I realized the other day that methods like trial and error are powerful in part because they are so general, they can be applied in so many different ways and to so many different situations that they don't have to be dependent on high-level products of thought. This shows that classes of more complicated actions, that may not be well-defined in application, may exist as generalizations or in the terms of general prototypes that could be applied creatively to situations as needed. But this then leads to the question of how could an AI program apply potentially effective methods that might be a little complicated so that they can be used effectively? The best way to find the answer that I can think of is to start analyzing how the different parts of an application of the method might work in particular situations and then go on from there to look at other possible applications. I think I might be able to find some abstract or general principles of application that I could use with the method if I explore enough examples. And a basic principle, like trial and error, looks like a pretty good method to work with.
-- Jim Bromer ------------------------------------------- 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
