h a l f b a k e r yI CAN HAZ CROISSANTZ?
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
InQ
Intellect Quotient: An alternate means to measure intelligence | |
In the field of AI, this would be a way to measure
progress.
IQ tests are useless for this. The simplest computer can
get
perfect scores on any IQ test but we know they're not
even
in
the same realm of humans or even lower life forms when
it
comes to accumulating, processing and utilizing
data. An
ant
can out walk, hunt, fight and survive a silicon based
computer many times its size. It can smell food, plan a
strategy of attack, kill and run the digestive system after
eating it, that is after it's calculated how much to eat
and
how much to take back to the colony. All this with a
brain
not much bigger than the head of a pin.
An intellect quotient would measure an artificial brain's
ability to perform particular tasks taking its size into
account.
The purpose of this metric would be to reach milestones
such
as having a computer that does everything an ant's brain
can
do that's the same size as a ant's brain or smaller.
So you'd pick the animal: dog. Create a computer that
does
everything a dog's brain can do, something we haven't
done
yet I might add, but I'll guess we could. Then we
calculate
the size of the artificial system necessary to replicate
that
biological brain. In this case we might be talking
"Dog:1,200
to 1 Intellect Quotient" in that the system would have to
be
1,200 times the size of a dog's brain to do the same
tasks.
An
ant? 800:1 intellect quotient.
We've been patting ourselves on the back with AI when
we
really should admit it's quite dumb compared to it's
biological
forebearers. This measurement would clarify the goals,
which
are to match and eventually surpass naturally occurring
data
processing systems on every level including size and
energy
requirements.
Pursuing AI from this direction might lead to some novel
new
ways to get the job done, and perhaps getting there
incrementally, by being able to say "By going down this
route,
we upped our ant InQ to 400:1 would be the way to do
it.
Recognising artificial intelligence
[xaviergisz, Nov 30 2015]
[link]
|
|
//I don't think its particularly important that an AI
could pass itself off as a human, but it would be a
good category.// |
|
|
Well, I think it would be the end game goal.
An Intellect quotient of "Human:1:1" |
|
|
We hit that, it's a very different universe we live
in. |
|
|
The main idea though is that approaching AI this way
might get some heretofore unconsidered approaches
looked into. Analog biochemical data processing vs
electronic digital for instance. |
|
|
//Speed another one or some metric of watts per
challenge// |
|
|
I was wondering how to approach that. I believe, and
I may be wrong, that biological data processors all
use
about the same amount of power per bit of data
processed. I could be wrong, but there you go. That
would be an area of study using this approach. |
|
|
See? Even posing questions about parameters of what is
intelligence gets the inquiry going. |
|
|
Having a hallmark like this to act as a point of reference
might be useful. |
|
|
//The simplest computer can get perfect scores on
any IQ test// |
|
|
I actually think that's very unlikely, but if you have
any evidence... |
|
|
Well, assuming you have a database of information in that
computer, it just searches it and gets the answer. It can
also do things like calculate Pi to as long as you want to go
where a human can't. |
|
|
Point is, there are things modern computers are very good
at. Running a complex system like an ant using the same
power and space available as the biological computer
currently doing the job, an ant's brain, isn't one of them. |
|
|
// it just searches it and gets the answer// Only if
that particular IQ test question is in its database.
Have you ever sat an IQ test? |
|
|
For instance, I am not sure it would be a simple
matter for a computer to look at a sequence of
five shapes (triangles and squares, with or without
different shading, numbers of dots etc) and come
up with the next one in the sequence (eg, spot the
fact that the number of dots is 1,2,4... times the
number of corners in successive shapes). |
|
|
My guess is that it would be almost impossible (at
present) to build a computer that would score
more than 50-70 on an IQ test, without extreme
overfitting to the particular questions in a
particular test. |
|
|
LOL, well, ok but that's not the point. |
|
|
The point is that computers are stupid, which as you point
out, is true. Ok, maybe stupider than I suggested but my
whole contention is that they're dumb. This is a
measurement that can be used to un-dumb them. |
|
|
The "dumb" computers we have at present tend to serve human agendas of one sort or another. Do we actually want computers to have their own agenda, which will inevitably conflict with ours sooner or later? If not, then why would we make it a research goal to mimic biological behaviour? |
|
|
To advance the human condition and continue down
the path of evolution rather than de-evolution. |
|
|
Computers have no agenda and never will unless we
put it into them. Life at its molecular level has
programming: to survive and expand, to render order
out of chaos. The programming at the molecular level
of computers, and all inanimate objects, is to go the
opposite direction, from order to chaos. A skyscraper
is "programmed" to turn to dust, not make more
skyscrapers. |
|
|
The consciousness that computers will achieve is
when
they meld with man. Unless there's an impediment to
this that I'm unaware of, eventually our consciousness
will be able to be transferred into "hard copy" like
any
other computer program. This is the most likely
scenario of how life conquers death, by bypassing the
limits of the biological consciousness host, the human
body. |
|
|
This goal is so programmed into our collective psyche
that all cultures, as far as I know, have the concept
of heaven or eternal life. This is because the life
drive is so strong it longs to continue. With
technology, the melding of man and machine, this
goal will be realized. |
|
|
What will we do with eternal life? Expand into the
cosmos because watching re-runs of reality shows and
looking at gifs of funny cats on the internet will get
boring after the twenty millionth millennia or so. |
|
|
Anyway, that's where this is heading in my opinion. I
know when it comes to boldly going into the future
it's a little intimidating to contemplate but when has
the future not been? |
|
|
The future belongs to those with the most life force,
the most drive and determination. Always has
been, always will be. Even when the body is cast
aside for membership in the
"Neuronet". "Eternalnet"? "Eternet"? "Outernet"? Naa. |
|
|
The Neuronet. I like that. |
|
|
This is probably what Quinetic is going to
called next year. |
|
|
Neuralnet? Neuropsphere? Galactic Neurosphere?
Intergalactic Neurosphere? |
|
|
How about getting plane in there? "Trans Galactic
Neural-plane" |
|
|
Neuralweb... Intergalactic Neural Web. |
|
|
Naa, it's gotta sound like someplace you'd want to
visit. A happy place. Lots of fun, bring the kids.
Ladies free, half off Wednesdays, casual Fridays,
come as you are. Hmm. |
|
|
Intergalactic Neural Continuum...mmmm nope. |
|
|
New Cortex sounds quite enjoyable as a destination. |
|
|
Neural Cortex maybe? Oh, I guess there's probably
already one of those in the human brain. |
|
|
Does putting "Infinite" in there someplace make to
sound like too much of a commitment? Eternal is
right out, that's just sounding like a retirement home
you'll never leave. |
|
|
Stars. Stars are nice. "Star-Net" "Stellernet" Eh. |
|
|
Intersteller Neural Cortex. Ok, I should probably go
do something more productive with my day than type
random words together to see how catchy they can
sound. |
|
|
I thought of that. Too deathlike. Plus good luck
changing the image of that place, it's pretty much
ingrained in the public consciousness. Angels,
everybody wearing sandals. At least in the west. Not
sure what other heavens look like. |
|
|
"New Improved Heaven?" "I Can't Believe It's Heaven"? |
|
|
Ok, that's my last dumb post for this idea. |
|
|
I like Skynet. I think by the time these systems are
ready people will have forgotten the movie. |
|
|
Although, maybe the free market will kick in and
people will offer different brands of eternity. |
|
|
"Tired of slogging around the universe at the speed of
light? Ask about Remulon's patented "Time Space
Bypass" technology. Mention this subnet broadcast
and get your first million years free! |
|
|
Remulon TSB... just because you have an eternity in
front of you doesn't mean you need to spend it on a
data transferrence particle beam." |
|
|
Revelation Space series: I'll check it out. |
|
|
Wonder if the "I'll put it off till tomorrow" problem
that's a small foible of limited life folks would
become a real problem with somebody who had
forever to get the job done. Would time slow down
for them? For instance, would a regular limited life
entity see their actions as glacially slow? Supposedly
a fly sees us as moving very slowly compared to how
sped up their life is. If there was no
rush, would your relative time constraint perception
be so warped that you might take a thousand years to
reach for a piece of toast for instance? Not that you'd
eat toast, but you get the idea. |
|
|
Right now our psyche, our entire consciousness
is designed to fit with the bodies we have now. There
would need to be a fundamental different wiring to
start living thousands or millions of years or beyond. |
|
|
And what about interaction with other entities? What
would drive that? Currently we have sex and
reproduction and the tending of the physical body to
drive us together. (Helping each other survive with
food, shelter etc) Would that be cut out of the
equation? Would there be some kind of analog to
biological interaction in the "Pan Galactic Ultranet" or
whatever? |
|
|
Clearly it would be a very different existence. I think
it's fascinating stuff to contemplate. |
|
| |