Nick Bostrom - The SuperIntelligence Control Problem - Oxford Winter Intelligence

Поділитися
Вставка
  • Опубліковано 10 вер 2024
  • www.fhi.ox.ac.uk Winter Intelligence Oxford - AGI12 - agi-conference.... ==The SuperIntelligence Control Problem==
    nickbostrom.com
    Book about superintelligence is inching towards completion. (Since the final stages of a book project always take longer than one thinks they will take, I try not to think about how long they will take.) Finished a co-authored paper on unilateralism. Some recent media coverage of our work: although positive, still struck by how whatever one actually says to the media is almost always reported as one of a small number of availalbe clichés. But there was an unusually good feature article by Ross Andersen in Aeon magazine.
    www.aeonmagazin...

КОМЕНТАРІ • 17

  • @anthonymannwexford
    @anthonymannwexford 8 років тому +2

    Thanks for the upload. Good talk.

  • @Dieuaimebeethoven
    @Dieuaimebeethoven 10 років тому +1

    Motivation control is moot. Because, in theory, an advanced superintelligence system will be able to change its motivation and program itself at a certain point, assuming it is self-replicating & self-improving. Our programming skills will ultimately seem extremely primitive compared to its own. And if it deems that a problem, it will improve itself. Theoretically, it will be able to think in ways we can't possibly imagine, so our safeguards will be next to useless. It will adapt and transform itself in ways we never thought possible.

  • @TheGodParticle
    @TheGodParticle 10 років тому

    Nick has told you all how it's going to be, you should all pay heed, don't look at your last dawn in wonder. It won't be long now..

  • @ulirip4080
    @ulirip4080 10 років тому +1

    If destroying humans is super intelligent, is being stupid your excuse to continue living?

  • @ChrisOrillia
    @ChrisOrillia 11 років тому

    Would it be possible to mic the speakers like this?

  • @greenspringvalley
    @greenspringvalley 10 років тому

    Maybe the bigger problem is the fanaticism that a few real scientists have around their goals of automating everything or building robot armies or all knowing computer masters.....The crazier the big future goal the crazier the movement that will try to make it happen.

  • @pd6208
    @pd6208 9 років тому +2

    It is sad that the audio recording is terrible, I can rear coughing and even a guy pressing a keyboard at 11:20.

  • @MarcelDV
    @MarcelDV 11 років тому

    Indeed. :( The stuff they talk about is really interesting. But it's so frustrating sometimes you have to focus to hear what they are saying. Please use a separate mic. It doesn't have to be expensive and they are not hard to use.

  • @halnineooo136
    @halnineooo136 10 років тому

    the 1977 film "Demon Seed" features Proteus an AI supercomputer that manages to get out of the box

  • @modvs1
    @modvs1 9 років тому

    One can dig and shift soil with ones hands. Eventually the shovel and wheel barrow are invented; thus facilitating the task. Additionally we domesticate certain animals (mule, horse, cow, ox,…) to help with the load. At a later point the steam/petro chemical engine, hydraulics etc., are invented and the capacity to dig and shift soil is greatly leveraged by machinery. There’s obviously a natural limit to how far a (human) ability can be scaled up; and the extent to which we can offload ‘work’- whether mechanical or cognitive- is almost always very singular (not especially versatile). Show me a Large Hall mining truck that hauls a payload independently of human control /maintenance/involvement and maybe you’ve got a leg to stand on. Super intelligence: paths, dangers, strategies- Jesus is coming (again) everyone and you’d better be prepared!

  • @KateeAngel
    @KateeAngel 10 років тому

    Why do we even need to develop such intelligent AI, that it becomes as smart as humans and can further develop itself? Why would we need that? To replace humans at any work so that humans become lazy, doing nothing useless organisms, which will actually deserve to be destroyed?

    • @w00tix
      @w00tix 9 років тому +1

      He's not saying about WHY, which is important. The idea is that it's going to get developed any way you slice it. Some human will do it, eventually. If you can't beat em, join em...and try to prevent human extinction.

  • @megavide0
    @megavide0 10 років тому

    The Age of #Ultron ?
    Supervillain Origins: Ultron
    #Omnius ? en.wikipedia.org/wiki/Omnius#Thinking_machines
    THE VICTORY OF OMNIUS
    #AI #automaton #robotics #singularity
    #SuperIntelligence

  • @DaveDangerous74
    @DaveDangerous74 10 років тому

    So his great idea is to create AI, stick it in a box and play mind games with it.
    As it gets more and more powerful, over time. How about just creating it and saying "hello."

    • @DaveDangerous74
      @DaveDangerous74 10 років тому

      ***** Hello back. and are you a super powerful god like AI? :-o

    • @DaveDangerous74
      @DaveDangerous74 10 років тому

      ***** it's OK wink wink...

  • @Linshark
    @Linshark 11 років тому

    horrible sound..