From: Billy Brown (bbrown@transcient.com)
Date: Mon Nov 08 1999 - 11:36:39 MST
Pvthur@aol.com wrote:
>> "Computing power doubles every two subjective years of work."
>
> Ok. Given this idea, if my brain were reprogrammed to run twice as fast,
how
> would my backhand be effected?
At the moment the quality of your backhand has nothing to do with the speed
your brain runs at, so there is not likely to be any change. However, if
you were an AI (or an upload, or any other sort of entity capable of doing
internal re-wiring) things would be a little different.
A tennis-playing AI would presumably contain software for predicting the
flight of a tennis ball, controlling the motion of its arm, observing the
actual state of the game via its sensors, making abstract strategy decisions
on the fly based on the situation and the rules of the game, and relating
all of these different modules together so that it can actually act in its
decisions. If the AI plays on the same level as a human it is likely that
this whole assemblage has trouble keeping up with actual events in a real
game - that is, after all, one of the criteria with which we separate
"interesting" games from "boring" ones.
If you double the available computing power, you can go back and look for
bottlenecks in all of these modules. You can speed up the flight-prediction
module so that it gives an updated result the instant your opponent's swing
is complete (or even before then, based on the trajectory of his racket).
You can run the strategy module faster, so you actually have a chance to
think about what to do. You can improve the precision of the whole motor
control subsystem, so that your swing is less likely to miss and more likely
to actually cause the trajectory change you want. You can try out all these
ideas, along with dozens of others, and then pick the ones that turn out to
have the biggest effect on your game and keep them.
If you think of intelligence enhancement in terms of improving some
abstract, nebulous unit of "intelligence", it can seem rather improbable.
If, OTOH, you think in terms of improving your ability to perform some
concrete task, the way to proceed tends to be obvious for any field that we
have even a modest understanding of. The contention of the "strong-SI" camp
is that this trend is likely to hold for those aspects of intelligence that
we do not yet have any understanding of, at that we should therefore expect
an exponential increase in processing power to make possible a rapid
increase in the effective intelligence of an AI.
Billy Brown, MCSE+I
bbrown@transcient.com
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:05:44 MST