From: Ben Goertzel (ben@goertzel.org)
Date: Wed May 03 2006 - 15:44:10 MDT
> Kurzweil's TSIN is not a tiger. It is a book. Flaws that hit people
> who read books are relevant - regardless of whether the biases arise
> from heuristics shaped more by regularities of the ancestral environment
> than by cognitive limitations.
Regarding this point, I would need a lot more convincing to agree that
Kurzweil has succumbed to the conjunction fallacy in his book.
I understand your point that he argues for "AI will be realized soon"
via arguing for "AI will be realized soon via understanding and then
mimicking human brain structure" ---- but taking this line of argument
does not immediately imply making a conjunction fallacy. The book
had to have a finite length, and so he chose to present detailed
arguments for only some of the possible ways that "AI will be realized
soon" might be validated. I personally would have made different
choices regarding which possible pathways to AI to present, but I
don't think it was wrong of him to choose to present **some**
particular representative ways that AI might be realized in moderate
depth, rather than giving a lighter and shallower treatment to a
broader variety of possible routes. In any event this was an
expositional choice rather than an inferential error, so far as I can
recall from my reading of the book...
Of course, you may be able to point out places in the book where he
made actual inference mistakes, but I don't happen to recall any...
I found that book frustrating in some ways, but not due to any actual
erroneousness, just due to his choices of emphasis. However,
Kurzweil's choices were likely made based on sound marketing-style
thinking, a kind of thinking he is better at than I am....
-- Ben
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT