From: Cliff Stabbert (cps46@earthlink.net)
Date: Sun Dec 08 2002 - 18:34:07 MST
Sunday, December 8, 2002, 1:08:38 PM, Ben Goertzel wrote:
BG> http://users.rcn.com/standley/AI/immortality.htm
BG> Thoughts?
BG> Can anyone with more neuro expertise tell me: Is this guy correct as regards
BG> what is currently technologically plausible?
INABS but the "replace a neuron at a time" strategy has been raised
before (I think _The Mind's I_ features one such scenario). I don't
think isolating single "brain functions" is relevant with that method;
besides, as Michael Roy Ames points out there's more to mind than io.
I'm a lot less sanguine than the author when he states that
We have the technology to monitor and record all output and input
of a single neuron in vivo.
This has been done, and preliminary work with this technique has
shown that accurately modeling its responses to a given input is
not the daunting task many thought it would be.
but as I said, INABS. I just suspect that there may be subtle and
less subtle effects left out by such an approach, e.g. don't neural
connections grow/change when we "learn" stuff? And how does the whole
hormonal / neurotransmitter thing figure into neuron-level brain
simulation?
I'm also unclear on how this equals uploading. If successful, the
procedure would result in having your brain running on somewhat more
durable hardware, but still interfaced with all that goopey organic
stuff so prone to decay and rot. (And replacing that is a lot
trickier: I don't think we'd be able to handle the "body shock"
involved in moving to the much starker proprioceptive interfaces of
something on the level of say, ASIMO, we'd need to figure out that
nasty visual interface thing, etc.)
So you end up walking around with a laptop in your backpack with a
thick cable plugged into your spine, or a Beowulf cluster on a
handle-free Segway rolling along behind you...while your body
continues to age and no drug can get you high ;)
Overall, I suspect the author is overoptimistic in regards to what
current tech can do (in a real world full brain scenario, not a few
neurons in a lab), and there's nothing new about the basic strategy
he outlines. His implication that amnesia is a 100% hardware problem
somehow typifies something about the whole approach that I don't know
quite the right word for, Technocratic?
-- Cliff
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:41 MDT