> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On
> Behalf Of Marvin Long, Jr.
...
> In my cycicism I would expect advertisers and lobbyists to try to use such
> technology to identify and suborn, er, pay-to-shill, influential people.
> Then we'd want to develop software designed to detect those who have been
> "turned" based on patterns of conversation....
The ones who were most interested were Wall Street, politicians and
Hollywood. Politicians got the idea faster than anyone else. Of course,
there's a problem with any kind of forecasting for politicians -- in the
long run, they will only pay for forecasts that tell them what they want to
hear. That's apparently well documented, at least according to a social
networks researcher with whom we collaborated.
The problem with any business that targets on-line influencers directly is
that no matter how it would approach them, it would be spam. And then
there's the Heisenberg uncertainty principle.
What's more intriguing is the idea of creating artificial participants in
on-line discussions, using what's known about influence to make them as
influential as possible. One of these days, I'm going to write a gizmo that
learns how to be influential by posting messages and watching what happens.
Once the identity becomes influential, the question becomes how to use the
influence.
Nick
