|
|
Re: Brainstorming about STEM (was About Functions)
Posted:
Dec 18, 2011 3:47 PM
|
|
On Sun, Dec 18, 2011 at 7:43 AM, Joe Niederberger
<niederberger@comcast.net> wrote:
> We have many examples of humans doing computations -- they did them by hand and with tablets and paper before before calculators and computers. We have no examples of humans performing correctly and repeatedly set tasks that are provably non-computable. And that, folks is the way it is.
>
You've offered no criteria for what would constitute provability, nor
has anyone proved the converse, that everything people do is
computable in the first place. That's the way it is, most definitely.
People pass the Turing Test every day, whereas no Turing Machine has
yet managed to pull it off.
I'd say until the "everything computable" camp has more credible
evidence for their pseudo-science / science fiction, there's not much
to argue about.
>
> I suspect Ramanujan had his own methods that
> unfortunately remained a trade secret. Riemann was
> similar. They may have had certain brain centers that
> developed in ways quite out along the long tail. ...
>
>>>To write him off as some kind of snake oil guy
>
>>I said nothing of the sort. Can you read?
Yes I can. You said "Ramanujan certainly strikes
me as Edgar Cayce of math" amidst other snide
remarks about UFOs and ESP, as if it's all of a
piece, Edgar being the paradigm "psychic" of the
kind I'd presume you would never accredit as such:
http://www.edgarcayce.org/are/edgarcayce.aspx
meaning you think him a fraud and/or snake oil
salesman of some kind. Ramanujan reminds you
of him.
Just connecting the dots a little, hoping to help
you articulate your somewhat incoherently expressed
world view. Since you like to dribble out these
associations and then not take responsibility for
them, I thought we might circle back to those
dribbles and get some more insights into your
beliefs.
>>If you do something non-computable, then
>>you're just a circus act, so why should we bother with
>>you?
>
>
> Excuse me, who did something non-computable?
Ramanujan came up with the expression for 1/pi.
I hope you looked at it. He had no explanation for
where it came from and there are no proofs of
why it's true, just verifications that it works great
(converges quickly).
You shared your belief in some behind the scenes
computational / symbolic reasoning process that
just had to be there, interpolating evidence where
there is none. I drew attention to your sloppiness
(again). Ramanujan said things like they came to
him in a dream or whatever. Muses.
> Since you won't say really what you mean by that I'll have to guess: Non computable to you mean something that hasn't been done yet with computers?
>
And what definitions have you offered, for computable?
Something computers cannot and/or have not yet done?
Because you seem, for some reason, to indulge the
fantasy that everything humans do is "computable"
(no definition provided) without proof or even much
evidence.
People who argue the contrary are somehow engaged
in some uphill battle against your poorly articulated
pseudo-science. I guess I don't see it that way.
>
>>So now the AI crowd musters all these ducks and puts
>>them in a row that supposedly proves their 'Terminator'
>>scenario: the on-rush of machine intelligence is soon
>>to sweep you sorry humans aside.
>
> I've never been concerned with that - why do you
>keep injecting into the discussion. I'm only concerned
>with your claims that humans do "non-computable" things.
>
"Chess is showing that with continual development chess programs are
now beating the very best humans - with more development I expect they
will completely dominate (I'm being generous - many would say they
completely dominate already. Statistically they certainly do - most
PCs still in service can run a chess program that will beat most
humans.)?" -- that's you writing. Then you said something about why
aren't people more agape about the miracle of modern technology.
I extrapolate that you're one of those people who wants us to imagine
computers are on the brink of being intelligent in the way humans are,
when nothing could be further from the truth.
What's true is take the best chess playing computer "off road" into
some chess puzzle, and suddenly their brute force algorithms fall
apart, where children can easily see a solution.
> What started this whole nonsense was my pointing out that for every effectively generated theory you do have an algorithm that automatically produces all the theorems. Its a theoretical result of theoretical interest. It was apparently mis-interpreted by you and DC that I was saying something or promoting something that I'm not.
>
I did start out by ridiculing this little tautology, that any
automatically generated domain of so-called proofs thereby qualifies
as effectively generated.
Lets just not be sloppy and confuse "effectively generated theory"
with something important, like geometry or number theory or any other
domain in math where computers have yet to effectively generate all
the theorems and show no evidence of being able to do so.
Quoting myself:
"""
The ability to generate proofs is quite different from the ability to
work backward from a conjecture still in need of proof.
In the tiny worlds created by tautologists, "generating proofs"
becomes a simple operation indeed. Child's play.
Don't confuse this with the idea the Euclid's corpus might've been
"generated" by one of today's IBMs running software written by Harvard
grad students. Not so.
"""
> - ---------------------------------------------------------
> That's the syndrome in a nutshell: some desire to assert that humans
> are a subservient subset of intelligent being and those who have
> mastered computers are mastering the secret of that intelligence.
>
> It's all sleight of hand hooey of course. There's no reason to feel
> inferior to your PC or grovel at the feet of your Mac (were your Mac
> to have feet).
>
> Your attitudes typify an ethnicity I would say. Thanks for reminding
> me of these memes, as I don't encounter this type of thinking every
> day.
> - ----------------------------------------
>
> You are WAY off base Kirby, pay attention to what's said
> and stop injecting your own fantasies.
>
It may not have occurred to you that this archive has other readers
than yourself and a thread is a place to elaborate many themes (more
than one anyway).
If you're not the AI-brand positivist you seem to be and show all the
signs of being, then perhaps you will give some indication in what way
you've been wrongly pigeon-holed.
In the meantime, I will assume I'm on target about your thought
patterns / ethnicity, which are on the predictable side, I admit -- a
good candidate for machine automation.
You've met Eliza. Meet Joe.
Kirby
>
> Joe N
|
|
