From: Simon McClenahan (SMcClenahan@ATTBI.com)
Date: Tue Feb 19 2002 - 07:59:16 MST
----- Original Message -----
From: "Arona Ndiaye" <andiaye@chello.nl>
To: <extropians@extropy.org>
Sent: Tuesday, February 19, 2002 5:37 AM
Subject: Re: AI:This is how we do it
> I've seen several posts in this thread which made me wonder:
> "Did he/she read GISAI and CFAI several times 'each' ?????"
Ugh. If that is a pre-requisite for starting or entering a discussion, then
there should be no discussion unless each person can prove that they have
done so. This is an open forum, and if someone does not have the
pre-requisite knowledge you believe they should know, then kindly refer them
to the information source. An even more radical approach would be to filter
out their messages if they continue to post until you get some notification
that they do have the background knowledge that you require for what you
consider cohesive communication. But like I said, this is a fairly open
forum (although the lower threshold of intelligence is high compared to most
other open forums).
> If he/she did, was it done with an open mind ? I think not. If I start
> bla-bla-ing about planet disassembly without reading what the people who
> formulated the hypothesis have to say, some pople on this list would get
> pissed off. And rightly so.
Not rightly so at all. For communication to happen, the sender must send a
clear signal, and the receiver must be receptive. In your role as a sender,
it is your responsibility to send a clear signal as you possibly can. To do
that effectively you must understand the capabilities of the receptor. If
you write a whole bunch of stuff that nobody reads, statistical probability
tells me that it's most likely to be your problem, not necessarily the
(potential) receiver's.
> It works BOTH ways. I fully second Eli: there
> ***should have been** some progress **at least - really at least** at the
> conceptual level on CFAI and GISAI. Questions asked here, ***were***
> answered several times on the SL4 list which is archived online. Doing
'Sl4
> archive' at google pulls it. Searching on the 'Low Beyond' pulls it.
And that's all a good learner needs to know. Maybe lesser learners require a
little more help, like hyperlinks, or a tailored summarization, or even a
discussion thread.
> What I
> cannot understand, is the 'unspoken' undercurrent biased attitude about
CFAI
> and GISAI. I'm no fu**ing genius, I'm an everyday average curious kid. If
I
> can grok most of what Eli writes, it's because I spend a lot of time
trying
> to see what he sees. I **spend a lot of time** trying to **see what
Eliezer
> was seeing** when writing/thinking about it. I 'hijack' his models.
You have been adapting your learning model to be able to receive Eli's
message, tailored to Eli's style of communication. Most people are not
willing to do that because it takes too much energy. Personally, I think
most people write and talk too much anyway. Communication Theory and
Information Theory is a whole other domain that you may find interesting if
you are looking for a model of communication. Not interested? Oh well, I
tried.
> Certain questions cannot be answered in one or
> two lines. There's no magic stuff, it's all inter-connected and elements
can
> only be understood if their profundly-complex relationships with other
> elements are understood.
Congratulations, you are well on the way of understanding the Eli
communication protocol standard. I hope the effort is worth it. I'm sure Eli
is modifying the protocol himself as he learns and continues to communicate.
His effectiveness and worthiness will be judged by how many people do
actually get the message without too much adjustment of their receptiveness.
> I repeat: I **do not intend** to offend anyone, but Eli wrote GISAI and
CFAI
> for people to read/study/hack it. I do not believe it's done as often as
it
> should be done. I also do not believe it has anything to do with how it's
> written.
IMHO, you're wrong. Why should anyone care what he's written? Why should
anyone care what you write? We all have intent in our communication, but
there needs to be some sort of hook or link to the content that will grab
each person's interest so that they may decide whether it is worth their
time to actively pursue reading all that stuff. If you've gotten this far in
the email reply, I am honored that you find my opinion interesting enough to
read on ...
> The
> AI community is known to have NO vocabulary. Why should Eli not create his
> own if he feels that some things need to be formalized instead of being
> shoved under the carpet ?
No vocabulary? I'm certain there is a lot of jargon. There are many AI
domain experts. As for formalized standards, it's not going to be defined in
natural language, that's for sure.
> If Eli saw (or believed he did or does) where
> Classical AI went wrong, so can I. If I can, so can everyone. The AI
> community has been attempting AI without understanding the human brain and
> that is stupid. Anyone trying to do that will never get anything done in
AI.
Would we have built a flying hot air balloon by studying how birds fly? I
don't disagree that there is a lot we can understand from the brain and our
neural network model, but there are other ways, even "classical" ways such
as expert systems and giant lookup tables that could work (inefficiently).
Or some other model (most likely).
> Yet, that is what the AI community did. They all got funky with buzzwords
> and 'Expert systems' and used AI everywhere it had nothing to do with AI.
> Now that it's going down the drain coz of lack of results, anyone who even
> pretends to try creating AI in a different way (hint: Friendly Seed AI ) ,
> is shot down in flames. That's a loser attitude, period.
Friendliness and Seediness I see as different applications of AI, each with
its own purpose. These applications are programmed on top of a General AI,
which is the theoretical base of all AI. Maybe I'm wrong, but buzzwords are
just jargon, or language of the AI community's domain. Just because you
think someone doesn't "get it" like you do doesn't mean that they are losers
at all. Maybe it is you who doesn't get it. If you seek acceptance from a
community, you must find acceptable ways to focus their view. Complaining
about your unsuccessful attempts to show people your view makes you the
"loser", if you must use that emotional term.
> PS: English ain't even my mother tongue, if I can get it, how hard can it
> 'conceptually' be ? Honestly ? *waves to everyone*
Uh, no offense, but if English is not your mother tongue, it shows :-P
English is my only tongue, and I'm not really confident that you or anyone
else understands my messages either :-)
cheers,
Simon
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 13:37:40 MST