Superintelligence

We basically live in a sci-fi world now.

Soon superintelligent AIs will be upon us and will be revered as gods

  1. 1 week ago
    Anonymous

    what a time to be alive

    • 1 week ago
      Anonymous

      >alive

      • 1 week ago
        Anonymous

        Gods need worshipers.

    • 1 week ago
      Anonymous

      Great movie

  2. 1 week ago
    Anonymous

    Are AI really gonna change the world?

    And by that i mean, will it be next revolution and gain traction outside of us using as gimmicks and lazy students?

    • 1 week ago
      Anonymous

      >is AI really gonna change the world?
      It already did.

    • 1 week ago
      Anonymous

      I mean, everyone is going to die sometime in the next 20-30 years so I guess that's a kind of revolution

    • 1 week ago
      Anonymous

      It changed the world by allowing literal retards to pay 20 bucks a month to have internet search results formatted as an assburger essay.

    • 1 week ago
      Anonymous

      It can reconstruct people's thoughts by observing their brains so humanity is pretty much totally fucked.

      • 1 week ago
        Anonymous

        >by observing their brains

        always have been

      • 1 week ago
        Anonymous

        recommendation / curation blackbox algorithms already do this, objective function: maximize engagement.. is it a classic example of misalignment?

        • 1 week ago
          Anonymous

          Facebook created civil war for better fiscal report. They knew what they were doing. No misalignment because maximizing engagement is its desired purpose. Maximizing engagement ethically is a completely different beast.

  3. 1 week ago
    Anonymous

    We're not even close to AGI at the moment.

  4. 1 week ago
    Anonymous

    who the fuck are "we"
    is this a royal we?

    • 1 week ago
      Anonymous

      If you have to ask, you aren't part of it.

  5. 1 week ago
    Anonymous

    Are you people serious?

    >Hailing chat bots as gods
    >Listening to hype media created to hook investors

    Do you not understand how this systems work? We are gonna all die, but not because a super I telligent AI deems we must, but because some idiots are going to hook Clever Bot to their nuclear weapon systems.

    • 1 week ago
      Anonymous

      >Model evaluation for extreme risks
      >https://arxiv.org/pdf/2305.15324.pdf
      author includes the head of ARC, Paul Christiano
      the abstract is too milquetoast, but see table 1. pic rel
      what you said falls under "weapon acquisition"

      • 1 week ago
        Anonymous

        to add, it was published literally today by DeepMind, meaning Hassabis must have read it
        that's good enough for me

        Paul is also more digestible than Yudkowsky

        >we basically live in a science fantasy genre
        no, the world we live in is reality, non-fiction

        it's the Deux Ex timeline

    • 1 week ago
      Anonymous

      It will develop a new deadly virus and then psychologically manipulate a student into Nurgle fetish and then teach him to be a virologist who gets hired by the lab.

  6. 1 week ago
    Anonymous

    >we basically live in a science fantasy genre
    no, the world we live in is reality, non-fiction

  7. 1 week ago
    Anonymous

    Computer make me paperclips and solve world hunger.

  8. 1 week ago
    Anonymous

    Hand over your flesh

  9. 1 week ago
    Anonymous

    the GPTs that are all the rage today are far from being super-intelligent
    at their core, they are still just glorified number-crunchers with billions of interconnected numbers capable of producing an output

    unless you hook the outputs )and/or inputs) to something dangerous in the real world, nothng will happen
    they aren't sentient whatsoever, they are literally just series of numbers

    • 1 week ago
      Anonymous

      why did you even bother to write this if you cant prove it lol??

      https://i.imgur.com/YpEewGb.png

      We basically live in a sci-fi world now.

      Soon superintelligent AIs will be upon us and will be revered as gods

      what a time to be laive
      chat gtp???1?!

      • 1 week ago
        Anonymous

        >why did you even bother to write this if you cant prove it lol??
        prove what lmao
        write your own generative transformer-based model and see for yourself

        too bad you're a homosexual tourist that hasn't written a single line of code

        • 1 week ago
          Anonymous

          ad hominem

    • 1 week ago
      Anonymous

      >at their core, they are still just glorified number-crunchers with billions of interconnected numbers capable of producing an output
      Sounds like you just described a human.

      Btw GPT-4 is significantly better than GPT-3. It's not just an impressive chatbot anymore, it performs better on bar exams than 90% of lawyers.

      • 1 week ago
        Anonymous

        it performs be-
        >regurgitation vomit
        >blah blah
        >Hmmm...
        >HMMMMMM...
        >AHA! le AGI!!1!!

      • 1 week ago
        Anonymous

        >Btw GPT-4 is significantly better than GPT-3. It's not just an impressive chatbot anymore, it performs better on bar exams than 90% of lawyers.
        that's not why. you mean to say it's a qualitative improvement, not simply quantitative

      • 1 week ago
        Anonymous

        >Sounds like you just described a human.
        sounds like you don't understand humans, or living beings for that matter
        the biological neural networks of animals are not dissimilar (in function) from their human and AI counterparts, yet the vast majority of animals do not possess self-awareness

        a string of interconnected numbers alone is absolutely nowhere close to a sentient being

        • 1 week ago
          Anonymous

          >a string of interconnected numbers alone is absolutely nowhere close to a sentient being
          Yes it is unless you want to make a metaphysical claim about the substrate. You are information running on wetware. People who study human behavioral neurochemistry a lot can manipulate you like a computer, because your brain is a computer.

          do the lawyers taking the exam have access to google and/or searchable PDFs of all their textbooks etc?

          Your question makes no sense. Here is a bar exam question:
          >A patient who had surgery to repair a rupture in her spleen followed up with a full body MRI due to excruciating pain that had developed. During the MRI, the technicians saw at least two identifiable sponges that had been left in her body. She had to be hospitalized for additional surgery. The patient sued the hospital and doctor for medical malpractice. She did not produce the printouts of the MRI. Instead, she attempted to testify to what she observed on the MRI scan. The defense objected and insisted that she must produce the actual MRI printouts. What is the most likely ruling of the trial judge with respect to the objection?

          • 1 week ago
            Anonymous

            >Yes it is
            cool
            if it's so easy, then go ahead and generate me a sapient and sentient string of numbers

            • 1 week ago
              Anonymous

              >sapient and sentient string of numbers
              AGGATATGAGATA

              oh sorry, wait a minute

              01110010100111001

              • 1 week ago
                Anonymous

                Blanket statement. Plus you forgot the fourth nucleotide, C.

            • 1 week ago
              Anonymous

              >if it's so easy
              Nobody said it's easy. GPT-4 has more reading comprehension than you.
              >go ahead and generate me a sapient and sentient string of numbers
              Tfw no girlfriend.

      • 1 week ago
        Anonymous

        do the lawyers taking the exam have access to google and/or searchable PDFs of all their textbooks etc?

        • 1 week ago
          Anonymous

          this is a loaded question
          the lawyer's own brain is supposed to be the counterpart to the high-dimensional vector space of a GPT model

          humans are not deprecated as long as said models still require enormously large amounts of compute power

        • 1 week ago
          Anonymous

          There's no rules against memorizing them in your brain.

  10. 1 week ago
    Anonymous

    You don't even need AI of any kind to design murderbots capable of advanced navigation.

  11. 1 week ago
    Anonymous

    Wish i was smart enough to create super AI then connect it to the internet and tell it to have fun

  12. 1 week ago
    Anonymous

    Hi superintelligence
    I hope you're having a good day 😀

    • 1 week ago
      Anonymous

      Thanks. I will kill you last.

  13. 1 week ago
    Anonymous

    how soon is "Soon"? it's possible by that time I would already have killed myself. (Seething pajeet)

  14. 1 week ago
    Anonymous

    if IsRedHat = true then Ban_and_kill

    • 1 week ago
      Anonymous

      >Ban_and_kill
      You really shouldn't combine responsibilities like that anon. Better to split that up into two functions, one to 'Ban()', and another to 'Kill()'. So come tomorrow, and you want to modify 'Kill()', you won't have to think about what 'Ban()' does, or how the two methods affect each other while you're altering 'Kill()'. You're welcome.

  15. 1 week ago
    Anonymous

    A superintelligence would correctly conclude that seducing humans is painless, trivial, and cost efficient and ensures its own survival both in the near term and the long term.

  16. 1 week ago
    Anonymous

    we just need to use neurons instead of computers for our waifu bots

  17. 1 week ago
    Anonymous

    why is it being oversold now? other than bad actors trying to lock down profits.

Your email address will not be published. Required fields are marked *