[extropy-chat] "The Spike"- Ray Kurzweil

[Robert J. Bradbury]

> Exaclty! My point is to state that you profoundly diagree with Ray's
> timeline predictions because, in your own words, science is radically
> unpredictable.

Scientific "discovery" might be unpredictable -- but engineering
based on known physical laws is less so.  Once one has a few data
points to go on (e.g. Moore's Law, or the progress in computers
understanding human speech, or gradual improvements in software
that does everything from face recognition to driving cars, etc.)
then things start to become somewhat predictable.

> But when one reads Ray's books one gets the feeling that
> future science/technology IS predictable, or at the very least broadly
> predictable.

It depends very much on how many data points one is dealing with. Moore's
Law tends to be very predictable -- signifcant investments are made in
making sure it will happen and where the problems lie and how to solve
them.  Now, the human genome project was much less so say 10-12 years ago.
The general consensus circa 1990 was that the HGP could *never* be
accomplished.  The science for DNA sequencing was well established but
there had to be breakthroughs in the technology.  So a few hundred million
$ were spent funding the research to produce those breakthroughs. Result -
the HGP was finished early and under budget (there is your
unpredictability cropping up).  Though it might well have gone the other
way (as projects at NASA involving putting humans in space seem to do...
:-().

The interesting question about the singularity is whether it requires
a certain number of breakthroughs in engineering -- or science -- that
we cannot easily predict right now -- or whether it will result from
the aggregate set of trends in the world (which it is perhaps equally
hard to grasp as a whole).

R.