RE: Implementation

From: Gilles HENRI <Gilles.Henri.domain.name.hidden>
Date: Wed, 4 Aug 1999 10:40:39 +0200

>Gilles HENRI <Gilles.Henri.domain.name.hidden>:
>> ... HLUT. As defined by Hans, it is ONLY programmed to
>> handle language, through exchange of ASCII characters
>> (am I wrong, Hans?).
>
>That's as far as the last posting went. But it could have gone
>further to define the "robotic" HLUT, organized exactly like the
>synchronous ASCII version, but with a finer time resolution (maybe a
>microsecond cycle) and connected to a robot instead of a teletype.
>It would be fast enough to absorb serially-multiplexed data bytes
>from digital cameras, microphones, force and other digital sensors
>distributed on the robot, and output control bytes to digital
>controllers for the robot's motors. Instead of encoding every
>possible typed conversation its inhabitant could possibly have, the
>table would encode all possible physical life experiences the robot
>could have (using only 10^(10^16) bytes for a century-long life).

only 10^(10^16) bytes...hum. However, it will still be insufficient because
some of our inputs/outputs come from our own analogic structure (our own
body). If your robot has another structure and has a correct representation
of its environment, he must abandon the idea that he is a human.What if I
ask him "how far can you spit?" or "Try this newly dicovered drug and tell
me how you feel?". (Contrary to all HLUT or UD-type machines, this kind of
questions is very usual and common among human beings!!).
I think you will agree that your robot should be programmed to answer
following its own structure rather the human one (answering "I have no
saliva to spit" or "I cannot absorb any drug!") . But how do you imagine
the physical structure of a machine handling 10^(10^16) bytes?

Gilles
Received on Wed Aug 04 1999 - 01:45:04 PDT