Well, at the end of the day, it's going to be far more than a "computer program" in the sense that this phrase currently evokes...
But even if it's tremendously superhuman, its nature may still be shape-able by the nature of the computer programs that were its ancestors... And if these ancestor programs are created to help humans, the odds that the ultimate superhuman programs will be helpful to humans would seem to me to be increased... ben > > I agree with this. I'm not sure why people think it cannot be a service > machine. At the end of the day, its still a COMPUTER PROGRAM. Albeit a > rather sophisticated one.. > > If we don't build it to serve us, than why in the hell should we > build it at > all? Technology serves us every single day without having a self > interest. > An AGI could do the same exact thing, only better.. > > Kevin > > > > > ----- Original Message ----- > From: "Ben Goertzel" <[EMAIL PROTECTED]> > To: <[EMAIL PROTECTED]> > Sent: Thursday, January 09, 2003 11:36 PM > Subject: RE: [agi] Friendliness toward humans > > > > > > > > I tend to agree with Damien. > > > > I see no intrinsic reason why a service-driven AGI system could > not become > > as intelligent as humans and then more intelligent. > > > > Suppose an AGI is given an initial motivational structure that > rewards it > > for > > > > * serving people effectively > > * discovering and creating new patterns > > > > Why are these not challenges enough to spur the evolution of > ever-increasing > > intelligence within the AGI system? > > > > Human intelligence arose largely out of the need to compete for > survival, > > but AGI's need not have the same motivators. > > > > Survival is a suitably complex goal to drive the emergence of > intelligence, > > but so, perhaps, is service. > > > > The service-oriented motivational structure is very close to what > > Novamente's initial motivational structure will be. Heck, the Novamente > > software system right now -- 25% complete and > non-generally-intelligent as > > it is -- is *already* being used to serve people. (Biomind LLC's alpha > > Biomind Toolkit product, based on parts of the incomplete Novamente > system, > > will be done in March... its goal is to serve biologists with great data > > analyses based on a broad integrative view of biological data... see > > www.biomind.com). > > > > - Ben Goertzel > > > > > > > > > -----Original Message----- > > > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On > > > Behalf Of Damien Sullivan > > > Sent: Thursday, January 09, 2003 10:27 PM > > > To: [EMAIL PROTECTED] > > > Subject: Re: [agi] Friendliness toward humans > > > > > > > > > On Thu, Jan 09, 2003 at 10:23:07PM -0800, Alan Grimes wrote: > > > > > > > You _MIGHT_ be able to produce a proof of concept that > way... However, > a > > > > practical working AI, such as the one which could help me > design my my > > > > next body, would need to be quite a bit more. =\ > > > > > > Why? Why should such a thing require replacing the original > > > service-driven > > > motivation with something riskier? > > > > > > -xx- Damien X-) > > > > > > ------- > > > To unsubscribe, change your address, or temporarily deactivate > > > your subscription, > > > please go to http://v2.listbox.com/member/?[EMAIL PROTECTED] > > > > > > > ------- > > To unsubscribe, change your address, or temporarily deactivate your > subscription, > > please go to http://v2.listbox.com/member/?[EMAIL PROTECTED] > > > > > ------- > To unsubscribe, change your address, or temporarily deactivate > your subscription, > please go to http://v2.listbox.com/member/?[EMAIL PROTECTED] > ------- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED]
