVA
🧵 Stats + process below 👇
Tomorrow: Full Reddit breakdown with workflow, tool comparisons, lessons learned, and what nearly broke me.
Coming from traditional film/VFX background - this taught me AI production works exactly like traditional, just solo.
More AI filmmaking: @phyofox 🎬
Concept: Dark fairy tale inspired by Cruella de Vil + Pinocchio
Woman orchestrates romantic dates. Each man transforms into a pony for her collection.
Gothic aesthetic. High fashion. Horror.
Twist ending you won't see coming.
My latest AI experiment: a 30s live-action Akira teaser made with Kling 2.1, Flux Kontext, and OpenArt.
Started as a test for two scenes—ended up reimagining Kaneda, his bike, and neo-Tokyo with full cinematic vibes.
#Akira#KlingAI#FluxKontext#AIVFX#AIanimation