finally got around to combining my threejs / webgpu fluid shader, ascii shader, and crt shader

WebGPU OpenLayers sounds simple, until pixel drift ruins your map. I broke down the root cause and the exact method we used in stratum‑map to achieve pixel‑perfect GPU rendering. shorturl.at/xGdja #WebGPU #OpenLayers #GIS #Geospatial #WebDev #TypeScript #OpenLayers #WebDev

WebGPU OpenLayers sounds simple, until pixel drift ruins your map. I broke down the root cause and the exact method we used in stratum‑map to achieve pixel‑perfect GPU rendering. shorturl.at/xGdja #WebGPU #OpenLayers #GIS #Geospatial #WebDev #TypeScript #OpenLayers #WebDev

先日リリースした vrm-game-starter を活用してみた！！楽しい！！！ #MadeWithVRM #threejs #webgpu #vroid

I got WebGPU support working so now my dumb little social MMO game runs fully in the browser (and on Newgrounds hopefully)

I added animation support to my gltF importer. However, I'm starting to feel more motivated to work on my 3D modeler again. So for now, I think this terrain/WebGPU game idea will be put on hold. #gamedev #indiedev

arxiv.org/abs/2602.11685 Case in point. This was benchmarked against DRACO. Fable-level intelligence for research tasks. That's not surprising - research is one of the simplest and most obvious use cases for cheap models. Even Claude Code delegates to Haiku when you use fable for agents What level of comparable intelligence does it give on complex coding tasks? Converting a SIGGRAPH paper to a working Three.JS WebGPU demo? Porting an application from one language to another without bloat? I'd hazard a guess it's incredibly far off

I gave Fable 5 one job: write custom WebGPU kernels for Gemma 4 inference. It climbed to 84 tok/s, then hit a wall, insisting further optimization was impossible. Hours later, Anthropic rolled back invisible LLM development safeguards, and it hit 255 tok/s. The next day, access to Fable 5 was suspended globally.

I gave Fable 5 one job: write custom WebGPU kernels for Gemma 4 inference. It climbed to 84 tok/s, then hit a wall, insisting further optimization was impossible. Hours later, Anthropic rolled back invisible LLM development safeguards, and it hit 255 tok/s. The next day, access to Fable 5 was suspended globally.

📰 AI in je browser, zonder server: hoe WebGPU en kleine modellen de inferentie verplaatsen Client-side AI dr 🔗 hetlaatsteainieuws.nl/achter… #AI #ArtificieleIntelligentie #AInieuws #AIInnovatie #ClientSideAi #Webgpu #OnnxRuntimeWeb

→ liquid-dom, WebGPU liquid-glass rendering with React and Three.js bindings: github.com/AndrewPrifer/liqu…

The clips are real, from my F1 team radio dataset: 14,681 of them across 149 Grand Prix and 43 drivers (2018-2025). The model runs entirely in your browser! Tiny Aya ONNX Transformers.js WebGPU. A fun insight into the messiness of fast-paced, stressful communication!