Dana Blankenhorn

Where the Next Boom Will Start

Think of this as Volume 11, Number 43 of A-Clue.com, the online newsletter I've written since 1997. Enjoy.

---

Every economic recovery is different. What worked before may recover but it's never the same.

That said, note that every new boom starts slowly. Growth is never uniform, not between industries or within an economic cycle. Recovery will be slow. It will take years. But its pace will accelerate and at some point we'll forget again what hard times felt like. That's when we'll be ready for another bust.

Let me start by looking at previous booms I have lived through:

• The 1970s happened in the oil patch. I was in Houston then and it was a lot like Dubai has been the last few years. One bank put up what looked from the air like a huge green dollar sign. That's how things were. After the crash, which was hard, there was finally a recovery, but it was never the same as it had been.
• The 1980s happened on Wall Street. I was a business reporter then and I watched as supposedly brilliant people took apart companies, then put them back together to reveal value. Everything else went up in sympathy. After the crash the Wall Street game changed. You still read about the "heroes" of that era, like Carl Icahn, but it's not the same. Nor will it be.
• The 1990s  happened on the Internet. I was an Internet reporter. I started warning of the crash in 1997, but the boom happened despite me. The bust took down a ton of companies, and I was out of work completely for over two years. But it did come back. Amazon and eBay came back. Other sites came back. Google barely existed at that time. But, again, this decade, online, has been mainly about consolidation and steady change, not revolution.
• The 2000s were all about housing. Whether the boom was real or artificial, the product of lax regulation or an attempt to cover up for the Iraq War's costs, is irrelevant right now. We created enough Confederate Money to paper the world, backed by houses that were only worth a fraction of what we paid for them, and it will take years to unwind that. When housing does "come back" it won't be the same. It can't be. But, in time, it will come back.

Moore's Law Gets Another 10 Years

I wrote The Blankenhorn Effect, my book on Moore's Law, over five years ago.

Then, as now, people underestimated the impact of Moore's Law, and predicted it would soon meet its demise.

In conventional terms -- circuits per square inch of silicon -- it's possible that limit was reached in 2004. But parallel processing and low-power designs have kept the improvements coming, and exponential improvements have also come in other areas -- hard drives, fiber optics, radios -- just as the book predicted.

Now it looks like chips themselves have another 10 years of Moore's Law improvements in them, thanks in part to design changes which replace the silicon semiconductor with metal conductors, and an insulator, hafnium oxide, between the conducting lines.

This let Intel put down lines 45 nanometers apart, against the 65 nanometer distances in other chips.

Rice Science Thursday: Palem's Project

The news release says a joint venture between Rice University and Nanyang Technical University of Singapore will launch a joint venture in very low power embedded chips, under Rice's Krishna Palem (right).

In fact, we've got something much bigger, something which demonstrates just how powerful Rice is becoming.

Palem, who co-authored a report on probabilistic CMOS chips last year, which will be at the heart of the effort, is actually a very new recruit to Rice.

He had been, until this semester, at Georgia Tech, where he had founded its Center for Research on Embedded Systems and Technology.  He's now a joint fellow of both Rice and CalTech, where he is listed as a Most Distinguished Scholar.

 

What the Financial Press Never Gets

Everything at CNBC, and in the rest of the financial press, is focused on personalities and numbers.

Never operations.

Yet operations are the real key to business success or failure. Once your business is built, it's the day-to-day operation of the business which determines your success. Not your marketing. Not your financial wizardry. Not the personality of your CEO. Not the food at your press events. Not how sexy your people are.

Operations. Nothing else.

This is easy to see by the performance this decade of Hewlett-Packard and Dell.

Quantum Dot Reality

In my online novel The Duke of Oil, I'm taking some liberties with the science of quantum dots. (Illustration from NASA.)

It's true that Professor Michael Wong and his team at Rice have made an important breakthrough in the production of quantum dots, but it's not true that we're about to see a flood of cheap, highly efficient solar panels, powered by dots, on the market any time soon.

One reason for this is that quantum dots, like Buckytubes (called carbon nanotubes by non-Rice people) are useful in many other applications, besides solar panels.

In fact, solar panels are a sort of low-end use, dependent on mass production of a particular type of dot, made of cadmium selenide. In his research at Rice, Wong made two breakthroughs in the recipe for these dots.

First, he used a cheaper, safer catalyst,  cetyltrimethylammonium bromide (it's sometimes used in high-grade shampoos), rather than the caustic alkylphosphonic acids (used in flame retardants) previously used. Second, 90% of the dots created by his method were tetrapods, with four molecular "legs" reaching into the solution. These happen to be the most efficient types of dots for use in solar panels.

While some quantum dots have photovoltaic properties, most scientific work with them over the last decade has been in medicine, where they can be used as chemical sensors in medical tests.

Computers Haven't Changed Much

Since the dawn of the PC age it's amazing how little computing has changed. (Picture from Hal Licino. Thanks to Grungy.)

We work with bigger files. We can access big files remotely. We do TV and radio and film. But the basic computing paradigm remains that of the Dynabook, first conceived by Alan Kay in the 1970s. The rest is bells and whistles and details.

In terms of my own personal productivity, the best portable machine I ever had was my TRS-80 Model 100. Wikipedia says this is the last computer for which Bill Gates personally wrote firmware, and he did good. The TRS-80 was light, just four pounds, it ran for hours on batteries, it stored up to 16 pages of notes (more with a 32K memory expansion), and its 300 baud modem let you move those files fairly quickly (given their size).

Moore Must Pay The Price

My recent piece called The Unpaid Price of Moore drew a link from my old friends at CMP Media, specifically EE Times.

Rick Martin's piece focused on dwindling supplies of rare metals used in PC manufacture, metals for which there are no substitutes.

It's a fair point, one I agree with.

But hasn't this industry yet figured out recycling? All the rare metals we need to make tomorrow's PCs are sitting, right now, in landfills, inside old PCs.

It's true that harvesting these metals is difficult. But that doesn't mean it's impossible. And it doesn't mean we can't re-design future systems to make recycling materials easier.

This is just one area in which the supposedly clean electronics industry is, in fact, anything but. As I pointed out in my original piece, the real reason why the chip-making business fled the U.S. wasn't our high labor costs, but the environmental damage done by chip plants, damage the vendors simply refused to fix.

The excuse was that if they paid the price, then competitors in other markets wouldn't, and so gain an advantage. This was then the reason why those same companies moved their facilities to those same countries.

Intel, a company I have the greatest respect for, is the leader of this awful trend. They continue to build new plants -- in China.

Intel Does Its Bit In The War Against Oil

Thanks to Moore's Second Law, which holds that as chips get exponentially more complex they get exponentially more expensive to make, chip companies like Intel have become huge, of necessity. (Moore's Second has also been steadily shrinking the pile of chip companies, with AMD the latest victim.)

Huge companies are hard to turn around. Intel announced back in 2004 it would concentrate on low-power designs. That effort is only now starting to bear fruit.

The fruit is in the form of what's called a Quad Core Xeon chip. As the name implies, this is four chips on the same sliver of silicon, a design advance pioneered by rival AMD. The news, for those who follow The War Against Oil, is that these are server chips, meant to go into powerful computers  that are on all the time, and they draw just 50 watts each.

How Michael Will Get His Groove Back

On the corporate Richter scale, where 1 represents the trouble at Microsoft and 10 represents the trouble at Ford, the problems at Dell Computer represent about a 5.

Mark Hurd of H-P and other competitors have begun copying its production techniques. The battery recall cost billions. Michael Dell's hand-picked replacement apparently got lazy and cooked the books, moving sales so jagged numbers would look smooth.

Michael Dell's personal reputation has been hurt. People assume Dell needs to play a new game, and question whether Dell has a new game in him.

Well he does, and he will. He has already made a good start, pushing incentives down the chain of command, cutting off the stock option fire hose, reducing direct reports, and trying to shoot straight with Wall Street and regulators.

What happens next?

The Unpaid Price of Moore

In all the excitement over Intel and IBM breaking through the 45 nm barrier, important facts are being ignored.

First, the breakthrough does mean the end of Moore's Law, in the sense Gordon Moore originally expressed it, and wanted it to be understood. Getting to 45 nanometers between circuit lines required the creation of a new type of material, dubbed High K.

The reason for the cute acronym? The new substance is based on hafnium,  an element similar to zirconium that is  also used to make control rods for nuclear reactors, because it absorbs neutrons so readily. It's not simple stuff and it's not cheap. Remember all those old commercials talking about how computer chips were made of sand? No more.