REFLECTIONS AND REVIEWS OF PAST PREDICTIONS
By ROBERT LUCKY
Bell Labs., New Providence, NJ 07974 USA
In October 1961, I drove up the entrance drive of the sprawling Bell Labs
complex in Murray Hill, NJ, to begin my career as an electrical
engineer. I can still remember how frightened I was, and at the same
time how thrilled. It was an awesome place I was entering, the Mecca of
my world, peopled by the legends of my imaginationVShannon, Pierce,
Shockley, Bode, Rice, Tukey, and Darlington among many others.
I thought this famed institution would be forever, and of all the
revolutionary events that have happened in these many years since that first
day at Bell Labs -- events that I shall speak of presently -- that this building now
bears a different company name and that the Bell System no longer exists are at
the top of my list of happenings that I could not have imagined.
In 1962, shortly after I began work at Bell Labs, the PROCEEDINGS OF THE
IEEE featured a series of future predictions authored by IEEE Fellows. I do
not remember if I had read the predictions when they had been first
published, but 25 years later I
reviewed those predictions in one
of the regular columns I was writing
for the IEEE SPECTRUM magazine.
Those thoughtful predictions from
our most famous engineers had by
then become the perfect subjects for
humor and derision. But in my
opinion the most serious failing was
the authors' failure to recognize and
appreciate the revolution in electronics
that would come about
because of the integrated circuit.
My favorite prediction, written as
if an engineer in the year 2012 was
looking back at the field in 1962, was
as follows:
After a competitive race in
the 1960s to produce the smallest
units, reason had prevailed.
While components were small
by earlier standards, the ultimate
sizes were such that costs
were reasonable and servicing
practicable. For example, whole
receivers were the size of pound
candy boxes, rather than cigarette
packs.
Jack Kilby had made his first
integrated circuit in 1958, so it had
existed when the predictions had been
published in 1962, but it was not until
1965 that Gordon Moore had made his
now famous prophecy about the exponential
progress that would occur in the
density of integrated circuitry. Moore's
prediction was the one prediction that
withstood the test of time, and the
prediction that drove our profession
into relentless acceleration.
It was amazing to me that our most
famous engineers in 1962 had had
such an impoverished idea of future
happenings. But in truth I would have
done no better, and probably even
worse. Of course, I was a callow,
inexperienced engineer then. But now
I have no such excuse, and even so
there is a high degree of probability
that any predictions I would make
now would be subject to equal derision
by engineers in the future.
I. THE WONDER AND
CURSE OF EXPONENTIAL
CHANGE
For some years now I have been
intrigued by the idea of exponential
progress. Moore's law has been a
quantitative measure of this in the
fabrication of semiconductors, but
perhaps it extends to all of technology,
and the whole concept of exponentials
is somehow incompatible
with the way we reason about the
future. Exponentials imply breaking
points and future shocks -- a murky,
unknown future driven by forces
seemingly out of our control. So we
revert to our usual behavior of linearly
extrapolating the present, and
this works for awhile -- and then it
does not.
When I look back on my frame of
reference in 1962, I realize that I had
no concept of the inevitability of
constant change, let alone exponential
change. It was if the world was
static, frozen in time at my graduation.
I remember how I had saved all
of my college texts in the basement,
sure that I would be referring to them
for the rest of my career. Alas, I do not
think I ever looked at them even once
afterwards, and I do not even know
whatever happened to those decaying
boxes of books rotting in the basement
of my parent's house.
Why was I unprepared for change?
Was it my fault or the fault of the
education system? There was a kind of
finality associated in my mind about
what I learned in college. I felt a
timelessness and constancy about
mathematics, and even in the technology
courses I had the idea that if I
learned the material being taught, I
would be able to apply it for the rest of
my career. No one ever told me -- or if
they did I did not listen -- that all of
what I was learning would soon
become obsolete and it would be
necessary to, in effect, be a college
student the rest of my career. But who
can forget the feeling of graduation,
throwing back the tassel on the funny
hat and feeling the exhilarating freedom
of being finished with formal
education?
II. HEATHKIT AND
SHANNON
In 1962 I was attracted and influenced
by two almost opposing poles --
Heathkit and Shannon. Heathkit
both symbolized and embodied the
physical world of electronics. I can
still remember the thrill, and even the
smell, of opening a new box with a kit
from that company in Benton Harbor,
MI. I have since on occasion looked
around my present accumulation of
possessions and asked myself which I
have had the longest? Perhaps it is the
Heathkit oscilloscope that I built in
high school about 60 years ago. I still
have it, and I think it still works. Alas,
there is no longer any use for it.
Building a Heathkit was mindless
and you did not need to be an
engineer to do it. It was like paint-by-the-numbers. Nonetheless, there
was a feeling of pride and accomplishment
when you turned on the
final product and it worked. I thought
the world of electronics would always
be like that Heathkit -- different assemblages
of tubes, transistors, capacitors,
and resistors. There were a
myriad of ways these fundamental
components could be interconnected,
so the future was rich in one sense,
and yet impoverished in another sense
that I had yet to realize.
Some years later when I wrote a
column in the IEEE SPECTRUM magazine
about the death of Heathkit I got
more mail about that essay than any I
had written before or would write
afterwards. For older engineers it was
as if their childhood had been stolen
from them. The smell of soldering
resin was gone forever. The physical
world had receded and had been
replaced by an intangible world of
software and of impenetrable and
unfathomable chips. The intricate,
paint-by-the-numbers wiring of a
Heathkit had become one big featureless,
colorless blob.
If I had imagined the future then,
and I doubt that I really did, it would
have been a world increasingly filled
by a stream of incrementally improving
Heathkit-like appliances -- better
televisions, high-fidelity amplifiers,
and so forth. And perhaps in some
sense this is what happened after all.
Fifty years later television sets still
have the same function. They got
bigger, as we anticipated, and like the
popular science magazines envisioned,
we got flat panel displays
that you could hang on the wall. But
they are still recognizable as TVs.
The
revolutionary changes lay elsewhere.
Shannon's theory of information
was the other pole influential in my
life at that time. While Heathkit
symbolized the physical world of
electronics, Shannon symbolized the
virtual, theoretical world of mathematical
analysis. His paper on information
theory published in the (now
defunct) Bell System Technical Journal
in 1948 was to me an object of beauty
comparable to a great painting or
symphony. Like Moore's law, it too
has withstood the test of time, and
like great works of art and music, its
beauty still shines forth from those
yellowed pages of a journal filled
otherwise with forgotten trivia.
Shannon's paper exemplified the
idea that it was possible to sit in an
office and dream on paper with
mathematics about things that would
change the world. It did not need to
be changed with a soldering iron in a
laboratory; it could be done with
imagination, pencil, and paper. It
inspired me, but at the same time
intimidated me. This level of work
was denied to us mortals, and as
subsequent years went by, I saw that
there were few jobs in our profession
that were supported by pure mathematical
thought.
In college, those two realms -- the
physical and the virtual -- were interleaved.
In one course, I would learn
about rotating machinery, and in
another, differential equations. Perhaps
the engineering profession at that time
was ecumenical in this regard, but as
the decades went by we were cleaved
into distinct disciplines. We had engineers
who were primarily physicists,
who worried about such things as
semiconductors and photonics, and
we had systems engineers whose primary
tools were mathematical analysis
and computer simulation. Another
category grew from almost nonexistence
in 1962 to a dominant category
today --that of software and computer
science. While there is some passage
through these boundaries, almost akin
to quantum tunneling, they have become
ever more distinct in education
and practice.
III. THE WORLD IN 1962
There were two big projects in Bell
Labs in 1962 -- the Picturephone and
the millimeter waveguide. It seems
incredible to me now in retrospect
that we were all so certain of this
vision of the future. Of course, the
network and its future at that time
were controlled by the Bell System, so
AT&T and Bell Labs had the liberty to
plan the long-term evolution -- and
this was it. But if the AT&T planners'
vision was flawed, they had a lot of
company. Just about every futurist
and science fiction writer believed
that after voice communication, the
next frontier was video. After all,
what else was there?
The Bell System demonstrated the
Picturephone with considerable fanfare
in New York at the 1964 World's
Fair, and Arthur C Clark, having been
enchanted by the Picturephone in a
visit to Bell Labs, featured it in the
famous movie, 2001. But by 1982 the
Picturephone on my desk at Bell Labs
lay unused; there was no one left to
call.
To support the 6-MHz analog
signal required by the Picturephone
a broadband transmission channel
was required. The plan was for the
bandwidth to be provided by hollow
pipes of about 5-cm diameter, with
millimeter wavelength signals being
propagated in a circular mode whose
electric field went to zero at the inside
edge of the pipe in order to minimize
transmission loss.
A lot of excellent engineering
went into the project, but it was never
deployed. The market failure of Picturephone
removed any urgency,
while the dramatic progress in optical
fiber transparency rendered the millimeter
waveguide obsolete before it
was even fully born.
In retrospect, these projects are
seen as missteps -- the Picturephone
because of the difficulty in predicting
market acceptance, and the waveguide
because of technological upheaval.
However, when I consider the
technological landscape in 1962 it
looks now like a time of great fertility.
We did not know it then, but we were
on the cusp of so much dramatic
potential. Consider: the integrated
circuit was just beginning its evolution,
the laser had recently been
invented, the digital computer was
just coming into being, and the first
modems were just being marketed.
Moreover, in fields like communication
we were far from the known
limits of optimization. So much was
just sitting there waiting for us.
IV. FAST FORWARD
50 YEARS
Sometime in the mid-1970s I was
visiting my parent's home, and I
remarked on the ancient Bakelite,
rotary-dial telephone that had occupied
a nook in the hallway since I was
a child.
"Why don't you get a new phone,
Dad?" I asked. "They're now relatively
inexpensive."
"What for?" he replied. "They do
the same thing as this one does."
That shut me up, and I knew
better than to pursue the conversation.
I lamented the fact that he was
mostly right -- a phone was still a
phone. But in 2011, as I write this
essay, my wireless smartphone lies
sleeping in front of my laptop computer.
I think my father would have
agreed that it no longer resembles his
old telephone. To me, this smartphone
exemplifies all of the unexpected
and dramatic progress that has
occurred over these recent decades.
My father kept his old phone for
more than 60 years. Possibly it still
exists somewhere even now. But the
cell phone represents an entirely
different understanding of technology
and the pace of change. In the United
States alone, about a half million cell
phones are thrown away every day. It
has become the most ubiquitous
device in the history of the planet,
and in a number of countries there are
more cell phones than people.
If I were to open up this phone
(which I am not about to do!), I would
first observe how little is really inside
the small outer case. The incredible
function it empowers is seemingly
built upon almost nothing. All of the
processing power is handled by a few
integrated circuit chips containing
millions of transistors. Much of the
space is taken up by a lithium battery.
There is a brilliant LCD display that is
so detailed that I cannot make out
individual pixels, and an internal
camera with an optical sensor CMOS
array chip with five million distinct
pixels. None of this was known, or
even considered remotely feasible in
1962.
But this engine under the hood is
only the beginning of the story.
There is a vast virtual and physical
infrastructure beyond in support of
this little phone. In the phone itself
there are very sophisticated processing
algorithms to implement orthogonal
frequency division multiplexing
(OFDM). These adaptive algorithms
are the culmination of decades of
research into communication theory,
and they approach closely the theoretical
limits that we now know for
the wireless channels they inhabit.
The phone has speech recognition
capability, again only made feasible by
voluminous, cheap processing and
much research on speech recognition.
It does not even have to be trained,
and it seems to do a good job.
A small chip in the phone implements
GPS, and in the skies above more
than two dozen satellites circle Earth,
streaming down time and position. The
clock in the little cell phone is corrected
to extreme accuracy by rather
complicated algorithms. Location and
altitude are determined through a
number of sophisticated calculations
even correcting for general relativity.
Voice conversations are digitized
and organized into packets for wireless
transmission to any of the ubiquitous
base station receivers that now
blanket Earth. The packets are constructed
and transmitted according to
the protocols of an entirely new kind
of digital network -- the Internet.
From the base stations the bits flow
over optical fibers carrying terabits
per second of capacity.
Many of the packets are encrypted
using public key cryptography dependent
upon the intrinsic difficulty of
factoring large numbers, while pictures,
music, and videos are compressed
using clever algorithms
conceived by information theorists in
these last few decades.
Using the programmable keypad
on the phone's display, the world's
information can be searched and
accessed. Friends can be found and
connected through social networks.
Products can be bought and sold. And
there are so many other things that we
have become accustomed to expecting
from that little phone that the magic
has somehow become ordinary.
And, by the way, you can also
make voice calls using a keypad for
dialing. Perhaps my father would say
that it is still a phone after all.
Just thinking and writing about
all the technology exemplified by
this little phone fills me with pride
for our profession. Most of these
achievements were created by people
that I have known through my career
these last 50 years. I take a vicarious
pride in our joint accomplishments. I
never expected it, but we engineers
did indeed change the world.
V. LOOKING AHEAD
Earlier I observed how technology in
1962 was fertile with potential. Could
we say the same thing about technology
today? Frankly, I do not sense that
same potential, but perhaps such
fertility is only evident in retrospect.
Moreover, in some fields -- perhaps in
conventional electronic circuitry and
communications -- we are approaching
known limits of technology. The
recent decades have been characterized
by the explosion of information
technology, but perhaps diminishing
returns have set in and the torch of
progress will pass to another field.
Conventional wisdom today believes
that new field will be biology.
Nonetheless, I do believe in
exponential progress as a rule driven
by economics and, possibly, simply
by the increasing number of engineers
and scientists in the world. In
many cases, the progress is one of
functionality itself, regardless of how
that functionality is achieved, rather
than necessarily being achieved
through increased transistor density
according to Moore's law. For example,
today much of the perceived
functionality of the smartphone has
been through social and business
invention. Companies like Google,
Facebook, eBay, and many others
would be in that category, in spite
of the fact that they are built upon
the underlying technology platform.
Even the Internet itself, the browser,
and the Web seem more due to social
invention than to the technology on
which they reside.
In order to ride the curve of
exponentially increasing functionality
as we use up the capabilities of
current technology, we will have to
open up new fields of exploration, just
as optics opened up new capabilities
for the engineers of 1961. I really have
no idea what these new fields will be,
but to give the engineers of the future
something to laugh about, I will
suggest two possibilities. The first is
based mostly on wishful thinking, and
it is significant breakthroughs in
batteries and energy storage -- both
in the small for portable devices and
in the large for the electrical grid. The
second possibility is based on a hope
for something magical, and it is that
we find some fantastic way to use
quantum physics in processing and
communication. I do not particularly
mean quantum computing, which is
now going through the traditional
hype cycle, but something else. Exactly
what this is, I could not guess,
but whatever it is, it will be seen as
magic.
Following some variety of the
second law of thermodynamics, the
world gets continually more complicated
as time passes. A young engineer
starting his or her career today
faces a world vastly more complex
than I did in 1961. At first look that
would seem to make great achievements
much more difficult to attain,
but perhaps overwhelming complexity
also portends rich unrealized potential.
In any event, I feel both sorry
for and jealous of those young engineers.
We took the low-hanging
fruit. I have no idea what is growing
further up the tree.