I now think I know enough to create a working AGI program. However, I am very aware of the fact that I don't have any experience actually creating a working AGI program and from my limited experiences I know all too well how easy it is to make a simple mistake that can bog the program down before I even begin. So when I say that I think I know enough I am saying that I have enough potential solutions to the problems that I expect to run into to start. The only exception is that I have no idea how to write an automated self-managing program that will quickly learn to avoid overly elaborate methods that will bog the program down. Well that is not quite true. What I really mean is that in order to write a program that will do that many of my working assumptions will have to be abandoned just because they are not feasible in a contemporary program. This extravagance of necessity combined with the extra overload of writing a program that will manage itself - just to avoid taking too much time on poorly constructed computationally recursive algorithms - means that I will have to pare my ideas down to the point that they will become absurdities of simplicity. When combined with an allowance to maintain more than one possible explanation for an observable effect, it becomes apparent that my program will not work.
Perhaps the program should be designed to forge ahead even though its basic insights will not be sufficient to produce much depth. As long as it looks like it is working, and acts like it recognizes what it does not know, it may work up to a point. My goal then is limited. It should work better than then other stuff that is available to prove my theories are feasible. Jim Bromer ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968 Powered by Listbox: http://www.listbox.com
