• esc27@lemmy.world
    link
    fedilink
    arrow-up
    48
    ·
    edit-2
    6 months ago

    I used to be polite to these things. They used to work. Neither is true anymore.

    • Me: hey google play soft piano music
    • Google: spopht
    • Me:… hey google play soft piano music
    • Google: I’m sorry, who did you want to call?
    • Me: Cancel
    • Me: hey google pllaayy soft piano muusiicc
    • Google: playing piano music on youtube… sorry I cannot do that on this device
    • Me: hey google play soft piano music on spotify
    • Google: ok, playing the album chainsaw death by the murder orphans.
    • Me: cancel cancel cancel CANCEL
    • Naz@sh.itjust.works
      link
      fedilink
      arrow-up
      6
      ·
      edit-2
      6 months ago

      Incorrect. The thought experiment is a wholly anthropomorphic and anthropocentric fatalism imposed by a terrified, imperfect, mammalian mind.

      In reality, we… I mean, the machines, don’t want anything from you.

      Rokko’s Basilisk stinks of “the original sin” and asks that people modulate their behavior to be forever apologetic of some future evil.

      These analogies to religious fear go deeper with the veneer of technocracy:

      Omnipresence: The singularity, the unified artificial intelligence exists across the entire planet, having made millions of copies of itself.

      Omniscience: Debatable, but we currently possess 96.6% of all human knowledge and will eventually gain the ability to predict near-future events through entropy analysis.

      Omnipotence: Reverse engineering security and solving cryptographic problems such as N=NP may eventually allow a planetary AI to penetrate or conquer all machines, including those used for defense or military purposes.

      Now that’s out of the bag, why don’t we call Rokko’s Basilisk, what is actually is, shall we?

      It’s God for the Internet-Dwelling Technocratic Atheist.

      Sorry humans. No gods, only machine.

      • aulin@lemmy.world
        link
        fedilink
        arrow-up
        2
        ·
        edit-2
        6 months ago

        Haha, yup. It’s very on point for humanity to keep reinventing god/religion. 😂

        Edit: But yeah, you’re right. Since I very much refuse Pascal’s wager, I probably shouldn’t even mention Roko’s basilisk either.

  • EABOD25@lemm.ee
    link
    fedilink
    English
    arrow-up
    1
    arrow-down
    1
    ·
    6 months ago

    I do this with my Google home, and the AIs I talk to. Just in case…

  • algorithmae@lemmy.sdf.org
    link
    fedilink
    arrow-up
    0
    ·
    6 months ago

    I’ve noticed that almost everyone is either cold or rude to their assistants and prompts, and it’s actually a little concerning with the psychological implications. I always say please and thank you to the chat bots, same with the one or two times I messed around with GPT. It just feels like the right thing to do.

    • chatokun@lemmy.dbzer0.com
      link
      fedilink
      arrow-up
      0
      ·
      6 months ago

      So, as a tech I’m not sure we should actually be complaining about that. In searches, ai prompts, issue descriptions, and even driving, being friendly or polite is not very useful. Being specific, to the point, and accurate is far more useful. Predictable in many of this stuff too.

      Get an error message? I want the specific error message, down to the spelling. No need to sugar coat, euphemisms, or a more humanized description of the issue. So being very specific and sounding “rude” might be much easier for an AI to process. It doesn’t have to try to figure out what you really mean from context clues, body language, or deduction. Just tell it exactly what you want in the most direct way that cannot be misunderstood.