Re: Singularity definitions

From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Wed Apr 16 2003 - 13:18:40 MDT

  • Next message: gts: "RE: evolution and diet (was: FITNESS: Diet and Exercise)"

    Anders Sandberg wrote:
    > On Wed, Apr 16, 2003 at 10:58:42AM -0700, Charles Hixson wrote:
    >
    >>Spudboy100@aol.com wrote:
    >>
    >>>The Singularity, as I understand it, is when technology becomes
    >>>powerful and complex enough to become self-aware, and have "internal
    >>>conversations with itself" and become asymptotic in intelligence. Most
    >>>on this list believe it will approach amazing and overwhelming...
    >>
    >>As I understand it, that's the way we predict the Singularity will
    >>happen. But from my understanding of the definition, the Singularity is
    >>that point at which the past becomes useless for predicting the future.
    >
    > Both of these definitions are different from Vinge's original definition,
    > although they are close in the general sense the word is used. He suggested
    > the singularity as a strong feedback loop in technological development
    > causing a fast transition to a state we could not at present predict. Note
    > that it does not have to involve AI, it could be IA or something else. The
    > feedback loop is the precondition, the unpredictability is the consequence.
    > It does not say anything about indeterminism; even a deterministic
    > singularity would likely be computationally unpredictable since so much
    > happens that can not be predicted without the same level of technology.

    Actually, Vinge's way-back original definition was that the Singularity
    was the breakdown in our PRESENT MODEL of the future occurring when we,
    ourselves, attempt to extrapolate our models past the point where they
    predict the rise of greater-than-human intelligence. The laws of physics
    don't break down at the center of a black hole, but our *model* of physics
    does. The more I think about this the more I realize that Vinge got his
    original definition exactly right on the very first try and it is superior
    to all later misinterpretations including my own. Though the
    superintelligent transition Anders refers to runs a close second (often
    referred to as "the Singularity" even in Vinge's own later work).

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


    This archive was generated by hypermail 2.1.5 : Wed Apr 16 2003 - 13:27:43 MDT