Next.js Hacker News
Delta – Cut LLM inference costs 30-60% with lossless compression
6 points by nicksec 1 hour ago | discuss