• CriticalResist8@lemmygrad.ml
    link
    fedilink
    arrow-up
    14
    ·
    30 days ago

    It’s always the same story with the west when news of chinese tech hits us.

    First it’s “they’ll need at least ten years to catch up! By then we’ll be soaring a century ahead!”

    then “well okay they made a huge leap earlier than expected, but it’s not actually good. they still have this or that bottleneck to close. I don’t see them doing it any time soon, it took us 50 years to do it.”

    finally “oh, it’s only been 2 years and they’re already at our level. well um, chinese quality! ban imports! taiwan!”

    • ☆ Yσɠƚԋσʂ ☆@lemmygrad.mlOP
      link
      fedilink
      arrow-up
      6
      ·
      30 days ago

      I’m really curious to see how they’ll go about banning open models in the US. Banning services from China will be easy enough I imagine, but it’s going to be really hard to justify why people can’t run models themselves, and now there are a bunch of companies relying on them too, so you have a powerful business lobby invested in using them. And Microsoft already tried doing something similar with Linux back in the day where they wanted to basically ban open source because Linux was eating their lunch in the server market.

      • PeeOnYou [he/him]@lemmygrad.ml
        link
        fedilink
        arrow-up
        2
        ·
        29 days ago

        yeah i hear a lot of good things about it… is there another harness like claude cowork that you can plug glm or any other llm into? cowork has so much of my customization set up and it has access to most of the apps i use… i don’t know if there’s another tool out there with comparable functionality that you can use with other models. it would be awesome if there were

        • ☆ Yσɠƚԋσʂ ☆@lemmygrad.mlOP
          link
          fedilink
          arrow-up
          3
          ·
          29 days ago

          I haven’t used cowork, so can’t compare with it, but Pi seems to be really popular for customizing. Although, I do find it too minimal myself. I find OpenCode is still the best general purpose harness for open models.