>You dont need 10^(10^16) bytes as one response is approriate to an enormous
>range of inputs. A subsystem between the sensors and the HLUT converts
>inputs into one of say 10^6 possible input-classes which is checked against
>a much smaller HLUT. Or better yet, the input bounces around a huge array of
>tiny HLUTS until one picks up on it and responds by modifying the input (an
>internal qualia-like experience) or requesting an output. Funnily enough,
>that's how your brain works.

I totally agree with you. What I suspect is that the mapping (in fact
compressing algorithm) between the initial 10^(10^16) initial possibilities
and the much lower number of actual inputs of the brain is made very
quickly and routinely by the brain because it is largely analogic. Analogic
systems are very efficient to solve instantaneously the particular problem
corresponding their own structure, for example a flow of water solves
instantaneously a particular solution of a the Navier-Stokes equation! Of
course they are completely useless to solve another problem; What I think
is that if you want to solve the compressing problem EXACTLY in the same
way as a particular brain, you have only two solutions:
* look at this brain itself
* construct so a huge digital machine (HLUTs) or wait such a long time (UD)
that it is totally in fact totally impossible to realize in our world. In
fact I suspect that solutions (2) require also a total knowledge of the
physical laws and the state of the Universe because of the possibility of
experiments concerning still unknown laws.
The two solutions are so far from each other that I see no reason to accept
automatically that solutions (2) have the same properties of consciousness
as solution (1). This would have the consequence of forbiding the
duplication of conscious machines, as I already argued : a conscious
machine could only be conscious to be THAT machine at THAT location and
THAT time, much like you know you are James Higgo, being in your office at
the time you are reading this mail, and nobody else can consistently think

Reply via email to