Re: How Robots Will Steal Your Job

P

Programmer Dude

Wojtek said:
BTW, what is (-: ?

Aren't smileys etc usually looked at by leaning the head to the
left? If that is a smiley, then I must lean my head to the right.
Are you in Australia??

Wish! I was told recently that all the smilies on amUSENET have
caused a world-wide shortage of right-parentheses. I'm just doing
my part in trying to help conserve a vital resource and restore
balance.

Also, I relish the opportunity to NOT do things the usual way.
Keeps the mind agile and young (and pisses off the fogies). (-:
 
M

Michiel Konstapel

Consider a naive sort algorithm:
1) Rearrange the data in a random order
2) Are they now in sorted order? If not, go to 1
3) Success

Randomness harnessed to solve a problem.

According to some theories on quantum mechanichs, it may even be the
perfect sort algorithm: it runs in linear time.

1) Rearrange the data in a random order using a quantum process. This
splits the timeline into parallel universes, in which each possible
outcome is represented.
2) Are they now in sorted order? If not, destroy the universe.
3) Success

As http://www.catb.org/jargon/html/B/bogo-sort.html puts it,
"Implementation of step 2 is left as an exercise for the reader."

Michiel
 
P

Programmer Dude

Wojtek said:
But now you have broken all the code which tries to convert :) to
a cute (and hated by me) graphic.

Oh, well. Life is hard sometimes. :-\
Yet a human can still interpret it. Which brings this thread back
on topic....

[BWG]
 
W

Wojtek

At the least, this is pattern recognition. When the day looks thus,
master returns. Animals (and children) are often much more aware
of the details of their surroundings than adults (I suspect our
brains have gotten too full).

There was a British show a few years ago that tried to prove that dogs
have ESP.

It seemed that a particular dog would start barking and running at the
door when the owner _decided_ to come home from a shopping trip. The
trial used two video cameras with the times synchronized. One was at
the home filming the dog. The other was with the owner.

After some hours of shopping, the owner suddenly decided that she
wanted to go home. At the same moment, the dog started barking. This
was repeated over several days with different shopping lengths/ times.
 
R

Richard Heathfield

Roedy said:
No, I see that assumption as an unnecessary addition that complicates
and explains or predicts nothing.

It would, at least, explain the concept of directed evolution. If there is
no director, there is no direction - merely chance.
You certainly don't need it to
explain evolution.

Quite. That's my point. Your point seems to have lost its direction.

<snip>
 
R

Roedy Green

1) Rearrange the data in a random order using a quantum process. This
splits the timeline into parallel universes, in which each possible
outcome is represented.
2) Are they now in sorted order? If not, destroy the universe.
3) Success

I wish I could remember the details but someone invented a molecular
computer. Basically fragments of DNA combined in all possible
combinations and something combed the combinations filtering out the
wrong solutions to the problem it modeled. It was massively parallel.
It would have been in Nature, SciAm, or Discover some time in the last
few years.

I have always been a bit nervous about using standard encryption
logic. Seems that when it cracks everyone's protection crumbles all
at once. Even a little fillip added on top gives you protection from
the tidal wave.
 
R

Roedy Green

As you've seen, the word "direction"
can be a little ... problematic. (-:

I mean direction in the sense of vector, not direction in the sense of
guidance. And please don't try to tell me that I don't.
 
R

Richard Heathfield

Programmer said:
Thereby demonstrating that a general definition of "solve" is not
an indicator of intelligence.

Right. Having said that, intelligence is certainly associated with
problem-solving to at least some degree. But there are other factors, too.
But see nearby post about a narrow
definition, which might suffice.

Um, life's too short. If I bump into it, I'll try to recognise it. :)
(you == me? I understood it first read.)

Sorry. s/you/whoever it was who misunderstood it first time round/

(I just tried to find out who this was, but didn't manage it.)
 
H

Hans-Georg Michna

Roedy Green said:
The place we may first see such creatures is pornography. It requires
only minimal intelligence to simulate a willing female. They may
appear first as interactive cartoons, then 3D cartoons, and gradually
more lifelike simulations.

Roedy,

can't wait. :) But honestly, the idea is not new to me.

The interesting part is when we create computers or robots that
surpass our intelligence. Fairly obviously, if they can reach
human level intelligence, they will soon surpass it.

This may create entirely new problems, not all of which are even
foreseeable. Most people won't notice in time anyway and will
never really know what hit them.

Hans-Georg
 
R

Roedy Green

It's not too hard (so far) to quickly establish whether there is
a mind or a machine responding. To date, nothing passes the Eliza
test.

For a human intelligence, you can tell, but for a non human
intelligence, or a low level intelligence with consciousness, I don't
know how you would tell if there was anything home --
something inside actually experiencing.

If you went back in time and tried to talk to medieval peasants, they
would presume you mad or a fool.

In like manner we could easily miss intelligences superior to ours.
We are good only at judging creatures similar to ourselves with
similar fascinations.

Humans are quite vain. They keep bending the rules on what counts as
intelligent to be able to view themselves as superior. For example, we
are no longer as impressed by feats of memory as we were. Arithmetic
skill no longer wows. We now make a virtue of rusticness -- making
furniture that does NOT quite fit together.
 
R

Roedy Green

Don't have much direct experience with that, so I can't comment.
I *have* always wondered how one sees during an OOBE. If one is
incorporal and invisible, how can eye lenses focus, and how can
the retina react?

One way it could work is this. You have a ton of information coming
in about your surroundings, via all your sensory channels. You are
quite capable of creating a 3D model of your surroundings. It is not
that much of a leap to project what it would look like from any point
of view, e.g. from the rafters of a theatre in my case. You do this
all the time in dreams, even constructing the entire reality as well.

You can do it in your imagination. It is not that big a leap to
inserting that imaginary projection as the actual. People do that
routinely in hypnosis ( incidentally also a very boring process ).

After all, normal viewing is just a special case of that. You don't
directly see objects. You process binary data from your eyeballs,
millions of little twitchings and build a simplified internal model
your experience.

Since you don't usually interact with your environment during an OOBE,
your model can be quite wrong, and still appear to function perfectly.
 
E

Eray Ozkural exa

Hans-Georg Michna said:
I think the messages about current programming languages all
miss the point. The problem of all current programming languages
is that the programmer ultimately has to define everything, down
to the last bit. He can occasionally use other peoples'
programs, but even then those other people have had to define
everything in their pieces.

Yes, but still I don't see how you could use the highly ambiguous
natural language to write the next parallel AI system. Maybe you are
referring to mundane applications such as database client apps?

I think there is a great deal to be done for a really visual language.
By visual I don't mean those stupid UI builders like Visual C++. I
mean a visual syntax and corresponding semantics. I believe that can
be quite different from text based languages.

Formal semantics all the way!

Cheers,

__
Eray Ozkural
 
R

Roedy Green

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

Or they miss their companion in the present.

To do that they would have to remember they once had a companion would
they not?

One thing we tend to associate with intelligence is the past having a
longer effect on present behaviour.

The other thing, intelligent beings look farther into the future when
considering their options.
 
R

Roedy Green

Creativity not a function of intelligence?

No. That whales humpback compose a long complex song each season, with
theme and regional variations, is not seen as evidence that whales are
intelligent.

I could imagine whales discovering we can't as proof we are hopeless
barbarians.
 
R

Richard Heathfield

Programmer said:

Well, it's quite clear that animals can remember. Therefore, they can
"sense" the past, presumably in a similar way to us. (Evidence: maze
experiments on rodents; dogs recognising their so-called "masters"; cats
being able to find their way home; and so on.)

It's also quite clear that animals can predict what /might/ happen. This is
especially true in predator/prey relationships. A less dramatic (but more
intriguing) example has been posted elsethread w.r.t. the cat that checked
its food supply before letting its owner head off for the day.
In one sense, you're right. Memories are adjustments of our NN, so
merely remembering something is not proof of a time sense. But the
ability to *order* those memories in sequence probably is. And I'd
think the ability to order expected *future* events would clearly
demonstrate it.

Sure. But proving the absence of that ability in all animals everywhere
sounds impossible to me.

No, but animal behaviorists do study and test these matters. MY
evidence is purely anecdotal, but theirs is more rigorous. All
I've read equates such behavior with instinct.

Perhaps "human behaviourist" squirrels form similar theories about humans.
Perhaps they misguidedly attribute our behaviour to instinct, in the same
way that we (perhaps) misguidedly fail to understand their own complex
behavioural dynamics.
Of course. SO FAR, all evidence suggests one thing. When that
changes, so will my point of view.

Well, that's the trouble. You say "all evidence suggests one thing", but I
haven't seen any evidence of the kind that looks even remotely convincing.
How can we know whether animals are conscious? Self-aware? Intelligent?
Reasoning? The only even remotely plausible way to answer these questions
is to /ask/ the animals. And even if we found a way to do that, we could
never be sure - if they didn't answer us - whether they were incapable of
answering, or merely unwilling.
Sure, but curiosity is not necessarily abstract behavior. Babies
are curious.

Please give me an example of an observation of an animal that would lead you
to deduce abstract thought. I can think of no particular reason why animals
would feel obliged to play our research games for us, unless there's
something in it for them. (And when there /is/ something in it for them,
some animals show dizzying ability to solve relatively abstract problems.
This can make for very entertaining TV documentaries.)
Yes! Delightful cartoon (from a favorite cartoonist)!!


Absolutely. As I said, when there's useful evidence to consider,
I will consider it.

I don't think we'll ever get any useful evidence in either direction.
Right. But again, the evidence strongly suggests one thing.

Please demonstrate this.
From my window, I can see lots of [animals...]. I've spent a fair
amount of time observing them. [Their] behaviour reminds me /so
much/ of people. There is much about them that is curious, amusing,
playful, caring, weak, strong, cruel, and deadly.

Animals often tend to act as a Rorschach test. We see ourselves
in them very, very easily (witness this thread).

Very true. But we are also very quick to shun the idea that they might be
"like us". Why?
Sure, but... I've owned and loved dogs all my life. My dog is a
member of my family in many ways, and we're very closly bonded.
Given all that, if she had higher brain functions, I'm *quite*
sure she'd have made some attempt to communicate with me ON that
higher level (rather than the simplistic communication we do
enjoy). Likewise those who've studied animals all their lives.

I lack your confidence. The most intelligent thing an intelligent animal
could do would be to hide its intelligence from human beings.
If the higher functions existed, don't you think we'd have found
them by now? Our history with animals is longer than our history
(so to speak).

No, I don't think we would. We are blind to the possibility. Well, some of
us are, anyway.
Instead we have what we have. Even sign language with apes is
controversial and (according to some) debunked as wishful thinking
on the part of loving owner/handlers.

I have, of course, read about humans teaching apes to use sign language.
What I have read seems reasonably solid, especially since the apes were
able to create sentences of their own, according to the grammatical "rules"
of the language they had been taught, and were able to hold meaningful
conversations with their teachers. There is a limit to how meaningful a
conversation you can have with a random number generator. Therefore, I am
inclined to believe the research you've cited until overwhelming evidence
to the contrary comes to light.
 
R

Roedy Green

Well, you do, anyway. You postulate a problem, and that assumes an
"intelligent" agency for whom, or for which, it is a problem.

No I don't. I also include the implied problem even a virus could have
of survival.
 
C

Corey Murtagh

Roedy said:
Come on. I corrected that ambiguity. You were TRYING, like a Hollywood
lawyer to break a contract, to misinterpret what I was saying.

I was expressing my interpretation, and the reasoning behind it. If
that makes me a 'Hollywood lawyer' then so be it.
I find that sort of tactic HIGHLY annoying, trying to tell someone
what they MEANT to say, despite their protests.

I suspect there is no reply I could make to this that would satisfy the
both of us.
It is fair game to point out ambiguities, but it is childish to nya
nya like that.

*sigh*

You made an apparently self-contradictory statement. I called you on
it. Subsequently other people have responded to /my/ post, and I have
replied to explain my position. How is that being childish?

And how's this for childish:

If you can't accept criticism, **** off.
 
P

Programmer Dude

Roedy said:
I wish I could remember the details but someone invented a molecular
computer. Basically fragments of DNA combined in all possible
combinations and something combed the combinations filtering out the
wrong solutions to the problem it modeled. It was massively
parallel. It would have been in Nature, SciAm, or Discover...

I read a clip in SciAm, but IIRC it was quoting, um... Nature?
It was interesting, but of limited application (as most special
design computers are).
 
P

Programmer Dude

Richard said:
I thought you said (elsethread) that you understood my point
straight away? Here, it is clear that you didn't. Is that what
you call "blowing smoke"?

[BWG] Nah, just me being confused in a busy thread. I mixed up
your other reference to gravity with Arthur's gravity/direction
thing.

I'd be more inclined to listen to a squirrel than a human...

Quoting out of context for fun and profit.... ;-|

Perhaps horses count when we're not looking. When we /are/ looking,
they only pretend to try to count. Subtle misdirection. No, I have
no more evidence for this theory than you have for yours - but no
less, either.

[shrug] Could be. I think Occam's Razor suggests that if it looks
like a non-counting duck, it probably is. If you want to believe
animals conspire to keep it all secret .... okay! (-:

I would be fascinated to learn how you would set about proving the
presence, or the absence, of this level of abstraction in animals'
minds.

No evidence in all of recorded history? That must mean *something*.
I am not convinced that it is possible to persuasively demonstrate
the ability of some /humans/ to count, let alone animals.

[BWG]
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
474,444
Messages
2,571,709
Members
48,796
Latest member
Greg L.
Top