[extropy-chat] I'm back.

[Eliezer S. Yudkowsky]

Dan Clemmensen wrote:
> 
> Actually, I tried to say that the probability density function has a 
> sharp peak somewhere in the region where the technology curve moves 
> through the phase transition region. This, coupled, lying on top of the 
> increasing slope of the technology curve, makes it very hard to predict 
> the crossover.

Given that probabilities express our lack of knowledge about events, a pdf 
that would be sharply peaked if we possessed knowledge of a key parameter 
theta, can still be very broad if we do not know theta.  As you know I 
agree with you about rapid transitions (most of which will instantly kill 
you, btw), but predicting "May 1st, 2006" is the wrong way to express it. 
  Rather it could be any thirty-minute period between now and 2020.

> So my real prediction is "any time now."  I'm willing to listen 
> respectfully to anyone who can support a more precise theory.

This is my prediction as well, I'm just disagreeing with your possessing 
such a good prediction and yet refusing to use it.  :)

-- 
Eliezer S. Yudkowsky                          http://singinst.org/
Research Fellow, Singularity Institute for Artificial Intelligence