Pinned Tweet
MarsCycler3D is an completed app for @spaceAaps based on @babylonjs that shows Earth-Mars travel. Can be used as a presentation or flight simulator. Paths were calculated by astrophys. M.Quartly from @QMUL. @elonmusk как тебе такое?
neuromodern.github.io/MarsCy…
#spaceX #webgl
You've all seen this Astra 6 case. Using LLM to transfer pixel color is like cracking nuts with a sledgehammer. It's sad because people actually think AI is drawing portraits, and they're burning tokens when a simple algorithm could do the trick.
Building #Web3 Escrow Smart Contract with third side Arbiter medium.com/@neuromodern/buil…
For a small benchmark, I asked AI models to complete my code of random procedural-generative city, written before 2022, with a detailed prompt about realism. He discarded all the little details on the source like signs, balconies, and fences! Here's what came out of it:
The bottleneck of agent development is that LLM is a text-based model. When you stream a screen or screenshots to it, it first converts them to text. Where work with space, volumes, animation , and thousands of 3D objects is required, it won't detect errors.
The GLSL shader that do a cyberpunk style volumetric pixel particles trail.
shadertoy.com/view/f32XDw - extruding Z
shadertoy.com/view/fX2XDw - extruding -Z
#glsl #shader #postFX #shaderToy #WebGL
NeuroModern retweeted
"MeshFlow: Efficient Artistic Mesh Generation via MeshVAE and Flow-based Diffusion Transformer"
TL;DR: learns a continuous mesh latent space and generates vertices and connectivity in parallel with flow matching, producing quality 3D meshes up to 18× faster than autoregressive.
NeuroModern retweeted
Microsoft has released a 4B parameter model that turns any image into a 3D asset in 3 seconds.
It uses a new geometry format called O-Voxel that converts to a textured mesh in under 100ms on CUDA.
Outputs GLB files with full PBR textures, ready for Blender, Unity, and Unreal.
100% Open Source.
NeuroModern retweeted
D-Rex by NVIDIA.
photorealistic digital humans that move and look correct under any lighting condition.
- built on EVA +Cosmos DiffusionRenderer
- supports facial expression,3D pose control under novel HDR lighting.
so, Hollywood-grade relightable characters without the usual technical headaches of traditional 3D rendering pipelines.
vcai.mpi-inf.mpg.de/projects…
NeuroModern retweeted
【革命】動画から3Dモーションを抽出！ComfyUIの新機能が凄すぎる
動画から人物の動きを抜き出し、高品質な3Dアニメーションに変換できる「Mocap Surgeon」が登場しました！✨
注目の神機能はこちら：
・Jitter Filtering：動きのガタつきを数理的に除去し、滑らかさを実現
・Slerp（球面線形補間）：関節のねじれを自然にブレンドして手動修正が可能
・Time-Travel Onion Skinning：前後の軌跡を表示し、完璧なポージングをサポート
作成したモーションデータは.glb形式で書き出してControlNetへ流用できます。実写ベースのAIアニメ制作や、3Dキャラクター動画の量産が驚くほど効率化されます！🔥
#ComfyUI #MocapSurgeon
NeuroModern retweeted
GPT2 + Seedance 2.0 → Capoeira Sequence.
Testing for more complex choreography.
(Full workflow + prompts in thread)
PROCESS:
+ GPT2: Choreography diagram
+ GPT2: Camera direction diagram
+ Midjourney: Setting base image
+ Seedance 2.0 Omni-Ref: Video gen
Didn't follow 100%.
Probably needs more than a 15s gen.
NeuroModern retweeted
📢📢📢introducing 𝐏𝐨𝐰𝐞𝐫 𝐅𝐨𝐚𝐦
A 3D representation that can be ray traced or rasterized in real time, with NO COMPROMISE in quality.
- Project: powerfoam.github.io
- arXiv: arxiv.org/abs/2604.24994
Rasterized at 3DGS-class FPS
Ray traced at Radiant Foam speeds
NeuroModern retweeted
A bunch of folks have been building machine learning models that turn a photograph into a 3D environment made of Gaussian splats (read: blobs of color floating in space).
Cool technology & a very admirable effort. But marketing these as "world models" seems wrong.
More accurate would be to say that they are a riff on the broader class of image-conditioned 3D generators, with a somewhat different flavor of condition image and output representation.
As far as world modeling, they don't make great predictions about how the natural world looks or behaves. (Even for, say, a chair behind a table.)
Again: I love the technology. Super cool creative stuff. I don't love the marketing and hype around it.
NeuroModern retweeted
Microsoft presents World-R1
Reinforcing 3D Constraints for Text-to-Video Generation
paper: huggingface.co/papers/2604.2…
NeuroModern retweeted
Took it further with @omma_ai, (@splinetool ) now it works on video, and everything working in a Web Browser.
Drop a video, Depth Anything v2 calculates the depth map frame by frame, and Three.js renders a fully lit 3D mesh in real time.
Depth estimation on video in the browser
+Automatic baking pass
+ Dynamic lighting reacting to the geometry
#vibecoding #threejs #webgl
NeuroModern retweeted
Neural motion synthesis in Threejs/WebGPU,
even runs on my 7y old Iphone!
try it here: motionsynth.sweriko.com
sauce: github.com/sweriko/ai4anim-w…
NeuroModern retweeted
🚀We just released Asset Harvester, an image-to-3D model and end-to-end pipeline that extracts real object assets from autonomous driving videos!
🌐 Website: research.nvidia.com/labs/sil…
💻 Code: github.com/nvidia/asset-harv…
[1/5]
#AssetHarvester #AVSimulation #WorldModel #AutonomousDriving
NeuroModern retweeted
The @playcanvas team has solved collision for 3D Gaussian splats. Install splat-transform via NPM to get a CLI tool + library that can output high quality voxel-based collision. Here you can see a splat navigated in first person mode with voxel rendering toggled on/off. 🧵