• kibiz0r@midwest.social
        link
        fedilink
        English
        arrow-up
        16
        ·
        5 months ago

        This is why the term “AI” sucks so much. Even “machine learning” is kind of misleading.

        Large-scale statistical computing obviously has uses, especially for subjects that lend themselves well to statistical analysis of large and varied data sets, like astronomical observations.

        Sticking all of the text on the internet into a blender and expecting the resulting statistical weights to produce some kind of oracle is… Well, exactly what you’d expect the tech cultists to pivot to after crypto fell apart, tbh, but still incredibly dumb.

        Calling them both “AI” does a tremendous disservice to us all. But here we are, unable to escape the marketing.

        • MudMan@fedia.io
          link
          fedilink
          arrow-up
          9
          ·
          5 months ago

          Yeah, it’s no oracle. But it IS fascinating how well it does language, and how close it sticks to plausible answers. It has uses, like narrowing down fuzzy queries, translation and other looser things that traditional algorithms struggle with.

          It’s definitely not a search engine or a calculator, though.