_______________________________________________
- THE FRIDAY LETTER -
(emailed weekly,
from Gilder Publishing,
for friends and subscribers)
_______________________________________________
| http://www.gilder.com/ | Issue 373.0/February 20,
2009
SIGN-UP A FRIEND FOR FREE!
HEADLINES:
- The Week / Gilder: An Israeli Entrepreneur
- Friday Feature / Forbes: Job
Un-Creation
- Friday Blogger Bonus / Government
Spending Sprees
- Readings /
The Week / An Israeli Entrepreneur
GEORGE GILDER, Gilder Telecosm Forum (2/17/09): [EZchip CEO] Eli Fruchter is a kindly, tough, humble,
inspiring man, with sandy hair above a broad, blunt weather-beaten face. You
would not recognize him as a miracle worker. He does not make grand claims. He
is not an agile debater on a panel. He is not full of artful analogies and
elegant prose or riveting details or luminous power points. He does not have
avian or angular features or dark hair or other prototypical Israeli
characteristics. He tells his story lucidly but without embellishment in
careful Hebrew accented English. Take it or leave it.
If you believe his glib rivals with their claims of chips that will excel and
eclipse his own, he does not seem to care. He knows what the customers say,
what he has accomplished, and he seems indifferent to the hyperbole of others.
Wall Street, the journals and magazines, the tech blogs, they will learn in
time. His competitors will learn in time. In an unimpressive gray glass-clad
multistory building by a pitted road on a hill in Yokneam, far from the centers
of Israeli enterprise, with no architectural distinction or flourish, Fruchter
has performed a miracle. But Eli does not preen as a miracle worker. He is
embarrassed by prophetic language. When I informed him of my plans for this
book, describing Israeli entrepreneurship and technology as the consummation of
the Jewish science of the Twentieth Century, he balked, waving me aside.
“I am not important,” he said.
Then he asked me about Einstein.
“You are going to put me in a book with Einstein?” the entrepreneur of EZchip
asked incredulously.
“Yes,” I said, “Einstein, and Bohr, and Pauli, and Von Neumann, and Feynman.
All those guys were just preparing the way for you, Eli, providing the
theoretical foundations for network processors that can compute at the speed of
fiber optic communications, at the speed of light.”
Eli peered back at me full of skepticism.
I tried to explain.
Science finds its test in engineering. If scientific theories cannot be
incorporated in machines that work, they are a form of theology. Throughout
most of the history of science, the pioneers actually built the devices that
proved their theories. Faraday, Hertz, Michelson, all those guys described by
George Johnson in his book on the great experiments, they all proved their
mastery of their ideas by creating the apparatus that tested them and embodied
them. If you cannot build something that incorporates your idea, you cannot
fully understand it and you probably cannot build on it.
The regnant physicists today are mostly mythopoeic metaphysicians reifying
math: string theorists exploring dozens of mythical dimensions; exponents of
mythical infinite parallel universes, with anthropic principles to explain us
and our ideas as mere random happenings; nanotech evangelists who imagine a
mythical reduction of all engineering to pure physics and its replication;
cosmologists with their black holes and myriad particle types and unfathomable
dark matter and dark energy dominating the universe. These guys cannot begin
the construct anything that proves their increasingly fantastic theories.
You, Eli, take the best work of twentieth century science—quantum chemistry and
solid state physics and optical engineering and computer science and
information theory—and make it into an entirely new device: a network processor
that can apply programmable computer intelligence to millions of frames of data
and packets of information traveling at rates of a hundred billion bits a
second. That’s one hundred gigabits a second. Equivalent to 100 thousand 400
page books, with each page or so scanned and addressed and sorted, and all sent
in one second to the right destination.
When conditions on the network change, the network processor can be
reprogrammed. With as many as eight “touches” of the data per packet,
classifying the packet, looking up addresses, finding the best route—parsing,
searching, resolving, modifying, resetting the packet headers. That means
trillions of programmable operations per second. You make the most efficient
computers on the planet.
You build things that the world has never seen before. In fact, even so, you
and your team—Gil Koren, Amir Ayal, Ran Giladi and the rest—may well not fully
understand everything that is going on in your machines. No one has fully
fathomed the quantum mysteries underlying modern electronics. But I believe
that von Neumann was the paramount figure of Twentieth Century science because
he was the link between the pioneers of quantum theory and the machines that
won World War II, that prevailed in the cold war, and that enabled the
emergence of a global economy tied together and fructified by the Internet. The
entire saga is one fabric. And you are the current embodiment of this great
tradition, mainly a Jewish tradition.
Von Neumann was the man who outlined the path between the new quantum science
of materials and the new computer science of information. “You Eli are a
leading figure in the next generation of computer technology: the creation of
parallel processors made of sand that can link at fiberspeed with the new
optical communications technology.”
“But there are thousands of entrepreneurs in Israel more important than me,”
Fruchter insisted. “Thousands. You should speak to Zohar Zisapel. He and his
brother created RAD in 1981, put the first modem on a single chip and then
started five companies that emerged from RAD. Today they have 2,500 employees
in Israel and hundreds more around the world.” I looked it up. They do
signaling for high speed trains, electronic messaging to motorists seeking free
parking spaces, communications for remote surgery across the globe.
“Zohar’s an Israeli entrepreneur,” says Eli. “He laid the foundations of
Israeli technology. EZchip is still just a small company…”
I first heard Eli describe his plans for a network processor at a forum in
Atlanta called InterOp 2000. At the time, EZchip was one of at least fifty
companies pursuing the technology. Linking the network to computers around the
world, it was the most challenging target for the next generation of
microchips. A network processor has to function at the speed of a network
increasingly made of fiber optic lines. For most of the decade of the 1990s,
fiber optics—light transmitted down glass threads—grew in bandwidth and
capability at a pace at least three times as fast as the pace of advance of
electronics. Called Moore’s Law after Gordon Moore of Intel, named and
researched by Carver Mead of Caltech, this law of the pace of advance of
computing capabilities ordains that computer technology doubles in cost
effectiveness every 18 months to two years. During the first decade of the 21st
century, fiber optic technology has been advancing nearly twice as fast as
Moore’s law.
The network processor has to bridge this gap. Just as the Pentium is the
microprocessor that makes the PC work, the network processor has become the
device that makes the next generation Internet work—that does the crucial
routing and switching at network nodes on the net.
I first encountered Eli Fruchter not in person but in a series of tapes.
Intrigued by the promise of network processors, I ordered them from a major
technology conference called InterOp that was holding a network processor
forum. At InterOp, engineers have to prove that their technologies can
interoperate with other networking technologies and standards. In
communications, systems must interoperate or they are useless. Interoperation
between systems that are rapidly changing requires devices that are
programmable. In the 1990s the fastest changing technology in the world was the
network. In those days, nearly anyone who was anybody in networking showed up
at InterOp and made his interoperability pitch.
With the market tumbling and my own company in chaos, I had missed InterOp1999.
But I was interested in network processors and InterOp hosted a two day forum
on the subject. I ordered the tapes and drove around the Berkshires where I
live, listening to all the vendors of new network processor designs.
At the time, the leaders were Motorola, Intel, IBM, Trimedia (now part of
Alcatel), Cisco, Lucent, Texas Instruments, AMCC, Broadcom, and Agere. You name
your technology champion, they were investing billions of dollars apiece in
network processor projects. The largest electronics and computer companies in
the world put more than $20 billion into network processor design and development
over the last decade.
I listened to seven or eight hours of tapes, and I decided that the most
plausible, scalable design for a network processor was presented by Eli
Fruchter of EZchip. Alone among the presenters, Fruchter seemed to grasp that
network processors would have to scale faster than computer technology.
Ordinary arrays of parallel RISC (reduced instruction set computing)
microprocessors might perform the role for a couple years. But within five
years they would be obsolete. Fruchter saw that a new architecture would be
needed.
This meant moving beyond the von Neumann computer architecture that had
dominated computing since the beginning. The von Neumann model was based on the
successive step-by-step movement of data and instructions from memory to a
processor. With scores of homogeneous RISC machines requiring data and
instructions at once, the performance of the system depended on the bandwidth
to memory. It seemed to me that none of the existing network processors had
addressed this challenge in a way that would scale with the constant
acceleration of dataflows across the Internet.
Most router and switch companies, such as Cisco, Lucent, Juniper, Alcatel and
others, had contrived specialized machines. These application specific devices
could perform network processing at tremendous speeds for particular protocols
and datatypes. But these processors could not change with the changes in the
network. They could not adapt. They could not scale. Every time the network
changed, the network processing function would have to change. That, it seemed
to me, would not be a successful solution.
None the less, at InterOp, Motorola, Intel, AMCC, Agere, Bay Microsystems, and
IBM among others were presenting programmable processors. Their devices were
available in the market and were being produced in volume in workable
programmable silicon devices.
As I said at the time, Eli Fruchter had developed a leading edge device,
alright, and it met the challenge of changeability and scalability, because it
was inscribed upon the easily adaptable and programmable substrate of
PowerPoint slides.
Now I reminded Eli: “You had at least 50 competitors and no customers, and no
product, and you invested maybe a hundredth of the money that they did.
“Now, just eight years later, you have more than 50 customers, six industry
leading products, and virtually no serious competitors. All the large
players—Intel, Motorola, IBM—have essentially left the field. That is stunning.
How did you do it, Eli?”
Eli finally gave up and told his story….
Read George Gilder’s complete post, as
well comments posted by the Gilder Telecosm Forum members, by logging on with
your member password today: http://www.gildertech.com/
|
The Gilder Telecosm Forum To
learn how to join this powerful network of talented, tech-savvy investors and
thinkers online daily to debate, discuss, and decode new and emerging
technologies and share valuable and actionable investment advice, visit www.Gildertech.com today. |
Friday Feature / Job Un-Creation
STEVE FORBES, Forbes.com “Fact and
Comment” (2/20/09): For all his talk of fighting unemployment, President
Obama has hurt future job creation by signing into law the Lilly Ledbetter Fair
Pay Act. This legislation eliminates the current statute of limitations (180 or
300 days, depending on the particular state) on discrimination suits. A worker
can now sue employers for alleged pay discrimination based on gender or race 20
years after leaving a company.
No wonder personal injury lawyers are drooling. This opens the floodgates
to frivolous suits over supposed wrongs committed many years ago, even if
memories have grown cold and some of the parties are dead.
Take the Ledbetter case itself. Only after leaving Goodyear Tire & Rubber in
1998 and collecting a pension--and having received some poor job evaluations
while employed there--did Ledbetter claim that her supervisor was guilty of
gender discrimination, going back to the early 1980s.
Ledbetter initially won a big jury award, but appeals from both sides
eventually took the case to the highest court. The Supreme Court ultimately
ruled that the then existing statute of limitations actually meant what it
said. Thus the congressional legislation to throw out those limits.
Can the President please show us how this incentive for costly lawsuit abuse
aids future job creation?
Comment on this post:
http://www.forbes.com/global/2009/0302/009_fact_and_comment.html
__________________________________________
Friday Blogger Bonus / Government Spending Sprees
Steven Kates (2/20/09): A few days ago, I found the following in the latest issue of the Claremont
Review of Books. This is a quote from George Gilder as part of a symposium
on government and economic crisis. In it he wrote something that should be kept
in mind as governments go about spending on their favourite projects. He, of
course, was speaking of the United States. His point is just as valid here.
"Meanwhile, the profession upholds the phantasmagorical models of
demand-side economics. Because these models find no confirmation in reality —
as Jean-Baptiste Say proved centuries ago, demand is always and only a side
effect of real supply — established economic theories are extremely difficult
to learn and remember. You get Nobel prizes for minor and obvious insights in
economic geography. Thus the exponents of the standard model are deeply
threatened by any reality-based economics."
What Gilder is discussing is something called Say's Law, the very
cornerstone of pre-Keynesian economics. A properly educated economist before
Keynes published his General Theory in 1936 would have understood that
only goods buy goods, with money as a mere intermediary. When all was finally
said and done, each person could only buy with the value added they had
created, which had then been converted into money….
Read on:
http://business.theage.com.au/business/government-spending-spree-has-no-realworld-benefit-20090219-8clf.html
__________________________________________
Readings /
Wave of Selling Spans Globe
http://online.wsj.com/article/SB123513190945132543.html
High-def Video Over Wi-Fi
http://www.technologyreview.com/computing/22195/
Wanted: Lousy Job, Low Pay
http://money.cnn.com/2009/02/18/news/economy/lousy_job/index.htm?postversion=2009021910
Where is Your Money Going
http://www.recovery.gov/?q=content/investments
__________________________________________
ADVERTISING INFORMATION
The Friday Letter is mailed each week to 40,000-plus subscribers
and friends of Gilder Publishing, including industry leaders, financial
professionals and individual investors. For information about advertising,
contact Lauren Klopacs at lklopacs@forbes.com.
PLEASE NOTE: The appearance of an advertisement in the Friday Letter
does not indicate an endorsement for the product and/or service by George
Gilder, Gilder Publishing LLC, or the Friday Letter staff.
FEEDBACK AND PROBLEMS
For technical problems, or to send letters to the editor, please
e-mail info@gilder.com.
MAILING ADDRESS
Gilder Publishing, LLC
ATTN: Friday Letter
291A Main Street
Great Barrington, MA 01230
_______________________________________________
The Friday Letter is published weekly for subscribers and
friends of Gilder Publishing. If someone you know would enjoy it, please feel
free to forward a copy.
Gilder Publishing makes the Friday Letter available for free. To
help defray some of the costs of producing this information on a weekly basis,
we will from time to time be sending you offers from companies we think you'll
be interested in. These offers will not come more than once a week. If you do
not wish to receive this related information, please opt out of this process at
the link below and we will not share your name with companies outside of Gilder
Publishing.
To SUBSCRIBE please visit http://www.gilder.com/
To UNSUBSCRIBE please go to http://www.gilder.com/fridayletter/unsubscribe.php
Trouble subscribing or unsubscribing?
Email info@gilder.com
http://www.gilder.com/unsubscribe/specialproducts.php
To SUBSCRIBE please visit http://www.gilder.com/
To UNSUBSCRIBE please go to http://www.gilder.com/fridayletter/unsubscribe.php
Trouble subscribing or unsubscribing?
Email info@gilder.com
_______________________________________________
Copyright 2009 Gilder Publishing LLC