On 17/10/2007, Edward W. Porter <[EMAIL PROTECTED]> wrote:
> IN RESPONSE TO
>
> "Hmm, how then is a modern PC valuable? It has no representation of the
> type advocated by AI designers interested in that sort of things.
>
> BUT IT DOES HAVE REPRESENTATION IN THE FORM OR CODE AND DATA.  DOES IT
> HAVE THE TYPE OF REPRESENTATION ADVOCATED BY AI DESIGNERS.  NO! BUT CAN IT
> COMPUTE IN THE WAY ADVOCATED BY AI DESIGNERS. NO!
>

Pedantically speaking, it can[1]! But no one knows the correct code
[2] to start it off with. Which just brings us back to the machinery
giving rise to the semantics.

A representation is nothing special, all it does is make some
transformations easier than others for us to see. With the correct
machinery, as long as one representation contains the same information
as another, they can both be searched as efficiently (as in number of
trials rather than computational efficiency) or compared easily.

Why do I say this? Because if they contain the same information you
can always translate one to the other and do the
manipulation/comparison and then translate back. See for example
storing the date and time as micro seconds from a certain date vs
storing as a string.

 Will Pearson

[1] Unless you are suggesting that intelligence can't be implemented
on a PC of course.

[2] If you allow self-modifying code (which the modern PC can), the
differentiation between code and data becomes near meaningless.

-----
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/?member_id=8660244&id_secret=54751111-b0e83b

Reply via email to