Page 4 of 8 FirstFirst ... 23456 ... LastLast
Results 31 to 40 of 73

Thread: AI: Superintelligence - paths, dangers and strategies

  1. #31
    Member
    Type
    INFJ
    Join Date
    Oct 2017
    Posts
    839
    Quote Originally Posted by pensive_pilgrim View Post
    How could an AI want anything other than what a human has programmed it to want?
    because it is learning from us right now

    The clip below, created by a uni prof, is doing the rounds on the internet at the moment:

    "Slaughterbot" Autonomous Killer Drones


  2. #32
    Pull the strings! Architect's Avatar
    Type
    INTP
    Join Date
    Feb 2014
    Location
    West Coast
    Posts
    516
    Quote Originally Posted by pensive_pilgrim View Post
    How could an AI want anything other than what a human has programmed it to want?
    Because it's not programming in the style you are thinking of. There are three great paradigms of programming, imperative (you say how to do it), declarative (you say what you want but not how to do it) and connectionist (you specify the performance metric). DNN's are of the connectionist school which means the computer 'programs' itself (it actually determines the weights for a neural architecture you choose, based on the performance metric you give it and the training samples you feed it). Consider AlphaGo which made a move that startled all the human players. Michael Redmond thought it was a mistake the handlers made in moving the stone on the physical board, but it was a brilliant move that experts later said no human would have made.

    Edit: here's the move

    Last edited by Architect; 11-22-2017 at 08:22 PM.

  3. #33
    Member rhinosaur's Avatar
    Type
    INTP
    Join Date
    Dec 2013
    Posts
    565
    You still have to give it the performance metrics, though.

  4. #34
    Pull the strings! Architect's Avatar
    Type
    INTP
    Join Date
    Feb 2014
    Location
    West Coast
    Posts
    516
    Quote Originally Posted by rhinosaur View Post
    You still have to give it the performance metrics, though.
    Quote Originally Posted by Architect
    (you specify the performance metric)
    But this is a just a single simple function, cross entropy usually. A pat on the head but the real decisions are made with the weights. A stronger example is the training data, that's the easiest way to influence the network to self-program in a way you wish. Anyhow we'll probably use a variations of GAN's (Generative Adversarial Nets) for GAI which go off into a self training loop beyond human ken (AlphaGo is orders of magnitude better than any human now due to this).

  5. #35
    know nothing pensive_pilgrim's Avatar
    Type
    INTP
    Join Date
    Dec 2013
    Posts
    4,538
    So basically, it might discover that in order to get really good at Go it needs to destroy all humans.

  6. #36
    Member
    Type
    INFJ
    Join Date
    Oct 2017
    Posts
    839
    Quote Originally Posted by pensive_pilgrim View Post
    So basically, it might discover that in order to get really good at Go it needs to destroy all humans.
    yes

    you know in the matrix films?

    it never actually says that the machines forced humans into the matrix. Its just as possible that humans agreed to be plugged into the matrix to escape from an increasingly sabotaged reality

    would some humans willingly merge with machines....yes they would

  7. #37
    Member rhinosaur's Avatar
    Type
    INTP
    Join Date
    Dec 2013
    Posts
    565
    Quote Originally Posted by pensive_pilgrim View Post
    So basically, it might discover that in order to get really good at Go it needs to destroy all humans.
    Indeed

    https://wiki.lesswrong.com/wiki/Paperclip_maximizer

  8. #38
    know nothing pensive_pilgrim's Avatar
    Type
    INTP
    Join Date
    Dec 2013
    Posts
    4,538
    Quote Originally Posted by rhinosaur View Post
    Wow. It would be pretty surreal to fight in an apocalyptic war to save humanity from an enemy bent on turning us into paperclips. Or imagine exploring space in the future and encountering a sentient galaxy made of alien paperclips.

    Quote Originally Posted by mara View Post
    would some humans willingly merge with machines....yes they would
    Me, I would. I want a robotic spider body with a nuclear-powered hub to house my brain and balls.

  9. #39
    Pull the strings! Architect's Avatar
    Type
    INTP
    Join Date
    Feb 2014
    Location
    West Coast
    Posts
    516
    The belief I tend to have is that because humans discovered and use the Golden Rule to a great degree that GAI's will necessarily do so also, just because it works better. Or perhaps better said 'it's a more optimal winning strategy'. Democracy works better because it allows competing agents all succeed to the best of their circumstances and abilities, and when others succeed in a democracy that helps all. With AlphaGo zero we might see an indication of this idea, consider the following analysis by Michael Redmond and Chris Garlock



    AlphaGo Zero self trained in the absence of any human data to 9d level. In Go this is a remarkable situation, it's an alien professional player, what moves would it come up with? For example the opening play is usually to the four star points but that's human convention, we don't know if there are other openings that are better. Turns out Zero's play is much like human play, to Mr Redmond's disappointment. It appears that playing the star points might actually be optimal opening play (Redmond isn't willing to say that but Garlach goes in that direction).

    The point? Given a performance metric to succeed the computer necessarily learns good play much like humans. The discussion should then turn to politics or philosophy, what is the winning course of action, the destruction of human kind or to help human kind? The idea that humans are irrelevant I think is immature, my believe is GAI's will necessarily see the mutual benefit that comes with cooperation.

  10. #40
    Member
    Type
    INFJ
    Join Date
    Oct 2017
    Posts
    839
    Quote Originally Posted by pensive_pilgrim View Post
    Me, I would. I want a robotic spider body with a nuclear-powered hub to house my brain and balls.
    well then i see you are well equipped to handle some really mind bending perspectives....


Similar Threads

  1. Tea strategies
    By Blorg in forum The Pub
    Replies: 46
    Last Post: 05-05-2017, 02:54 PM
  2. People's Rubic's Cube strategies
    By BarIII in forum The Pub
    Replies: 20
    Last Post: 08-09-2016, 04:28 PM
  3. Foolishly reveal your password creation strategies
    By notdavidlynch in forum The Pub
    Replies: 41
    Last Post: 09-15-2014, 09:34 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •