Richard Feynman and Bill Gates ---An imaginary encounter
By Hasan Zillur Rahim
Millions of kids around the world dream of becoming the next Bill Gates, the whiz kid who made good by anticipating the digital future and creating one of the world's most successful companies around it. That he has acquired a fortune in the process is almost incidental; the Microsoft chairman seems to have transcended the software industry to become a cultural icon that young minds find irresistible.
But who is a hero to Bill Gates? Who does he look up to?
In a syndicated column that he writes for The New York Times Special Features, Gates has identified individuals he respects and admires, "individuals who achieve something inspirational or who possess extraordinary character." Of these, one name comes up more often than others: the late great Nobel prize-winning physicist Richard Feynman.
Gates had planned on meeting with Feynman in 1988 but didn't get a chance. Feynman died of cancer in February of that year. "It’s an opportunity I’m sorry I missed," wrote Gates in the Times in 1995. "His book, Surely You’re Joking, Mr. Feynman, is a favorite of mine."
Feynman was a hero because, as Gates put it, "he was incredibly inspirational. He was an independent thinker and gifted teacher who pushed himself to understand new things. I have enjoyed everything I’ve read about him and by him. I admired him deeply…"
In 1964, Feynman gave a series of lectures at Cornell University, the Messenger Lectures, under the title "The Character of Physical Law." His topics ranged from symmetry, probability and uncertainty in physical laws to techniques by which physicists seek new laws. The lectures were recognized for their extraordinary quality. According to Gates: "I have videotapes of physics lectures Feynman gave at Cornell decades ago. They are the best lectures I’ve seen on any subject. He shared his enthusiasm and clarity energetically and persuasively.
During an interview with CIO magazine in September 1997, Gates was asked: "Who would you invite to a dinner party?"
Feynman was on the list, along with Einstein and Leonardo da Vinci.
At the time of his death, Feynman had become everyone’s favorite physicist, thanks to the popularity of Surely You’re Joking, Mr. Feynman and What Do You Care What Other People Think. With these books, transcribed by his friend and drumming partner Ralph Leighton from taped conversations over a period of years, Feynman captured the public imagination as no other physicists had before him, with the possible exceptions of Albert Einstein and Enrico Fermi.
The Feynman Lectures on Physics, a set of lectures Feynman gave to undergraduates at Caltech in '62-63, is now a classic. (For a description of how the lectures came about, see the definitive article by Feynman's colleague Matthew Sands in Physics Today, April 2005.) Feynman’s fame grew when he was appointed to the Rogers commission in 1986 to investigate the Challenger shuttle explosion. His dramatic demonstration on television of the loss of resiliency in O-ring at freezing temperature as a principal cause of the Challenger accident made him a national celebrity. In applauding his performance, the physicist Freeman Dyson said: "The public saw with their own eyes how science is done, how a great scientist thinks with his hands, how nature gives a clear answer when a scientist asks a clear question."
Since he passed away in 1988, Feynman lore has continued to grow. Several books have been published, including Genius: The Life and Science of Richard Feynman (James Gleick, 1992), Most of the Good Stuff: Memories of Richard Feynman (American Institute of Physics, 1993), No Ordinary Genius (Christopher Sykes, 1994), The Beat of a Different Drum (Jagdish Mehra, 1994), Feynman’s Lost Lecture: The Motion of Planets Around the Sun (W.W. Norton, 1996), The Meaning of It All (Helix Books, 1998), The Pleasure of Finding Things Out (Perseus Books, 1999) Feynman's Rainbow (Leonard Mlodinow, 2003), and Perfectly Reasonable Deviations from the Beaten Track: The Letters of Richard Feynman (Edited by Michelle Feynman, Basic Books, 2005).
Feynman has even made it into the billboards! When Apple Computer began its "Think Different" series of ads featuring great scientists, artists, humanitarians and the like, the company chose Einstein and Gandhi among its first examples of the uncommon rewards awaiting those who dared to follow the beat of a different drum. "Can Feynman be far behind?" I wondered.
In November '98, I was driving in San Francisco's Mission District when I suddenly saw that familiar face with the knowing grin inviting commuters to ponder the mysteries of ... what? The photograph showed Feynman wearing the corporate T-shirt of Thinking Machine, a Boston-based company where he had briefly worked as a consultant in 1983. The shirt bore a schematic representation of Connection Machine - a cube of cubes - that he helped design for Thinking Machine. (It's the same photograph on the cover of What Do You Care What Other People Think?) Then, in April '99, Feynman "came" to Silicon Valley where I live. Anyone driving along Highway 101 in the South Bay could "see" Feynman teaching quantum mechanics at the California Institute of Technology, in front of a blackboard on which he had written matrices and differential equations. According to Caltech archives, the photograph was taken on May 2, 1963, during his "Lectures on Physics" period. (Feynman has a large number of fans in Silicon Valley, so there was disappointment when Apple "replaced" him with an image of an iMac five months later.)
What may have pleased Bill Gates most, however, were the publications of The Feynman Lectures on Computation (edited by Anthony J. G. Hey and Robin W. Allen - 1996) and Feynman and Computation (edited by A. J. G. Hey - 1998.) The first is a collection of lectures Feynman gave at Caltech from 1983 to 1986 as part of an interdisciplinary course called "Potentialities and Limitations of Computing Machines." The second contains contributions by distinguished computer scientists and physicists who were guest lecturers in Feynman's interdisciplinary course. It also contains reprints of Feynman's prescient articles on the physics of computing: "There's Plenty of Room at the Bottom" (1959!) and "Simulating Physics with Computers" (1982). Anyone reading these two books will agree that Feynman's insight and ingenuity make his lectures on computation almost as timeless as his physics lectures.
And now Feynman has his own 37-cent first-class stamp! The US Postal Service has honored four American scientists - physicists Richard Feynman and Josiah Willard Gibbs, mathematician John Von Neumann and geneticist Barbara McClintock. The stamps were issued on 4 May, 2005. The Feynman stamp shows the physicist in his 30s, framed by the unmistakable Feynman diagrams. The stamp came about thanks to the decade-long lobbying effort by Ralph Leighton, who organized a celebration on May 11, 2005, at the post office in Far Rockaway, the New York City neighborhood where Feynman grew up. On the same day, the street in Far Rockaway where Feynman lived - 2 blocks from the post office - was renamed in his honor, from Cornaga Ave to "Richard Feynman Way." The date was appropriately chosen: May 11 is Feynman's birthday.
Gates never met Feynman but it is fascinating to imagine a meeting between the two. Here is the whiz kid transformed into a wide-eyed pupil, marveling at the master’s facility with ideas and insights, wondering at the source of that magical genius that’s uniquely Feynman’s. What does Feynman think of the current state of computing? How does he envision its future? Are any architectural breakthroughs in software imminent? Where is the limit and why?
An autumn afternoon. Gates is at the Feynman house in Altadena, Southern California. Feynman introduces Gates to his menagerie - one horse, two dogs, one cat, and five rabbits. Gates smiles as Feynman addresses each animal by name and inquires of its health.
Afterwards, they settle down in the book-lined living room to talk.
Gates: When did you first take an interest in computers?
Feynman: My interest in computers really grew with my interest in physics, which is to say, very early. I recall reading in high school about mathematical machines, tide predictors, area measuring devices, and all kinds of wonderful things about computing in the Encyclopedia Britannica.
Gates: How have computers helped you in physics?
Feynman: When I try to solve a difficult problem, I ask myself: What is it that I can compute that will explain how this particular physical system behaves? This approach has served me well. I am interested in useful results, not abstractions, whether it’s in understanding how light interacts with matter or why helium behaves so strangely at low temperatures. I get useful results by coming up with numbers that can be experimentally verified. Computers play an important role in this verification process.
Gates: But isn’t your kind of computing different from the computing most of us are used to?
Feynman: It’s true the kind of computing I am interested in is based mostly on physical insights and mathematics - mathematics always in the service of physics, I must add - but computers are a big help. They can perform millions of important calculations a scientist may never have the time for. Sometimes computers can even suggest ideas one hasn’t thought of before. I find this exciting. For anyone curious about how nature works at the deepest level, a computer can be a valuable tool. I certainly find them so.
Gates: Back in 1942, when you and other scientists were working on the Manhattan Project, digital computers were several years away. What was computing like then?
Feynman: We had these Marchant and Monroe computers - hand calculators with numbers, really - that were good for adding, multiplying, dividing, and so on. They were about a foot across and several inches high, with all kinds of levers on them that you pushed to get results. Unfortunately they broke down often. Metal parts wore thin and came out of alignment because of the pounding they took, and had to be sent back to the factory for repair. We just couldn’t afford the downtime - this was a wartime effort after all - and so some of us began to tinker. We would take the covers off and try to figure out ways to fix them. Pretty soon we got good at it and kept things going.
Then the calculations became complicated, way beyond the capacity of the Marchants. We had to get IBM machines - multipliers, tabulator, verifier, keypunch, sorter, collator and what have you. They were the best machines at the time. We managed to assemble them ourselves and came up with results that turned out to be very important.
Gates: I am curious about some of those calculations…
Feynman: The biggest challenge was to figure out how much energy would be released from various designs of the bomb. Then we had to narrow it down to how much energy would be released from specific designs that would be used in the actual bomb, and how much fissile material would be needed in each case. This was very complicated, nonlinear equations and all, and we had so little time! The experimentalists couldn’t help; they needed our results to carry on their work. We computed by simulation, using a primitive form of what you would now call parallel processing. But we rose to the challenge. Our calculations told us what we could and couldn’t do. Lots of important results, very accurate.
Gates: Fifty years later, how do you view your wartime efforts?
Feynman becomes pensive. "At Los Alamos, we were doing what we had to. We started for a good reason. We worked hard. It was exciting. We discovered, invented, and pushed the limits of science and technology to create a bomb to help us win the war. We won the war but afterwards many of us weren’t so sure about the bomb itself. We had second thoughts. The bomb took on a life all its own. I remember sitting in a restaurant in New York shortly after returning from Los Alamos to teach at Cornell, wondering: what if New York City were to become another Hiroshima! Everything around me would be smashed. What was the point of life, of all this creativity? It didn’t make any sense at all.
"It took me a while to shake off this feeling. Eventually I got busy with physics and moved to Caltech, but that’s another story."
Gates: What about your experience with computers, though?
Feynman brightened. "There’s no second thought about that," he replied. "The main idea I came away with from Los Alamos was that even simple, primitive machines could be used to calculate important results. And it keeps getting better!"
Gates: I just finished reading Feynman Lectures on Computation. I am intrigued by a remark you made at the end of one of the chapters: "In 2050, or before, we may have computers that we can’t even see!"
Feynman: What I was investigating in those lectures was the answer to a fundamental question: What is it that we can and cannot do with computers today, and why? One issue was, how small could you make a computer? Was there any physical limitation to its size due to laws of physics? That led me to investigate the characteristics of a computer operating according to the laws of quantum mechanics. If we want to make extremely small computers, no more than the size of a few atoms, we would have to use the laws of quantum mechanics, not classical mechanics. So I began to analyze what you would call quantum computers. People thought that the uncertainty principle would be a limitation: that is, you wouldn’t be able to make a computer as small as you wanted because of the way time and energy, for instance, was related. I found to my surprise that quantum mechanics didn’t impose any limitations on the size of a computer, over and above those due to statistical and classical mechanics.
Gates: So you don’t have to worry about any unavoidable limitation arising from natural laws when you are trying to build the smallest possible computer?
Feynman: Exactly. Nature is quantum, not classical, so if you are trying to simulate nature, it had better be built on quantum laws. And if there’s no restriction there, you’re on solid ground! Of course, you have to consider the second law of thermodynamics, reversible computing and so on.
Gates: How would you write software for such a computer?
Feynman: That’s for you to figure out! I did my part!
Gates: I noticed in your lectures that you derived Shannon’s Theorem in three different ways, using concepts from statistics, geometry and physics! Why?
Feynman: I am an explorer. I like to find things out for myself. That’s how I understand anything. What I cannot create, I do not understand. If I can derive a theorem or a result independently, even when I know it has been discovered before, it means I understand it. That’s the only way I know how to learn. In the case of Shannon’s theorem, each method I used to derive it taught me something new.
"Besides," Feynman lowers his voice conspiratorially as Gates instinctively draws nearer, "what one fool can do, so can another!"
Gates: The topics of your lectures - coding and communication theory, Shannon’s theorem, quantum computers and such - are fundamental research topics. We have a growing research department at Microsoft but our focus is somewhat different. We are primarily interested in such things as: How do you increase people’s creativity through software? What will make computers easier to use, more responsive to the needs of the user, more natural? Can computers extend human cognition by assimilating speech and linguistics? These are the issues that interest us. Do you have any interest in these aspects of computing?
Feynman: Of course I do. Any tool that can make computers easier to use and more natural, as you say, is important. My own work in physics reflects this. I invented something called Feynman diagrams that allowed me to make complicated quantum calculations in one evening that used to take physicists six months! It was my moment of triumph, to realize that I had succeeded in working out something worthwhile.
So if you are inventing tools and products that simplify computing and at the same time open up complex problems for intelligent analysis, you are doing right by me.
Gates: Is there anything we should look out for?
Feynman: You need to make sure that the tools you create do not become more complex than the problems they are designed to solve.
Gates: One thing that concerns me very much is trying to anticipate the nature of computing ten or twenty years from now. We want to be as intelligent about it as we can, so that we can begin laying the foundation for it now, if that’s possible. Personally, I am looking for some good ideas to take us there…
Feynman: Well, it seems to me you need a new model for writing software, considering how important software has become in everything we do. I think that the next generation of software ought to be modeled after natural objects that are defined by natural laws. If you can model software objects after objects of nature, I think you will have moved on to something new and significant.
Gates: How so?
Feynman: For one thing, you can be sure that the software will do what it is meant to. If you push this idea further, the same software should be able to transform itself appropriately if the boundary conditions were to change. It seems to me no matter how well thought-out a computer program is, there’s always some unforeseen error in it. It’s not the fault of the designer or the developer, it’s the model on which the program is written. A large software system now seems to me to be like an elaborate sandcastle one builds at the shore. Suddenly a big wave hits and it’s gone!
Gates: So we have to change the foundation?
Feynman: I think so. You need to bind software development to criteria higher than standards and protocols. Numerical data modified by computers should be treated much as laws of nature govern the way objects behave in the real world.
Gates: But doesn’t that imply that people who write software have to be physicists as well?
Feynman: Not really. If you are talking about physical laws, the fundamental laws of nature are simple. That’s where their power and beauty come from. You go through all these complicated calculations and what comes out in the end is unbelievably simple. That’s what I talked about in those videos you have of mine. Besides, it can’t hurt to know a little physics. It’s a part of our culture!
Gates: That is true. Perhaps the new model could help in the area of testing too. Software testing has become very critical at Microsoft. We are always fighting testing deadlines against product release deadlines!
Feynman: Exactly. How do you test software against all possible failures? I don’t think you can, using current methods. But if software objects can be modeled after natural objects, testing becomes more straightforward. You have more confidence in the result. If the test fails, you may end up discovering something new and unexpected. That’s how it is in physics. There’s no fooling natural laws!
Gates: I remember the last sentence in your personal report on the Challenger accident: "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled."
Feynman (obviously pleased): You get the idea!
Gates: Any example modeling software after natural objects?
Feynman: Some years ago, a bunch of guys were trying to build super fast computers using parallel processors. The company was called "Thinking Machines." My son Carl, who is interested in such things, joined them. Since the kids running the company didn’t know any better, they ended up hiring me too!
Anyway, the problem was to design a router that delivered messages from one processor to another. There were a million processors in that machine and it wasn’t practical to connect every pair of them. We chose a model where each processor needed to talk directly to only a few of its neighbors. The problem came down to figuring out the minimum number of buffers to hold messages for the router to operate efficiently.
I analyzed the problem by treating the router circuit diagrams as if they were objects of nature, the kind of stuff I do all the time in physics, and came up with a set of partial differential equations. The equations said five buffers per chip would do. Others predicted seven but in the end, it had to be five from a practical standpoint. My equations apparently saved the day!
Gates: We use certain types of object models at Microsoft for writing software. Coming to think of it, you can say these objects are quantized software!
Feynman: That means you are already using concepts from physics, only you didn't know!
Gates: What you are suggesting, then, is that future software should include ideas and concepts from physics as well as from computer science.
Feynman: Yes, but I think ideas should also come from biology. In fact, I think the intersection of biology and computer science could prove even more fruitful for developing software than physics.
Gates: It’s already happening. Bioinformatics is an emerging field that brings together ideas from biotechnology and computers. You studied biology for a while, didn’t you?
Feynman: Yes. I actually worked in the field during my sabbatical year at Caltech. This was after Watson and Crick’s discovery of the DNA spiral. My big moment came when Watson himself invited me to give a seminar on my work at Harvard. I am convinced biology has a lot to offer to computer science, especially in writing software.
Gates: In what specific ways?
Feynman: Well, insights can come from understanding how living organism function, from their adaptive, fault-tolerant, error-handling traits. It can come from studying how human genes are laid out, how the 'book of life' actually reads … There are thousands of possibilities, really.
Gates: That’s exciting! If I were not in computers, I would most certainly be working in biotechnology. I think we are only scratching the surface here. Have you been following the Genome project?
Feynman: Yes, and I think what scientists have accomplished is remarkable. It's a historic milestone to have mapped the genetic blueprint for human life and make it available to researchers!
Gates: The way we treat diseases and prevent them will be revolutionized!
Feynman: I certainly hope so. We can end suffering, at least to some extent, only if we know what disease really is. And when that happens, perhaps we can start talking more about health and less about disease.
Gates: One of my goals in life is to help eradicate diseases like smallpox and malaria and cholera and polio from the world. It's terrible that so many people still die from these diseases in this day and age. The success of the Genome project should certainly help.
Feynman: That will make it all worthwhile, won't it? In many ways, the Genome project reminds me of the Manhattan Project. I feel the same sense of excitement, the same anticipation. I wish I could start over again!
Gates: I am sure there's much you can contribute still.
Feynman: Is there any Microsoft database that can store the Genome and sift through its data quickly?
Gates (momentarily taken aback): We are working on it. All of us at Microsoft consider it extremely important to create a database system that can meet this type of challenges.
Feynman: That should be a milestone for Microsoft. If biological data managers find widespread use, database research will pick up speed too!
They ponder the implications of the coming revolution in genetics. Both see beneficial possibilities but recognize that there are important moral and ethical issues to consider too.
Gates: It seems to me from our conversation that a really good software engineer should be able to derive inspiration from different disciplines.
Feynman: Yes! I made the point in my Nobel lecture that a good physicist might find it useful to have a wide range physical viewpoints and mathematical expressions available to him. If everyone follows the current fashion in expressing and thinking about the generally understood areas, then understanding the open problems is limited. It’s possible that the truth lies in the fashionable direction. But if it is in another direction, who will find it?
I would make the same point to the new generation of software developers. Don’t limit yourself to what you know or what already exists. Be an explorer, not a tourist. Look across disciplines. Dare to follow the beat of a different drum. Your inspiration may come from the dance of molecules on a wave in the sea, the complexity of a beehive or an ant colony, the march of stars across the heavens, the nature of memory and language, the symmetry of a snowflake, and so on. There’s no end to it! After all, nature’s imagination is richer than ours, so why not use what we know of it to our advantage?
Gates: It all goes back to childhood curiosity, doesn’t it?
Feynman: Right! And it’s a tragedy we can’t hold onto some of that curiosity as we grow older!
Gates: Who are your scientific heroes?
Feynman (after a pause): There are three, really. Sadi Carnot, James Maxwell, and Paul Dirac.
He explains: "Carnot obtained a general principle of nature from the nuts and bolts of the thermal efficiency of steam engines. In one stroke, Maxwell unified electric, magnetic, and optical phenomena. And Dirac, a hero of mine ever since I read his book The Principles of Quantum Mechanics, discovered the relativistic equation for the electron."
Gates: What do you consider your most interesting discovery since winning the Nobel Prize?
Feynman: I liked my work on the theory of liquid helium. Another was discovering the laws of weak interaction with Gell-Mann. I also worked out something called the theory of partons to explain some of the properties of protons. Right now I am standing back. I am playing around with some ideas in my mind and I don’t have a clue where I'll come out. I guess that’s what makes it exciting, not knowing what strange territory one will end up in!
Gates: Any disappointments?
Feynman: Certainly. I’ve spent years trying to solve some difficult problems without success. The theory of turbulence is one. In fact, it is still unsolved. Another was my inability to understand superconductivity in which I worked for a couple of years. I should have grasped that one after my success with liquid helium but I didn’t.
Gates: Do the disappointments linger?
Feynman: Not at all! Even where I failed, I worked very hard and had a terrific time. People only hear of successes but not of failures. The important thing is to decide which problems are important and give them your best shots. If you succeed, fine. If not, you almost always end up learning something new!
Gates: You mentioned turbulence. Is that a part of the study of complexity?
Feynman: Yes. There are many phenomena we do not understand yet, from the flight of a swarm of bees to the self-organizing properties of neural networks in the human brain. How is it that a few basic rules can lead to such extraordinarily complex behavior? This is the fundamental question any theory of complexity must answer. And it’s proving to be a real challenge!
Gates: Does that reflect a failure of conventional methods for understanding complex systems?
Feynman: Probably, or else we would have solved these problems long ago! It’s the same idea in software: How do you make sure a software system consisting of 20 or 30 million lines of code works coherently? You have to retain a certain amount of skepticism about accepted ideas and keep an open mind about ideas that appear flaky. It also means any serious study of complexity will require us to explore the fundamental relationship between physics and biology to computation.
Feynman: Meaning whether or not there is a computational model of the universe and of biological systems that we need to consider and understand. The field is wide open and it’s undoubtedly going to require interdisciplinary research.
Gates: During the Manhattan Project you came in contact with some of the most powerful minds of the twentieth century - Enrico Fermi, Niels Bohr, Hans Bethe, John Von Neumann, Stanislaw Ulam, Robert Oppenheimer and others. Do you think there will be such a gathering of "monster minds" again?
Feynman: That was a different situation at Los Alamos. The project came about because we had to win a war. Besides, the science was also good, very good. We were discovering and inventing as went along. I don’t know that such a situation will reappear. Times have changed. It’s more structured now, more what you call "market-driven." But it really doesn’t matter, because there are many great minds about, brilliant people working in your field who are pushing the limits of technology, worrying about how people interact with computers, how software is written, and so on.
This was the moment Gates was waiting for, the real reason he had come to Altadena. Everything else was a prelude to this moment.
Gates: Will you consult for Microsoft?
Feynman’s response is instantaneous and unequivocal. "That’s the wackiest idea I ever heard!"
Gates is relieved. He has read enough about Feynman to know that the response really means Feynman is interested!
Gates smiles and suggests that his son Carl can perhaps join him too.
Feynman’s face glows. "Ain’t a bad idea at all," he says in his best Brooklyn accent.
Gates doesn’t want to leave any loose ends behind. "How about next month?" he asks. "We will give you a tour of the Microsoft campus and you can meet the people working on new ideas. I think you will like what they are doing."
But Feynman stops him. "Not possible," he says. "I am going on a trip to," he pauses dramatically and then says with a flourish, "Tannu Tuva!"
Gates is vaguely familiar with the planned trip to this place deep in Central Asia, outside of Outer Mongolia, with a capital named Kyzyl. One reason Feynman is interested in going to Tuva is because, in his own words, "any place that’s got a capital named K-Y-Z-Y-L has just got to be interesting!"
For a moment Gates thinks of asking Feynman if he can come along too. Then he sighs. Too many commitments, too many keynote addresses. With a shock he realizes he too is a prisoner of schedules and deadlines, just like other worker bees!
"After you return from Tuva, then?" he asks, almost plaintively.
Feynman looks out the window. The day has about an hour of sunlight left. Shadows are lengthening and a small wind is stirring the sycamore leaves.
"Why not?" he says, grinning, and extends his hand. Gates gratefully shakes it.
Gates: There’s a revolution coming in software and I think you will enjoy being there when it arrives.
Feynman: I’m sure I will.
Gates: Information is our most important asset. We plan to create great stuff out of it.
No, replies Feynman softly, imagination is.
On the flight to Seattle, Gates is restless. He picks up an airline magazine and is immediately repulsed by its banal content. He wishes he can run up and down the aisle to work off the intellectual energy that has gripped him.
He reclines as far back as the seat allows and closes his eyes. "Must quickly put a team of good people together," he thinks. "Feynman is coming to Microsoft. Seek inspiration from across disciplines. Learn to tap into nature’s imagination. It’s time to take Microsoft to another dimension."
Copyright: Hasan Zillur Rahim (2006)