Creative AI
Sora's exit opens the image race
Disney walked away as the Sora app closed; Luma and Google answered with stronger image models and Gemini passed 750M users.
Orbit Roundup · 23 Mar – 29 Mar
Week 13in XR, AI, 3D & creative tech
20 stories · 8 min · out 30 Mar 2026. Picked by Tom Martin-Davies.
What's in the week
What happened this week:
- Sora's exit opens the image race. Disney walked away as the Sora app closed; Luma and Google answered with stronger image models and Gemini passed 750M users.
- Handmade craft wins the week. A hand-animated short took SXSW, a miniature New York was built by hand, and LAIKA showed how it animates water frame by frame.
- ComfyUI grows into a pipeline. Vector output, a 3D pipeline in ComfySketch and new Blender helpers: node graphs are turning into production tools.
- Snap's AR keeps scaling. Two trillion snaps, AI Clips in Lens Studio and image-to-3D for AR: Snap keeps making AR creation cheaper.
01
Trending
What happened this week
2D, 3D & Design
Handmade craft wins the week
A hand-animated short took SXSW, a miniature New York was built by hand, and LAIKA showed how it animates water frame by frame.
Stashmedia Tv ↗Designboom ↗Beforesandafters ↗Itsnicethat ↗Stashmedia Tv ↗
Creative AI
ComfyUI grows into a pipeline
Vector output, a 3D pipeline in ComfySketch and new Blender helpers: node graphs are turning into production tools.
XR & Spatial
Snap's AR keeps scaling
Two trillion snaps, AI Clips in Lens Studio and image-to-3D for AR: Snap keeps making AR creation cheaper.
02
Inspiration
6 projects · tap to watch03
Your feed
All the week's newsFollow topics to build your feed
Designr/creativecoding
Still Life in p5.js
Visual artwork created through creative coding with the p5.js library. Simple but effective use of code to generate digital art.
AIThe Decoder
Luma Uni-1 challenges Google Nano Banana
New Uni-1 model from Luma Labs handles both image understanding and generation in one architecture. It is holding its own against Google's flagship and is cheaper to run through the upcoming API.
AITech Insider
Google Gemini hits 750M users
Gemini is expected to reach 750 million users by Q3 2026. Growth is being driven by deeper Android integration and the global rollout of AI Mode across Workspace.
AIr/comfyui
Raster to Vector SVG ComfyUI Workflow
Beginner-friendly workflow using the Flux.2 Klein model. It generates raster images and converts them to vector SVG outputs on the fly, including a prompt saver node.
AIForbes
Solo founders scaling via vibe coding
Founders are using AI to handle technical implementation while they focus on creative decisions. This shift is allowing solo makers to ship revenue-generating products faster than ever.
XRr/augmentedreality
AR GenAI: Image to AR 3D Model
Generate immersive AR experiences and 3D models from a single photo using AR Code's new generative AI tool. This approach is gaining traction for making spatial creation much simpler.
XRWebpronews
Snapchat hits 2 trillion snaps
Snap continues to grow its user base despite struggles with monetization. The report notes that while AR investments are impressive, hardware like Spectacles remains a niche part of the business.
DesignDesignboom
Ambitious handmade miniature New York City
Artist Joe Macken built an incredibly detailed miniature model of New York City. The piece is currently on display at the Museum of the City of New York.
AIThe Decoder
Disney exits OpenAI deal as Sora app closes
OpenAI is shutting down the Sora app and API to refocus compute resources. Consequently, Disney has withdrawn from their billion-dollar partnership and character licensing agreement.
DesignStashmedia Tv
BOL and Device explore memory in new short
Directors Pol Solà and Guille Comin teamed up with Device for a universal short film. It’s a beautifully animated exploration of solitude, memory, and light.
Designr/creativecoding
Interactive Dark Cubes Reflections
Creative coding project using interactive cube-based reflections. Includes a live demo and source code available on CodePen for inspection.
AIr/comfyui
ComfySketch Pro adds 3D pipeline and sync
Major update for the ComfyUI drawing node. Now supports 3D model imports (GLB, OBJ, FBX), viewport sync with Blender and Maya, and new spot healing AI tools.
AIBlog Google
Gemini adds chat and memory import tools
Google is making it easier to switch from other AI apps by allowing users to import their entire chat history and memory. A clear move to centralize user AI context in their ecosystem.
XRDevelopers Snap
AI Clips launches in Lens Studio
Snap's new plugin lets creators generate 5-second AI videos directly within lenses. It transforms a single user photo into a dynamic clip using embedded prompts and predefined concepts.
04
