Comment on Nvidia loses $500 bn in value as Chinese AI firm jolts tech shares

UnderpantsWeevil@lemmy.world ⁨1⁩ ⁨week⁩ ago

• Womble@lemmy.world ⁨1⁩ ⁨week⁩ ago

  It’s not a joke, it wont:

  Image

  source

  • Not_mikey@slrpnk.net ⁨1⁩ ⁨week⁩ ago

    It’s even worse / funnier in the app, it will generate the response, then once it realizes its about Taiwan it will delete the whole response and say sorry I can’t do that.

    If you ask it “what is the republic of china” it will generate a couple paragraphs of the history of China, then it’ll get a couple sentences in about the retreat to Taiwan and then stop and delete the response.

    source

    • Womble@lemmy.world ⁨1⁩ ⁨week⁩ ago

      In fairness that is also exactly what gpt clauude and the rest do for their online versions too when you hit their limits (usually around sex). IIRC they work by having a second LLM monitor the output and send a cancel signal if they think its gone over the line.

      source

      • JasSmith@sh.itjust.works ⁨1⁩ ⁨week⁩ ago

        Okay but one is about puritanical Western cultural standards about sex, and one is about government censorship to maintain totalitarian power. One of these things is not like the other.

        source

        • -> View More Comments
  • Bronzebeard@lemm.ee ⁨1⁩ ⁨week⁩ ago

    You missed the entire point of their comment

    source

    • Womble@lemmy.world ⁨1⁩ ⁨week⁩ ago

      Maybe they should have been clearer than saying people were joking about it doing something that it actually does if they wanted to make a point.

      source

      • Bronzebeard@lemm.ee ⁨1⁩ ⁨week⁩ ago

        People caring more about “China bad” instead of looking at what the tech they made can actually do is the issue.

        You needing this explicitly spelled out for you does not help the case.

        source

        • -> View More Comments
    • Eyekaytee@aussie.zone ⁨1⁩ ⁨week⁩ ago

      I’m slow, what’s the point? how does people joking about the fact China is censoring output explain

      why the US tech sector is absolutely fucked going into the next generation

      source

      • bennieandthez@lemmygrad.ml ⁨1⁩ ⁨week⁩ ago

        Because they care more about the model not parroting US state dept narratives than the engineering behind it.

        source
  • Smokeydope@lemmy.world ⁨1⁩ ⁨week⁩ ago

    Try an albliterated version of the qwen 14 or 32b distills. it will give you a real overview.

    source

    • Womble@lemmy.world ⁨1⁩ ⁨week⁩ ago

      Oh I hadnt realised uncensored version had started coming out yet, I definitely wil look into it once quantised versions drop.

      source

      • Smokeydope@lemmy.world ⁨1⁩ ⁨week⁩ ago

        huggingface.co/…/DeepSeek-R1-Distill-Qwen-14B-abl…

        source
  • Scolding7300@lemmy.world ⁨1⁩ ⁨week⁩ ago

    That’s just dumb. It at least doesn’t suppress that when provided with search results/refuses to search (at least when integrated in Kagi)

    source
  • UnderpantsWeevil@lemmy.world ⁨1⁩ ⁨week⁩ ago

    What training data did you use?

    source

    • Womble@lemmy.world ⁨1⁩ ⁨week⁩ ago

      ??? you dont use training data when running models, that’s what is used in training them.

      source

      • UnderpantsWeevil@lemmy.world ⁨1⁩ ⁨week⁩ ago

        DeepSeek open-sourced their model. Go ahead and train it on different data and try again.

        source

        • -> View More Comments
• bennieandthez@lemmygrad.ml ⁨1⁩ ⁨week⁩ ago

  deepseek bad because it doesn’t parrot my US State Dept narrative 😞

  source