Jim,

If you can’t deal in actual examples, and can only make with wordplay, 
discussion is pointless.

Find an actual example of a pattern-generating algorithm, and you will find it 
is extremely limited – and then you will also be able to see the difference 
from patchworks.

Every member of a collection of patchworks is a new configuration of new 
shapes, and yet every one, looked at overall [but not at a local level] bears 
some family resemblance to the others. At one extreme, a collection of 
patchworks may have ZERO elements/shapes in common. Look at a collection of 
figurative Google logos, and there will be literally zero shapes in common.
There are NO new shapes OTOH in a set of variations on a given pattern., All 
variations share the SAME basic set of shapes, and configurations of shapes .
Decimal calculations – 22*33, 44*51 etc -  all adhere to the same basic 
pattern, and set of shapes/numerals.
If you introduce one new shape into a pattern, you BREAK THE PATTERN – screw up 
the maths/logic/algo.
If you introduced one arbitrary new shape into the present decimal system, you 
would screw the system up.
If you think there are patterns that introduce new shapes/elements, show us 
such a pattern example and its variations.
You guys really are talking nonsense.
Examples, examples. Have you never heard of them, Jim?
From: Jim Bromer 
Sent: Wednesday, November 21, 2012 1:50 AM
To: AGI 
Subject: Re: [agi] Randomness: Mathematics as Perceptual Bias

Mike,
If I assume that you are making sense then your entire two year long argument 
is based solely on a difference of opinion about what the term pattern means in 
computer science. Since you are not a computer programmer or a computer 
technician it is very unlikely that you know what the term means and the rest 
of the agiers (as you call us/them) don't.
When we talk about patterns we are including the new patterns that can be 
created and about hidden patterns that can be abstracted. Since you have never 
mentioned hidden patterns that can be abstracted it is a pretty good guess that 
your reliance on your use of the term "patchworks" is more naive than the 
concepts that we are thinking about. Furthermore, you have repeatedly used the 
term "element"
in a relativistic way so again, since you have never acknowledged that you do 
so it would seem that you are the one who is struggling to come to grips with 
the fact that this is not something that we have never thought about.
Arguing with you guys always ends up providing me with some insight that I had 
never quite possessed before. But that is not because I am catching up to you 
but because when we are talking about problems that are extremely complex (as 
in complexity) the relativistic problems have a funny way of becoming relevant 
all over again. So if there is no such thing as a true 'element' then what is 
the difference between a pattern and a feature. What is different between an 
element and a patchwork to use your term? What is the difference between a 
patchwork and a pattern?  A patchwork-pattern can an element to some more 
complicated patchwork-pattern. And since we can abstract relations that cannot 
be immediately sensed in a patchwork-pattern we can find novel elements that 
are less complicated then the patchwork-pattern that is under consideration.
Jim Bromer


On Tue, Nov 20, 2012 at 5:41 PM, Mike Tintner <[email protected]> wrote:

  Jim,
  You’re doing it again – arguing logically as opposed to scientifically – and 
seem deeply confused about the difference. Hence:
  “if I could write a program that did not react to Input but which created an 
endless variety of patterns then would you (Mike) accept that it verified the 
idea that rational computational methods had the *POTENTIAL* to be intelligent”
  That’s not evidence, Jim, as you proceed to argue. That’s a hypothesis about 
a fictional program of your invention. “Evidence “ can only be derived from 
****ACTUAL*** programs. I told you: you’re lost in logical definitions. The 
difference between your fictional hypotheses about what programs **may** be 
able to do, and what programs **actually** do, escapes you – and not just this 
time, but always, as far as I know you.
  And so you get lost in your imaginings – here, for example, you posit a 
program that can produce “an endless variety of patterns” – and proceed to take 
that as a given, a virtual fact of life.
  There is no such program (or algorithm), and there is no reason whatsoever to 
think that an algo can create such an endless variety.
  An algo can be designed to produce "a “very large variety” of patterns 
(numerically large), just as a chess program can produce a very large number of 
chess games – but only a very limited (not “endless”) range of patterns, with a 
very limited range of elements. 
  You would have an AGI if your program (and it would have to be 
non-algorithmic) could endlessly create new **kinds** of patterns with new 
elements – and formally that would be equivalent to endlessly creating new 
patchworks – because, just a tad paradoxically, a new pattern – a new kind of 
Escher, say – Is in a sense a new patchwork when considered alongside previous 
kinds of patterns. It breaks the principles of previous patterns – breaks the 
pattern in a sense. When s.o. introduced random numerical variations into 
patterns, as in cellular automata, they were introducing a new kind of pattern 
with a new element.
  The potentially infinite class of patterns, BTW, – taken as a whole – 
constitutes a patchwork and not a patterned affair. There is no common pattern 
to the class of patterns. People can and do keep inventing new kinds and 
principles of pattern.
  So we add a new injunction to you: “stick to EVIDENCE of actual programs 
doing actual things” - and not your fictional programs doing fictional things.
      AGI | Archives  | Modify Your Subscription   



-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968
Powered by Listbox: http://www.listbox.com

Reply via email to