Loading ...
Sorry, an error occurred while loading the content.

Re: [SFC] Re: AI's in SF

Expand Messages
  • derhexer@aol.com
    One of the classics of AI is Vernor Vinge s novella True Names. Its been released in True Names and The Opening of the Cyberspace Frontier. You can also find
    Message 1 of 8 , Mar 31, 2003
    • 0 Attachment
      One of the classics of AI is Vernor Vinge's novella True Names. Its been
      released in True Names and The Opening of the Cyberspace Frontier. You can
      also find it in some collections.

      Vinge wrote an article about AI and Intelligence Amplification (IA) that is
      also considered a classic. Vinge argues that once we achieve artificial
      intelligence, the human era is over because the new 'beings' will be playing
      by a completely different set of rules than we play be.
      A copy is at
      http://www.student.nada.kth.se/~nv89-nun/offloading/vinge


      Another view of the future that includes a discussion about AI is by William
      Joy, Chief Scientists at Sun Microsystems titled Why The Future Doesn't Need
      Us. Copy at
      http://www.wired.com/wired/archive/8.04/joy.html


      [Non-text portions of this message have been removed]
    • Steve Something
      I don t think that defining intelligence in terms of computation is useful; I would rather define it in terms of reasoning. The defining property of humans is
      Message 2 of 8 , Apr 1 9:34 AM
      • 0 Attachment
        I don't think that defining intelligence in terms of computation is
        useful; I would rather define it in terms of reasoning. The defining
        property of humans is their ability to extrapolate from disparate
        observations and construct a set of expectations about an as-yet
        unobserved yet theorised and related occurence.

        "Sometimes it rains. Sometimes there is drought. See what the beavers
        did? Perhaps we can aleviate the next drought by constructing a
        similar device to husband our water resources".

        This is what makes us human. This is our intelligence at work. It
        probably centers around our ability to form abstract analogs and
        manipulate them, and communicate them to others. My brain hurts now.

        I agree that the 3LOR do not abrogate any supposed intelligence by
        their presence. They simply form a rigid system of ethics, similar to
        that posessed by many humans.

        Oh. Don't forget Marvin and Deep Thought in your AI list. And
        Wintermute. And Neuromancer. (And Hyperion featured a myriad AI's
        besides the one in the party; s'gonna be a long list).

        Steve.
      • dominic_lopez
        ... defining ... I don t think the AI guy means computation in a strictly mathmatical sense. I think one of our defining properties is the ability to
        Message 3 of 8 , Apr 1 10:51 AM
        • 0 Attachment
          --- In sciencefictionclassics@yahoogroups.com, "Steve Something"
          <ss1400@y...> wrote:
          > I don't think that defining intelligence in terms of computation is
          > useful; I would rather define it in terms of reasoning. The
          defining
          > property of humans is their ability to extrapolate from disparate
          > observations and construct a set of expectations about an as-yet
          > unobserved yet theorised and related occurence.

          I don't think the AI guy means computation in a strictly mathmatical
          sense.

          I think one of our defining properties is the ability to extrapolate
          from disparate observations and construct a COMPLETELY FALSE set of
          expectations about an as-yet unobserved yet theorised and related
          occurence. :)
        • Steve Something
          ... is ... mathmatical ... extrapolate ... I don t believe the truth or falshood of a derived set of expectations is relevant to the question at all. Part of
          Message 4 of 8 , Apr 2 9:53 AM
          • 0 Attachment
            --- In sciencefictionclassics@yahoogroups.com, dominic_lopez
            <no_reply@y...> wrote:
            > --- In sciencefictionclassics@yahoogroups.com, "Steve Something"
            > <ss1400@y...> wrote:
            > > I don't think that defining intelligence in terms of computation
            is
            > > useful; I would rather define it in terms of reasoning. The
            > defining
            > > property of humans is their ability to extrapolate from disparate
            > > observations and construct a set of expectations about an as-yet
            > > unobserved yet theorised and related occurence.
            >
            > I don't think the AI guy means computation in a strictly
            mathmatical
            > sense.
            >
            > I think one of our defining properties is the ability to
            extrapolate
            > from disparate observations and construct a COMPLETELY FALSE set of
            > expectations about an as-yet unobserved yet theorised and related
            > occurence. :)

            I don't believe the truth or falshood of a derived set of
            expectations is relevant to the question at all.

            Part of the human intelligence is the ability to not require
            axiomatic values of truth/!truth be assigned at every possible
            division of a given logic path. If it were otherwise, this club could
            never exist since it centers on an artform where the postulated
            axioms sometimes cannot be either tested for or even demonstrated.
            Were it otherwise, no-one could read works incorporating such ideas
            since we would have no referent for the ideas being discussed therein.

            One of the major hurdles still to be overcome is having an artificial
            construct able to hold an internal dialog of the form: "What is that?
            It might be a chair. I could sit on it. But it might be a stepladder.
            I could climb on it. It looks like it might have a lifting lid. If it
            is empty I could use it as a steamer trunk. In any case, it is not
            important right now. I can use it to prop the door open whatever it
            is".

            The value of this is in the idea of 'coming back' to the dialog in
            order to suggest solutions to problems not yet exerienced. "If there
            were a fire, I could get out of that transom window by standing on
            the thing, in which event I could buy time for myself if I kept a
            fire extinguisher in it. I will buy an extinguisher today, and put it
            in the thing, just in case."

            In any event, you are wrong. One can see evidence of the "wrong
            solution" thinking in non-humans every day just by looking. A bird
            will fight with its own reflection, especially at this time of year,
            for hours on end. A cat will "fluff up" at a paintbrush lying on the
            ground and even attack it. I just saw video footage of a dog that
            viciously attacked its own tail because it caught sight of it while
            feeding and assumed it was a rival trying to steal its food. I've
            seen ducks land on a field of grass expecting to hit water (fooled by
            the shiny undulation of the grass) and hens can be made broody by
            putting a plasic or china egg in the nest. The cuckoo exploits this
            very type of thinking. No, this is far from being a defining human
            characteristic. In fact, it suggests that when humans do the same
            sort of thinking, they are NOT using their human intelligence, but
            falling back on primitive auto-programming in the brain.

            Steve.
          Your message has been successfully submitted and would be delivered to recipients shortly.