• cyd@lemmy.world
    link
    fedilink
    English
    arrow-up
    43
    arrow-down
    21
    ·
    edit-2
    1 year ago

    Ideally, they’d just blow the entire $330M training an LLM, and release the weights. In reality, much of that money will probably go into paying salaries, various smaller research projects, etc.

      • AeroLemming@lemm.ee
        link
        fedilink
        English
        arrow-up
        18
        arrow-down
        1
        ·
        1 year ago

        I think what they mean is that ideally, the money wouldn’t be going towards a few executives’ multi-millionaire salaries.

      • cyd@lemmy.world
        link
        fedilink
        English
        arrow-up
        12
        arrow-down
        1
        ·
        edit-2
        1 year ago

        The context is that LLMs need a big up front capital expenditure to get started, because of the processor time to train these giant neural networks. This is a huge barrier to the development of a fully open source LLM. Once such a foundation model is available, building on top of it is relatively cheaper; one can then envision an explosion of open source models targeting specific applications, which would be amazing.

        So if the bulk of this €300M could go into training, it would go a long way to plugging the gap. But in reality, a lot of that sum is going to be dissipated into other expenses, so there’s going to be a lot less than €300M for actual training.

        • interceder270@lemmy.world
          link
          fedilink
          English
          arrow-up
          4
          ·
          1 year ago

          Is there any way we can decentralize the training of neural networks?

          I recall something being released awhile ago that let people use their computers for scientific computations. Couldn’t something similar be done for training AI?

          • Mahlzeit@feddit.de
            link
            fedilink
            English
            arrow-up
            4
            ·
            1 year ago

            There is a project (AI Horde) that allows you to donate compute for inference. I’m not sure why the same doesn’t exist for training. I think the RAM/VRAM requirements just can’t be lowered/split.

            Another way to contribute is by helping with training data. LAION, which created the dataset behind Stable Diffusion, is a volunteer effort. Stable Diffusion itself was developed at a tax-funded public university in Germany. However, the cost of the processing for training, etc. was covered by a single rich guy.

          • Sanyanov@lemmy.world
            link
            fedilink
            English
            arrow-up
            1
            ·
            1 year ago

            Btw yes! Why not include such project in something like BOINC and let people help training free AI?

          • Dojan@lemmy.world
            link
            fedilink
            English
            arrow-up
            2
            arrow-down
            6
            ·
            1 year ago

            Folding at home.

            I dunno. I wouldn’t lend my spare power to put people out of a job.

      • 5BC2E7@lemmy.world
        link
        fedilink
        English
        arrow-up
        3
        arrow-down
        9
        ·
        1 year ago

        ?? Do you know people with enough time qualifications and money that are willing to work for free? I haven’t.

              • 5BC2E7@lemmy.world
                link
                fedilink
                English
                arrow-up
                1
                arrow-down
                6
                ·
                1 year ago

                I don’t know him but I do know that people have been banned for criticizing Lemmy administration wrt potential monetization.

                I don’t want to be banned so I will not comment further on the admin.

                I don’t deny that there are software engineers that would work for free but they are not common.

                • iByteABit [he/him]@lemm.ee
                  link
                  fedilink
                  English
                  arrow-up
                  3
                  arrow-down
                  1
                  ·
                  1 year ago

                  The potential monetization being the donations to give them a living wage? What exactly is the criticism about, that they shouldn’t get a living wage through donations and should rather make the platform paid or ruin it with ads?

                  Why do so many people get off on attacking the idea of FOSS, much of the software that is running our everyday lives is supported through FOSS (and lots of them are also being donated to so that the devs can afford to put food on their tables).

                  There are also many devs (on Lemmy as well) that contribute a lot without being paid, simply because they like the project, want to make it better, and want to learn by doing.

                  • 5BC2E7@lemmy.world
                    link
                    fedilink
                    English
                    arrow-up
                    1
                    arrow-down
                    2
                    ·
                    11 months ago

                    you must be pretty dense if you expect me to respond after I said I’ll not comment further.

        • interceder270@lemmy.world
          link
          fedilink
          English
          arrow-up
          10
          arrow-down
          1
          ·
          1 year ago

          I don’t think many people were paid to work on the Fediverse.

          Or emulators.

          Or most free software.

          • Honytawk@lemmy.zip
            link
            fedilink
            English
            arrow-up
            2
            arrow-down
            6
            ·
            1 year ago

            That isn’t a job to them though, it is more like a hobby.

            If you want peoples undivided attention, you will have to pay them, no matter how utopian your vision.

            Which you can easily afford with 330 million funding.