Via @rodhilton@mastodon.social

Right now if you search for “country in Africa that starts with the letter K”:

  • DuckDuckGo will link to an alphabetical list of countries in Africa which includes Kenya.

  • Google, as the first hit, links to a ChatGPT transcript where it claims that there are none, and summarizes to say the same.

This is because ChatGPT at some point ingested this popular joke:

“There are no countries in Africa that start with K.” “What about Kenya?” “Kenya suck deez nuts?”

    • sibbl@feddit.de
      link
      fedilink
      English
      arrow-up
      21
      arrow-down
      4
      ·
      1 year ago

      This sounds like “Hmm, maybe calculators won’t replace mathematicians.” to me.

      Not sure why it should replace them. They’ll co-exist. Sometimes you can do the math in your brain and for other things you use calculators. Results of calculators can still be wrong it you don’t use them properly.

      • Hiccup@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        4
        arrow-down
        1
        ·
        1 year ago

        What a blast from the past! AI gives me second hand embarrassment for the people that work and get paid on this/for this shit. It’s the second (or third) coming of crypto and NFTs. Just junk software that fixes nothing and that wastes people’s time.

        • regalia@literature.cafe
          link
          fedilink
          English
          arrow-up
          3
          arrow-down
          3
          ·
          1 year ago

          LLMs have absolutely tons of actual applications that it’s crazy, and it already changed the world. Crypto and NFTs were just speculative assets that were trying to solve a problem that didn’t exist. LLMs have already solved a huge amount of real world problems, and continues to do so.

          • ALostInquirer@lemm.ee
            link
            fedilink
            English
            arrow-up
            3
            arrow-down
            1
            ·
            1 year ago

            LLMs have already solved a huge amount of real world problems, and continues to do so.

            Would you happen to have some examples? I don’t disagree that LLMs have more of a use case and application than the cryptoNFT misapplications of blockchain, but I’m honestly not familiar with where they’ve solved real world problems (and not just demonstrated some research breakthroughs, which while impressive in their own respect do not always extend to immediate applications).

            • regalia@literature.cafe
              link
              fedilink
              English
              arrow-up
              1
              ·
              1 year ago

              They are augmenting search engines. They write and can digest articles. GitHub co-pilot has been a pretty big deal. It can act like a personal tutor to walk you through math problems, code, language, whatever. Building trust LLM search for medical information without hallucinating. It can do financial analysis and all sorts of stuff with that. It’s replacing a huge amount of jobs. This stuff is like all over the news, I’m not sure if you’ve lived under a rock this whole time. For very little effort you can find an endless more amount of examples. It’s creating real world use cases daily, so fast that it feels impossible to keep up.

              • ALostInquirer@lemm.ee
                link
                fedilink
                English
                arrow-up
                1
                ·
                1 year ago

                This stuff is like all over the news, I’m not sure if you’ve lived under a rock this whole time.

                Oh, no, I’ve heard it, I’m just skeptical of their accuracy and reliability, and that skepticism has been borne out by the numerous reports of glitching (“hallucinations” as they insist on calling them, in furtherance of their inappropriate personification of the technology). Moreover, I’ve found their mass theft of others’ work to further call into question the creators’ trustworthiness, which has only been compounded by their overselling of their technology’s capabilities while simultaneously suggesting it’s just untenable to log & cite all the sources that they push into it.

                It can supposedly do all you describe, but it can’t effectively credit its sources? It can tutor but it can’t even keep basic information straight? Please. It’s impressive technology, but it’s being overblown because the markets favor exaggeration to facts, at least as long as people can be kept enamored with the fantasy they spin.

                • regalia@literature.cafe
                  link
                  fedilink
                  English
                  arrow-up
                  1
                  ·
                  1 year ago

                  With all that silicon valley, it’ll probably be pushed more to do those things regardless of its hallucinations and accuracy lol.

    • mindlight@lemm.ee
      link
      fedilink
      English
      arrow-up
      12
      arrow-down
      8
      ·
      1 year ago

      I’m pretty sure a lot of people said something like “Hmm,maybe the automobile won’t replace horses.” after reading about the first car accidents.

      • regalia@literature.cafe
        link
        fedilink
        English
        arrow-up
        15
        ·
        1 year ago

        Finding sources will always be relevant, and so will finding links to multiple sources (search results). Until we have some technological breakthrough that can fact check LLM models, it’s not a replacement for objective information, and you have no idea where it’s getting its information. Figuring out how to calculate objective truth with math is going to be a tough one.