[EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] wrote:
>
> This
> endogenous information (as apposed to perceptual data such as
> vison and sound
> which give us info on the world "out there") would be
> difficult to emulate
> in a computer I would think and what would be the point?
If you're aiming at either generic consciousness or human-type
consciousness (either from scratch or emulating a scanned person), and
you've determined that this is vital to the goal, then that's your
point.
Difficult just means it would take a while to do. Even right now we
could hook a gazillion simulated neurons up to a Quake-type environment,
which provides visual and auditory sensation. Adding proprioreception
and some limited tactile feedback would be trivial. The components I
think we're missing are (1) the ability to simulate the right number of
gazillion neurons in real-time, (2) knowledge of the basic structures
and interconnections all humans develop, (3) a deeper and more complete
understanding of how the various non-neuron based elements affect the
operation of the brain (e.g. hormones, etc).
(1) Will come with time - I've heard estimates of 2020 to 2050. With
special-purpose hardware (e.g. 1 billion neurons on a chip acting in
parallel vs. 1 n-th generation Pentium chip simulating 1 billion neurons
in sequence) it could come much more quickly. According to a recent
SciAm article, a 1GHz processor can simulate the human retina -
something like 0.02g of neural tissue. The brain is (something like) 2kg
- 100000 times heavier. A lot, but not implausible.
(2) & (3) So get on it, neuroscientists. :)
> The
> machine would
> need to monitor its own internal state which would key to its
> survival rather
> than emulate an organic body.
For it to be practical as a robot or as anything other than a brain in a
similated environment, yes. However, don't discount a sim resident -
they'd be remarkably useful in certain circumstances, and depending on
how critical environment & body is to consciousness, it might be the
only type of non-human consciousness we could have a meaningful dialog
with. A conscious robot, with its own self-awareness, might be too
different from us to have anything in common or any ability to
communicate with us.
> The hard part of the modeling of all the processing
> that goes on
> between optic nerve and conscious mind that's difficult.
Some argue that this processing is the bit that's conscious - there's no
pipeline leading from the eye through some processing to a Cartesian
"theater of the mind" where the little conscious homonculous is sitting
and being the conscious for the body. Rather, its the processing itself
that's the place where consciousness arises.
Joshua
