With LLaMA V2, Meta may be trying to benefit from the open-source community, similar to what Google has done with Android.

The Financial Times, citing three sources familiar with the project, reports that Meta wants to launch a commercial AI model to compete with OpenAI, Microsoft, and Google. The model is said to generate language, code, and images.

It may be a new variant of Meta’s LLaMA, a large language model used in numerous open-source projects. LLaMA v1 has only been released under a research license and therefore may not be used directly for commercial purposes. However, replicas exist.

Meta CEO Mark Zuckerberg has already announced that a new AI model is in the works, which could be LLaMA v2 or under a different name. Meta wants to use the model for its services and offer it to external interested parties, according to Zuckerberg. Special attention is safety.

      • Lukecis@lemmy.fmhy.ml
        link
        fedilink
        English
        arrow-up
        3
        ·
        1 year ago

        Exactly, everyone’s goal is to ‘be the best at X’ so saying that LLaMA wants to usurp GPT-4’s dominance is just stating the obvious.

        • Martineski@lemmy.fmhy.mlOPM
          link
          fedilink
          English
          arrow-up
          1
          ·
          1 year ago

          I want an ai tool that automatically chooses the best titles for posts with no hype stuff. Current models are not capable to understand the full context of articles/papers to then choose the title that will work well in the real world.

          • Lukecis@lemmy.fmhy.ml
            link
            fedilink
            English
            arrow-up
            2
            ·
            1 year ago

            There will never be a “best title” for things, since its all subjective and you never know what will gather more clicks or interaction than another- an algorithm for choosing the most likely relevant title may exist however its never guaranteed to perform better than another one.

            • Martineski@lemmy.fmhy.mlOPM
              link
              fedilink
              English
              arrow-up
              1
              ·
              1 year ago

              Nope, titles have qualities that could be considered universally good or universally bad so if you know your shit then you can make consistently very good titles.

              • Lukecis@lemmy.fmhy.ml
                link
                fedilink
                English
                arrow-up
                1
                ·
                1 year ago

                Sure but it’s never going to be possible to know the absolute perfect title for any given topic, since public opinion and click through rate is a nebulous unknowable thing which could just as likely respond to a perfectly algorithmically selected title negatively as opposed to a joke title given on a whim could garner far more interaction.

                • Martineski@lemmy.fmhy.mlOPM
                  link
                  fedilink
                  English
                  arrow-up
                  1
                  ·
                  1 year ago

                  never going to be possible to know the absolute perfect title for any given topic

                  And I’m not asking for that