Wilson's mental hierarchy

Eliezer S. Yudkowsky (sentience@pobox.com)
Thu, 10 Dec 1998 15:40:07 -0600

Paul Hughes wrote:
>
> "Eliezer S. Yudkowsky" wrote:
>
> > I have to agree that Wilson's hierarchy is total gibberish, cognitively
> > speaking. Any attempt to characterize one mental state as "higher" than
> > another usually is.
>
> This a perfect case of the "pot calling the kettle black", as I
> quote from your 'Coding a Transhuman AI':
>
> "The modules sum together to form abilities; the abilities sum
> together to form intelligence."

And please quote the disclaimer!

" *This principle is not useful for hierarchical design* [italics in original]; it is just a way of viewing the AI. You can't design abilities independently of each other; some modules, perhaps even the majority, will be useful for everything - mathematical ability and causal analysis ability and combinatorial design ability. Some modules will be so basic as to constitute a part of the architecture itself."

> If that is not another way of saying -- "lower" modules sum
> together to create "higher" abilities, those "higher" abilities
> sum together to create an even "higher" intelligence -- I don't
> know what is. It is saying *EXACTLY* the same thing.

My bogosity detectors go off whenever I read "X levels of consciousness" arranged into ascending order, because I know perfectly well that the "good guys" will be operating at level eight and the "bad guys" will be operating at level one. Wilson is making a moral distinction; I'm making a design distinction.

Furthermore, as has been pointed out, anything designed in for "future evolution" is a flaming lie. If you could find one single example of that it would overturn the entire field of evolutionary biology.

> I'm aghast that you and Damien (both whom I respect highly by the
> way) can't see past the 1977 hippie prose

Any author who wants to make me wade through 1977 hippie prose to find out what he's saying can take a hike. (Similar considerations apply to artists, social scientists, psychoanalysts, and professors of literature.)

> to see the obvious fact
> that manipulating genes (circuit 7) which created the neural
> structures to begin with, is more flexible than simply playing
> with those neurons themselves (circuit 6). Or that manipulating
> genes is inherently less flexible compared to
> nanotechnology's (circuit 8) ability to manipulate the atoms that
> compose those genes.

So, if I dig through the hippie prose, I'll find a claim that the multilevel laws of biology and physics are somehow mirrored in the brain? I don't think so.

> > Where does genius come from? From computation. What kind of
> > computation? See
> > "Coding a Transhuman AI".

> One of Tim
> Leary's greatest wishes before he died was to see his 8-circuit
> model re-written
> by someone like you, with a more articulate and precise command
> of scientific and technological language. As I suggested
> earlier, this thread is a perfect example of why such a revision
> is increasingly necessary for a more cyber-savvy generation.

Go right ahead. Don't let me stop you. Let me know when I can read about distinctions between neurons, genes, and atoms instead of extraterrestrial ear lobes or whatever Wilson/Leary was gibbering about. 

-- 
        sentience@pobox.com         Eliezer S. Yudkowsky
         http://pobox.com/~sentience/AI_design.temp.html
          http://pobox.com/~sentience/sing_analysis.html
Disclaimer:  Unless otherwise specified, I'm not telling you
everything I think I know.