Re: Penrose (Was: Infinities)

Hal Finney (hal@rain.org)
Thu, 13 Nov 1997 16:40:42 -0800


Leevi Marttila, <lm+extropians@sip.fi>, writes:
> By STFW (searching the fine web) with keywords penrose, goedel, random
> I found this:
>
> "REVIEW OF: Roger Penrose (1994) *Shadows of the Mind*."
> by Drew McDermott
>
> http://bion.mit.edu/ejournals/b/n-z/.unindexed/Psyche/by_filename/2/psyche-95-2-17-shadows-9-mcdermo

Another URL is:

http://psyche.cs.monash.edu.au/psyche-index-v2.html#som

(maybe the same thing; Psyche is an Australian on-line journal)

This has critiques of Penrose by such as Moravec, Chalmers, John McCarthy,
and other AI philosophers, including McDermott. Penrose has a rebuttal as
well. Quite technical, especially in terms of the discussions of Godel.

> Why not program computer in a similar way? I think program that uses
> guesses could many times finds proof faster than program that thinks
> formally. Conjectures would be zero order uncertain proofs.

I believe there have been various AI programs which worked on heuristic
reasoning (smart guesses). Doug Lenat's Eurisko was a well known example
from the early 1980's.

Hal