The new global study, in partnership with The Upwork Research Institute, interviewed 2,500 global C-suite executives, full-time employees and freelancers. Results show that the optimistic expectations about AI’s impact are not aligning with the reality faced by many employees. The study identifies a disconnect between the high expectations of managers and the actual experiences of employees using AI.

Despite 96% of C-suite executives expecting AI to boost productivity, the study reveals that, 77% of employees using AI say it has added to their workload and created challenges in achieving the expected productivity gains. Not only is AI increasing the workloads of full-time employees, it’s hampering productivity and contributing to employee burnout.

  • Lvxferre@mander.xyz
    link
    fedilink
    English
    arrow-up
    86
    arrow-down
    3
    ·
    3 months ago

    Large “language” models decreased my workload for translation. There’s a catch though: I choose when to use it, instead of being required to use it even when it doesn’t make sense and/or where I know that the output will be shitty.

    And, if my guess is correct, those 77% are caused by overexcited decision takers in corporations trying to shove AI down every single step of the production.

    • bitfucker@programming.dev
      link
      fedilink
      English
      arrow-up
      12
      arrow-down
      1
      ·
      3 months ago

      I always said this in many forums yet people can’t accept that the best use case of LLM is translation. Even for language such as japanese. There is a limit for sure, but so does human translation without adding many more texts to explain the nuance in the translation. At that point an essay is needed to dissect out the entire meaning of something and not just translation.

      • Lvxferre@mander.xyz
        link
        fedilink
        English
        arrow-up
        6
        ·
        3 months ago

        I’ve seen programmers claiming that it helps them out, too. Mostly to give you an idea on how to tackle a problem, instead of copypasting the solution (as it’ll likely not work).

        My main use of the system is

        1. Probing vocab to find the right word in a given context.
        2. Fancy conjugation/declension table.
        3. Spell-proofing.

        It works better than going to Wiktionary all the time, or staring my work until I happen to find some misspelling (like German das vs. dass, since both are legit words spellcheckers don’t pick it up).

        One thing to watch out for is that the translation will be more often than not tone-deaf, so you’re better off not wasting your time with longer strings unless you’re fine with something really sloppy, or you can provide it more context. The later however takes effort.

        • bitfucker@programming.dev
          link
          fedilink
          English
          arrow-up
          2
          ·
          3 months ago

          Yeah, for sure since programming is also a language. But IMHO, for a machine learning model the best way to approach it is not as a natural language but rather as its AST/machine representation and not the text token. That way the model not only understands the token pattern but also the structure since most programming languages are well defined.

          • Lvxferre@mander.xyz
            link
            fedilink
            English
            arrow-up
            3
            ·
            3 months ago

            Note that, even if we refer to Java, Python, Rust etc. by the same word “language” as we refer to Mandarin, English, Spanish etc., they’re apples and oranges - one set is unlike the other, even if both have some similarities.

            That’s relevant here, for two major reasons:

            • The best approach to handle one is not the best to handle the other.
            • LLMs aren’t useful for both tasks (translating and programming) because both involve “languages”, but because LLMs are good to retrieve information. As such you should see the same benefit even for tasks not involving either programming languages or human languages.

            Regarding the first point, I’ll give you an example. You suggested abstract syntax trees for the internal representation of programming code, right? That might work really well for programming, dunno, but for human languages I bet that it would be worse than the current approach. That’s because, for human languages, what matters the most are the semantic and pragmatic layers, and those are a mess - with the meaning of each word in a given utterance being dictated by the other words there.

            • bitfucker@programming.dev
              link
              fedilink
              English
              arrow-up
              2
              ·
              3 months ago

              Yeah, that’s my point ma dude. The current LLM tasks are ill suited for programming, the only reason it works is sheer coincidence (alright, maybe not sheer coincidence, I know its all statistics and so on). The better approach to make LLM for programming is a model that can transform/“translate” a natural language that humans use to AST, the language that computers use but still close to human language. But the problem is that to do such tasks, LLM needs to actually have an understanding of concepts from the natural language which is debatable at best.