From: Michael Vassar (michaelvassar@hotmail.com)
Date: Sun Apr 23 2006 - 11:22:20 MDT
I read the first 70 pages or so, prior to it's mention on this lis. Decided
it was a waste of time.
Acceleration skepticism is worth while, but this book was simply not well
thought out, credible, or well written.
The Singularity is Near is full of poor reasoning, and clear thought would
be served by a book of better reasoned skepticism and analysis, but such a
book hasn't been published yet. I keep track of the field well enough that
I'd know. The closest thing is the excellent "Human Accomplishment: 800 BC
to 1950 AD".
Phil Goetz could probably write an intellectually honest and competently
argued version of "Future Hype" if he chose to, and it might contribute
signficantly to the quality of low SL3 discussion. I'm not at all convinced
that well reasoned arguments can sell enough, or even attract enough
attention from high status people, to make them worth the effort of writing.
Anyway, none of this discussion is SL4. Popular conceptions of the
"singularity", grounded on trend extrapolation for either poorly designed
metrics like "change" or narrow metrics like "transistors per dollar" are
squarely SL3 because they inspire a belief in discontinuous change without
any real understanding of the reasons for discontinuous change or it's
likely nature/dynamics. At SL4 people don't believe in "the singularity"
because they have extrapolated the past, but because they see in rough
outline how it can be brought about. They don't believe it will happen in
2037 because that's when it's due, they believe that it will happen when
someone makes it happen, if anyone ever does. At SL4 "the singularity"
isn't seen as a technological trend, like Moore's law or more powerful car
engines, it's seen as a specific class of technology, like "heavier than air
flying machines". We see how the laws of physics and information allow for
such things, and how contemporary technology is sufficient to allow credible
attempts, and how near future technologies, which are predicted
substantially as the outcomes of trend projection (but trend projection
stepping forward another decade or at most two decades on trends that have
proven themselves for many decades) should make it slightly easier to make
credible attempts to bring about singularity.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT