At 12:05 28/12/2012, Ed wrote:
Krugman's piece in this morning's NYTimes
appears to take us well into the realm of
science fiction. But then maybe it isn't fiction any more?
(KH) For those who want to read Krugman's latest
in the here-and-now I've copied it after my comments:
Surprise! Surprise! Paul Krugman might actually
be waking up to reality. That we might now be in
a period of no-growth. This is something I've
been saying for years and before the 2007/8
crunch, too. but Krugman is now being equally
naive about the future -- if he thinks that,
somehow, automation will soon produce miraculous
economic growth. Or perhaps it's only Prof Gordon
who believes that. Perhaps Krugman will blow
Gordon's prognostications into the sky in the
sequel he promises to write. If so, I would
welcome that because I'd be able to praise Krugman for a change.
I also believe that automation will take away all
repetitive work away from humans. But it won't be
anytime soon. Ever since so-called "5th
generation" computing -- the massive effort by
the Japanese government in the 1980s -- to
develop super-computing, artificial intelligence
(AI) and the like, full realization
of automation is no nearer being achieved today that previously.
The reason is (IMHO) that automation software is
uni-directional. It simply goes from A to Z. It
may, in the course of it, be temporarily directed
into sub-sets, and even into sub-sub-sets, but,
sooner or later, the instructions rejoin the main
track. This is why AI has got absolutely nowhere
in the last 30 years Outside Japan many
researchers were working on AI many years
beforehand by building neural circuits that were
copies of the dense networks in the human cortex,
and hoping that the act of cognition would somehow follow. Well, it never did.
The reason is that cognition and decision-making
of the human variety seem to require two separate
inputs, not just one. For example, in daily
decision- making (mental or physiological) our
own software, instruction from our genes, also
require quite independent feedback from thousands
of different sorts of chemical agents which also
lie along our DNA. These are called epigenes.
My guess is that mathematicians who are involved
in AI will have to invent a double software
system. If this is of the same mind-boggling
nature as the discovery of epigenes was then
no-one can possibly say when it might
occur. Epigenes were suspected as existing for
over 50 years (150 if we count Lamarck and
Wallace) but the dicovery had ti wait until human DNA was finally sequenced.
Keith
<http://www.nytimes.com/2012/12/28/opinion/krugman-is-growth-over.html?hp&_r=0>http://www.nytimes.com/2012/12/28/opinion/krugman-is-growth-over.html?hp&_r=0
Ed
IS GROWTH OVER?
Paul Krugman
The great bulk of the economic commentary you
read in the papers is focused on the short run:
the effects of the fiscal cliff on U.S.
recovery, the stresses on the euro, Japans
latest attempt to break out of deflation. This
focus is understandable, since one global
depression can ruin your whole day. But our
current travails will eventually end. What do we
know about the prospects for long-run prosperity?
The answer is: less than we think.
The long-term projections produced by official
agencies, like the Congressional Budget Office,
generally make two big assumptions. One is that
economic growth over the next few decades will
resemble growth over the past few decades. In
particular, productivity the key driver of
growth is projected to rise at a rate not too
different from its average growth since the
1970s. On the other side, however, these
projections generally assume that income
inequality, which soared over the past three
decades, will increase only modestly looking forward.
Its not hard to understand why agencies make
these assumptions. Given how little we know about
long-run growth, simply assuming that the future
will resemble the past is a natural guess. On the
other hand, if income inequality continues to
soar, were looking at a dystopian, class-warfare
future not the kind of thing government agencies want to contemplate.
Yet this conventional wisdom is very likely to be
wrong on one or both dimensions.
Recently, Robert Gordon of Northwestern
University created a stir by arguing that
economic growth is likely to slow sharply
indeed, that the age of growth that began in the
18th century may well be drawing to an end.
Mr. Gordon points out that long-term economic
growth hasnt been a steady process; it has been
driven by several discrete industrial
revolutions, each based on a particular set of
technologies. The first industrial revolution,
based largely on the steam engine, drove growth
in the late-18th and early-19th centuries. The
second, made possible, in large part, by the
application of science to technologies such as
electrification, internal combustion and chemical
engineering, began circa 1870 and drove growth
into the 1960s. The third, centered around
information technology, defines our current era.
And, as Mr. Gordon correctly notes, the payoffs
so far to the third industrial revolution, while
real, have been far smaller than those to the
second. Electrification, for example, was a much
bigger deal than the Internet.
Its an interesting thesis, and a useful
counterweight to all the gee-whiz glorification
of the latest tech. And while I dont think hes
right, the way in which hes probably wrong has
implications equally destructive of conventional
wisdom. For the case against Mr. Gordons
techno-pessimism rests largely on the assertion
that the big payoff to information technology,
which is just getting started, will come from the rise of smart machines.
If you follow these things, you know that the
field of artificial intelligence has for decades
been a frustrating underachiever, as it proved
incredibly hard for computers to do things every
human being finds easy, like understanding
ordinary speech or recognizing different objects
in a picture. Lately, however, the barriers seem
to have fallen not because weve learned to
replicate human understanding, but because
computers can now yield seemingly intelligent
results by searching for patterns in huge databases.
True, speech recognition is still imperfect;
according to the software, one irate caller
informed me that I was fall issue yet. But its
vastly better than it was just a few years ago,
and has already become a seriously useful tool.
Object recognition is a bit further behind: its
still a source of excitement that a computer
network fed images from YouTube spontaneously
learned to identify cats. But its not a large
step from there to a host of economically important applications.
So machines may soon be ready to perform many
tasks that currently require large amounts of
human labor. This will mean rapid productivity
growth and, therefore, high overall economic growth.
But and this is the crucial question who will
benefit from that growth? Unfortunately, its all
too easy to make the case that most Americans
will be left behind, because smart machines will
end up devaluing the contribution of workers,
including highly skilled workers whose skills
suddenly become redundant. The point is that
theres good reason to believe that the
conventional wisdom embodied in long-run budget
projections projections that shape almost every
aspect of current policy discussion is all wrong.
What, then, are the implications of this
alternative vision for policy? Well, Ill have to
address that topic in a future column.
------------------------------
_______________________________________________
Futurework mailing list
[email protected]
https://lists.uwaterloo.ca/mailman/listinfo/futurework
