Explicitly set subgroup sizes in WGSL compute shaders with WebGPU. Fine-tune AI workloads on specific platforms using the new subgroup_size attribute. See the full release details →

Cyberpunk 2077 is now running directly in a web browser. An unofficial fan project has recreated the game for the browser using WebGPU and WebAssembly. Unlike cloud gaming, the game is processed on your own device not streamed from a server. It can run on some PCs and even certain Android devices The project is still experimental, with long loading times and issues such as freezes or missing audio reported by some users. This not an official CD Projekt RED port, they likely will shut it down soon

Brave is adding stronger protection against a type of online tracking called browser fingerprinting. The update focuses on WebGL and WebGPU, technologies that websites use for things like 3D graphics, games and other demanding visuals. The problem is that they can also reveal details about your device, including its graphics hardware and capabilities. Websites can combine this information with other data to create a unique “fingerprint” that can be used to recognize your device. Brave is changing how this information is shared, making it harder for websites to create a reliable fingerprint

Shareholders, Phase 1 of the Net Forge public release is now live. The complete game engine is open source and live on npm. 16 packages published straight from a public repo with npm provenance. What is in it: - the game engine (three.js, WebGPU with a WebGL2 fallback) - the shared simulation that runs the same in the browser and on the server - an asset library of 530 CC0 models and sounds across six art styles - a runtime for card, idle, tower and collector games It also ships the 10 templates every Net Forge game starts from: survivor co-op, arena shooter, tower defense, dungeon crawl, idle, platformer, card battler, monster collector, kart racer and party game. Each one is a complete game you can read, change and run. Developers can start today. Clone the repo, build it, and any template plays on your own machine with a local game server and play money. Fork a template, change the rules, swap the art style, and see it running in a few minutes. What comes next: the Kit, which connects a game to its token, store and rewards, goes out once the game contracts are live. Until then everything here runs with play money, while we fine tune the studio for the public launch. If you build something on it, show us. Issues and pull requests are open on the repo. More Soon.

Google open-sourced the GPU engine running Local AI across millions of devices. It’s called ML Drift. It gives developers one GPU layer across: 🍎 Metal + 🌐 WebGPU + 🤖 OpenCL / OpenGL ES + 🖥️ Windows + Linux via Dawn And it’s built for LLMs. ML Drift changes its kernels depending on what the model is doing such as 📥 prefill → compute-optimized ✍️ decode → bandwidth-optimized 🧠 custom KV-cache layout 💾 memory reuse Google reports up to 12% lower memory overhead in its Gemma tests. Google says ML Drift runs today across millions of devices in Chrome, YouTube Shorts, Photos and Meet. Chrome uses it to accelerate Gemini Nano locally. it’s Apache 2.0 so developers now have one open GPU foundation for running AI across phones, browsers, Macs and PCs. This is something to test! Link in ALT.