Re: Keeping AI at bay (was: How to help create a singularity)

From: Damien Sullivan (phoenix@ugcs.caltech.edu)
Date: Fri May 04 2001 - 15:15:59 MDT


On Fri, May 04, 2001 at 01:26:40PM +0200, Eugene Leitl wrote:
> On Fri, 4 May 2001, Damien Sullivan wrote:

> > At the moment. I'm not inclined to believe it's inherent, until we
> > know a lot more about the human brain.
>
> Well, it's a mess. Whatever it tells us, it doesn't tell us how to build
> cold, hard, clean designs, to appeal to the inner anal-retentive in us. If

Yet the brain is modular, or at least specialized. There's a part key to
recognition of human faces. There are parts key to different parts of
language.

Of course, the APIs are probably non-existent. But so far the brain doesn't
really tell us much of anything, because we've barely started figuring it out.
We're still developing the tools to measure with.

E.g. Pinker's _Words and Rules_, exploring how language works in the brain
just from mistakes in using the past tenses of irregular verbs. Proving to my
current satisfaction that we do have rules, an unconscious syntax and use of
variables not well captured by current neural network models.

> > And I'd avoid this fetishism of low-level evolutionary processes.
>
> The world *is* made from nails. Look, if you disagree, I have this neat
> little argument here, to pound my point home...
 
But you accept some sophistication with your hammer, you find that some
"nails" go in much more quickly and securely with a screwdriver.

> > thoughts. But a child learning chess may try to move anywhere, and be
> > swatted away from illegal moves. A chess program only explores legal
> > moves. A human grandmaster only explores good moves. If she gets
>
> I have no idea what a human grandmaster explores, actually. (As opposed to
> what he tells us he thinks he's exploring). I would like to see an fMRI of

We do know that normal players, when they make mistakes in replicating a
board, will shift a piece one square or something like that. So the board
looks similar physically, but may be rather different in game terms. A
grandmaster who makes a mistake makes a bigger difference in where pieces are,
but preserves more of the game balance. Clearly she's thinking differently
than the normal player; she's got a high-level conceptual _sense_ which the
player doesn't.

I just don't like to see this defeatism about the human brain, and early
turning to the blind idiot god of natural selection[1]. Yeah, selecting on
bitspray can work, but selecting on higher-order stuff is much cooler and
faster.

-xx- Damien X-)

[1] Lovecraft reference. "Azathoth, the blind idiot suzerain of the universe,
gibbering madly at the center of all things." I've thought that's a wonderful
metaphor for natural selection. The gibbering can be the random mutations.

I don't think Yog-Sothoth maps to anything real. Laws of physics, maybe.



This archive was generated by hypermail 2b30 : Mon May 28 2001 - 10:00:02 MDT