Inspired by @shuding to try WebGPU in the browser.
Built an in-browser fuzzy search engine: on 500k–1M+ items, it runs ~5x–10x faster than CPU (~6.9ms vs 63ms on 1M) without locking the UI main thread.
Local-first compute is really exciting.
I trained a small model to do syntax highlighting in the browser with GPU.
Meet gpu-lexer from Vercel Labs: Small (27.5KB), fast (runs on WebGPU), and language-agnostic (model guesses the syntax).
gpu-lexer.vercel.app
It is experimental and built for learning!
