[ExI] Consequentialist world improvement

[spike]

-----Original Message-----
From: extropy-chat-bounces at lists.extropy.org
[mailto:extropy-chat-bounces at lists.extropy.org] On Behalf Of Mike Dougherty
...

>>... If one often gets the feeling one should spend less time downloading 
> one's brain into ExI-chat, a cheerful thought is that it might not be 
> such a terrible waste.  You might be creating your future posthumous 
> self with each paragraph you write.  spike

>...Not just Exl-chat; the Internet in general...

Ja.  ExI-chat gets most of my best ideas and real thoughts, along with some
of my best cut-ups.

>...btw, are you serious about building an archive-powered impersonator bot?
_______________________________________________

I would be serious about it if I were a better programmer, and I am willing
to work towards being a better programmer to get this done.  I am open to
suggestion here from *real* programmers, but I thought the way I would start
is to collect a particular person's archives.  I will use my own, to get
around the ethics questions.  Then I drop all that into (I suppose) an excel
file, although I do not know the storage capacity limits of that software
(it is crazy big, but might not be big enough.)  I might need to break down
the archives into subgroups somehow.  Then I need to think about a macro
that will filter away everything that is a quote from someone else.  In my
case that might be relatively easy, because I have made a habit of preceding
other peoples' comments with >... as shown above.

So, here's a vague outline of the algorithm:

1.  I filter away all those quotes, separate the remaining material into
paragraphs, along with the subject line of the post from which it was
extracted.  

2.  The subject line text goes into a second column.  

Once I have all that, I need to figure out a way write a script to:

3.  generate a list of words that would trigger this paragraph as a reply.
I don't know how to do this yet, but I am willing to work on it.  It needs
to take the text in each quoted text before each paragraph.  For instance,
in this example, we look at the comment >...btw, are you serious about
building an archive-powered impersonator bot?  From that comment, I then:

4.  extract the nouns and assign a score of five points.  I extract the
adjectives and assign a value of 4 to each, adverbs are worth 3 and so on.  

5.  Then when a new post comes in, we evaluate each sentence, dividing
through each by the number of words in the sentence.  

6.  When a super high score is seen, our macro takes that sentence, deletes
the rest, inserts >..., quotes the sentence, then inserts the paragraph that
was originally associated with the comment.

Mike are you a code guru?  Anyone else with ideas on how to create a
you-bot?  For the first time in history, we have enormous text samples of a
person's thought space, readily available in a form easily manipulated.
This is a cool time to be alive, ja?

spike