News for hackers. Stuff that matters.
Re:Meta disbanded its Responsible AI team by l33t7332273 on Sunday November 19, 2023 @01:08AM (#38329633)
If I publish a massive quantity of source code — to the point that it’s very expensive to compile — it’s still open source.

If the training data and model training code is available then it should be considered open, even if it’s hard to train.

Re:Meta disbanded its Responsible AI team by nextaccountic on Sunday November 19, 2023 @02:06AM (#38329990)
> the training data

This will never be fully open

Re:Meta disbanded its Responsible AI team by l33t7332273 on Sunday November 19, 2023 @02:35AM (#38330132)
Maybe not for some closed models. That doesn’t mean truly open models can’t exist.
Re:Meta disbanded its Responsible AI team by earthnail on Sunday November 19, 2023 @04:00AM (#38330658)
I doubt you’d say that if one run of compiling the code would cost you $400M.
Re:Meta disbanded its Responsible AI team by Ericson2314 on Sunday November 19, 2023 @10:08PM (#38341787)
If it was only feasible for a giant corporation to compile the code, I would consider it less than open source.