I want to return to what seems to me the high-school-naive idea of how an AGI's or any body of knowledge can and/or does grow - i.e. linearly, mathematically and logically.
Correct me, but I haven't seen any awareness in AI of the huge difficulties that result from the problem of : how do you test acquired knowledge? Such difficulties are fundamental to all current philosophical and sociological treatments of the growth of knowledge - and arguably they apply to the growth of personal as well as formal, organized knowledge. Consider this blurb for an example of the real complexity of knowledge's growth:: "Thagard analyzes the development and acceptance of scientific ideas. At the heart of the book is a case study of the recent dramatic shift in medical understanding of peptic ulcers, most of which are now believed to be caused by infection by the bacterium Helicobacter pylori. When this explanation was first proposed in 1983, it was greeted with intense skepticism by most medical experts, but it became widely accepted over the next decade. Thagard discusses the psychological processes of discovery and acceptance, the physical processes involving instruments and experiments, and the social processes of collaboration, communication, and consensus that brought about this transformation in medical knowledge. How Scientists Explain Disease challenges both traditional philosophy of science, which has viewed science as largely a matter of logic, and contemporary science studies that view science as largely a matter of power. Drawing on theories of distributed computing and artificial intelligence, Paul Thagard develops new models that make sense of scientific change as a complex system of cognitive, social, and physical interactions." ------------------------------------------- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=98558129-0bdb63 Powered by Listbox: http://www.listbox.com
