I built a pure WGSL LLM engine to run Llama on my Snapdragon laptop GPU

(github.com) by Beledarian | Apr 11, 2026 | 1 comments on HN