One could use a semi-virtual domain for an AGI: i.e., a domain where the avatar of an AGIis embedded in a virtual domain, but the avatar also has interaction with a human user in the real world.
~PM. Date: Sat, 20 Oct 2012 22:14:26 -0400 Subject: Re: [agi] Robot-Based AGI is Not a Necessity to Make Intelligence to Emerge From: [email protected] To: [email protected] On Sat, Oct 20, 2012 at 12:30 PM, Alan Grimes <[email protected]> wrote: Jim Bromer wrote: There is no experimental basis for concluding that robotic interactions are necessary for genuine AGI to emerge. And, based on insight about how computers actually work the belief that computer-robotic connectivity is essential for AGI to emerge is not insightful. Computer programs react to data whether they are in robots or not. Its all just data. Strictly speaking, that's true. However, I would argue that some kind of closed cybernetic loop is required for procedural learning to take place. I would also argue that it is much easier for the monkey programming the thing to think in terms of a robot than in some abstract domain. We are able to take symbols further than monkeys. So we do not have to work with robots in order to avoid thinking of abstract domains. We can deal with abstract ideas. We can abstract based on our knowledge of things to determine what is necessary and essential and what may enhance but is not necessary. Drinking water is not necessary for computer intelligence but it is necessary for almost all kinds of living intelligence. Just because knowledge of the world is much more intimately involved in our thinking it does not mean that it is absolutely necessary for intelligence. If straight interaction with the world through sensors that are comparable to the senses that we have then why hasn't any sensory AI project actually succeeded to procuce general intelligence. We might imagine IO data worlds which are different the world that is revealed to the human senses. And we can think of sensors that humans lack without concluding that anything with those kinds of sensory modalities must be capable of far greater intelligence than us. Although this sounds like it is a little close to sophistry, it isn't because it makes sense (so to speak). We can load any kind of AI program up with all kinds of sensors and you are still not going to get genuine intelligence without first discovering a way to do it. Jim Bromer AGI | Archives | Modify Your Subscription ------------------------------------------- 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
