Re: Jaron Lanier Got Up My Shnoz on AI

From: J. R. Molloy (jr@shasta.com)
Date: Wed Jan 16 2002 - 13:08:11 MST


From: "John Clark" <jonkc@worldnet.att.net>
> It took bumbling evolution nearly
> 4 billion years and a astronomical amount of death to build a brain as
> complex as an earthworm's, it took intelligence only about 50 years to
> do the same thing. I conclude it would be far far easier to make a
> sentient creative computer than a non sentient one.

I agree that it would be easier for an intelligent computer to create
worthwhile products than for an unintelligent computer to do so. But
intelligence is not sentience, and responding to sense impressions (sentience)
doesn't necessarily entail intelligence (the ability to solve problems and
correctly answer questions).

> >The poetry-writing programs do not *claim* to be sentient.
>
> I can rectify that oversight with one line of code.

Of course *you* can. The question is, can *they*.

Incidentally, we don't have to "guess" that we are aware and responsive. We
can confirm this by interacting intelligently with other people. Solipsism is
for kids, the scientific method is for adults.

----------------------

From: "Dossy" <dossy@panoptic.com>
> Please define "creativity" in a non-circular fashion, if possible.

Please remit $10.00 for each requested definition, federal reserve notes if
possible.
To create means to make, to put together in such as way as to render a new
thing (as every schoolchild should know).

> Otherwise, it seems that you define creativity merely as the ability
> to create (as the dictionary does), in which case: are bacteria
> "creative"?

Obviously bacteria can be creative, in the strict sense of the word.

> Is reproduction (sexual or asexual) considered a
> "creative" ability?

I'd consider it an adjunct of a creative process, yes.

> Would you consider offspring the product of
> it's parent (or parents) creativity?

Some parents are more creative about their offspring than are others. G.
Gordon Liddy, for example, went to considerable and creative lengths to find a
woman who could produce the kind of children (strong, athletic, intelligent,
and healthy) that he wanted. (He settled for a tall math teacher, IIRC.)

> Can one have sight without being sentient?

You mean like a camera? Sure, many people don't know what they're looking at.

> Does having senses
> imply being sentient, or asked differently, do machines that
> purport to have senses (vision, smell, hearing, etc.) that
> are clearly not "senient" by common agreement, truly not have
> the senses that they claim to have?

What machine claims (or purports) to have senses? Or do you mean that people
claim that some machines have senses?

> Isn't sensory perception composed of two things: sensory datum
> placed in a frame of reference of one's self-awareness and awareness
> of one's surrounding environment?

Perception is always sensory. Non-sensory perception is meaningless (or only
meaningful circularly). Perception isn't a "thing," it's a process whereby a
representation of a component (or components) of reality are rendered into a
model which constitutes understanding. "One's self-awareness" unnecessarily
fragments one. Notice, if you will, that when "one" is divided into itself and
its awareness, it is no longer "one." In reality, one IS awareness.

> Sensory data without the ability
> to contemplate (at least at a sub-conscious level) what the data
> implies is simply data.

Good. Now consider the ability to contemplate without sensory data. Without
sensory data, no contemplation occurs. (Imagination re-orders previously
garnered sensory data.) You are what you think. All that you are arises with
thought. Those last two sentences comprise the first two sentences of the
Dhammapada. Pure awareness, with no content whatever, constitutes the mirror
that reflects reality and is sometimes referred to as "nirvana" -- the
elimination of all contamination from the perfect mirror of awareness is
enlightenment (which is another of the ninety-nine names of nirvana or
nothingness), and in this process, the observer is the observed. Consequently,
the pure awareness abiding in your brain is identical to the pure awareness
abiding in Buddha's brain, and from this has arisen all sorts of
misconceptions about reincarnation, because total awareness is total
awareness, regardless from what brain it emerges. Something goes beyond words,
and to find it the brain needs to leave language behind and to discover
meditation, which means direct experience of reality, the perception of
immediate existence.

> An algorithm which influences a machine's
> behavior based on it's sensory data would appear to give the machine
> some kind of awareness of the physical world in which it's
> interacting -- isn't this a form of low-level sentience?

You can think of it as levels of sentience, or you can consider sentience as
an absolute with varying degrees of purity. As water can be pure H2O or it can
be mineral water saltwater, alkaline water, lemon-flavored water, and so on.
So human perception can be colored by emotional content or by its connection
to other layers of sensory input and tandem cognitive systems. One of the
advantages of parallel processing is fault tolerance. Serial processors crash
or halt when the series is disrupted or interupted, but parallel processors
can keep chugging along even in very chaotic environments. By superlative
sentience I mean to indicate perception beyond which no greater accuracy is
possible. It's perfect perception whose topography exactly coincides with and
maps upon reality.

> > Superlative sentience (total awareness) reveals that the highest use
> > of thought is to allow us to stop thinking (while being wide awake),
> > so that we can appreciate the epiphenomenon of self-awareness.
>
> Don't all machines currently enjoy superlative intelligence, then?

Machines have superlative intelligence perhaps (until they break), but not
superlative sentience (and their "enjoyment" is purely mechanical).

> Just because they don't express their appreciation for self-awareness
> doesn't mean they aren't appreciating it.

Indeed, when one expresses appreciation, the appreciation is slightly
diminished by the effort of expression, and by the noise that expression
interjects into the gestalt. Two people can appreaciate a magnificent sunset
over cocktails... until one of them spoils it by commenting on how beautiful
it is.
Nevertheless, I don't expect that machines appreciate their self-awareness
merely because they remain mute, because even while remaining silent,
enlightened humans demonstrate their awareness by the way they live, by the
way they pass knowingly through the noise and haste of the world, and by the
way they touch life with compassionate discipline/disciplined compassion.

> The problem is that we
> try to anthropomorphise machines and try to insist that because
> the sentience they enjoy isn't the same as the sentience we enjoy,
> therefore it's not sentience.

No, the problem is that we make problems out of everything. Pure sentience is
by its very nature spontaneous, non-contrived, unplanned, a by-product of
evolution.
While spontaneously playing with his young son, a man serendipitously noticed
that the words "now" and "here" when combined spell "nowhere," and he suddenly
understood how the immediate is connected to the timeless. A machine could do
that only by transcending the mechanical, predictable, constrained limits that
define machines. Humans stop being machines when we overstep all the habits of
our thought, and come to our senses.

> Just as for a long time, we couldn't
> accept that animals could be just as sentient as humans are

But of course humans are animals too. What makes us different is that we don't
know what we're going to be when we grow up. Kittens predictably become cats,
and puppies become dogs, colts become horses. In contrast, children become
firefighters, lawyers, doctors, chefs, scientists, and diverge into other
specialties. Our big brains make us different, and knowing that makes even
more difference. Since we can't objectively/quantitatively measure sentience
that is not correlated to behavior, and since behavior is connected to
genetics, we might misinterpret this to mean that genetics and evolution
determine sentient levels. However, the variability and flexibility of human
behavior provides a buffer between biology and psychology. We've decided that
the human animal is more sentient than other animals primarily because it's
humans who will fabricate other robotic animals rather than other animals that
fabricate humans.

> probably for the same reason (as well as the fact that they lacked
> the proper language to convey their sentience to us humans).

I think the differences between humans and other animals are less than the
differences between humans and machines. What makes machines special is that
they can become more like humans quicker and more completely than other
animals can.

"You'll learn to like robots.
They'll be nicer than human beings."
--Hans Moravec

From: "Dossy" <dossy@panoptic.com>
> On 2002.01.15, J. R. Molloy <jr@shasta.com> wrote:
> > and of course, evolution is not sentient,
>
> Would you care to defend that assertion?

If you don't think it's a true assertion, then let evolution defend itself.

> mechanical clocks are also
> "creative" when they display the time they are instructed to
> keep, which is nonsensical

However nonsensical it may be for clocks to create the time they are
instructed to keep, it is nevertheless what they do. Clocks do, indeed, create
the time of day. Clock time is entirely an abstract invention of the human
brain, and without clocks, there would be no time measurement. In the same
way, numbers are abstractions invented by the human brain.

> Creativity implies (to me) an external
> influence that affects the product that is created.

Sounds like a good description of evolution (to me).

> Worse, lay-people would find it very hard to agree that a
> robot in an automobile manufacturing plant, "creating" the same
> component over and over, posesses creativity.

Such a robot may "possess" creativity while not exercising it, if such a robot
is programmed to function below its potential abilities. Likewise, some people
never realize their creative potential.

> I don't think sentience is easier to create than creativity, but
> I don't think that creativity necessarily precedes sentience.

Look at what you've written... "easier to create than creativity"
Earlier you were complaining about "circular" definitions, and now you've
written a perfectly circular self-contradiction. That which creates is
self-evidently already creative, and so creativity already exists whenever
anything at all acts to create. Hence, by definition, creativity precedes all
of creation. To posit sentience which precedes creativity is to invoke the
supernatural.

> The reason why it appears to us (humans) that it is easier to
> achieve creativity than it is to achieve sentience is because
> we know so very little about sentience.

I think the reason we know so little about sentience is that it's so easy to
create (fabricate and confabulate) explanations based on imagination rather
than to do the hard science necessary to discover the underlying principles of
sentience.

> I think until we get a much firmer grasp of what sentience
> really is and what it consists of, we'll never actually produce
> truly creative machines.

I don't know about "we" but according to experts in the field, truly creative
machines have already emerged. Google on "Automatic Creation of
Human-Competitive Programs"

--- --- --- --- ---

Useless hypotheses, etc.:
 consciousness, phlogiston, philosophy, vitalism, mind, free will, qualia,
analog computing, cultural relativism, GAC, Cyc, Eliza, cryonics, individual
uniqueness, ego, human values, scientific relinquishment, malevolent AI,
non-sensory experience, SETI

We move into a better future in proportion as the scientific method
accurately identifies incorrect thinking.



This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 13:37:34 MST