3 ms·

Google's TurboQuant AI-compression algorithm can reduce LLM memory usage by 6x

• redanddead 7mo ago
  You'd think it'd be bigger news on hn
  • axiologist 7mo ago
    See https://news.ycombinator.com/item?id=47513475 https://news.ycombinator.com/item?id=47513475 from two days ago.
    • redanddead 7mo ago
      thank you sir, I missed it