From: Carl Feynman (carlf@abinitio.com)
Date: Wed Aug 01 2001 - 10:31:36 MDT
The distinction between a dumbass and an AI researcher consists of both
intrinsic smartness and education. From the tenor of the rest of your
message, I assume you're interested in the intrinsic smartness part of the
distinction. Psychologists like to reify 'smartness' by something called
the 'g factor', which is supposed to differ between people. IQ tests
measure the g factor, and the amount of g factor is commonly called IQ.
Here's a semi-popular article on the topic, most of which I agree with:
http://www.sciam.com/1998/1198intelligence/1198gottfred.html
The higher the g factor, the better a person is at all kinds of mental
functioning, including 'non-reasoning' tests like response time and accurate
perception of small stimuli.
Here's a quote from the article:
Moreover, research on the physiology and genetics of g has uncovered
biological correlates of this psychological phenomenon. In the past decade,
studies by teams of researchers in North America and Europe have linked
several attributes of the brain to general intelligence. After taking into
account gender and physical stature, brain size as determined by magnetic
resonance imaging is moderately correlated with IQ (about 0.4 on a scale of
0 to 1). So is the speed of nerve conduction. The brains of bright people
also use less energy during problem solving than do those of their less able
peers. And various qualities of brain waves correlate strongly (about 0.5 to
0.7) with IQ: the brain waves of individuals with higher IQs, for example,
respond more promptly and consistently to simple sensory stimuli such as
audible clicks.
<End quote>
I want to respond to an unspoken assumption that some posters to this list
have been making: that a person with an IQ of 50 is "half as smart" as a
person with an IQ of 100. This is an unwarranted assumption. An adult with
an IQ of 50 can walk around without hitting things, ride a bike around town,
interpret facial expressions, use and understand more than a thousand words,
and use and recognize hundreds of ordinary objects. All these tasks, in
aggregate, use far more of the brain than the kind of abstract reasoning it
takes to do AI research. The difference is that takes a person with an IQ
of 50 much longer to learn these things than it does someone with an IQ of
100, and they never get as good as it. So I think a person with an IQ of 50
is something like 90% as smart as a person with an IQ of 100. (It could be
80% or 99%-- this is assigning a precise number to an imprecise concept.)
--CarlF
Dani Eder wrote:
> If we knew more about what distinguishes a
> dumbass from an AI researcher in humans, it
> could help us figure out how to design a better
> AI.
>
> We know that the gross size of the brain is not
> what distinguishes a dumbass from an AI researcher.
> The variables I'm aware of that could be relevant
> are (a) number of synapses and how well they are
> organized, (b) different allocations of the
> brain's regions to tasks, and (c) neurochemical
> differences like production rate of neurotransmitters
> or number of receptors per synapse.
>
> Are any of you aware of research on the above, or
> have additional variables I didn't mention?
>
> Daniel
>
> __________________________________________________
> Do You Yahoo!?
> Make international calls for as low as $.04/minute with Yahoo! Messenger
> http://phonecard.yahoo.com/
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT