Smaller, faster, safer: running Kimi and GLM at scale

hackernews 2026-08-04 원문 보기 ↗