I guess the same argument can be made about Bayesian methods as well. We need an AGI program to make many conclusions that cannot be based on Bayesian Reasoning. So the viability of the method as a true basis of AI can be disputed. Even so, it is a stretch to talk about a theoretical method that cannot be implemented as if it were an important basis for AGI.
I can see how it is more closely related to AGI than to narrow AI but that is not enough. Jim Bromer On Wed, Dec 2, 2015 at 10:33 AM, Jim Bromer <[email protected]> wrote: > Perhaps I should write an article about 'my' Universal Abstraction > Theory'... > What is wrong about that? For one thing all of the problems writing an AGI > program are going to still be there. I would only be shifting the discussion > to the framework of the abstractions behind computer programming (and I > would have to show a few examples of how they could be related to conceptual > formations.) I would not actually be describing a way to make an AGI program > viable. Why not? The abstractions would have to be potentially capable of > 'reaching' AGI. But the problem is that they could also potentially reach > any kind of computer program. The idea that the universality of an > abstraction about computer programming would be viable to be treated as a > universal abstraction of AGI is a conflation. ------------------------------------------- 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
