On Wed, Sep 24, 2014 at 9:15 AM, John Rose via AGI <[email protected]> wrote: >> I agree. An AGI needs to be able to model human minds in order to >> communicate effectively with people. If the model didn't claim to be >> conscious then I would consider that a bug. >> > > They don't matter UNLESS engineering a real p-conscious entity verses an > ersatz Google-like behavioral regurgitation, is easier to build and requires > significantly less computational resources to run AND results in > significantly more intelligence and human interactive assistance capabilities > on said resources. > > IOW which one is easier to build and which one runs better. My argument for > p-consciousness AGI design is also based on engineering and runtime estimates.
A p-conscious AGI and a zombie AGI would have identical behavior because that is how a zombie is defined. (See http://en.wikipedia.org/wiki/Philosophical_zombie ). Therefore both AGIs would have identical designs and identical costs. Unless you believe (like Penrose) that human behavior is not computable. I don't think anyone on this list believes that a sufficiently powerful computer couldn't at least in principle do what our 86 billion neurons do. BTW, do you agree with my cost estimate? https://docs.google.com/document/d/1Z0kr3XDoM6cr5TgHH0GXQTjyikr7WpCkpWFn9IglW3o I realize it is tempting to look for some magic shortcut like consciousness or quantum computing or P = NP to get around this $1 quadrillion problem that we have been working on for the last 60 years. -- -- Matt Mahoney, [email protected] ------------------------------------------- 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
