I feel very strongly that discovering the abstract principles which have made AGI so elusive is a primary necessity to advancing toward the goal. If computers keep getting more and more powerful and faster faster then someone could discover how to create an effective AGI program without a thorough grounding of the abstract principles that are necessary to make AGI feasible in the near future. Peter's philosophical statement on his company's website make it clear that he recognizes that full robotic implementation is not necessary for feasible AGI. But on the other hand, the other extreme that he denies is that language alone is not enough to make AGI feasible. I believe that (conversational) language probably provides the best advantage to create the basic program that would be needed for true general learning just because it can contain such a variety of powerful relational structures between references. Failing to grasp this is, in my view, representative of a major grounding problem. Jim Bromer
On Fri, Sep 20, 2013 at 4:58 PM, Matt Mahoney <[email protected]> wrote: > Jim, does this mean you are turning down the job offer? > > On Fri, Sep 20, 2013 at 4:41 PM, Jim Bromer <[email protected]> wrote: >> An AGI program cannot be -fully implemented- on a system that uses >> language alone but the prototype of such an AGI program can be. I >> feel that the inability to grasp this theoretical foundation >> represents a major flaw in your corporation's theoretical foundation. >> I would not know how to prove this but it seems so obvious to me that >> it is difficult for me not to find ways to validate it. >> >> An AGI program has to have input output of course. The IO Data >> Environment has to be rich enough so that it contains a great variety >> of kinds of events that can be observed and interacted with. A >> conversational language has that sort of rich diversity. The problem >> is one of complexity (complicatedness) and not the rather minor issue >> of the constraints on the variety of IO modalities. This kind of >> complexity can be handled by designing systems that can handle it. >> However, this creates a dilemma. The solutions that are necessary to >> limit the complexity of the program will actually be adding to the >> complexity. This dilemma then becomes the essential problem. Jim >> Bromer >> >> On Fri, Sep 20, 2013 at 10:34 AM, Peter Voss <[email protected]> wrote: >>> Here’s a brief update on our new AGI (artificial general intelligence) R&D >>> company http://agi-innovations.com/newsFlash_1.html (please spread the >>> word) >>> >>> >>> >>> Things are moving! >>> >>> >>> >>> Peter >>> >>> >>> >>> AGI | Archives | Modify Your Subscription >> >> >> ------------------------------------------- >> AGI >> Archives: https://www.listbox.com/member/archive/303/=now >> RSS Feed: https://www.listbox.com/member/archive/rss/303/3701026-786a0853 >> Modify Your Subscription: https://www.listbox.com/member/?&; >> Powered by Listbox: http://www.listbox.com > > > > -- > -- Matt Mahoney, [email protected] > > > ------------------------------------------- > AGI > Archives: https://www.listbox.com/member/archive/303/=now > RSS Feed: https://www.listbox.com/member/archive/rss/303/24379807-f5817f28 > Modify Your Subscription: https://www.listbox.com/member/?&; > Powered by Listbox: http://www.listbox.com ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
