Re: JOIN: Joshua Fox

From: Richard Loosemore (rpwl@lightlink.com)
Date: Wed Feb 08 2006 - 08:49:27 MST


No, in my comment about Artificial General Dumbtelligence I was taking a
swipe at attempts to do AI research and development using techniques
that ignore learning mechanisms, the grounding problem and the role of a
complex systems disconnect between local and global. A little quibble I
have with [what is now called] conventional AI, that's all.

Richard Loosemore

Mike Dougherty wrote:
> I bet the General Dumbtelligence will be Actual rather than Artificial :)
>
> Are you saying that human-bootstrapped non-human intelligence is
> unlikely? (which would be strange for you to be on this list) I do
> agree that any >single< attempt to create AGI will probably run short of
> the critical mass needed to spark Intelligence. My bet is that it will
> take something more organic - in the sense that evolution has been
> successful using organic growth, and that the ingenuity we notice in
> nature seems to be motivated by the need for efficiency in resource
> management or optimizing the opportunities for growth/advancement. [the
> seemingly infinite resource of virtual hardware does not (imo) provide
> inherent survival urges that nature has used to prompt evolution]
>
> On 2/7/06, *Richard Loosemore* <rpwl@lightlink.com
> <mailto:rpwl@lightlink.com>> wrote:
>
> I also disagree with many about the idea of Unfriendly AI (UFAI). Some
> talk as if this is almost inevitable, and say that we will have to work
> like crazy to avoid it. I think they base this UFAI-inevitability idea
> on their own particular take on how to build an AGI, and I dispute their
> methods. In short, I am of the opinion that the approach to AGI they
> espouse is going to lead to an AGD (Artificial General Dumbtelligence),
> which will never reach the level of human intelligence even after many
> more decades of painstaking work, and hence will never be a threat.
>
> My position is (ahem) hotly disputed in some quarters.
>
>



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:55 MDT