Re: How Robots Will Steal Your Job

[Roedy Green]

By
answering that question carefully and accepting Roedy's definition of
intelligence, we can make just about *anything* appear intelligent.

You need somebody or something that wants or needs the problem solved.

There has to be a problem -- options on how to proceed, some of which
lead to more desirable outcomes than others.

This implies a definition of desirable by someone or something.

My definition is quite broad. If a virus "wants" to persist against a
drug, and develops ways around the drug, and it outwits the
scientists, it has exhibited a form of intelligence, even if mindless
intelligence. It has solved a problem that would severely tax a
terrorist.

The Madagar aye-aye needed an untapped food source. The process of
natural selection and genetic variation came to its aid gradually
extending its middle finger into a long probe to go for grubs buried
in trees. Evolution solved the problem for the aye-aye, albeit very
slowly, yet with perfect upward compatibility. That could be looked on
as a form of problem solving -- how to survive better.

Life is a sort of universal intelligence game. Either your kind play
to survive better or you die. By natural selection, the game forces a
built in motivation to play well on all players.

A highly intelligent problem solver is able to model and discard
various solutions without having to actually realise them to test
them. That's what we are learning to do now in developing new drugs.
The deeper your understanding of the fundamental laws, the better able
you are to model and predict. Viruses don't have a deep
understanding. All they can do is randomly mutate and "hope" for the
best trusting their sheer force of numbers to come up with a solution.

Part of intelligence is speed at coming up with solutions.

 

[Roedy Green]

I've observed that animals /do/ have a sense of past and future.

There is the famous story of the dog that went to wait for his dead
master at the evening train each day.

That suggests the dog remembered the master, and that he had a notion
of time of day.

Dogs don't usually have a way of distinguishing being abandoned from
the master dying.

Even corals have a sense of timing that lets them all release sperm
and eggs on the same day relative to the moon.

Elephants grieve their dead, handling their bones.

Dogs grieve if a companion dies. They must remember the past.

Do bears worry about being fat enough to get through the winter? What
would it take to find out?

 

[Roedy Green]

If someone later solves it and lets you peak in the black box and you
see it is a brute force search of likely regions of solution space
using special purpose hardware to help, then you decide this no longer
counts as intelligence. You move the goal posts.

I look at intelligence perhaps the way a pragmatic businessman would.

Imagine a black box. You give it a problem it comes up with a
solution, the faster the better. If it comes up with a good solution,
I call it intelligent. It does not matter what is inside the black
box.

It might be a mess of DNA molecules combining in random ways and being
culled for the solution.

It might be microscopic life forms being pushed to the limits of
natural selection to produce some drug.

It might be a little Jewish man with a long white beard.

It might be a neural net.

It might be a Cray.

It might be a team of Indian Java programmers and some PCs.

It might be a TV Show offering $1million for the best solution.

Yet to count as intelligence, we often want the intelligence to come
in a familiar package, preferably with unkempt hair and a slightly
foreign accent.

It is cute to see how people adjudicate paintings done by elephants
and cats differently when they know the author. I was quite impressed
by some cat paintings. They were far from random.

 

[Airy R Bean]

It is not a deception, it is the underlying survival "instinct"
but expressing itself in a group, or social, form. (An example
of group action but without intelligence on the part of the
individual is the ant colony)

 

[Airy R Bean]

We have the same model, that of an underlying reaction to
a life-threatening situation. it is only the self-aware body that
interprets this as "pain" because we might otherwise proceed on
a course of action that does not minimise the threat.

"Pain", therefore, is a necessarily evolved mechanism for the
self-aware to survive, and only exists in the self-aware.

 

[Airy R Bean]

If it runs on a computer, then it is being simulated by machine
code. How much easier to simulate it in Java or Forth!

 

[Airy R Bean]

A few random thoughts, not well presented because I've
something else pre-occupying me and don't have the
time, follow.....

-----oOOOo-----

A good question. It seems that most of the people contributing
to this thread haven't got a clue about the AI (or even I) topic
being discussed. As (I think it was you) was said, there seems
to be a quasi-religious protest about the whole thing; which
causes emotional responses which prevent cold calculated
thought. If we term ourselves as the meta-self-aware, then
most of the contributors to this thread are to us as they are to
cats and dogs - any apparent contribution is only apparent because
it shares the same characteristics of words about the topic
but contains no real contribution; such as the fuss about
bottom posting. I am not a top poster, for example, I
am just a poster. (That last is interjected as a trap to
disambiguate the self-aware from the meta-self-aware)

Most seem to be skimming about on the surface, when you
need to get under the skin. Propose your engineering model
of brain action from your deep understanding of what is there.

It is necessary to dissociate the various elements of the brain
and their manifold interactions to get to the heart of the matter.
Emotions (which are elemental chemical stimuli), memories,
intelligence, stasis mechanisms.

Try to dissociate one's experience of "self" from the
experiences around us, sight, sound, smell, touch, electric fields
and it seems difficult. Take this with the non-self-aware
"experiences" (for want of a better word) of sleep, (a
dissociation from external stimuli, and in the case of
dreaming, associating with memories of previous stimuli)
anasthaesia, epileptic fits, pre-birth, infancy and death....

My current thinking on self-awareness is that we are seeking
something that is not there; because we cannot define to
ourselves the nature of the phenomenon. We only experience the results of
a super-capable pattern matcher, that is able to correlate
the external world from within the organism that does the sensing
and correlation.

An early maxim on the "science" of psychology suggests that
complex claims should be set aside if simpler mechanisms can
account for behaviour exhibited.

Therefore - my answer to your question is that I do not know (yet)
but that is not to say that the answer must lie within one of the
other explanations for then we fall into the world of make-believe
and lunacy that is the world of the religionist.

 

[Don Stockbauer]

Yes, I imagine they all talk about the same thing. However there is a
tendency for people to blather about how the 'artistic' or
'creative' side is undervalued by society, when the truth is more likely
that it is overvalued by a lot of people who don't have much of a bump
on either side. The problem with the 'creative' side is that it's not
so good at evaluating the worth of its own efforts...

Gerry Quinn

The either/or fallacy? Holism? Synergism? Local vs. global? "Not
invented here"? Cooperation? Or extinction?

 

[Bent C Dalager]

Let me guess, you're posting from the philosophy group right? :>

Attempting to paraphrase your statement in (slightly) less ambiguous terms:

"You can use random input plus selection to produce, or increase
progress toward, a desired outcome."

Is that close to what you intended to say?
Yes.

What I have a problem with is the sophistry exercised by Roedy -
specifically his mixture of two different meanings of the word 'direction.'

Ok. I must say that on the face of it, it seems like a disagreement
over semantics, but I must admit to jumping in without having read the
post of contention so I can't really comment intelligently

Cheers
Bent D

 

[gswork]

You can go to Disneyland and play with the animatronics that can play
that same game. There does not necessarily have to be anyone home.

I have a pet south american catfish on my desk. He has learned my
actions when I am about to feed him, and he gets quite excited and
races up and down. He is obviously aware of me, and has computed some
way of guessing when feeding time is. But is there anybody home
inside? I presume there is. Most people would presume not.

I can imagine a machine being able to closely emulate human behaviour
(including quite complex motivations and emotional reactions) if
programmed sufficiently well, having enough input and information.
Such a machine would, at a glance, be the same as a living person -
yet it would be a 'mechanical creation'. It may be more complex an
entity than your catfish yet we could argue it was less 'alive',
indeed not 'alive' at all.

What is also interesting is that humans appear to be able to 'break'
in ways not dissimilar to machines/computers. Humans can suddenly
become irrational in their reactions and behaviours for reasons
overidingly related to brain chemistry - the mechanics of thinking.
We would (hopefully) not think of such a person as less human.

Indeed, IMO, the emergence of genuine 'AI' is more likely to come from
'guided' evolution by virtue of manipulation of DNA and upbringing
than [directly] via the technologies of computation. And here the
'grey' area is even more grey.

Then there is this whole business of out of body experiences. I know
people are telling the truth, since I had one myself. Now this does
NOT prove your consciousness can live WITHOUT a body, just that you
can have an experience that feels exactly like being outside your
body. However it suggest that perhaps consciousnesses inhabit bodies.

It suggests a physical [brain] existence too.

I thought out of body experiences were now more understood, in that
studies of pilots in high-g stress situations found they had similar
experiences, and actually the experience is somewhat replicable 'in
the lab'. That they are likely to do with oxygen deprivation /
extreme stress and the hormones therby released.

(for many the notion that it has a physical cause takes the wonder out
of an awe inspiring experience)

The question that intrigues me is not can computers or other creatures
compute things that man can, or compute things he can't, but is there
anybody home in them, and if so how is it different from the thing
that is home in me that experiences pain and pleasure and worry.

Does this experiencer just materialise wherever there is a "brain" or
maybe it takes much much less a built-in side effect of Quantum
Mechanics. Or are they like Buddhist souls seeking incarnation?

It seems to me as if most people act as if there was nobody home in
other beings human or otherwise. They seem to treat them like objects
in the way, or giant playdolls. I think happen partly because of the
limitations of language. Everyone else seems so dull and stupid
compared with yourself. Inside your mind is going a mile a minute.
Everyone else's minds appear stopped, uttering a clumsy fragment every
once in a while. Yet if you look at your own inadequate utterings,
how far they are from what you really meant...

A lack of empathy too, in failing to perceive other people to be
equally valid to oneself in the sense of 'being alive', and seeing
them only on the surface in terms of their immediate actions.

 

[Hans-Georg Michna]

Indeed, IMO, the emergence of genuine 'AI' is more likely to come from
'guided' evolution by virtue of manipulation of DNA and upbringing
than [directly] via the technologies of computation.

I think the opposite will happen. The only advantage of the
biological way is that it could evolve in small steps through
natural evolution. Otherwise it is not optimal at all. It is
slow, unreliable, and fragile, and it requires a biosphere.

Once natural intelligence (human intelligence in the case of
this planet) grows high enough to explore other ways, better
ways will be found and, I think, have been found already.

Hans-Georg
http://www.michna.com/transition.htm

 

[Hans-Georg Michna]

On Mon, 25 Aug 2003 13:13:13 +0200, Hans-Georg Michna

Raw size is just a cheap trick. We humans like to pretend we have
some complex circuitry unlike anything else in the animal kingdom.

Changing size is something evolution handles with ease. Devising new
structures is something else.

In computers, there is quite a difference in what a 4K computer and a
4 gig computer can do. However, It would not be obvious if you had
only seen a 4K computer.

Up to now, human had a terrible speed disadvantage and a massive size
advantage over computers. That size disadvantage is rapidly being
whittled away. Computers also have the advantage of deliberate
design. The human is a kludge that sort of worked. If you don't
believe me, ask a woman in labour.

Roedy,

I agree entirely.

Hans-Georg

 

[Hans-Georg Michna]

Some observational differences I've noticed are: animals don't
really have a sense of past and future; and animals don't wonder
"why".

Fairly obvious for most, but not, for example, for elephants.

On the sidelines, elephants' brains are roughtly three times as
big as ours. This doesn't prove anything, but elehants appear to
have an extremely good and detailed memory, and the dictionary
of elephant vocalizations currently stands at around 73 and is
growing rapidly, as fast as research grants allow. One of these
vocalizations has been labelled "discussion" and could possibly
stand for a variety of meanings.

Hans-Georg

 

[Hans-Georg Michna]

Elephants grieve their dead, handling their bones.

Roedy,

very true. And they only do this with elephant bones. When they
come across a deceased gnu or some other skeleton, they don't
care at all.

They do some other mysterious things. They sometimes take the
tusks with them and put them down some distance away. Sometimes
they then cover the tusks with grass or leaves or soil.

We don't know what this means. Until we'll be able to ask them,
if ever, we may not get to know.

We're working on it, but talking to an elephant is difficult
because of the low frequencies, the strange sounds, and the huge
difference in their and our perspective of the world. They live
in a different world, really.

But some people understand many of their vocalizations and can
produce a semblance of a few. With special equipment we can
analyze and understand even more, but not on the spot. It has to
be done in the computer audio lab.

Elephants sometimes appear to show a sense of humor. For
example, an elephant bull standing next to a research jeep once
said "let's go", then walked straight into the swamp. Of course
the elephants know that jeeps can't go into the swamp, and we
can assume that the elephant knew from previous "conversations"
that the human understands some of what he says.

But we can't prove it. There's always a residual risk that we
interpret something into the elephants that isn't really there.
And a scientist has to be careful what to publish. But most
elephant researchers apparently believe that they know more than
they can publish.

Hans-Georg

 

[Hans-Georg Michna]

Instead of casting us aside, I would hope they would treat with the
respect we give doddering parents with Alzheimer's.

Roedy,

we seem to think along similar lines. The last sentence from
http://www.michna.com/transition.htm is:

If it turns out that our destiny is to create a superior,
posthuman race, then let us hope that they will treat us better
than we treat the mountain gorillas today.

Hans-Georg

 

[goose]

We are almost to the point where you could microslice and digitise a
brain so that you had the digital anatomy and connections of all the
cells. That is not that much further off than sequencing DNA seemed a
decade ago.

ok. so we know about its composition. we know (from experimentation)
that certain bits of the brain are in some way responsible for
certain functionality.

Maybe, some day we can take a persons brain and /reproduce/
it using software ... an exact replica of the persons brain
wrt to the neural net.

that still doesn't buy us /intelligence/, does it ?

that wont give us a thieving conniving bastard like a real
person would. That wont give us a machine that can lie like
a politician.

so, we get a machine that can play chess, or can /figure out/
through trial and error, how to roast a duck.

will that neural-net-inna-machine be able to learn to communicate
feelings, desires or passions?

with it even *have* a desire to /live/, to /love/ and to /procreate/ ?

will it learn to smile[1] at weddings and would it feel loss at
funerals ?

you cannot call it 'intelligence' if it bases all its actions
on rules (whether programmed or learned), we certainly dont ...

will it pass the turing test ?

[1] or the machine equivalent.

goose,
busy at work, wish I had more time to expound on this

 

[Bent C Dalager]

we seem to think along similar lines. The last sentence from
http://www.michna.com/transition.htm is:

If it turns out that our destiny is to create a superior,
posthuman race, then let us hope that they will treat us better
than we treat the mountain gorillas today.

If we are foresightful enough to ensure that this is a machine-based
superior race, chances are they'll let us have the biosphere and go
out to explore the universe for their essentials: metals and energy.

They might keep us around for much the same reasons we have museums
today, perhaps even seeding us to new biospheres they come across just
for the novelty value

"Upload to the new Tau Ceti #3113-B full service server farm, with
cutting edge sensor pods to rent and a human zoo planet within 2.3543
clicks" <g>

Cheers
Bent D

 

[Bent C Dalager]

Maybe, some day we can take a persons brain and /reproduce/
it using software ... an exact replica of the persons brain
wrt to the neural net.

that still doesn't buy us /intelligence/, does it ?

Impossible to tell at this point. Sounds like an interesting
experiment, certainly.

that wont give us a thieving conniving bastard like a real
person would. That wont give us a machine that can lie like
a politician.

Why not?

will that neural-net-inna-machine be able to learn to communicate
feelings, desires or passions?

It might lose out on hormonal effects, which begs the question of how
well a human-based neural net would cope without the hardware
interrupts it's evolved to interface with.

with it even *have* a desire to /live/, to /love/ and to /procreate/ ?

Presumably it would, but you might have to synthesize hormones and the
like in order to make it function properly.

will it learn to smile[1] at weddings and would it feel loss at
funerals ?

It would be reasonable to assume that a neural net copied from a human
would retain basic social behaviour, so yes it probably would. Or at
least it might try - difficult to say how the lack of body chemistry
would affect one's feelings and emotional life.

you cannot call it 'intelligence' if it bases all its actions
on rules (whether programmed or learned), we certainly dont ...

We don't know if we do or not. We have little to no idea what we
actually mean by the term "intelligent". On the surface of it, it
strikes me as little more than a term we've invented in order to make
ourselves look more l33t.

Cheers
Bent D

 

[Wojtek]

Computers are already better
than humans at designing electric transmission lines, diagnosing rare
diseases, face recognition with disguises, memory, assembly line work.

I beg to differ.

If a computer had created the ability to solve these problems, then
you would have a point. However, it is a human who created the
concept, designed the algorithms, designed the architecture, designed
the decision paths, and finally, wrote the code which solved the
problem.

Now, once the code is written, then the computer can "solve" the
problem, usually in a fraction of the time required by humans.

But is the computer better at it? I do not think so. To take the
designing of electric transmission lines as an example. A human could
make the effort to consider all possible factors, optimize the
solution, and then make a perfect design.

A computer will do this faster.

But.

Introduce a factor which the developer of the computer program had not
taken into account, and the computer will either not produce an
asnwer, or will produce an answer which is incorrect.

A human will use his "experience" and "intuition" to create a
solution.

We are a long way from computers having intuition.

 

[Evgenij Barsukov]

Perhaps you are an inexperienced programmer, or the
programs with which you deal are trivial?

An indented IF statement in a high level language, a FOR loop.
or a CASE statement in whatever syntax, indented and using
meaningful variable names is readable and understandable
with no more than a glance. The equivalent in Labview where
you have to follow traces to find the nested blocks is an
exercise in Maze solving and not conducive to commercial
pressures.

You might be quite out of the loop of LabView improvements.
Recent versions have a special "box" for IF and FOR where
you place everything which has to be repeated. Give it a try,
it is quite entertaining.

Regards,
Evgenij



--

__________________________________________________
*science&fiction*free programs*fine art*phylosophy:
http://sudy_zhenja.tripod.com
----------remove hate_spam to answer--------------