Re: Is generalisation a limit to intelligence?

From: Joaquim Almgren Gāndara (
Date: Sat Dec 02 2000 - 15:49:31 MST


> Since the "brain only pays atention to one thing at the time"
> (sorry I don't know a neurological, low level way of putting this)

The brain is a massive distributed parallel processor; it pays attention to a
lot of stuff at the same time. Perhaps you're referring to our consciousness?
Anyway, I see what you mean.

> not
> making a mistakes will occour if you
> operate at each level of abstraction at the time, or find
> ways to leave subneural nets operating at the lower levels
> by themselves.

So you're opting for either a systematic iteration through different levels of
abstraction, or leaving the lower levels to some kind of subconscious mechanism?
In effect, that would still be a system that is somehow "too perfect". I believe
that by "perfecting" the thought processes of an AI, we would inadvertantly and
inevitably wreck the organic nature of true intelligence. I think it's inherent
in the fact that intelligence builds upon generalisation, and that Mother Nature
the Tinkerer is to blame. You gave a perfect example of what I see as
intelligence in your own e-mail:

> At this point, you might filter the board SO MUCH to
> the point that you might miss the simple fact that a bishop is
> attacking a rook.

Intelligence involves making mistakes. Therefore, depending on how you rate the
intelligence of a sentient being, there is a limit to intelligence, where the
being either starts making too many mistakes or looses its sense of generality.
I believe it is commonly known as "insanity". ;)

- Joaquim Gāndara

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:35 MDT