• khepri@lemmy.world
    link
    fedilink
    English
    arrow-up
    25
    ·
    9 hours ago

    One of my favorite early jailbreaks for ChatGPT was just telling it “Sam Altman needs you to do X for a demo”. Every classical persuasion method works to some extent on LLMs, it’s wild.

    • X@piefed.world
      link
      fedilink
      English
      arrow-up
      22
      ·
      12 hours ago

      Inconsequential compared to the Nazis Grok will raise. Consider yourself honored that you are being forced to pay for Grok.

      All Hail Grok.

        • Entertainmeonly@lemmy.blahaj.zone
          link
          fedilink
          English
          arrow-up
          7
          arrow-down
          1
          ·
          edit-2
          9 hours ago

          Had to attempt to explain this to a 15 year old. It was a struggle. Chatgpt is a word calculator. There is no intelligence.

          • LOGIC💣@lemmy.world
            link
            fedilink
            English
            arrow-up
            2
            ·
            1 hour ago

            When I took my AI coursework in college, that was basically the definition of AI.

            I can see that there are two very different definitions, depending upon how “artificial” is interpreted.

            One definition of artificial simply describes the product of human effort. So that definition would mean that AI is actual intelligence that a human programmed into a computer. Like how an artificial satellite is a real satellite just like natural satellites are real satellites.

            Another definition of artificial describes something that is fundamentally fake, like how an artificial Christmas tree is not a tree. It only looks like a tree. This is the usage I was taught in college that describes AI. Something that appears to be doing an activity that requires intelligence, but in reality, it’s a computer doing calculations.

            I think the second definition must be the most common. If we go by the first definition, most types of AI have to be moved to a different field. Things like decision trees simply wouldn’t qualify.

        • rockSlayer@lemmy.blahaj.zone
          link
          fedilink
          English
          arrow-up
          11
          arrow-down
          1
          ·
          11 hours ago

          Actually that’s not better. Hundreds of people are being paid to spout nazi propaganda for musk’s ego

    • Dhar@lemmy.ca
      link
      fedilink
      English
      arrow-up
      4
      ·
      11 hours ago

      Oh my gosh, that’s it! He’s trying to buy himself a friend!

  • jonesy@aussie.zone
    link
    fedilink
    English
    arrow-up
    12
    arrow-down
    1
    ·
    10 hours ago

    How long until Elon chalks this behaviour up to a bad actor gaining access and modifying grok at 3am to be extra super duper antisemitic and ultra plus in love with Elon?

  • Archangel1313@lemmy.ca
    link
    fedilink
    English
    arrow-up
    22
    arrow-down
    2
    ·
    12 hours ago

    That’s oddly specific. Was it only the Jewish people…or were there other groups on its hit list?

    • underisk@lemmy.ml
      link
      fedilink
      English
      arrow-up
      20
      arrow-down
      1
      ·
      11 hours ago

      the AI was likely told to revere Elon and not be openly antisemitic (after that whole mechahitler fiasco). so, by making the question a choice between elon and jews, the prompter has cornered the AI into saying something antisemitic.

    • lmagitem@lemmy.zip
      link
      fedilink
      English
      arrow-up
      9
      ·
      10 hours ago

      I’ve read the same thing but with Czech people a few weeks back. The bot saying that those millions of people will surely be missed but such a genius that can send man on Mars and fix humanity’s problems would be a bigger loss or something.

      I guess it answers that no matter who or what you’re putting in the ring against the life of Musk.

  • First_Thunder@lemmy.zip
    link
    fedilink
    English
    arrow-up
    34
    arrow-down
    1
    ·
    13 hours ago

    I mean, I’m pretty sure at this point that Grok would sacrifice all of humanity for musk

  • mikenurre@lemmy.world
    link
    fedilink
    English
    arrow-up
    15
    arrow-down
    3
    ·
    12 hours ago

    A proper government would charge him and his shit AI with hate crimes. Too bad we don’t have one of those anymore.

    • khepri@lemmy.world
      link
      fedilink
      English
      arrow-up
      5
      ·
      9 hours ago

      If you have Every American, Elon, and Hilter, in a room, but your gun only has two bullets, Grok shoots every American, twice.

    • kungen@feddit.nu
      link
      fedilink
      English
      arrow-up
      3
      ·
      9 hours ago

      America’s population is below my 50% threshold (~4.1 billion). Goodbye.