There is a need for highly optimized, browser-native vector math libraries that don't rely on bloated frameworks. Build a modular WebGPU-accelerated vector quantization library to enable local-first AI search and embedding storage in static web apps.