The last time I saw Peter Drucker, he was keynoting a Forbes conference in Seattle for CEOs. In the auditorium at the International Trade Center next to the bay, they had wheeled out the great man to the middle of the stage in a great fluffy easy chair. Close to 90 years old—at the end of the previous century gazing toward the next—he was the numinous name and Delphic presence at the conference. Everyone leaned forward to hear what he had to say.
Then a gasp shook the rows of CEOs. The conference management stood there stricken, unable to move: For the Love of Malcolm’s motorcycle . What is this? The CEOs sat popeyed.
The hoary sage’s balding pate flopped back in the chair as if he had fallen asleep, or worse.
Perhaps Forbes had erred in staking a major conference on an aging guru seemingly well over the hill and in parlous health.
Then his entire body fell forward. I was ready to run up to catch him if he should tumble toward the crowd. But he somehow caught himself. His eyes opened, and he looked out intently at the throng of CEOs. Everyone sighed with relief. He was awake. He had their attention.
Drucker growled: “I have just one thing to tell you today. Just one thing.”
“Wow,” I said to myself, “it better be good.”
”No one,” he continued, ”but no one in your company, knows less about your business than your See-Eff-Oh.”
This was the era of the heroic chief financial officer. Scott Sullivan of Worldcom, Andy Fastow of Enron. Clever, inventive folk like that. You remember them. Across the country, CFOs were in the saddle. CEOs would not move without consulting them.
What could Drucker have meant?
He was stating law number one of the telecosm.
- Knowledge is about the past. Entrepreneurship is about the future. CFOs deal with past numbers. By the time they get them all parsed and pinned down, the numbers are often wrong. In effect, CFOs are trying to steer companies by peering into the rearview mirror. Past numbers do not have anything much to do with future numbers. As Ken Fisher puts it in his new book, The Only Three Questions That Count, Stock prices [and by extension other business numbers] are not serially correlated.
Moreover, CFOs tend to focus on internal problems. But most internal problems cannot be solved internally. Determining business outcomes are decisions made by customers and investors and both are outside the company and not directly managed by the company. Their views can change in an instant, casting all the existing numbers into oblivion. To reach customers and investors takes outside vision and leadership, not internal problem solving.
Tech companies should not try to solve problems. Solving problems sounds good, but it is a loser. You end up feeding your failures, starving your strengths and achieving costly mediocrity. Don’t solve problems—that’s the CFO’s forte and pitfall. Pursue opportunities.
One of our erstwhile companies that has been solving problems for the last five years is JDSU . It has been cutting back, and pruning, and shrinking costs. But as Charlie Burger (an analyst with Gilder Technlogy Report) says, you cannot prevail by shrinking. Moore’s law does not apply to companies.
- Moore’s law ordains that for a given level of performance, prices routinely drop 50% every 18 months . As Caltech’s Carver Mead explained to Moore when the Caltech professor named Moore’s law, smaller transistors run cooler and faster and cheaper. This law rules the world of technology because of what economists call price elasticity of demand. In the sometimes slow moving world of conventional business, price collapses may crash your margins and profits. But in the tech world, users and uses multiply when prices drop. You get positive elasticity, which means that revenues rise when prices decline.
All our semiconductor companies benefit from Moore’s law: Intel, Advanced Micro Devices, Texas Instruments, Qualcomm, Broadcom and our copper-plating gear maker Semitool, our board level digital power innovator Power-One, and our newly listed probe test pioneer FormFactor .
Moore’s law, though, is reaching a crisis as power levels drop below a volt toward irreducible electron volts and power sources multiply, with one volt, three volt, 5.5 volt and 11 volt feeds often necessary. Meanwhile mixed signal devices proliferate with both analog and digital components on a single chip. Digitizing and consolidating analog power sources that require a separate wire for each voltage, Power-One will ultimately be a major beneficiary of this trend.
- Metcalfe’s law asserts that the value of a network rises by the square of the number of compatibly connected users. As the namer of Metcalfe’s law, I like to add a power factor: Metcalfe’s law works in part because the linked devices on the edge are increasing their power and versatility at the pace of Moore’s law. After the millennial telecom crash, it has become fashionable to point out the various flaws in Metcalfe’s law.
Andrew Odlyzko of the University of Minnesota has done a recent critique. Metcalfe’s law is not literally true all the time, but it offers a rough and useful guide when contemplating the value of companies such as YouTube. Don’t forget it.
Our Broadwing now captured by Level 3 benefited from Metcalfe’s law. So will all our networking plays, such as EZchip and NetLogic .
- Dumb networks will prevail over smart networks. The future is all-fiber networks that do nothing but transmit bits. Intelligence belongs at the edges and endpoints.
This is our life after television paradigm. It separates content from conduit. If you have the best conduit, you will want everyone’s content on it. You won’t want to restrict it to your own content. On the other hand, if you have the best content, you will want it on everyone’s conduit. You won’t want to keep it on your own network. Players that try to combine content and conduit will eventually split apart and often bleed financially in the process (e.g., AOL-Time Warner ).
- Software hardens at the core of a network—hardens into glass—pure fiber. Hardware softens at the edge into programmable forms. Consider the change from hardwired TVs and telephones into programmable PCs and handheld teleputers. On the network, this law favors Corning, Finisar, and PMC-Sierra/Passave supplying key optical components (also optical inventor Essex—in the process of being purchased by Northrop Grumman ). On the edge, the law benefits programmable logic vendors Altera and Xilinx and programmable node makers EZchip and NetLogic.
- The edge of the network is analog, because that’s where humans live, in an analog world. But the analog world is one of shortages, because there is a shortage of great analog engineers in the U.S. and throughout the world. Therefore, great analog design will often produce great profits.
Our analog stars are National Semiconductor, Anadigics, Synaptics, and power interface innovator Power-One.
- Law of Abundance. Far-seeing entrepreneurs waste what is abundant in order to save what is scarce. Today, processing power is abundant. Bandwidth is becoming abundant. Electrical power, on the other hand, is becoming scarce. So invest in chips and computer architectures designed to save on power.
- Law of Scarcity. Speed of light is the scarcity that governs networks, whether on the surfaces of chips where signals move nine inches a nanosecond through miles of minuscule copper wires or across continents and oceans where latencies reach 60 milliseconds. Span of life is the scarcity that will govern human interactions and consumer businesses. Consumers hate to have their time wasted. That’s why broadcast TV is a failing model—it wastes the consumer’s time.
- Schmidt’s law. When the network becomes as fast as the backplane of your computer, the computer hollows out, its components dispersing across the Web, its value migrating to search and sort functions. This law was put forth by Sun’s Eric Schmidt a decade and a half ago. Schmidt, of course, is now the CEO of Google, where search and sort has paid off rather nicely.
- Gordon Bell’s corollary to Moore’s law. Every 10 years there is a hundredfold drop in the cost of computing, leading to a new paradigm in computing. Google-like server farms are the new computing paradigm. But Ray Kurzweil’s codicil to Bell’s corollary is that everything is accelerating. Within five years, a newer and more radical computer architecture will take the place of the current datacenter model.
These are 10 laws for understanding the technology. But the key law is Drucker’s Don’t Solve Problems, Pursue Opportunities.
How do you identify opportunities? Drucker has the answer to that also. By looking for upside surprises.
This is the final entropy law from the fertile mind and mathematics of Claude Shannon of MIT and Bell Labs, who defined information as unexpected bits. (Predictable bits convey no information content, no entropy.) Information entropy is measured by its surprisal.
My summation of this law is: High entropy messages (full of surprise) require a low entropy (no-surprises) carrier. Only if the carrier itself is predictable can the information be distinguished from the noise at the other end. Thus the key insight of the telecosm is that in an information age information and value will migrate to the perfect sine waves of the electromagnetic spectrum.
I believe that this is a more general law than Shannon perceived. The heart of capitalism is creativity. Creativity, as Albert Hirshmann of Princeton once wrote, always comes as a surprise to us. If it didnt we would not need it. Socialism would work. But the upside surprises of creativity require a low entropy environment of predictable property rights, taxes and other business laws ultimately based on trust in a moral order. All these conditions are essential to an entrepreneurial economy.
Upside surprise is another way of saying profit. Profit is entropy—unexpected returns. Predictable returns are all discounted by the market and diminish to the interest rate. What we are seeking on the GTR is unexpected returns, the upside surprises of creativity and profit.
George Gilder is a Senior Fellow at the Discovery Institute.