🎨 Exploring the latent space between creativity and Ai. Principal creative technologist at Google (through Yunojuno). Prev: Lumalabs, Media Molecule, LEGO etc.
denmark
Joined August 2009
- Tweets20.6K
- Following1.6K
- Followers33.6K
- Likes65.3K
Pinned Tweet
So the title might be a bit bombastic, but if this isn't one of the most incredible glimpses of a future way of creating and rendering 3D.. I don't know what is.
This was sculpted on my PS5 in mere minutes and transformed as I was sculpting, through a "Real-time Latent Consistency Model".
Sony, you should reconsidder your plans for Dreams and make a Dreams V2 with generative AI or sell it to someone that will.. just saying ♥️ No other 3D tool would allow me to sculpt this freely which turns out to be a super-power in the land of AI.
#madeindreams #ai
Martin Nebelong retweeted
Starting from a single Nano Banana image, Gemini 4 Argon created this finished 3D cathedral audio-reactive scene.
Driving Blender MCP in an autonomous loop in @Antigravity to:
🏛️ Sculpt intricate Gothic architecture
🎥 Align complex camera grids
🔊 Stage dynamic, audio-reactive lighting
Martin Nebelong retweeted
When a vfx pro like hector conjures up worlds in houdini using voice notes - it just mogs all other opus builds.
Seasoned creators can zoom out to create at an unprecedented scale without losing track of the details.
Because they go from playing the instruments to conducting the orchestra.
En esta prueba he querido poner Claude al límite con Houdini (modelo Opus 5.5). He montado toda la escena hablándole al iPhone para ver dónde está el límite, y mi conclusión es que no hay límite. Esto que veis es donde he puesto el freno, hay cosas que corregir aún pero quiero seguir experimentando con otros proyectos.
El render ha tardado 50 horas en mi 5090 dado lo compleja que es la escena, volumétricas etc.
El color grading lo he hecho en Davinci trabajando en Aces y también lo hice hablándole al teléfono con el MCP de Davinci.
Me cuesta imaginar un futuro donde no vayamos a trabajar con esto de forma masiva.
Martin Nebelong retweeted
Been quiet because I tore out the renderer (Three.JS), and rebuild it in custom high perf SDF.
Nook is clay all the way down now: landscape, objects, light. Early, but it finally feels like the thing I had in my head.
#gamedev #indiedev #screenshotsaturday
Martin Nebelong retweeted
Amazing visual code sketches made with Gemini 4 Argon by @kickingkeys
Gemini 4 Argon is the first model where It felt like an image, video and language model all at once.
Everything in this thread is made with code. @GoogleAIStudio
More examples + the full code for all 534 paintings below ⬇️
My colleagues at Google make stuff like this in their free time.. No pressure!
As a kid of the 80s who grew up reading Akira, Ghost in the Shell and seeing Blade Runner and having my world view changed.. this is the type meditative city builder I know I'll be spending way too much time playing 😁
Might be Dreaming..!
Dreams PS4 running on PC, not through emulation but a custom runtime that I built. A little side project I started some years back.
Multiverse Engine can also now load levels & models exported from Dreams, using an exporter a friend of mine made. The exporter isn't mine & isn't part of the engine. The engine only reads the exported files.
No code, assets or content from Media Molecule or Sony is built into Multiverse engine.
#MadeInDreams #Dreams #MultiverseEngine
Martin Nebelong retweeted
Retro anime sequence made with Scenario MCP:
>Opus 5.5 blocks every shot in Blender
>GPT Image 2.5 designs the pilot and the mech
>Seedance 2.5 animates it, ElevenLabs scores it
Martin Nebelong retweeted
Introducing Gemini 4 Argon – our new frontier model.
It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.
The Unbound team has been a favorite of mine for a long time ♥️. Pushing for that granular creative control that we need to turn gen ai into a truly useful creative companion.. the combination with Tripo segments that.
The Tripo API is now integrated with @unbound_io.
Bring Tripo P2's clean, editable quad topology directly into a real-time sculpting loop. Edit live 3D geometry and export production-ready assets instantly. No endless cleanup or Blender skills needed.
Generated with Tripo. Sculpted in real-time with Unbound.
Martin Nebelong retweeted
How to export Dreams models by Elca Gaming #MadeInDreams youtu.be/5dk3QHIu3Ow?si=191W…
Martin Nebelong retweeted
Oh hello Opus 5.5, please meet Unbund Loop's MCP:
"sculpt a sci-fi looking car with jets instead of wheels"
Low-poly outout model by @tripoai
Martin Nebelong retweeted
Old scene from @andreintg running in the browser with GI. He set a lot of things to minimally emissive for style back then and I did not change anything :-)
Gaussian splats in Blender 5.3! Had to take it for a quick spin ofc - here with a scene I did using a generated image, Minimax H3-Max camera control through fal.ai - converted to Gaussian splats with @KIRI_Engine_App.
It is possible to relight scenes with Gaussian Splats too, but it was pretty sloggish for me, even on a 5090. Could probably get better results with some optimization.
It is kinda crazy to see this unfold - from image, to video, to 3d to dynamic generative worlds.