[extropy-chat] 12,000 IQ and nothing on?

[Anders Sandberg]

Lee Corbin wrote:
> Arthur Jensen specifically warns against trying to extend the notion of
> "cognitive ability" to other species, but what the hell, this is the
> Extropians list. Even among humans and our closest relations there
> are interesting results. Ashkenazi Jews have IQs on average of at
> least 112, and the Kalihari San sport, according to Richard Lynn
> in the 2006 book "Racial Differences in Intelligence", an average IQ
> of 56.  One might say that the average Bushman is half as smart as
> a Jew  :-)

I would be a bit cautious about those very low IQ scores found in Lynn &
Vahanen, just as Satoshi Kanazawa found that Mississippi had 62.7 as
average IQ:
http://www.lse.ac.uk/collections/methodologyInstitute/pdf/SKanazawa/I2006.pdf
There is a tricky interaction between education, culture and intelligence
where they can both support, hinder and hide each other. My guess is that
a lot of the very low scores are simply due to lack of education and
cultural misunderstandings (see the hillarious dialogue between the great
soviet psychologist Luria and an Uzbek farmer:
http://cscs.umich.edu/~crshalizi/weblog/484.html ). Still, intelligence as
measured early in life does predict educational success, income and
professional success fairly well, even if IQ measured in late life may
have become a mixture of innate capacity, learned abilities and ability to
conform to tests.

>  Also, the genius bonobo chimpanzee Kanzi seems to
> understand about as much as a European four-year old, and so we
> might go on to say that he's almost half as smart as a human being,
> or nearly a quarter as smart as an adult European.

The problem is that IQ is not a ratio scale (no natural zero), it is at
best an interval scale. So it doesn't make much sense of calling somebody
twice as smart in the IQ sense. And general intelligence is at best
ordinal: we might measure greater or lesser intelligence, but it is no way
of measuring one unit of intelligence.


>> One could imagine putting various beings into competition where they
>> play randomly selected games neither has ever seen before. Their
>> ELO scores would to some extent general intelligence over the domain
>> of these games. The problem is that that domain may be rather narrow.
>
> Another idea I've had is to just pit them against each other in life and
> death struggles.

That would be one of the possible domains. Very motivating, but
destructive on the participants.

I think there is a potential test in Heinlein's quote: "A human being
should be able to change a diaper, plan an invasion, butcher a hog, conn a
ship, design a building, write a sonnet, balance accounts, build a wall,
set a bone, comfort the dying, take orders, give orders, cooperate, act
alone, solve equations, analyze a new problem, pitch manure, program a
computer, cook a tasty meal, fight efficiently, die gallantly.
Specialization is for insects."

Some of the entries might be more about social or emotional intelligence,
but ideally they would correlate with general intelligence.


> Yeah, I'm afraid you're right.  Jensen says "Cognitive ability is a lot
> like
> money; it doesn't really matter how much you have so long as you have
> a certain enough."  I take him to mean that insofar as ability to
> accomplish
> goes (among humans today), an IQ of 130 or 150 or something is all
> you need.

My research suggest that low ability is indeed the biggest problem. Once
you go below 100 IQ points, problems start to rise rather quickly. Whether
there is an advantage in going from 130 to 140 is less obvious. However,
at least one study demonstrated that even among the top 1% performers
there were significant differences in professional success (PhDs, tenure,
income) and number of patents between the top and bottom quartiles. The
patent part is interesting, because that is non-competitive: it just
represents crystalized creativity and signifies that these people actually
do contribute significantly to society.


>
> So in the far future, a maximal entity at a certain point in time will
> extend
> his control over his environment (probably incorporating it into himself)
> at a rate proportional to his intelligence. Standing back a ways, this
> velocity will probably be the speed of light, since the entity would
> doubtless
> resort to the Von Neumann probe approach.   On the other hand, if we
> restrict its resources temporarily to, say one cubic meter, or say, to one
> Jupiter of material, then we may claim that an upper limit of ability must
> exist.
> But these levels will be maximal---as you are essentially suggesting above
> ---and no one of them a maximum.

Communications limitations likely keep expansion limited. It is somewhat
uneconomical to have to wait for coordinating one's subsystems. My guess
is that entities will instead distribute along some power law
distribution: a few really godlike ones, many more godlings, lots of
human-level entities and hordes of smaller "animats".

-- 
Anders Sandberg,
Oxford Uehiro Centre for Practical Ethics
Philosophy Faculty of Oxford University