Most AI/AGI methods of learning or of figuring things out are either based on applying a method on a particular (typically narrow) kind of situation or on using a crude method to adapt to a narrow kind of situation. Using reason-based-reasoning I expect to be able to discover new possibilities that are not used in mainstream AI/AGI methods as far as I can tell. Of course sometimes the nature of the aspects (or parts) of a situation can be appreciated by simple association but in those cases those kinds of associations (or correlations) can then be considered to be (part of) the reason we can explain familiar observations. We probably noticed that we could see an airplane in the sky before we could hear it before we were able to understand the explanation. But the difficulty of insightfully understanding the relations of noticing things and then discovering an explanation then applying the explanation and confirming it through similar situations (including other kinds of observations) is the key to intelligence. Finding thorough discussions of this kind of situation in AI discussions has been pretty thin for many years because the initial attempts to develop computational methods of integrating observations and ideas failed in the early years of AI. So now instead of recognizing the relation that observation plays in forming and generalizing superficial theory-like descriptions and then trying to discover the details of how the theory-like descriptions can then be integrated with explanations and how the explanations can be confirmed (to some extent) through a variety of other observations which can almost isolate the central effect, the discussions on groups like this get lost on some banality like the question of whether robotics is necessary for AGI or not.
It is my insight into this that gives me so much confidence. Writing an AGI program is going to be complicated. That is a challenge. But that just takes time. 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
