[iDC] if AI were a reality, which, currently, it is not

[subbies at redheadedstepchild.org]

We aren't at the point where they can be replaced outright.  They are a 
temporary solution, in light of the fact that computers cannot yet 
understand what they read.  Truthfully, I don't believe there is a decent 
solution that can exist short of this event happening.  All solutions 
will be kludges as long as we use stupid computers to display and 
manipulate our data.  But, here are at least a few things I think that 
need to happen before this occurs (obviously, not a complete list):

1) Computers need to be capable of contradicting themselves.  To steal is 
wrong.  To allow someone to die is wrong.  What to do if the only way to 
save someone is to steal?  Which is less wrong?  Are there still other 
factors that can change the answer?  Computers must be able to make such 
calculations.

2) Computers - and PEOPLE - need to recognize that there is not a single 
entity called "truth."  There are multiple solutions that appear to work 
for any given problem.  This doesn't mean that there aren't certain 
solutions that are more useful than others, insofar as 
testing/implementing them consistently exhibit positive results, make 
progress, match what is expected, etc.  Nonetheless, when our knowledge 
base and set of assumptions changes down the road, we will discard these 
successful methods for more successful methods.  The more recently 
successful method is not more true, however, just more successful at 
answering a given set of questions.

3) Given 1 & 2 above, we should stop expecting all computers to come up 
with the same solution to the same questions.

4) We should provide computers with "experiental" knowledge bases.  These, 
by definition (at least as far as I understand them) can never be 
complete.  This is okay.  See 1-3.  Each computer can have a different 
knowledge base and weigh the likelyhood of any event based on its 
personal knowledge base.  Should they be programmed by the computer owner?  
Can the computer add its own information to its own knowledge base?  How 
would it go about learning?  Choosing things to add?  What would be 
weights to give things?  A weight for positive association?  Links to 
other events? Percent similarity?

In essence, currently, computers are quite literally CALCULATORS.  We 
cannot ask a calculator to read for us.  All we can do is make it 
tabulate probabilities. I don't think any effort to make 
a calculator read will ever be successful.  So any solution that 
might be dreamt up would have to start with making computers more than 
adders and subtractors.  Is this actually possible?  Tough to say.  To 
what extent are people calculators?  Does a person calculate through their 
memory when confronted by a new situation (this is X% similar to this 
event that happened to me once before, and that event made me y% happy and 
b% unhappy, and it was related to my mother, for whom I have g% of trust, 
so I now understand this situation to be...)?  Or do they understand 
things differently?  We don't even know how people understand the world 
around them, or how a particular sentence will affect a specific 
individual, so how to build a computer that can understand a sentence? 
-Alexis




+ --------
   redheadedstepchild.org
        ------- +

On Mon, 17 Mar 2008, Sandy Klausner wrote:

::Date: Mon, 17 Mar 2008 10:56:15 -0700
::From: Sandy Klausner <klausner at coretalk.net>
::To: subbies at redheadedstepchild.org
::Cc: Paul Prueitt <psp at ontologystream.com>, idc at mailman.thing.net,
::    susan.turnbull at gsa.gov
::Subject: Re: if AI were a reality, which,  currently, it is not
::
::What do you propose to replace them?
::www.coretalk.net
::Sandy
::
::On Mar 17, 2008, at 10:48 AM, subbies at redheadedstepchild.org wrote:
::
::Two points:
::
::One:
::Actually, we don't disagree about the likelihood of AI becoming a reality.
::We disagree about *why* it is unlikely to become a reality.  I disagree
::with the physicality argument.  Unless you believe in magic, it seems
::that the gist of that position is that people can *experience* - ie, they
::have senses and their existence in space and exposure to things - via
::their senses - informs their ability to comprehend the information they
::come across. This is the only thing that separates them from
::machines.  I don't believe in magic, and I don't believe it is impossible
::to replicate the senses (as to the latter point, I suspect pattern
::recognition is more than enough to accomplish this feat).  Instead, I
::believe that there is nothing mystical about humanity, or
::special about the senses, but that what makes us human is simply our
::complexity - an extremely non-trivial problem, especially when
::tackled using concrete logic.  If nothing else, people are successful at
::being people  in their willingness to embrace contradiction, something
::which computers are unable to do at this time, and at a very fundamental
::level.  There is more difference than this, of course, but certainly this
::is one of the biggest stumbling blocks to modern efforts at AI.
::
::Two:
::As for my "Church" - my "Church" is the one that says a person ought to
::find realistic solutions to problems, not sit around crying like a damn
::baby that someone else ought to solve their problems for them.  My church
::is the church of construction, not deconstruction.  Deconstruction is for
::hothouse flowers that can't be bothered to work for what they desire
::and/or who are too afraid of making something that visibly doesn't work to
::even try.  It's easy to tear down an existing system.  It's preposterously
::hard to build a new one that works, but nonetheless we must try lest we
::end up doing nothing but sitting proudly atop our rubble.
::
::I don't believe all people should learn to program.  But I do believe that
::if you truly feel that you are in thrall to an evil semantic overlord,
::there is a solution that is *legitimately* within reach to rectify the
::problem.  Your comparison to television was weak because there is no
::legitimately realistic solution to changing what is on the channel, not
::one that the average person can take upon themselves.  Programming, on the
::other hand, while specialized, is nonetheless within reach and therefore
::practical to my mind.  If data doesn't trouble you, don't learn to
::program.  If democracy doesn't interest you, don't vote, or vote along
::your party line instead of doing your research.  On the other hand, if you
::feel strongly enough to believe you should retain a level of control over
::these things from the powers that be, then, yes, I am afraid you must take
::the matter into your own hands at some point.
::
::So, to that end, you think semantic web standards are broken.  What do you
::propose to replace them?
::-Alexis
::
::
::+ --------
::   redheadedstepchild.org
::        ------- +
::
::On Mon, 17 Mar 2008, Paul Prueitt wrote:
::
::::Date: Mon, 17 Mar 2008 11:06:21 -0500
::::From: Paul Prueitt <psp at ontologystream.com>
::::To: subbies at redheadedstepchild.org
::::Cc: Sandy Klausner <klausner at coretalk.net>, idc at mailman.thing.net,
::::    susan.turnbull at gsa.gov
::::Subject: if AI were a reality, which,  currently, it is not
::::
::::I am thankful for this note, as it clarifies some things.
::::
::::The issue that we seem to disagree over is seen in the word "currently:
::::
::::> The only way that would be possible is if AI were a reality, which,
::::> currently, it is not.
::::
::::The argument that Penrose and Robert Rosen and others make, is that, like
::::creating life from abstractions; abstractions may only be a part of
::::intelligence.  One need actual substance, and the silicon computing paradigm
::::does not have that substance.  In fact Pribram and I and others regard the
::::necessary substance as something that deals with emergence, locality and
::::non-locality in a way seen in the emergence of function in biological
::systems.
::::
::::There is no future to AI.  What there is is a recognition of the things you
::::are saying about the nature of the problem in data interoperability and the
::::aggregation of information into a synthesized presentation (un touched by a
::::centralized control).
::::
::::The nature of the Church that you may have membership in is illustrated by
::the
::::statement:
::::
::::>  We do, however, have the ability to
::::> program computers using whatever algorithm we feel is most meaningful and
::::> most able to send us the information we want.  Perhaps you are mad because
::::> the other people are not doing it for you in the way you most desire.  In
::::> which case I say learn to program.
::::
::::First, I am not mad, not upset and not insane.  I am insistent about a
::::specific point of what I and others regard as science.  I am also insistent
::::that the programmer class should not be forced on all of us .  Suppose we
::take
::::your statement at face value and evolve a culture where the right to vote
::::depends on one's programing ability?
::::
::::Why not let them eat cake?
::::
::::
::::
::::
::::
::::
::::On Mar 17, 2008, at 10:43 AM, subbies at redheadedstepchild.org wrote:
::::
::::> Sigh.  I'm no disciple of anyone but myself, thanks.
::::>
::::> As for "the point is that interpretation by a human in context is where
::::> meaning is defined" - I would completely agree with this, which is
::::> precisely why I think you're tilting at windmills.  In spite of that
::::> pretty statement you made, the ROOT of your argument is in fact a
::::> complaint that machines are not currently capable of human nuance, that
::::> they are not able to glean meaning from data.
::::>
::::> To glean meaning is not the purpose of machines.  To glean meaning is the
::::> purpose of humans.  The purpose of machines is to assist humans with tasks
::::> that are too herculean for them to complete unaided.  Given the vast
::::> amount of data presented to us, interpreting it all, unaided, in its naked
::::> glory, is quite impossible.  To that end, machines are needed to sift.
::::> But they are not, nor should they be, employed to derive meaning from the
::::> data.  The only way that would be possible is if AI were a reality, which,
::::> currently, it is not.
::::>
::::> Your statement that machines will push information to us like TV is
::::> ludicrous.  We have no control over what the television gives us beyond
::::> changing the pre-programmed channel.  We do, however, have the ability to
::::> program computers using whatever algorithm we feel is most meaningful and
::::> most able to send us the information we want.  Perhaps you are mad because
::::> the other people are not doing it for you in the way you most desire.  In
::::> which case I say learn to program.  You are only as much a slave to the
::::> W3C as you allow yourself to be.
::::> -Alexis
::::>
::::> + --------
::::>    redheadedstepchild.org
::::>         ------- +
::::>
::::> On Mon, 17 Mar 2008, Paul Prueitt wrote:
::::>
::::> ::Date: Mon, 17 Mar 2008 09:50:49 -0500
::::> ::From: Paul Prueitt <psp at ontologystream.com>
::::> ::To: subbies at redheadedstepchild.org
::::> ::Cc: Sandy Klausner <klausner at coretalk.net>, idc at mailman.thing.net,
::::> ::    susan.turnbull at gsa.gov
::::> ::Subject: Re: [iDC] please make comments regarding semantic overlay term
::::> ::
::::> ::
::::> ::On Mar 17, 2008, at 7:49 AM, subbies at redheadedstepchild.org wrote:
::::> ::
::::> ::>  Semantic standards are not designed
::::> ::> to aid humans in understanding the data that passes across the web -
::::> they
::::> ::> are designed to assist machines in parsing the data and delivering it
::to
::::> ::> humans.
::::> ::
::::> ::
::::> ::Oh my.   So in the near future we will be the receivers of massive
::amounts
::::> of
::::> ::data, that the system pushes on us; but which is not organized to be
::::> ::meaningful to us.
::::> ::
::::> ::Hummm.. sounds like we already have that, it is called TV.  The Internet
::::> could
::::> ::be and is more.  The potential is for true social networking driven by a
::::> ::demand economy.  The demand side is what we do not now have fully
::::> operational.
::::> ::Balance balance balance.   Supply side by itself is ruining American,
::and
::::> the
::::> ::world.
::::> ::
::::> ::There is an organization to the ads and the TV programming, and there is
::::> value
::::> ::sometimes; but mostly this just feeds the addiction that is wasting the
::::> Earth,
::::> ::its resources and its people in the pursuit of consumption.
::::> ::
::::> ::The point is that interpretation by a human in context is where meaning
::is
::::> ::defined; unless you are one of the disciples of the Current Church of IT
::::> ::(CCIT) Incorporated.
::::> ::
::::> ::
::::> ::
::::> ::
::::> ::
::::> ::
::::> ::
::::> ::
::::> ::
::::
::::
::
::