From: David Hart (dhart@atlantisblue.com.au)
Date: Sat Aug 10 2002 - 19:27:05 MDT
I agree that advanced AGIs and SIs that follow them will probably not
use human-brain-signal-based, human-visual or human-language type
programming paradigms as their fundamental software substrate, but they
will eventually need to be deeply acquainted with and fluent in all
three (and probably others as well).
An interesting application of intimate knowledge of human-brain systems
is the translation of uploaded brains to more flexible substrate
(whether software, hardware or both). While uploading is a nice, mundane
SL3 concept, what happens after uploading quickly ventures into SL4.
Simply making uploaded brains run faster, or interface with existing
computers is still SL3. I suspect that shortcomings of the signal-based
nature of the brain (now emulated in software, or running in a new
hardware equivalent) will become quickly apparent to anyone trying to
drastically reprogram or augment it. Will one of the first SL4
human-relative feats of an SI be to engineer a new computing substrate
and translation system for uploaded human brains? Could it eventually
manage, through iterative translation, to migrate a human mind (or
whatever it has become) to its own substrate, thus using a unified
substrate?
David
Ben Goertzel wrote:
>I agree with that -- advanced AGI's will develop nonhuman ways of
>programming. Our programming paradigms are based on the linear-syntax
>nature of human language, whereas AGI's won't communicate using linear
>syntax in the human-language sense.
>
>But I consider it very very unlikely that the new prog. paradigms introduced
>by AGI's will be based on either
>
>-- human brain metaphors, or
>-- humanly simple visual metaphors
>
>as I think the first real AGI's will have neither humanlike brains nor
>humanlike visual systems...
>
>ben g
>
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:40 MDT