From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!uunet!tdatirv!sarima Wed Dec 18 16:02:28 EST 1991
Article 2218 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!uunet!tdatirv!sarima
>From: sarima@tdatirv.UUCP (Stanley Friesen)
Newsgroups: comp.ai.philosophy
Subject: Re: Scaled up slug brains
Message-ID: <330@tdatirv.UUCP>
Date: 17 Dec 91 19:39:43 GMT
References: <12723@pitt.UUCP> <60372@netnews.upenn.edu> <349@idtg.UUCP> <45031@mimsy.umd.edu>
Reply-To: sarima@tdatirv.UUCP (Stanley Friesen)
Organization: Teradata Corp., Irvine
Lines: 52

In article <45031@mimsy.umd.edu> harwood@umiacs.umd.edu (David Harwood) writes:
|	As I said before, despite the neuron doctrine, I guess that the
|neuro-morphogenesis of different species (and of individuals, for that
|matter) is very specific, and sensitive to small genetic variation. 
|(How many genes do slugs have anyway? They don't look like they have many.)

They have aproximately the same number as we do.  All multicellular animals
have DNA contents that are of the same order of magnitude.  The variation
is at most about triple.  (Compared to ricketsias and viruses, this is
totally insignificant).

Brain structure, at least in forms with complex brains, tends to be *very*
conservative.  The brains of the apes even have the same sulci as ours do,
and *that* is just a topological feature to allow better packing!

Our brainstem is still very little changed from our reptilian ancestors.

|Scaled-up slug-brain NN simulation ignores genetics, biochemistry, 
|neuro-morphogenesis, etc, as if all neurons were glorified transistors no
|matter how they were manufactured.)

Sigh, such a limited imagination on how to scale something up.
I have never said, and I can remember no one else saying, that a mammalian
brain was simply a isomorphically magnified slug-brain!  All we are saying is
the the same *operational* *principles* apply, albeit to a system with far
more interactions and abstractions.

|	So my guess is that no matter how you cut and paste slug-neurons,
|you will never get an intelligible bark out of neo-Fido; indeed, if you
|tinker very much, you won't even get an ordinary quack out of your neo-slug
|since the pieces of the puzzle simply won't fit.

Again, a very simplistic, and unintended, interpretation.

It is well known that there are many kinds of neurons, and that humans have
more kinds than slugs.  So, I would *never* expect to simply replicate slug
neurons and build a mammalian brain.  [The reverse might be possible though,
to select appropriate neurons from a dog and build a slug brain].

|	I'm not denying the possibility of creating artificial intelligence,
|but it probably won't be human intelligence, whatever it is.

Probably not, at least because it would develope in a very different environment
than human infants.

And it is not likely to be economically desirable to build exact duplicates
of humans anyway, we are likely to want entities that can do well what we
do poorly.
-- 
---------------
uunet!tdatirv!sarima				(Stanley Friesen)



