• archomrade [he/him]@midwest.social
    link
    fedilink
    English
    arrow-up
    0
    ·
    17 hours ago

    People keep complaining that AI is hallucinating information not in its source material, but here it is strictly limiting itself to the source and even citing its findings and yet everyone is still upset?

    Curious

    • droans@midwest.social
      link
      fedilink
      arrow-up
      0
      ·
      1 day ago

      Back when it was 90% junk, 5% okay, 4% great… But also that 1% that was absolutely worth it.

      Back when Schwarzenegger and T-Pain were basically the highlights. Victoria running the AMAs - I mean that had to be one of the biggest losses. That bird scientist dude and then the drama after he was caught alt upvoting.

    • CharmOffensive@lemm.ee
      link
      fedilink
      arrow-up
      0
      ·
      1 day ago

      Oh, you mean like “the narwhal bacon’s at midnight?” Or “Anne frankly I did nazi that coming” or “somebody summon unidan!”?

  • nthavoc@lemmy.today
    link
    fedilink
    English
    arrow-up
    0
    ·
    2 days ago

    Most AI’s are really just digital dumbasses persuading people it’s not a fancy chatbot.

  • mynameisigglepiggle@lemmy.world
    link
    fedilink
    arrow-up
    0
    ·
    2 days ago

    I know there is a lot of hate for AI here but it has its uses.

    However. I still haven’t found a use for Gemini. What a massive steaming pile of shit llm google have cooked up. It is utterly useless gutter trash that nobody could possibly love.

    Pull your finger out google. What the fuck are you doing, you stupid pieces of dog shit?

    • Buglefingers@lemmy.world
      link
      fedilink
      arrow-up
      0
      ·
      2 days ago

      One of the features I thought it could be wicked useful for is printer clarity. When scanning or copying bad/old papers, hand written, etc. It could clean up, focus, or turn handwriting to print. That would actually be a good use for it IMO, ofc it’ll fuck shit up sometimes but it’ll probably be better than nothing at all

      • Acters@lemmy.world
        link
        fedilink
        arrow-up
        0
        ·
        2 days ago

        My phone is starting to take better document pictures than my printer. Especially if I hold my phone steady at the right height. I bet if I used a 3D printer to create a stand to hold it for me. Heck I could buy a giant bucket or rack or whatever and bright high color index lights to further improve image clarity to ungodly levels. Still, the Samsung camera app with the AI features already being preloaded that will likely be improved in the next 5 years will just get better. Printers and scanners are completely a pita.

    • rumba@lemmy.zip
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 days ago

      Even Llama gives better results than Gemini. They’re just perpetually behind everyone else. It’s like they took GPT3 and tried to bolt one their search results, let someone forgot to make their search results not suck again for their own internal tool.

    • phoneymouse@lemmy.world
      link
      fedilink
      arrow-up
      0
      ·
      edit-2
      2 days ago

      Mods on c/politicalmemes took down a post I made saying that it was not a meme (spam was what they called it, it was not spam). When I pointed out that very few of the posts in that community are memes, they banned me.

      Anyway, be careful out there.

      • WrenFeathers@lemmy.world
        link
        fedilink
        arrow-up
        0
        ·
        2 days ago

        Political Memes mod here. You had a post removed because it broke the rules of the community, which is not a big deal at all… However, you chose to then get angry and report a bunch of other posts that did not break the rules and therefore you were banned for report abuse- which IS a big deal.

        If you had simple messaged one of us and pointed out where you thought you were wronged, I or any of the other mods would have had no problem hearing you out.

        Here is your modlog in the event that you have forgotten how things went down.

        Hope this helps!

    • leadore@lemmy.world
      link
      fedilink
      arrow-up
      0
      ·
      2 days ago

      I’ve been able to reproduce some, like the “how to carry <insert anything here> across a river” one where it always turns it into the fox, goose and grain puzzle.

      But generally on anything that’s gone viral, by the time you try to reproduce it someone has already gone in and hard-coded a fix to prevent it from giving the same stupid answer going forward.

    • dingus@lemmy.world
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 days ago

      Yeah, I never get these strange AI results.

      Except, the other day I wanted to convert some units and the AI results was having a fucking stroke for some reason. The numbers did not make sense at all. Never seen it do that before, but alas, I did not take a screenshot.

      • kadup@lemmy.world
        link
        fedilink
        arrow-up
        0
        ·
        2 days ago

        Those LLMs can’t handle numbers, they have zero concept of what a number is. They can pull some definitions, they can sorta get very basic arithmetic to work in a limited domain based on syntax rules, but it will mess up most calculations. ChatGPT tries to work around it by recognizing the prompt is related to math, passing it to a more normal Wolfram-Alpha style algorithm, and then using the language model to format the reply into something more appealing, but even this approach often fails because if the AI gets confused for any reason it will feed moronic data to the maths algorithm.

          • ImplyingImplications@lemmy.ca
            link
            fedilink
            arrow-up
            0
            ·
            2 days ago

            LLMs don’t verify their output is true. Math is something where verifying its truth is easy. Ask an LLM how many Rs in strawberry and it’s plain to see if the answer is correct or not. Ask an LLM for a summary of Columbian history and it’s not as apparent. Ask an LLM for a poem about a tomato and there really isn’t a wrong answer.

      • synae[he/him]@lemmy.sdf.org
        link
        fedilink
        English
        arrow-up
        0
        ·
        2 days ago

        The “sauce vs dressing” one worked for me when I first heard about it, but in the following days it refused to give an AI answer and now has a “reasonable” AI answer

        The original, if you haven’t seen it:

        1000020167

      • Zorsith@lemmy.blahaj.zone
        link
        fedilink
        English
        arrow-up
        0
        ·
        2 days ago

        At one point they were packing a shitload of usb ports onto the IO panel, 5 stacks of 4 ports wouldnt surprise me

  • MirthfulAlembic@lemmy.world
    link
    fedilink
    English
    arrow-up
    0
    ·
    2 days ago

    I mean, if you’re googling that without even providing a model number, I can excuse the AI choosing to show it. It’s not a mind reader.