Comment on Flailing OpenAI Calls for Ban on Chinese AI

tonytins@pawb.social ⁨5⁩ ⁨weeks⁩ ago

• uranibaba@lemmy.world ⁨5⁩ ⁨weeks⁩ ago

  Isn’t DeepSeek open source? Is there a need to reverse engineer it?

  source

  • Dran_Arcana@lemmy.world ⁨5⁩ ⁨weeks⁩ ago

    “Open source” in ML is a really bad description for what it is. “Free binary with a bit of metadata” would be more accurate. The code used to create deepseek is not open source, nor is the training datasets. 99% of “open source” models are this way. The only interesting part of the open sourcing is the architecture used to run the models, as it lends a lot of insight into the training process, and allows for derivatives via post-training

    source

    • cyd@lemmy.world ⁨5⁩ ⁨weeks⁩ ago

      Deepseek actually released a bunch of their infrastructure code, including the infamous tricks for making training and interference more efficient, a couple of weeks ago.

      source
    • Naevermix@lemmy.world ⁨5⁩ ⁨weeks⁩ ago

      It certainly is a lot more open source than OpenAI, that’s for sure.

      source
  • tonytins@pawb.social ⁨5⁩ ⁨weeks⁩ ago

    Yes, and no. Yes in that they’ve released the research papers, pretrained parameters or weights of the model itself, but the training data or other vital components. Which is more than I can say for “OpenAI.” But no in that it doesn’t include training data or other critical components. Luckily, they’ve shown how they did it which makes it easy for anyone else to reverse engineer the process. That’s what Altman is afraid of.

    source

    • cyd@lemmy.world ⁨5⁩ ⁨weeks⁩ ago

      They released the major components of their training and interference infrastructure code a couple weeks ago.

      source