A Look Inside Four Decades Of Breakthroughs At IBM Research
American industry has a rich heritage of top-notch corporate labs. Bell Labs created not only the transistor, but also other fundamental breakthroughs, such as the laser and information theory. PARC, developed much of the technology we associate with modern computers, such as the mouse and the graphical user interface.
Both labs have attained mythical status and rightly so. Yet IBM Research has been no less important, developing early breakthroughs such as the first computer language and the relational database, garnering 5 Nobel Prizes along the way. And unlike Bell Labs and PARC, it’s still going strong.
Today, it employs thousands of scientists in 12 labs across six continents and continues to make breakthroughs in areas such as cognitive computing, quantum computing and neuromorphic chips. To get a better sense of what makes IBM Research tick, I talked to Bernie Meyerson, IBM’s Chief Innovation Officer about his 35 years there.
The World In Fractals
When Meyerson was still a young researcher at IBM, he received a call from Benoit Mandelbrot, who was already a legend at the time. A thousand ideas rushed through his head, wondering how his still nascent work on Silicon-Germanium chips could have entered the great man’s orbit. As he picked up the phone he heard Mandelbrot’s accented voice, “Meyerson, I heard that you’re good with cars. Mine just broke down.”
The source of Mandelbrot’s legendary status can be traced to a visit he made to his uncle Szolem’s home in Paris in 1951. Asking for something to read on the train home, his uncle, a prominent mathematician himself, fished something out of the wastebasket and said, “Take this reprint. That’s the kind of silly stuff only you can like.”
The paper, about an obscure phenomenon known as the Zipf distribution, a strange scaling pattern in language, left Mandelbrot transfixed. When he arrived at IBM, he would find similar patterns in communication lines, coastlines, floodplains and even financial markets. Mandelbrot eventually generalized his findings to create the field of fractal geometry.
One consequence of Mandelbrot’s work is that the world is far more complex and turbulent than most assume, which is why he is also considered a prominent figure in chaos theory. It also led him, way back in 1964, to warn that the wizards of Wall Street were underestimating risk in financial markets. He was mostly ignored until 2008, two years before his death.
The “Force Of Nature” That Made Today’s Computers Possible
In his almost four decades at IBM Research, there is probably no one who made an impression on Meyerson like Bob Dennard, who he called “a force of nature.” “Bob’s focus at IBM,” Meyerson says, “was simply to do amazing things that would make the world a better place. He’s one of those rare, wonderful individuals who also happens to be remarkably bright.”
One of those amazing things came to Dennard as an epiphany one night while sitting in his living room. At the time, all computer memory was magnetic, which was neither flexible nor very dense. What Dennard realized was that you could store information with just one transistor and one capacitor representing each bit, an amazing revelation.
The result was DRAM memory, which runs on nearly every computing device today. It’s what allows you to pull applications and data from your hard drive and work with them in real time. Without it, you would have to wait for your computer to boot up each time you wanted to use stored information, which would make work as we know it today impossible.
But Meyerson thinks an even greater achievement was the laws he developed that explained how microchips must be scaled down to keep power density constant. It was Gordon Moore who showed that computing power would double every 18 months, yet it was Bob Dennard who wrote the recipe for how to modify each generation of transistors to achieve it.
Enabling Innovation At Atomic Scale
When Gerd Binnig and Heinrich Rohrer received the call that informed them that they would be receiving the 1986 Nobel Prize in Physics, they ran excitedly out of the building where a crowd of reporters awaited them. You can only imagine the journalists’ surprise when the two kept running. As it happened, the call had made them late for their soccer game.
It was that same iconoclastic spirit that led them to their achievement, the scanning tunneling microscope. As research scientists, they were constantly frustrated by their lack of ability to study atomic surfaces and so they created a tool to do just that. Using a quantum effect called “tunneling”, they found that they could create a small electric charge using a microscopic metal tip and, by moving the tip around a bit, they could scan an atomic surface.
Soon, it became clear that their creation could not only scan atomic surfaces, but actually move individual atoms around and arrange them. Later researchers at IBM’s Almaden Research Center used Binnig and Rohrer’s creation to make the world’s smallest movie, with a cast of just 65 carbon monoxide molecules.
Today, the scanning tunneling microscope has become essential for a wide variety of fields, such as electrochemistry, molecular biology and especially nanotechnology, which the National Science Foundation recently estimated to result in products with over a trillion dollars in revenue.
High Temperature Superconductivity
If you’ve ever been in an MRI machine or travelled in a Maglev train, you experienced the amazing things that a superconducting magnet can do. The same types of magnets are being used to contain fusion reactions at the International Thermonuclear Experimental Reactor, which many believe holds the key to a clean energy future.
Yet for a long time, such applications were impractical because superconductivity was only possible at temperatures near absolute zero. That all changed when Georg Bednorz and Alex Müller discovered high temperature superconductors made out of ceramics. That’s what made things like Maglev trains and MRI machines possible using liquid nitrogen, which is much cheaper and more abundant than the liquid helium that’s required to get close to absolute zero.
To make that happen, Bednorz and Müller had to try out thousands of different materials before they found one that exhibited the properties they needed. It wasn’t only a feat of genius, but one of dogged pursuit. The two persevered for fifteen years of seemingly thankless work, keeping at it long after most people would have quit and moved on to something else.
Years later, after they had won a Nobel Prize and become world famous, Meyerson asked them, “What kept you going?” Müller explained that when he showed his early results to a prominent theorist, the man told him that trying to make a superconductor out of ceramics—traditionally considered an insulator—was the dumbest idea he ever heard of. That’s when Müller decided he would not quit until he achieved what he set out to do.
The IBM Way
When I asked Meyerson how he accounted for IBM’s consistent success over such a long period of time, something that no other corporate lab has achieved, he had quite a bit to say. He told me, “First, we have had the incredible benefit of a long term view of the world held by our founder, T. J. Watson Sr., that has set the tone for all the generations of management to come
“We also have a very flat organization, with very little management stacking. Ideas have the ability to percolate both laterally and vertically with the same ease, and it is that free flow of ideas that has often led to the most striking outcomes due to collaboration across what in other organizations would have been nearly insurmountable barriers.”
Finally, he pointed to IBM’s collaborative culture. “I have met many scientists who simply were in their fields for their own glory,” he says “and they do not do well and frankly are ill tolerated in our culture, as they fail in the long term given that you can always do better with more pairs of brilliant eyes on the problem being tackled.”
To be sure, IBM has never suffered from a lack of competition. In the past, Bell Labs and PARC rivaled its achievements, just as today it vies for leadership in key areas with other top players, such as Google and Microsoft in quantum computing and artificial intelligence and Qualcomm in neuromorphic chips. Still, it’s hard to think of any company that is competing for leadership in as many areas of advanced technology as IBM is.
And that speaks to IBM’s unique DNA. Its research division was not set up merely to chase market opportunities. It was, in fact, founded by Thomas Watson in the depths of the Great Depression when few opportunities existed. Instead, he believed that when top notch minds are given the opportunity to make breakthrough discoveries, new markets will be created.
– Greg