Re: Recursive self-improvement curves

From: Eliezer S. Yudkowsky (
Date: Sun Jan 04 2004 - 16:14:48 MST

Mitchell Porter wrote:
> Eliezer:
>> Problem is, the equations you're offering can't be fitted to the case
>> of hominid evolution, or evolution, or to human culture, or to any of
>> the other classic cases of accelerating complexity offered by John
>> Smart, Ray Kurzweil, and the like.
> Processes leading to us may have an anthropic element - the
> critical path to sentience has to be traversed before the sun
> boils the oceans, without premature extinction of intermediate
> ancestral forms, etc. So they may be unrepresentative of
> self-improving processes in general.

True, but the anthropic element here should contribute to, at most, the
illusion of linear progress - "hard steps" should be roughly evenly
distributed across the historical record, not coming in quicker sequence,
unless something else is going on. That is, if it's extremely improbable
to go from single-celled to multi-celled life, and extremely improbable to
go from primate to human intelligence, then (if that's all that's going
on) we should expect to spend as much time between single-celled and
multi-celled life as between primate and human intelligence. Or so Robin
Hanson informs us; I confess that I have not done the math.
"Must Early Life Be Easy? The Rhythm of Major Evolutionary Transitions."

Eliezer S. Yudkowsky                
Research Fellow, Singularity Institute for Artificial Intelligence

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:43 MDT