Re: When Subgoals Attack

From: Samantha Atkins (
Date: Thu Jan 18 2001 - 20:46:04 MST

Dale Johnstone wrote:
> "Samantha Atkins" wrote
> > "Eliezer S. Yudkowsky" wrote:
> > >
> >
> > > > Your above analogy would blow your system architecture apart. Over
> and
> > > > over again we have found that individual freedom to pursue individual
> > > > goals does more for the overall good than any sort of top down
> > > > organization.
> > >
> > > Really? Okay, I demand the immediate liberation of your visual cortex
> to
> > > pursue the independent processing of whichever pixels it finds most
> > > interesting. I can no longer tolerate your enslavement of a living
> > > cognitive subsystem to your mere navigational requirements.
> >
> > That is a cheap and worthless shot.
> >
> > -s.
> Samantha, although I agree with your position about fuzziness vs rigid
> hierarchical schemes (and other things), your reply here didn't help move
> the discussion forward.
> Eli's point I think was a response to a perceived anthropomorphism in your
> line of reasoning. If that was the case, then it was a valid response.
> However, judging from your previous posts I doubt that was what you had in
> mind.

That is reading a lot into the comment that doesn't seem to be there.
That may be your perception or even his perception but that is not what
was said. I was simply referring to the apparent fact in human societies
(this isn't simply antromorphism on my part as those are the only
intelligent entities we have at hand to draw inferences from currently)
it is quite important to have multiple agents with variant agendas and
viewpoints that interact. This seems to be true even in non-intelligent
living systems where diversity of individual units becomes important
when the system faces stresses beyond its mainline patterns. If this is
a general pattern in complex adaptive systems then it is a valid
question how the SI will generate alternate
personas/viewpoints/goal-sets when faced with the limits of its mainline
responses or if it will or should.

I don't think an absurd analogy helped move the discussion forward. Do

> Personally I find it difficult to know if the architecture would fall apart
> simply because I have difficulty modelling it.

This is a good point. When will we see mock-ups of what is proposed?
Or is this possible?

- samantha

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:35 MDT