From: J. Maxwell Legg (income@ihug.co.nz)
Date: Thu Sep 03 1998 - 05:38:26 MDT
Dan Fabulich wrote:
> J. Maxwell Legg wrote:
> >see http://www.airtime.co.uk/users/station/m-worlds.faq
>
> I know about that. You still have to spell this out for me. I still don't
> see how many-worlds relates to AI in any way that it doesn't also relate to
> anything else that makes entropy.
Quite right, it doesn't. It was an aside that to show my confidence that the
existing ruler's Achilles heel is exposed by m-w as AI might allow a revolution of
power.
> >
> >To see this simple software idea - http://come.to/ingrid
>
> This software idea is so simple that I don't see how it relates to our
> conversation at all, except possibly to the extent that we might think of it as
> a primitive form of AI.
Maybe primitive today but my involvement with this software is my unique ticket
into uploading land. I'm just waiting for hardware to catch up with my own unique
view of how to develop a massively parallel version. If you can't see the Ingrid
software's potential to transform into a full blown brain, well that's not my
problem.
> >More to the point, look at the way my email is slopped around, not only on
> this
> >list but all throughout the course of unmediated human communication. Thread
> >meltdown occurs as easily as subject headings getting changed willy nilly.
> When
> >the Skull & Bones storm troopers come out of the woodwork to slay perceived
> >dragons, information loss is the result.
>
> <sigh> You're bringing up topics before you're explaining them; your use
> of metaphor isn't helping me here.
>
> Are you referring to the secret society "Skull & Bones" of which former
> president George Bush was a member? How does Skull & Bones "slay a
> perceived dragon," as you put it, and how does that result in information
> loss? What does the activity of Skull & Bones and/or the loss of certain
> information have to do with AI? I presume that it relates to m-w in that
> the loss of information represents an irreversible process. Moreover,
> beyond the fact that S&B has an agenda, what does any of THAT have to do
> with accounting?
It has to do with knowledge discovery by being able to have software parse the
actual email content. Just look at the actual format above as an example of email
that us humans can barely decipher and then think how hard it will be to extract
meaning ny getting an AI to read this thread. My comment was a lament that
software agents are not available during the actual writing process in order to
structure the knowledge on the fly. Secretly I hope that 'knowledge extractable'
email might be the nicest way to kick start my AI concept. Hey, I'm a dreamer.
> >With all this undisciplined communication
> >all that is likely to happen is that we will continue to jabber on while
> the super
> >intelligent AI networks surpass any efforts at human control. However this
> doesn't
> >have to be the case. As you know, I have an expressed interest in the future
> >development of feedback loops via continuously adjusted summaries of all
> >encompassing global activities.
>
> I didn't know that, nor do I know what sort of feedback loops you had in
> mind. Continuously adjusted summaries of all encompassing global
> activities? I have no clear idea as to what this is; I have even less of
> an idea as to why it would be useful.
I'm sorry I can't help you see the usefulness of my view of how and why AI should
best be implemented at a global level.
> >Don't like what's going on here? Need a soapbox? Have a beef about
> something or other
> >then this will be your chance to have a say in the making of new global
> politics.
>
> What will? What's "this?"
"this" is my view of how and why AI should be implemented at the global level.
> What exactly do you mean by a "qualitative statistical mediator?" What's a
> "super ordinate construction?"
"super ordinate constructions" is another term for "core constructs" from the
specialty field of Personal Construct Psychology where the Ingrid software was
developed. See George Kelly's theories. I foresee grids (neurons) dynamically
activating the levels of knowledge extracted according to Kelly's theories.
Qualitative or synthetic data includes such independent componentry that make it
applicable to this imagery type of processing. I would hope that an open structure
would allow enough variables to be processed that the contributors of the data
would trust the inferences displayed. Therefore a "qualitative statistical
mediator" is a software system capable of say presenting solutions to large
problems. I wouldn't be surprised if such a tool was used to design the Northern
Ireland peace proposal. (I know for example that Tony Blair was greatly influenced
by Kelly's work.)
> In short, super-AI will read everything, know everything we can know,
> process all the information, and deal with it accordingly. Yes?
To what point in the life cycle of the AI does your irrelevant question refer.
> Here again you lose me. How will super-intelligences get a self image?
> Where did common sense come in here? Why should we hope that it will
> "reflect human effectiveness?" What sort of daunting realities will remain
> as unconscious elements in a super-intelligent mind?
There are classified records of many failed experiments locked away. Read
"Ellsberg on secrecy" to get an idea of the sort of daunting realities that I
referred to. Please understand that my concepts for a global AI based neuronomy
are only possible if all records are available for knowledge extraction and
integration. It is after all only my hope for a hypothetical possibility of a
specific future implementation of AI and it isn't a certainty. To get a handle on
what constitutes the psychiatric term "ideal self image" you need to know that
it's like something that happens to young animals when getting their first
glimpses of their parents.
> Right, now, let me clear something up. Would humans in this case be
> external to the AI, or synaptic-like-units which comprise the AI?
I'd prefer that humans continually contribute to the AI's synaptic weights along
as humans are around. Wouldn't you?
> ...
> [needs and wants] are undertaken as different processes than something else
> which you have only intimated?
>
> And for that matter, how are needs and wants "undertaken" in the first place?
As a preference I would like to see this tech be as unobtrusive as possible as
regards needed transactions or undertakings - does the tech in the novel about the
Aristoi do this? Wants may operate using other structures but don't pin me to what
these subtleties should be as I don't know and nor could anyone ever address the
situation until it was modeled or in operation.
> >The present economic drivers may/will still be active but there
> >will also be so many more all using the same forms of abstraction and
> delving into those
> >hard to get at bionomic and personal areas that the methods of reporting
> the same and
> >the motivation for doing so will be so unlike what restricted capitalism
> offers, so much so,
> >that the means of transactional measurement should at least for clarity's
> sake not be
> >called money.
>
> Economic drivers? You mean things that drive the economy, like needs and
> wants? Or something different? And what other sorts of "drivers" might
> appear? What do you mean by forms of abstraction and delving into bionomic
> (do you really mean ecological?) and personal areas? How do you report a
> bionomic area? What WILL be the motivation for doing so?
AI reporting structures will hopefully evolve to suit the needs and before it's
built, like you, I really don't know what to expect or what will be the motivation
for doing so. (Ask me again when I get unlimited computing power and interfaces to
all the worlds knowledge bases:-) The differing sorts of drivers (factors) in each
network node are derived using Kelly's repertory grid technique from the actual
elements making up the node itself.
> >Don't forget that no land area on earth is presently exempt from the
> >totalitarian concept of money and that exchanging one concept for another
> at particular
> >moments in history has been done before.
>
> As a point of interest, this is a terrible abuse of the term
> "totalitarian." Most totalitarians throughout history have, in fact, been
> socialists. I suggest you stick to the tried and true term "tyrannical,"
> if you insist on calling it a system of domination. But at any rate...
Yes that was a bad term. I meant that money was in use across the totality of the
earth's surface so much so that it crowds out any discussion of an alternative.
But remember even an ice age can end.
> Here, lets try a thought experiment. Suppose you've got a two person
> economy, Alice and Bob. Alice has an apple which she happens to value less
> than Bob values the apple.
>
> Now, in capitalism, Bob will pay Alice for the apple, thus maximizing value
> and resulting in economic efficiency in the Marshallian sense. This
> process is scalable to all kinds of economic activity.
>
> Now, suppose Alice and Bob had chosen to reject capitalism and instead
> operate under a neuronomy. What would happen? What would they do? How
> would it work?
>
When thinking about a neuronomy it might be easier to consider that it couldn't
work with only two people and would need a highly developed interdependent global
economy where no element was independent. Furthermore it would have come into
existence in a similar way to the internet in that it couldn't be designed without
building it. Your whole question is simply unanswerable because I refuse to answer
that level of detail. This is a pity because you're also not going to find anyone
else to answer your questions as they are too afraid of upsetting their position
in capitalism.
>
>
>
> What I'm looking for is fewer buzzwords and more content. ;) You're using
> terms I know in ways that don't seem to make sense in this context. What
> do you mean here?
I don't think you want help to connect the dots but if I'm wrong you could do well
by taking my buzzwords and use a search engine to determine how I'm using them.
> What does this have to do with WINSOCK.DLL?
Just a humorous example of a point and click enabling something like a
hyperlink_with_responsibilities.
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:49:32 MST