Re: Singularity: The Moore's Law Argument

Eliezer S. Yudkowsky (sentience@pobox.com)
Fri, 18 Sep 1998 18:21:13 -0500

Robin Hanson wrote:
>
> If you think that post is a persuasive technical analysis on timing claims,
> rather than suggestive prose, I suspect you don't know what such analysis is,
> and so my requests for it are futile. Does *anyone* reading this other than
> Eliezer think that Eliezer's first post constitutes such a persuasive
> technical analysis?

First of all, my first post is a selection of a summary of a technically detailed speculation. The summary of the speculation is in the "Singularity Analysis" web page, and the speculation itself is in "Coding a Transhuman AI", which if posted here would bring the mailing list to its knees.

As for technical analysis, if you'd like the kind of elaborate equations beloved of economic forecasts, you're simply out of luck. But be comforted by the fact that such equations have never successfully predicted the course of the human race and never will, so it is utterly absurd to extend them to superintelligence.

I suppose I could fake up an elaborate power/intelligence/speed graph, using EURISKO's runs, past trends in computing power, the gradual drift in intelligence, and records of economic revolutions. But it'd be a statistical lie (and as we know, those are even worse than damned lies). I don't intend to do it because I don't think it'd be honest to my readers. I have enough confidence in my rough models. I shall not disguise them with impressive mathematical uniforms they are not entitled to wear.

-- 
        sentience@pobox.com         Eliezer S. Yudkowsky
         http://pobox.com/~sentience/AI_design.temp.html
          http://pobox.com/~sentience/sing_analysis.html
Disclaimer:  Unless otherwise specified, I'm not telling you
everything I think I know.