Re: Seed AI milestones (was: Microsoft aflare)

From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Wed Feb 27 2002 - 10:21:59 MST


Eugene Leitl wrote:
>
> General intelligence is not property of a simple system. Far from it.
> As a result I predict that human software engineers coding an AI
> explicitly (i.e. using not stochastic/noisy/evolutionary methods) are
> going to fall short of the goal.

As far as I can tell, stochastic/noisy/evolutionary methods buy you
relatively little. They buy you smooth fitness landscapes on the
code/modality level. That's it. Nothing to sneer at, but not worth the
cost of losing all hope of high-level design ("coding an AI explicitly").
Personally, I think the best course is to treat "smooth fitness landscapes"
as an explicit design requirement.

> Ben wrote:
> >
> > I agree that goal-directed self-modification is a specialized mental
> > function, similar (very roughly speaking) to, say, vision processing, or
> > mathematical reasoning, or social interaction. However, also like these
> > other things, it will be achieved by a combination of general
> > intelligence processes with more specialized heuristics.
>
> Am I correct to assume that we're talking about explicit codification of
> knowledge destilled from human experts? Is there any reason to suspect
> that we're going to do any better than Lenat & Co? The record track so far
> is not overwhelming.

I can't speak for Ben, but certainly SIAI does not advocate (and never has)
the explicit codification of knowledge distilled from human experts.
Knowledge is learned complexity; it is abstracted from experience, not
hardwired by the programmers.

> Um, implementing a runaway AI is certainly not my problem. I'm interested
> in modelling biological organisms, which does not involve such dangerous
> components. This is the wrong forum to discuss it, however.

An upload is just a seed human.

-- -- -- -- --
Eliezer S. Yudkowsky http://intelligence.org/
Research Fellow, Singularity Institute for Artificial Intelligence



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT