Nvidia's custom LLM called ChipNeMo aims to accelerate the design phase of GPU production.
  • wahming@monyet.ccEnglish
    6·
    10 months ago

    Welp, that was a load of clickbait headline

    TLDR: They trained a LLM to help with knowledge transfer to new junior engineers.

  • hightime@lemmy.worldEnglish
    4·
    10 months ago

    Are they limited by design speed or manufacture? I was always under the impression that production is the bottleneck

    • wahming@monyet.ccEnglish
      1·
      10 months ago

      They’re two separate bottlenecks. As long as new designs aren’t complete, they keep manufacturing the old one, which is by definition state of the art.