> llama.cpp via Vulkan (AMD / Intel / NVIDIA) or CPU fallback

I got excited about someone paying attention to intel. Oh well.