Re: Design patterns
In article <hi0in1$7pk$1@news.albasani.net>, Branimir Maksimovic <bmaxa@hotmail.com> wrote:
Alf P. Steinbach wrote:
* Nick Keighley:
I was trying to remember where I first came across the godel argument
"disproving" AI (Weinburg?). It sounded BS then and it sounds BS now.
p1) machines must operate by a fixed algorithm
p1a) and hence are bound by godels result.
p2) people do not have to operate by fixed algorithm and hence are not
bound by godels result.
conclusion: people can do things machines can't do
well, woopy doop. I didn't accept p1 and p2 originally. Now I'm not
convinced p1a is even applicable.
It's like winning a race by disqualifying the other contestants.
Hm, this is very OFF TOPIC, but p1 is false, and p1a is meaningless (it
doesn't follow even if p1 were true, it's a category error). p2 is
meaningless.
p1: is false : machines have to operate on algorithm,
since there is no algorithm for creativity...
follows that
p2 :) people's creativity is not based on algorithm...
Well, why not ask what is creativity to begin with?
By definition, it has to deal with something new.
Now, since it is new, there exists no information about it
in the system at the moment.
If there does not exists any information about it,
how can you possibly find a way to create something,
that you can not even describe in your present terms?
Algorithms refer to things that are known and describable.
If you throw some totally uknown thing at an algorithm,
how does it "know" where to plug it in, into what table
or what description of what? What category it belongs to?
Under what label do you stick it in? What does it realte to?
So, there is an inherent contradiction of adding something
that does not exist in the system yet. One one end, what you
have created is valid and new, and on the other hand, you
don't know where to plug it in.
Creativity is based on intuition and not a set of known "facts".
Intuition is the inherent ability of Intelligence to go beyond
all known "facts" and limitations associated with them.
It is largely contextual, even though revelations are not.
Intuition "works" when you are willing and courageous enough
to tune into things you don't know. It is a great trust.
Trust into validity of your own being.
Trust that you can go beyond your limitations
and it will be revealed to you eventually
when you are mature enough and ready enough to even allow
such a thing to enter into domain of that, which you already
know.
There is an argument in AI that Intelligence is something that
is created via randomness. You just flip some coin, and bang,
you created something new, sooner or later, which is utter
fallacy.
Intelligence is not based on some random thing. It is very
directed, very contextual, forever staying within the bounds
of existing information. Plus allowing a totally new information
to enter, thus expanding the domain of known.
Intuition is an opening within you that allows you to communicate
with domains of beyond.
Most of the authors of great discoveries claimed
"it was not me, who created this thing.
It was given to me."
Given to you by WHOM?
Algorithms, on other hand, can not deal with totally new information.
It is basically an exception situation.
When exception happens, it is obvious, you can not continue
processing something because there exists no logic within
your system to handle it.
At THAT point, all you can do is try to either restart some
operation, hoping that some wild parameter is going to go away
like in situations where you lost a network connection for
totally unknow reason, or rather a reason, not described by
your system, or abandon the whole operation and go to the next
item on the list if there is one.
The bottom line is intelligence is not algorithmic.
Never was and can not possibly be.
The very evidence of AI shows it in no uncertain terms.
Marvin Minsky very pointedly stated that all the great "progress"
and "achievements" of AI are so primitive that they basically
view intelligence in terms of ideas of backlash and hysterisis
level.
We simply have no idea what Intelligence is.
We simply have no clue what consciousness is,
without which, no intelligence is possible
pretty much by definition.
Everything that was "discovered" in AI field is nothing more
than immitation of the existing Intelligence, which is biological
in the physical domain at least, the domain of manifest and
embodied into matter.
There exists no algorithm or method to discover something new.
It is not a matter of random permutations of known things,
as of necessity.
We are profoundly limited by the very nature of physical domain,
forever groping in the darkness to see some opening, some light
"at the end of tunnel" and forever enjoying the tremendous
blessings of seeing beyond the ordinary, beyond known,
beyond the limitations of that which we know.
If it were algorithmic, we could just keep throwing the random
numbers at some algorithm and it would eventually find something
new, except it would not even know where to stick it, under
which category of what.
The idea of AI is fundamentally flawed.
There is no such a thing.
ALL we know of is biological intelligence
and that is the only reference framework available to us
in the physical domain.
Roger Penrose, the inventor of the above, is a genius (e.g. Penrose
tiles, his work with Hawkings, etc.), but he is also utterly mad --
No Penrose, just found all teachers of mathematical logic
new before him since 30's...
:--}
Cool. I wish I knew what you guys are talking about.
Greets!
--
Programmer's Goldmine collections:
http://preciseinfo.org
Tens of thousands of code examples and expert discussions on
C++, MFC, VC, ATL, STL, templates, Java, Python, Javascript, PHP,
organized by major topics of language, tools, methods, techniques.