Pool spare GPU capacity to run LLMs at larger scale

(github.com) by i386 | Mar 24, 2026 | 0 comments on HN