I think I have a better idea.

 

I believe the problem has to do with assumptions.  If I have an AI system
that assumes an AGI can be made without using the human model, most comments
by people who believe the opposite aren't worth much.  Instead of some
helpful criticism of the AGI I might be proposing, people get into a pissing
contest over the assumptions (basic beliefs that can't be proven one way or
the other) instead of helpful information about the idea being discussed.

 

My proposal is to use a link in our emails that states what assumptions that
poster is using and then all valid comments must take those assumptions into
account in their argument.  This wouldn't stop discussion of those
fundamental assumptions but it should stop people continuously talking past
each other because they are using a different set of assumptions that
further arguments are based on.  I think I could make contributions to
projects and ideas that I don't think will create an AGI but in thinking and
discussing others ideas, it might spark some good ideas that could help me.

 

If someone just wants to continually argue that AGI is impossible or even
that certain ways of going about AGI are worthless, why not just keep it to
yourself?  We don't all have to agree on the best way of creating an AGI for
this list to be useful.  Just stating that something is wrong when almost
everything we discuss hasn't been proven one way or the other isn't helpful.

 

If some threads aren't interesting then don't read them.

 

If we could set a ink to the assumptions for a thread and then comment using
those assumptions, I think the content/noise ratio would be increased
substantially.

 

David Clark

 

From: Ed Porter [mailto:[EMAIL PROTECTED] 
Sent: August-03-08 1:19 PM
To: [email protected]
Subject: RE: [agi] META: do we need a stronger "politeness code" on this
list?

 

I think my email about Richard has spurred some valuable discussion.

 

Here is a proposal for how to make this list work better. Unfortunatly it
would involve some web coding, something I am not sufficiently up to speed
to do, or even to estimate exactly how much work it would require.

 

The proposal is to have a participatory grading of posts.   To adapt the
system better to each user, a user could select whose gradings he or she
wanted to use as a filter.  Perhaps people could select under which of one
or more topics one wanted to assign a post, to create a more forum like data
base under which users could rank posts, or even extract of posts.
Hopefully, ultimately there would even be a participatory method for cutting
out the most worth while parts of the most worth while posts and crafting
them in to a forum in which the entries under each topic would be
participatorily ranked by user selectable rankings. 

 

To keep it as simple as possible for the coder, I would suggest is that if
you read a post you like, you could just forward it with a 0 to 10 ranking
to the email address of a web site that had a script to record the ranking,
from whom it came, and perhaps suggestions as to under what subject matter
they should go.

 

People who had less time to read the list could go to a page on that web
site, paste in a string of email addresses (perhaps followed by a weight)
and the number of posts he is willing to see for a given time period, and
the page would generate a corresponding list of posts.  If no name were
pasted in the weighting would be over all people who had forwarded emails.
One could also do the same under any topic for any selected time period.

 

People viewing the emails through this filtering site would also be able to
forward their rankings.

 

This is a simple form of what I call the "intelligent forum"  a web based
participatory architecture which I developed in the early '90s.  Obviously
many more features could be added and the interface could be greatly
improved.  I have an over one hundred page description of the idea I wrote
up in 2000.   I have discussed this and other related ideas with some people
at the MIT Center On Collective Intelligence and they have expressed an
interest in seeing more about it.

 

The problem with this system, as applied to our list, is it won't work well
unless there are a reasonable number of good people willing to read and rank
all the initial submissions.  They will act as the initial gate keepers, and
if they are only the trolls and the blowhards, the whole system would be a
waste.

 

Tell me what you think.  And if any of you would be interested in doing the
coding, ping me.

 

Ed Porter 

 

P.S. I think human collective intelligence is going to be one of the most
important technologies, if mankind, is to fare well in its transition
through the worm-hole of the singularity to the trans-human future.
Although a life-long Republican I have been very impressed by Barack Obama's
very forward looking policy statements on technology, the internet, and how
the internet can be used to make government more efficient and more
democratic.  In fact, July 28th I organized an Obama "Listening-to-America"
platform meeting, which included intendees from the MIT Center on Collective
Intelligence, Harvard's Berkman Center on Internet and Society, and
Harvard's Kennedy School's Network Governance Group and which was moderated
by the Berkman Center's Charles Nesson.

 




-------------------------------------------
agi
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=8660244&id_secret=108809214-a0d121
Powered by Listbox: http://www.listbox.com

Reply via email to