 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 14.12.2017 um 01:38 schrieb Bald Eagle:
> SDL is not intelligent.
> POV-Ray is not intelligent.
> The parser is not intelligent (just ask clipka!)
Intelligence as we know it is the result of errors. The parser contains
gazillions of errors. Therefore, the parser must be an AI.
QED.
"It's life, Jim, but not as we know it."
:-P
> The slime mold example seems to be a Clockwork Orange. There is no objective
> autonomy which consciously makes decisions of its own accord. There is a
> programmed array of cells that are acted on by its environment and responds in a
> certain manner.
> There doesn't seem to be much more intelligence there than a liquid adopting the
> shape of its complex container, or a clock, or a GPS navigation system
> calculating the fastest route given current traffic conditions, or a
> supercomputer arriving at a programmed solution to a fluid dynamics problem or a
> finite element optimization of a suspension bridge.
I totally agree. I'd say the slime mold is an /elegant/ solver of
certain problems, but not necessarily an /intelligent/ one.
"Intelligence" derives from the Latin verb "intelligere" - to
comprehend, or perceive.
So in its original meaning intelligence is not about /solving/ problems,
but /understanding/ them (or the world as a whole, or one's own existence).
That's not the same as the "I" in "IQ", of course, but that "I" isn't a
quality anyway, but rather a quantity. You could talk about a slime mold
having a non-zero IQ. But when used in a qualitative sense, we most
certainly have to turn to the "I" in "AI".
In that qualitative sense, "intelligence" is not about /if/ you can
solve problems, but /how/ you do so -- it is a certain /approach/ to
solve problems. And in contrast to the slime mold's superpowers, it is
one that is extremely versatile, and can do more than just solve Steiner
tree problems.(*)
(* Well, yes, slime mold can also solve other problems. But only if that
problem is re-formulated into an analogous Steiner tree problem by an
external entity.)
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I suspect the "look at how smart these slime molds are" hype is rooted
in the fact that people see how slime molds are employed(!) to solve
problems that can't be easily solved by a linear algorithm, which seems
to be the sort of problem solving we humens are (currently?) good at.
But people forget that those slime molds are /employed/ to solve the
problem, rather than actively solving the problem on their own -- much
like 1940s "computers" were (literally) /employed/ to design the atom
bomb rather than actively designing it on their own.
Remember that in the 1940s, "computer" was a job description, not a
technical device. But by the 1960s those jobs had been taken over
entirely(*) by electronic computers -- devices which, looking back,
nobody in their sane mind today would call intelligent.
(*Except in Japan, AFAIK, where human computers are still a thing for
some obscure reason.)
So maybe slime molds would make for a great /component/ for building
machines that, some time in the future, might qualify as intelligent.
But by themselves they're not.(**)
(**This is by no means intended to say that 1940s computers weren't
intelligent. They most certainly were. Their job just didn't (normally)
require that level of intelligence /per se/. It's just that intelligence
is pretty much the only tool we humans have at our disposal to train for
such a job.)
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 14.12.2017 um 07:48 schrieb clipka:
> to be the sort of problem solving we humens are (currently?) good at.
Ahem... humans.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 14/12/2017 00:38, Bald Eagle wrote:
> jr <cre### [at] gmail com> wrote:
>
>
> SDL is not intelligent.
> POV-Ray is not intelligent.
> The parser is not intelligent (just ask clipka!)
> The algorithm Thomas is using to arrange blobs in the form of a human is not
> intelligent.
> The people who write that code are. (Don't hate me, Stephen!) :)
>
Why would I do that?
I've never hated anyone, with the exception of my old school dentist.
And may she burn in Hell for the rest of eternity.
> The slime mold example seems to be a Clockwork Orange.
I had to look that up for the meaning.
> There is no objective
> autonomy which consciously makes decisions of its own accord. There is a
> programmed array of cells that are acted on by its environment and responds in a
> certain manner.
> There doesn't seem to be much more intelligence there than a liquid adopting the
> shape of its complex container,
All of a sudden Fluidics springs to mind.
https://en.wikipedia.org/wiki/Fluidics
With a little bit of imagination, I can see the Fluidic amplifier, being
organic.
--
Regards
Stephen
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
clipka <ano### [at] anonymousorg> wrote:
> I totally agree. I'd say the slime mold is an /elegant/ solver of
> certain problems, but not necessarily an /intelligent/ one.
Yes, I knew when this was first posited that there would much to say about - the
history of man and literature and philosophy have all been exploring this for
thousands or millions of years.
The researchers are intelligent, the slime mold is a mechanism - a tool that
they are using.
They could have used been using anything that would adapt to the proper shape to
solve a problem, slime mold was just particularly suited to this one - because
it's kinetic and can iterate.
> "Intelligence" derives from the Latin verb "intelligere" - to
> comprehend, or perceive.
>
> So in its original meaning intelligence is not about /solving/ problems,
> but /understanding/ them (or the world as a whole, or one's own existence).
I was thinking about that as well, and I thought it useful to comment that some
interesting aspects of understanding a problem are employing seemingly unrelated
phenomena to formulate a solution (soap bubbles, minimal surfaces, mathematics,
architecture), understanding that there may be a "best" way to solve the
problem, considering if it's worth it to solve it the "best" way or the most
practical way (isolated instance, or something to be solved millions of times -
cost/benefit of optimization), whether it's worth solving the problem at all,
and whether a solution can even exist.
There are also _levels_ of intelligence, and being able to abstract a situation
to think about the problem is an important concept.
I was also thinking about the attributes we might ascribe to "intelligence", and
in considering how an intelligence would interact with its environment
(otherwise how would we be able to assess its intellect?) I wondered about the
how the ability of something to perceive and manipulate its environment ties in
with its potential and real intelligence.
Taking away those things made me think that an interesting attribute is being
able to communicate with other intelligences to organize and accomplish tasks
that one might not be able to independently.
Ayn Rand famously explored this aspect when she wrote about the difference
between the work someone does with their mind vs what some can only do with
their hands (figuratively). By working together, the thinker is freed from
performing the tasks that don't require an elevated level of understanding,
while the workers who have the understanding of the organization and the
equipment, and the physical capability to do what the thinker might not be able
to strength-wise perform another set of tasks. As a result everyone benefits.
Someone designing a computer processor would not be able to build and run an
entire facility themselves, and most people don't have the amount of knowledge
and intelligence to design that kind of advanced electronic architecture.
> In that qualitative sense, "intelligence" is not about /if/ you can
> solve problems, but /how/ you do so -- it is a certain /approach/ to
> solve problems.
I would say that might tie more into the level of intelligence idea - the slime
mold isn't aware that it's solving a problem - it's just mechanically searching
for food and avoiding areas where it doesn't find any. A blind, brute force
algorithm driven by chemotaxis.
Every new maze will start with the searching and end with an optimized matrix.
An intelligence would be able to abstract the idea and optimize its method
_before_ being presented with a new maze.
This also brings up the question of the degree that that memory plays in
intelligence.
We make tools that help us perceive things we can't naturally, manipulate things
we can't, and make records to enhance our memory. We can discover problems that
exist but are not directly recognizable, and have solutions that take decades or
more to understand an implement the solutions for. We're even building machines
to "think" so that they can single "minded"ly focus on solving the given problem
with great speed and without distraction (food, cold, mate, shelter, etc)
I'd also like to say that one of the greatest things I've explored in thinking
is that thinking is a function of _asking questions_. "Why?" being the most
pervasive and important.
Also "why NOT?"
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
hi,
On 14/12/2017 11:54, Bald Eagle wrote:
> Every new maze will start with the searching and end with an optimized matrix.
> An intelligence would be able to abstract the idea and optimize its method
> _before_ being presented with a new maze.
how?
regards, jr.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
jr <cre### [at] gmailcom> wrote:
> On 14/12/2017 11:54, Bald Eagle wrote:
> > Every new maze will start with the searching and end with an optimized matrix.
> > An intelligence would be able to abstract the idea and optimize its method
> > _before_ being presented with a new maze.
>
> how?
>
>
> regards, jr.
Just look at how we do the same search optimizations.
subdivision, parallel logic, BST, quadtree, octtree, etc.
https://www.google.com/search?q=tree+algorithms&rlz=1C1CHFX_enUS633US634&oq=tree+algorithms
find a way to conceive of a 3rd dimension from your Flatland, grow up, branch
out, drop new seed colonies, and execute multiple simultaneous searches - divide
and conquer.
There's so much here to talk about... but alas, I have to go be a worker bee...
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
hi,
On 14/12/2017 13:09, Bald Eagle wrote:
> jr <cre### [at] gmailcom> wrote:
>> On 14/12/2017 11:54, Bald Eagle wrote:
>>> Every new maze will start with the searching and end with an optimized matrix.
>>> An intelligence would be able to abstract the idea and optimize its method
>>> _before_ being presented with a new maze.
>> how?
> Just look at how we do the same search optimizations.
> subdivision, parallel logic, BST, quadtree, octtree, etc.
> find a way to conceive of a 3rd dimension from your Flatland, grow up, branch
> out, drop new seed colonies, and execute multiple simultaneous searches - divide
> and conquer.
"how?" wasn't the right question (sorry).
imagine then you're the controller/observer of two mazes, one contains a
mould which is on its second maze, the other a human, naked no tools,
also on their second maze. previous maze layout was different.
observing their behaviour, how could you tell one is (supposedly)
"intelligent" while the other operates on "instinct"?
> There's so much here to talk about...
a fascinating topic.
regards, jr.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
jr <cre### [at] gmailcom> wrote:
> imagine then you're the controller/observer of two mazes, one contains a
> mould which is on its second maze, the other a human, naked no tools,
> also on their second maze. previous maze layout was different.
>
> observing their behaviour, how could you tell one is (supposedly)
> "intelligent" while the other operates on "instinct"?
Well, you see, therein lies another attribute - the tool-using.
(Which also plays into wealth, property rights, Liberty, and
self-determination.)
Humans have evolved to use tools - which requires them to capitalize on wealth
(amassed food and resources) to leverage the free-time that wealth buys to
design and make those tools. The conditions of your comparison are artificial.
A human would elevated himself above the maze, use a mirror, or a drone, or a
satellite photo.
Slime molds are still blindly oozing through mazes millions of years later...
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
hi,
On 14/12/2017 13:54, Bald Eagle wrote:
> jr <cre### [at] gmailcom> wrote:
>> imagine then you're the controller/observer of two mazes, one contains a
>> mould which is on its second maze, the other a human, naked no tools,
>> also on their second maze. previous maze layout was different.
>> observing their behaviour, how could you tell one is (supposedly)
>> "intelligent" while the other operates on "instinct"?
> Well, you see, ...
> Humans have evolved to use tools - which requires them to capitalize on wealth
> (amassed food and resources) to leverage the free-time that wealth buys to
> design and make those tools. The conditions of your comparison are artificial.
well yes, quite. to both.
thing is, when you're the subject of study you don't get to choose.
> A human would elevated himself above the maze, use a mirror, or a drone, or a
> satellite photo.
humour me then. and, to make it "easier" we could imagine
exo-biologists from Sirius II conducting the experiment, they obtained
one slime mould and one human and can configure two identical but for
scale mazes, the floor is uniformly 5m (human scale) wide, the walls 20m
high, an air condition distributes the smell of an appetising food
everywhere equally. the human is naked and has no tools but if it pulls
out a hair and sticks it (with spit) to the wall that'll hold (akin to
slime mould depositing chemical marker compound), but no mirrors, etc.
as before, second run for both.
how could the Sirians tell the difference between intelligent and
"instinctual" behaviour?
> Slime molds are still blindly oozing through mazes millions of years later...
which means /they/ manage. :-)
regards, jr.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
|
|
