Re: How Robots Will Steal Your Job

[Roedy Green]

Or why consciousness is an illusion.
-Steve

It has a lot in common with its subjective feel as a dream. The main
difference is the degree of coupling with the outside world.

There's another possible measure. If an animal dreams, then is it not
likely also conscious. By that measure dogs are conscious.

Dolphins do something strange to sleep. They alternately shut down
opposites sides of their brain. One side has to stay away to look
after surfacing to breathe and to look out for predators.

I could imagine an eminent dolphin pointing out the inferiority of all
land based mammals because they go totally unconscious for such long
periods.

 

[David Longley]

Why not just settle for a reasonable anomalous monism based on what we
are learning about behaviour - namely that even in humans it would
appear that what we call "consciousness" is compositional. There are
layers of behaviours and we can strip away many of them without
*devastating* effect. If we look at behaviour, and at what is required
for different levels of behaviour, we find some of the simplest
creatures (Drosophila with about 250,000 neurones) still capable of
operant and classical conditioning. Psychologists can present these
processes in "cognitive" terms if required, although we can't of course
really project ourselves into such simple forms of these processes. What
one tends to find is different emphasis being given to different parts
of functionally similar systems throughout the animal kingdom, albeit at
different levels of sophistication or elaboration in space and over
time.

Isn't that enough?

 

[Patty Cutman]

Why not just settle for a reasonable anomalous monism based on what we
are learning about behaviour - namely that even in humans it would
appear that what we call "consciousness" is compositional. There are
layers of behaviours and we can strip away many of them without
*devastating* effect. <snip>

Have you ever examined a patient with a brain injury who presents with
extreme loss of effect, yet who continues to perform well on
intelligence tests? What layer of behavior was striped?

Isn't that enough?

Enough for what? I won't "settle" for less than I possibly can be.

Sentences ending with the "?" mark are requests for *specific*
information from you.

Patty

 

[R. Steve Walz]

If they are smarter than us, why wouldn't they understand? Further, we
will be even more dependent on them than now. We won't be able to even
if the majority wanted to, which it won't. "That robot is such a
polite young man."

It would make about as much sense as killing all the women when they
first started asking for rights.

 

[R. Steve Walz]

It would come as a heck of a shock to you, and likely you would
suddenly start apologising all over the place, and start treating your
fellow creatures with much more respect.

-----------
Fairy tale.

You would in effect join one of the primitive cults where every tree
has a "spirit".

-----------
Not any time soon, thanks, unless my minds turns to jello.

It strike me as much simpler, and hence more plausible, if
consciousness were a continuum. Then, in a way it does not need an
explanation. It just is, like protons. Otherwise, you have to figure
out what makes humans so all fired special when anatomically they are
very close cousins of the mountain gorilla sharing 98.6% of the same
genetic code.

 

[R. Steve Walz]

It has a lot in common with its subjective feel as a dream. The main
difference is the degree of coupling with the outside world.

--------------
There is NO such thing as an "outside world".

There's another possible measure. If an animal dreams, then is it not
likely also conscious. By that measure dogs are conscious.

----------------
Your Venn diagram is inside out and bleeding. You are extorting "we have
REM and experience it as a dream and we are conscious" into "anything
that has REM is conscious". Nope, sorry. A cash register
being rung-out reiterates the events of the day, but it's NOT
"dreaming".

-Steve

 

[R. Steve Walz]

There's a very simple test for whether species X or system Y is
conscious: It's definitely conscious if it decides you aren't.

-Wm

 

[Mark I Manning IV]

This quote is from Airy R Bean. I see you still haven't decided to play
ball with attributions. Are you really trying to make life difficult
for the rest of us? Or does your newsreader not support threading?



I have no doubt that a Forth interpreter is smaller than a JVM. What I
/do/ doubt is that it is more complex.

From memory Forth has a core instruction set of about 20 instructions,
and everything is (or at least /can/ be, although common functionality
is often added as native code) built from there. A JVM has a whole lot
more core instructions, plus object lifetime management, plus garbage
collection, plus... the list goes on.

None of those things are /impossible/ in Forth, but they're not core
functionality of Forth. At best they're addins in the same way that the
various Java modules are... and thus the complexity is moved out of the
interpreter.

it is forths lesser complexity that is one of its BIGGEST plusses. the
english have a great saying... keep it simple stupid!! (erm please dont
take that quote as me calling you stupid, im absolutely posative your not

 

[Mark I Manning IV]

What do you want? Are you asking for 100 gigs of code in 10 gigs of space?



Yes, that's the point.



The idea isn't just to move the complexity out of the compiler. (Though
that's there. A typical (basic, not optimising) Forth compiler only
parses text, looks up routines, and compiles or executes them. Things
like + or ( or IF are all routines that have been moved out of the
interpreter.

But the important idea is to see how much of that JVM complexity you can
do without. If somebody writes an assembler routine in 100 instructions
(that include, say, checking for error conditions that will not happen)
and you can get the result in 10, your code is shorter but you wouldn't
agree that it's more complex. And it's an open question whether you
have just moved the error checking out of the routine and into a calling
routine. If you can move the error checking to the *program* inputs,
you might need say 1% as much error checking (in the final correct
program).

Similarly, for many projects you might find you don't need object
lifetime management or garbage collection etc. Maybe those things help
you manage complexity. Maybe they *are* complexity and they add to the
problem. If you can simplify enough you don't need them, so why put
them in?

I don't *want* ten times the complexity in 10% the code. Squeezing more
complexity in less code is probably not the answer. (Or maybe it is,
but it needs that code to be intuitive and correct. Otherwise you wind
up with 1% of the code and it takes 100 times as long to write and 100
times as long to debug and 100 times as long to read later.) Finding
out how to get the correct final results with 1% the complexity is much
better, when you can do it.

erm.. what he said!

 

[Hans-Georg Michna]

we have
not even see comparably intelligent machines yet. They are not due for
another 30 years or so.

Roedy,

could be much earlier. My best guess is 10 to 20. 10 if we
already have closed intelligence improving loops that I'm unable
to see yet.

Hans-Georg

 

[Hans-Georg Michna]

You would in effect join one of the primitive cults where every tree
has a "spirit".

Roedy,

while you mention tree spirits---have you or has anybody seen
the wonderful anime movie Princess Mononoke? Do see it.
Recommended to everybody as very good entertainment, a little
insight in Japanese culture, history, and thinking, and some
general food for thought.

Hans-Georg

 

[Roedy Green]

There is NO such thing as an "outside world".

Well I know that I am more than a figment of your creation, but of
course there is no way you can know that.

I did experience the world in your solipsistic way for a few hours and
then abandoned it once I realised it was just a point of view, and not
that pleasant a point of view. Why is your creation so damn
obstinate?

In much the same way, I experienced walking while standing still,
while the earth spun beneath me, jerking the earth about each time I
turned a corner. Einstein suggested that this was legit, just
changing the traditional frame of reference. This game is a bit of
heady wine for the ego.

 

[Constantinople]

In much the same way, I experienced walking while standing still,
while the earth spun beneath me, jerking the earth about each time I
turned a corner.

So that was you! Would you cut it out, please?

 

[R. Steve Walz]

You're saying that all people who espouse any faith system are either
ignorant, arrogant or both? Even those systems which place high value
on knowledge and humility?

 

[R. Steve Walz]

Well I know that I am more than a figment of your creation, but of
course there is no way you can know that.

------------
We are both figments of OUR Imagination, the nature of "you" is that
you are deluded into thinking the Mind is yours alone.

I did experience the world in your solipsistic way for a few hours and
then abandoned it once I realised it was just a point of view, and not
that pleasant a point of view. Why is your creation so damn
obstinate?

------------------
Because it changes a bunch of important things in higher philosophy
and sociology and politics.

In much the same way, I experienced walking while standing still,
while the earth spun beneath me, jerking the earth about each time I
turned a corner. Einstein suggested that this was legit, just
changing the traditional frame of reference. This game is a bit of
heady wine for the ego.

-----------
Yup. Try the universe as an egg, the inner surface of which is the
world, into which you look with your eyes and hear into with your
ears, etc., that simply changes to simulate motion.

-Steve

 

[soft-eng]

The most persuasive QM interpretation, and the simplest is the MWI,

What about the "they failed to account for the interference
from the electron gun's field leaking from two holes,
and made a storm out of a teacup wave" interpretation
at http://www.mukesh.ws/ ?

 

[Roedy Green]

could be much earlier. My best guess is 10 to 20. 10 if we
already have closed intelligence improving loops that I'm unable
to see yet.

What's expected in 2030 are machines with the computing power of the
human brain. We have already seen that computers with only the power
of a dragon fly can run circles around humans for certain classes of
problem.

It is a bit like being a competitor of Intel. We can shave the
circuits of Intel and study them layer by layer, except with the
brain, there are no copyright problems. We are already using designs
inspired by the what we learned about human vision to build optical
filters that do what the lower levels of vision processing do in the
eye and brain. We have been able to steal some of its algorithms and
improve them.

Most of us are still writing strictly procedural code. We are not
familiar with genetic algorithms that find solutions by allowing
tentative solutions to mutate and compete. We are not familiar with
software or hardware neural net programming. Programming in future
will merge these two other parallel paradigms.

In addition, there are at least two other sources of massively
parallel computing power, quantum computing where you let quantum
indefiniteness explore all possibilities simultaneously and realise
the solution, and DNA computing where you model your problem as a DNA
pattern, and filter out bad solutions. Professor Leonard Adelman of
UCLA claims to have a solution for the travelling salesman problem
done that way. I read a description in Kurzweil's book. It did not
seem to me to be a solution to the classical problem, just something
superficially similar, but at least it shows there is more than one
way to skin a cat.

The whole business feeds on itself. Every discovery accelerates the
process of evolution.

For a simple example, Intel now has more powerful computers on which
to simulate the next generation. They can afford to optimise to the
transistor level.

One of the other things to consider is that much of what is going on
in AI is secret. Companies using AI don't want their competitors to
know their techniques, neither does the military. The state of the
art is much more advanced than the general public realise.

For a few hints on what has been published, see Kurzweil's book
The Age Of Spiritual Machines:
http://www.amazon.com/exec/obidos/ASIN/0140282025/canadianmindprod

 

[Roedy Green]

We are both figments of OUR Imagination, the nature of "you" is that
you are deluded into thinking the Mind is yours alone.

The part that I experience seems to float about in a bubble with a
radius of no more than 30 meters. Even within that, awareness is
blocked by objects.

Oddly, the universe seems to center about one being, Roedy.

I presume your usual internal experience is similar.

The links between these two appear weak. Limited to coded tappings on
a keyboard.

I have had several experiences where I did not experience reality this
way. It was more is as I was everything, or perhaps more accurately
that "everything was". I talk about the moral implications of this in
an essay http://mindprod.com/jgloss/ccism.html

In workshops I get people to do a process that gives them a taste for
what this experience is like. I get everyone to sit on the floor. In
turn each person leads the group in some movement or sounds. The rest
of the group follow, but not necessarily by direct mimicking. They
can do theme and variations or something else that feels harmonious.
The effect is something like being allowed to conduct an orchestra.
You conduct the movement of a giant creature with many arms and legs.
It is fairly easy to imagine they are all part of one body.

It sounds silly, but people go for the whole rest of the day with
goofy smiles on their faces.

 

[Roedy Green]

We have already seen that computers with only the power
of a dragon fly can run circles around humans for certain classes of
problem.

Even dragon flies can run circles around humans.

 

[Programmer Dude]

It takes a /very/ long time for two people with no common language
between them to learn each other's language.

Depends somewhat on the people involved, doesn't it? It would
take *me* forever and a day, but I know people who pick up new
languages rapidly.

But then from what you've already said, if it can't come and make
itself intelligible to /us/, then you don't class it as intelligent.

(I would prefer the word, sentient. My *dog* is intelligent.)

I still think a key difference is that--if you meet another human
(or, presumably, alien), there would be obvious *attempts* to
communicate and find common ground. This is missing with regard to
our animal friends.

There is limited communication between my dog and me, but most of
it comes from my attempts to communicate with her. She has very
few messages other than, "Want food", "Want to go out", "Want food",
"Want petting", "Want food",... and "Want food."