Michael Acheson Isard
For a human or a bird, the task is trivial: visually track an object moving through a cluttered scene. But computer vision can’t do this- there are too many visual ambiguities for even the most advanced artificial intelligence programs. Michael Acheson Isard is working to get past these obstacles. For his doctoral thesis at Oxford, he devised an algorithm called CONDENSATION, for CONditional DENSity PropagATION. Isard’s thesis supervisor, Andrew Blake, says Isard’s technique promises a “revolution in the design of intelligent machines.” The premise is that as a computer watches a scene, it must continually reweigh alternative interpretations of what is signal and what is noise-mimicking human visual perception. Isard’s work, says Blake, has “sparked a whole new area of activity by other researchers.”
This fall, following a three-year postdoc at Oxford’s Magdalen College, Isard took a research position at Compaq’s Systems Research Center (SRC) in Palo Alto. According to Blake, this Anglo-American, born in England to American parents, is “set to have a major impact on the way machine intelligence develops over the next decade.”
Some computational problems, such as defeating today’s commercial encryption, strain even the most powerful machines. Adam Beberg has figured out how to tackle such challenges: Throw the unused time of 10,000 computers at them. Such “distributed computing” promises greater access to number-crunching power, possibly leading to scientific and technological breakthroughs. For example, SETI@home, a search for intelligent life in the universe, is following Beberg’s lead with a distributed computing scheme to analyze radio telescope data. In a realm with more commercial significance encryption Beberg’s ideas have already paid off. In 1997, he founded a nonprofit group called Distributed.net. During the group’s first year, it hosted an alliance of computers called the Bovine Cooperative, which won a prize by breaking a form of encryption known as RC5.Beberg left Distributed.net in April to work on Cosm, an open-source distributed computing project. Says former colleague Michael Labriola, now CEO of Invisible Web Publishing: “The ideas that came intuitively to him could literally change the world.”
The last decade has seen cell phones and wireless communication blossom from a tool for privileged executives to something close to a necessity for almost everyone. But the revolutionary impact of this wireless world will come only when more bandwidth is available to provide video, Internet and other services. One of the folks at the forefront of the search for bandwidth is Paul Bender. After completing his doctorate at the University of California, San Diego, he joined Qualcomm and started working on projects to improve the quality of wireless communications. Using a protocol called Code Division Multiple Access (CDMA),Bender developed new equipment to take maximum advantage of it, including a sophisticated processing system on a single chip that reduced the mass and cost of cell phones. He currently leads a project to develop high-speed wireless data access with only the amount of spectrum currently used for a few voice users. Movie studios could soon transmit digital versions of their films to theaters using such technologies. Qualcomm is impressed with Bender’s versatility. “What is remarkable is that he excels in four distinct areas of engineering: the design of hardware, the development of software, the creation of new and improved systems and the mathematical analyses of these systems,” says Qualcomm vice chairman Andrew Viterbi.
One of the “grand challenges” facing biologists is predicting the complex three-dimensional structure of proteins. The kinks and folds in these large molecules largely determine their behavior; certain combinations of coiled proteins in virus membranes, for example, allow the virus to insinuate itself into a cell. Bonnie Berger is leading a group of computational biologists to develop software that uses mathematical algorithms to predict protein folding based on the sequence of amino acids. Such insights could eventually lead to new drugs to combat viral disease such as AIDS. Her lab is also tackling the problem of gene identification: devising software that can help indicate where in the mind-numbingly vast strings of DNA sequences (most of which are random filler called introns) lie the bits and pieces of actual genes that carry the blueprints for proteins. Berger started out as a computer scientist and applied mathematician. But as an MIT postdoc looking for ways to apply the algorithms she was devising, she found that it was the biologists who gave her some of the most interesting problems. So she took some courses in biochemistry and, she says, “picked up biology on a need-to-know basis.” Moving from computers and mathematics to biology, she admits, has brought on “culture shock.” But she is thriving in this mixed milieu. “Her work has been highly innovative and important,” says MIT math professor Daniel J. Kleitman–all the more remarkable, he says, because “it is very unusual for a computer scientist to make a recognized impact in biology.”
The ballooning volume of information available through the Web and other media presents a problem: How do we make sense of all that data? Businesses are particularly hungry for tools to sift massive amounts of information to yield nuggets of insight about their customers’ behaviors and needs. Such “data mining,” which involves a blend of marketing techniques, innovative software, graphic design and other disciplines, is just now finding the strategies that work. It may be one of the most important economic realms of the next century-and David Blundin, founder and CEO of DataSage, is likely to be a key innovator.
Venture capitalist Duncan McCallum of One Liberty Ventures calls Blundin “the most visionary person I have ever met” in the field of decision-support software. “Dave has established an impressive track record of leading and commercializing groundbreaking innovations in data mining, particularly its application to customer intelligence,” McCallum says. Blundin, he adds, has “the ability to spot important trends and an ability to create novel solutions to problems.” As retailing becomes e-tailing, this is someone to keep an electronic eye on.
Matthew Brand studies mathematical approaches to learning and perception. That may sound dry, but, in fact, you may soon be seeing Brand’s work at a movie theater near you. The reason? His research leads to the creation of “digital puppets” that incorporate not just the appearance but also the mannerisms of movie stars. Brand moves easily between academic research and the private sector, but in both settings his persistent theme is designing computer systems that understand and interpret three-dimensional reality. In his research at MIT’s Media Lab he built an “artificial artist” that designed mobiles. This synthetic Calder combined computer vision methods for accumulating images with AI software that enabled it to pick out the most relevant features of each image. After studying at MIT and teaching at Northwestern, Brand moved to the Mitsubishi Electric Research Labs in Cambridge, Mass., where he focuses on entertainment applications. He has developed a puppet that uses a voice interface (a face animated by speech input) and one driven by shadows. “Early adopters will most likely use this technology to synthesize people–historical, contemporary, or even nonexistent,” says Brand. As such technology permeates Hollywood, it may create a new type of cinema for the 21st century.
One way for a search engine to cope with the Web’s explosive growth is to employ a cluster of cut-rate computers, since more machines can easily be added to keep pace with increasing demand. Eric Brewer took this scalable approach in forming Inktomi, a startup that went public (very lucratively) last year. Brewer straddles academia and entrepreneurship; the technology underlying Inktomi was developed with one of his computer science students at the University of California, Berkeley. Inktomi’s system not only scales up easily, it can also keep running in the face of massive processor and disk failures–qualities that have persuaded HotBot, FindWhat and other Web search engines to use Inktomi technology. Recently, Brewer has led Inktomi to develop the Directory Engine, which for the first time allows a Yahoo-like catalog to be built and maintained automatically, after humans set up categories and put sample documents in place.
Wearing his academic hat, Brewer designed a wireless system, GloMop, that enables handheld devices to draw computing power and network access from a stationary machine. “Brewer is one of the rare individuals who has the analytic, design and experimental skills required to make a truly great computer systems innovator,” says Berkeley computer science professor Randy H. Katz.
Internet search engines and wireless communications will be two of the hottest information technologies in the coming years, and Eric Brewer is at the leading edge of both.
God toppled the Tower of Babel, and prevented the human tribes from speaking a single language. Something similar might have happened to the equally presumptuous computer revolution if it weren’t for middleware software that enables disparate programs to talk to one another. Middleware is big business for companies such as IBM; hundreds of programmers labor at the UK Development Laboratory in Hursley Park to create Big Blue’s latest offerings for the enterprise market. One rising star is Mandy Chessell, a software developer who says she was first captivated at 17 by computing’s strange combination of “logic and creativity.” Those are terms that IBM Distinguished Engineer Tony Storey says apply to Chessell herself: “She’s a constant source of innovative ideas.” Chessell’s name appears on 14 filed or issued patents.
Chessell’s most important project is the code behind Component Broker, a new middleware system that has been bought by companies such as Charles Schwab. Chessell’s work is crucial as companies move to launch new services, often on the Internet, and need to connect their “legacy” systems to new software.
Computers have been an industrial miracle for a couple of decades–becoming exponentially more powerful and cheaper. But miracles aren’t forever, and it will soon become impossible to continue this trend with the current silicon-based technology. One possible solution is an entirely new model for computing, “quantum computing,” in which data are encoded in the quantum spin of atoms and molecules. But what is intriguing in theory can be difficult in practice; quantum computing is no exception. The creative leaders who can reduce the concept of quantum computing to practice may go down in the history of technology. Isaac Chuang might be one of them.
While still a graduate student at Stanford, Chuang was one of the developers of a basic, two-bit quantum computer. Since then he’s demonstrated that a quantum computer can run simple algorithms and perform database searches. He continues his work on the topic at IBM’s Almaden Research Center, trying to scale up quantum computing. Some think Chuang could be one of the ones to move from vision to reality in this field. “His omnivorous curiosity, mastery of so many levels of description, and profound contributions really are exceptional,” says a collaborator, MIT professor Neil Gershenfeld. “If anyone is going to turn quantum computing into a reality, it will be Ike.”
Proteins are the workhorses of biology. In their active form, they are folded up into complex three-dimensional molecules, and understanding how folding happens is one of biology’s enduring problems. Solving it could lead to safer and more effective drugs–even therapeutic proteins designed from scratch. In this search, an important new technology is David Clemmer’s method of sorting molecules, including proteins, according to their shapes.
Rather than working with proteins in their normal, liquid state, as others typically do, Clemmer observes these molecules in the gas phase as they travel across an electrically charged space. Travel times can be used to construct a theoretical model of how the protein is configured and of the forces that determine how a protein folds.
Clemmer’s method also could provide a screening technique for combinatorial chemistry–a drug discovery process that produces large mixtures of compounds. Methods for sorting mixtures of isomers (compounds that have the same mass but have different shapes and often different biological activity) are slow, relying on complex data that are difficult to interpret. Clemmer’s approach makes it possible to sort large mixtures of isomers in seconds. Combined with other new drug-discovery methods, the reward will be a tremendous increase in the rate of screening molecules that have therapeutic potential.
It’s hard to imagine a better raw material for plastics than carbon dioxide. It’s cheap, non-toxic and ubiquitous. There’s one problem, though: It’s very stable and difficult to engage in chemical reactions. But Geoff Coates has invented a zinc-based catalyst that can make polycarbonates (a common plastic usually made from petroleum-based chemicals) using carbon dioxide as a starting point. The reaction is far more efficient than commercial poly- carbonate processes, and the resulting plastic is biodegradable.
This particular reaction may not revolutionize plastics production.
After all, the low cost of oil makes it tough for other feed stocks to compete, even one as inexpensive and omnipresent as carbon dioxide. Yet there is little doubt Coates’ impact will be felt in coming years in the chemical industry. That industry has long relied on discovering catalysts through laborious trial-and-error methods. One of Coates’ ambitions is to rationally design catalytic structures to produce desirable polymers. “My dream,” he says, “is to be able to sit at a computer and design a catalyst, then go to the lab and make it.”
If he’s able eventually to make that dream come true, the result could be more efficient manufacturing processes–and at the same time a cleaner environment. The future of the chemical industry will turn on whether it can meet both criteria. Look for Geoff Coates to help it.
The search for as-yet-unknown reactions–and the accompanying technological potential–is an unending pursuit of inorganic chemistry. Despite years of exploration, there’s plenty left to be discovered, a fact that Kit Cummins quickly points out.
Cummins, an assistant professor of chemistry at 27 and professor by 30, has already forged a world reputation for keen intuition and a deft touch in exploring new chemical ground. His most notable success: finding a way, at room temperature and pressure, to break apart the extremely strong triple bond that holds the atoms of a nitrogen molecule together. Such feats have so impressed the research community that one of the several leading scientists who nominated Cummins for the TR100 described him as “definitely Nobel Prize material.”
For the moment, Cummins’ work remains in the realm of fundamental research. But Cummins suggests that his efforts could lead to ways to better utilize small inorganic molecules such as dinitrogen oxide in the manufacturing of everything from pharmaceuticals to plastics.
Miguel De Icaza
Geeks love Linux the “open-source” operating system that, amazingly, might pose a serious threat to Microsoft Windows. But the rest of the world–say, 99 percent of us–have little patience with Linux’s arcane command line interface, which seems like something out of the bad old days of DOS. The solution? Give Linux an easy-to-use graphical interface, designed, naturally, as open-source software. That project, already well under way, is led by Miguel de Icaza in Mexico City. De Icaza is coordinating development of a graphical interface called GNOME that makes Linux accessible to all by giving Linux the windows and icons that the masses have become accustomed to. (The system is free for the download.) In addition to spearheading GNOME, de Icaza has developed an open-source spreadsheet program called Gnumeric-impressive accomplishments for a mathematics undergraduate at the National University of Mexico who holds a day job at the school as a computer systems administrator. Richard Stallman, the MIT open-software guru, calls de Icaza “not only a capable software designer, but an idealistic and determined campaigner for computer users’ freedom.”
Combinatorial chemistry is a radical departure from the way researchers have traditionally identified new drugs and materials. Rather than painstakingly making and testing compounds one at a time, in combinatorial chemistry you make hundreds of thousands–even millions–of variations at the same time, then screen to find the winners. Michael Deem is working to improve the odds in this high-tech game of chance. By making an analogy between a computer simulation technique called Monte Carlo and combinatorial chemistry, Deem has provided a way to search more efficiently and broadly for new compounds. In general, Monte Carlo simulations are a powerful technique to sample data, using an algorithm that takes random “walks” among large data sets. Taking advantage of his chemical engineering background, Deem has developed “biased” Monte Carlo techniques that allow combinatorial chemists to greatly expand their searches, with the computer selecting the most promising paths.
Among other research projects, Deem is helping advance a new field called protein molecular evolution. “It’s basically combinatorial chemistry for proteins,” says Deem. The payoff could be a powerful new technique for finding protein-based therapeutics.
Virtual reality (VR) is hot these days, because of its potential in entertainment, learning and research. However, the most advanced systems render three-dimensional reality in two dimensions. Gregg Favalora thinks three dimensions are better. He has been working on “three-dimensional” VR since his undergraduate days at Yale, where he developed a prototype exploiting lasers, lenses and mirrors to render images that are truly three-dimensional.
Unlike conventional VR systems, Favalora’s fictions occupy a volume of space: They can be walked around and viewed at almost any angle. He refined his concept as a graduate student at Harvard, then co-founded Actuality Systems and went on to win the 1997 MIT $50K Entrepreneurship Competition. The company is working on larger versions of its 3-D display, with the goal of producing a display that can be connected to a computer for applications ranging from pharmaceutical research to computer-assisted design. “I think great entrepreneurs must have three values: integrity, intelligence, and initiative,” says Rob Ryan, founder of Ascend Communications and a mentor to Favalora. “Gregg possesses all three”
It’s an invention that forces you to rethink one of man’s most basic tools: the mirror. No, it’s not a new vanity item. The “perfect mirror” Yoel Fink invented last year as a graduate student at MIT could mean radical new ways of directing and manipulating light. Potential applications range from a flexible light guide for delivering laser light to a specific internal organ, to new devices for optical communications, to coatings for windows that efficiently reflect heat while being transparent.
Fink’s mirror combines the best property of the everyday metallic mirror–its ability to reflect light from all directions–with those of highly specialized dielectric mirrors, widely used in photonics. Like other dielectric mirrors, Fink’s devices can be tuned to reflect only certain wavelengths of light with high efficiency. But Fink found a way to layer the dielectric material so that the mirror can reflect this light from all angles; other dielectric mirrors can’t. What’s more, his techniques for building these “perfect mirrors” are so general the devices can be made from a wide range of materials, including polymers.
Indeed, Fink is trying to exploit a class of polymers, called block copolymers, to create self-organizing optical components. MIT materials science professor Robert Rose says enthusiastically that “Yoel’s approach using soft materials which can be processed inexpensively to form conformable reflectors may bring vast new markets into play.”
Young humans spend an awful lot of time playing with toys and, these days, an equally large amount playing computer-based games. The two areas would seem to offer a natural intersection. So far, though, there hasn’t been much crossover. Amy Francetic, co-founder of Zowie Intertainment, hopes to change that with “smart toys.” Zowie’s “Play-Zones,” due in stores this fall, are toys that connect to a PC running accompanying software. The link allows children to use the software by playing with the toys: Moving the telescope mounted on a pirate ship in one PlayZone, for example, changes the view displayed on the computer screen. Francetic believes these toys will encourage the cooperative play that’s common with conventional toys but rare for computer games. She brings to the company a background of developing CD-ROM games for Hasbro and Electronic Arts, as well as work at Interval Research,the research lab from which Zowie was spun off. “Her ability to inspire creative people to solve technical design problems will help her continually to uncover the strengths of emerging technology platforms,” says Stewart Bonn, who worked with Francetic at Electronic Arts.
The desktop metaphor that has dominated (and limited) computer software for more than a decade may face its first major challenge from Lifestreams, an operating system environment that began as Eric Freeman’s doctoral thesis at Yale. Lifestreams relies on a different visual metaphor, presenting e-mail, schedules, online feeds and so forth as a chronologically arranged stack of documents, all automatically captured and easily reorganized on demand. The idea was granted a patent this year. Freeman stands out for the boldness of his vision. “His choice of thesis topic showed considerable nerve,” says computer scientist David
Gelernter, Freeman’s advisor at Yale. “It was risky–a radical departure and not an incremental improvement. He thought he could bring it off and he did.”
In 1997, Freeman and Gelernter co-founded Mirror Worlds Technologies (the name plucked from a Gelernter book on future computer interfaces). Their goal: interfaces, software architectures and tools for managing electronic information. With Freeman as CTO, the company has introduced its first product: Lifestreams Office. Now comes the hard work of convincing information technology managers that this is indeed a better mousetrap. With a foot in the private sector, Freeman is maintaining academic ties; he was recently named a fellow at Yale’s new Center for Internet Studies.
The turnaround in IBM’s fortunes led By CEO Louis Gerstner would have been impossible without grass-roots agents for change like David Gee. Under Big Blue’s old regime, innovation was often stifled by a reluctance to embrace and exploit ideas from outside IBM. Gerstner let in new blood. Gee joined IBM in 1995 from Dun & Bradstreet, soon establishing himself as a proponent of an emerging Sun Microsystems technology: Java. This platform-independent computer programming environment, now the lingua franca of the Internet, has been a cornerstone of IBM’s successful push to make itself an e-business behemoth.
In managing IBM’s Java initiative, Gee was responsible for “building a team, working with customers, developers, analysts and influencers the world over, and truly establishing IBM as a leader in open standards and e-business,” notes Mike Lawrie, general manager of IBM’s operations in Europe. Gee also managed
Alpha Works, IBM’s online research lab, which promotes new Internet technologies (mostly Java-based). His evangelizing proved so effective that IBM now has more than 3,800 Java professionals developing applications–an effort on a par with Sun’s. Earlier this year Gee was given a vote of confidence with a new position in Paris as Lawrie’s executive assistant, a job title often on the fast track for joining IBM’s top management ranks. Could Gee be an eventual successor to CEO Gerstner? You read it here first.
Materials research used to be laborious. That was before Symyx–a company that is speeding up the discovery of new materials by applying the methods of combinatorial chemistry. Combinatorial chemistry involves synthesizing a large array of compounds simultaneously, then using innovative screens to pick out the winners–materials with desirable properties such as the ability to act as a catalyst. Symyx, a 1994 startup, is the first company specifically devoted to using this process to replace trial-and-error methods in materials discovery. In the last several years, the company has raised millions in private and venture backing and has signed partnerships with a who’s who of top chemical and materials companies. Much of the credit goes to Isy Goldwasser, Symyx’s co-founder and vice president of business development until he was promoted to president and COO last year.
The potential of combinatorial chemistry? Shortened timescales for finding new materials, and the ability to search a broader range of possibilities. When Goldwasser was hired as the company’s first employee, he says, “people thought we were crazy and we would never be able to do what we aimed to do, but we are proving them wrong.”
Remember high-temperature superconductors? These high-tech darlings of the late 1980s brought a Nobel Prize to their discoverers and generated endless hype about how their near-perfect conduction of electricity would revolutionize energy transmission. Well, it hasn’t happened–at least not yet. One big obstacle has been the difficulty of forming flexible, long superconducting wires that can carry large amounts of current.
Amit Goyal, an Indian-born materials scientist, may have found a way over this hurdle. His contribution: growing thin layers of ceramic superconductors on a polycrystalline metal template, using the highly aligned metal to line up grains of the superconductors. The resulting structure of the superconductor resembles a single crystal, and the method has allowed Goyal and his co-workers at Oak Ridge National Laboratory to form superconducting wires capable of very high current densities.
That might–might–be enough for superconductors to fulfill their delayed promise. Within several years, says Goyal, high-temperature superconductors in wires for transmission cables and transformers could be a reality. If you sold all your stocks in companies with the word “superconductor” in their name, Amit Goyal might make you regret it.
Those who devise efficient ways to cut through “data smog” will be much in demand in the future. Among that group, one name to keep in mind is Joe Hellerstein of the University of California, Berkeley. Hellerstein’s work lies in finding the best way to put information into databases and then get out what you want–and nothing else. Some of his database related contributions include CONTROL (Continuous Output and Navigation Technology with Refinement On-Line), an approach that exploits continuous user feedback to refine the action of a search engine, and GiST (Generalized Search Tree), a way of finding answers to questions without having to worry about the type of data in the answer. Jim Gray, manager of Microsoft’s Bay Area Research Center, calls Hellerstein “the most promising of his generation of database systems scientists.”
Although working at a research university, Hellerstein is also pursuing commercial applications. He has designed Cohera DFS, a system intended to manage the computing resources of a large enterprise. Hellerstein is also on the technical advisory board of MySimon, a Web-based shopping service that provides inventory and price information from a large array of Web sites. Work like Hellerstein’s offers hope that today’s data smog will dissipate in tomorrow’s information daylight.
Nicola Hill’s work shows how little space sometimes separates research and commercial application. Her work is in very basic research: fundamental physics and theoretical chemistry. Yet it centers on a growing field called “spintronics” that attempts to exploit the spin of electrons in magnetic fields as a means of information storage. Spintronics remains in its early stages, but one day it could have exciting applications for ultra-high density magnetic data storage, even powerful quantum computers. “But even if it didn’t turn out to be all that practical, the physics is so exciting, it would never be a lost effort,” Hill says.
Hill’s research skills, intellect, and the gracious way she serves as a role model for young women in physics and materials science combine to make her an innovator to watch. She has started an “ambitious new research program in the theory of magnetic nanostructures, a field which holds great promise for its potential to revolutionize technologies such as magnetic data storage, next generation computers and magnetoelectronic devices,” says Fred Lange, chairman of the materials science department at UC, Santa Barbara.
Until recently, neither computer developers nor computer users paid much attention to the exterior of personal computers: What mattered was inside. This created an industry “where there is an obsession about product attributes that you can measure empirically,” such as processor speed and hard disk size, says Jonathan Ive, vice president for industrial design at Apple. In 1997, though, Ive was charged by interim CEO Steve Jobs to design a radically different computer–with attention paid to style as well as content. The result of the work of Ive’s design team was the iMac, a computer whose abilities were not so different from other computers, but whose design set it apart from any previous PC. Its colorful translucent case captured the interest–and pocketbooks–of millions; the design has inspired the sincerest form of flattery from makers of computer peripherals and, more recently, rival PC makers. Apple recently unveiled their latest design, the iBook, a laptop version of the iMac. While Ive’s work helped Apple distance itself from the pack, that wasn’t the primary purpose for his group’s innovative design, he says. “Our goal wasn’t just to differentiate our product, but to create products that people would love in the future.”
Electronic books hold great promise as a better way to read, but because of the limitations of display technology, they have lower resolution than the printed page, and thus are tough on the eyes. That may change thanks to the work of the Media Lab’s Joe Jacobson. He and his group have developed a system using “microspheres”–two-toned particles about the size of grains of laser toner–embedded in a sheet of paper to display text and graphics. A conductive, transparent “ink” is used to flip the microspheres into the correct position, all controlled by a microprocessor printed directly onto the paper.
In 1997 Jacobson co-founded a startup, E Ink, to commercialize the technology. Its first product, large signs that can display changing messages, was put to use in J.C. Penney stores this year. In addition, the ability to print microprocessors on paper and other surfaces could drastically change computing. “This technology puts forth the possibility of completely remaking both the chip industry and the fundamental way in which we make high technology devices,” claims Media Lab director Nicholas Negroponte.
In selecting the TR100, we looked for folks who keep innovating even after they’ve had one triumph. Christina Jones fits that description. She has already had one major success in Trilogy, the $100-million-a-year front office software company she founded in 1996 with fellow Stanford student Joseph Liemandt and three other young partners. But Trilogy seems to have been just a start.
Some of Trilogy’s software coordinates the configuration of multitudes of interchangeable parts in computing systems. Jones realized that PC makers and distributors could use this technology to facilitate sales of made-to-order Pcs–a booming business now dominated by Dell. Jones thought there was room for others in that market, however, particularly with Trilogy’s software as a way to improve efficiency. To start her new company, pcOrder.com, she sold her Trilogy stock back to Liemandt in exchange for access to technology and some financial support.
Dell’s growing market share means its competitors will be looking for an edge–creating a niche for pcOrder’s products. Indeed, Compaq, Hewlett-Packard and IBM have already licensed pcOrder software.
His line of work is venture capital, but “adventure capitalist” might be a better way of describing the managing director of Draper Fisher Jurvetson, a Silicon Valley firm that’s at the heart of the Web frenzy. Jurvetson has repeatedly shown his ability to recognize young entrepreneurs who can deliver. Examples: Sabeer Bhatia (see p.106) and John Smith, co-founders of Hotmail, the free Web-based e-mail site. More than 20 other firms passed on the deal, but Jurvetson’s firm plunked down $300,000 on Hotmail. The move paid off when Microsoft bought the company for a staggering $400 million. Jurvetson achieved another notable success with his investment in Four11,a free e-mail and directory services company. A classic overachiever, Jurvetson graduated from Stanford at the head of his class in just two and a half years, and made partner at his firm after six months. He shows no signs of slowing. In mid-1999, Jurvetson was financing no less than 11 companies ready for public offering.
The next time you’re stopped at the new traffic light on the corner, give a little thought to Fred Kish. Chances are that the red light you’re staring at gets its brightness and color from the light-emitting diodes (LEDs) Kish played a lead role in inventing. Kish is one of the big reasons LEDs aren’t just for wristwatches and alarm clocks anymore.
When Kish joined HP in 1992, the company was struggling to convince customers to use its LEDs in a wide range of devices. The advantage was obvious: LEDs are solid-state devices that don’t burn out. The existing LEDs, however, weren’t bright enough to compete with conventional lighting. Kish changed that by bonding red-orange-yellow LED semiconductor wafers on a transparent substrate.
The results were the highest-performance red-orange-yellow LEDs ever produced. The invention propelled HP’s high-brightness LED products into a multimillion-dollar business. Most recently, HP and Philips Lighting formed a $150 million LED venture to compete head-to-head with conventional lighting. If LEDs do indeed light up our lives, it will be fair to say that Kish was one of those who struck the spark.
Computer security is a little like the War on Drugs–you catch as many perpetrators as you can, but you never catch them all. As more and more of the economy goes online, the stakes are rising and anyone offering a better form of Internet security deserves attention. As the founder and chief technology officer of Internet Security Systems (ISS), Christopher Klaus is at the forefront of those peddling peace-of-mind.
Klaus has been noodling with computers for most of his life–at 9, he was programming games on his Commodore 64. Klaus’most ingenious creation is the Internet Scanner,a software package from ISS that uses hundreds of tests to probe a network’s vulnerability, analyze security weaknesses and recommend solutions. Klaus released his first version as shareware while he was an undergraduate at Georgia Tech. Internet Scanner has won several awards as the best Internet security product of the year. The pervasive angst about computer security has been very good to Klaus: Forbes estimates his net worth at $187.5 million.
Can you keep a secret? Paul Kocher can. As a leading cryptographer, he is helping make the Internet safer for business transactions and personal information. As more business is conducted on the Net, the need for online privacy grows apace. As a result, cryptography–making and breaking codes–is becoming central to the Web’s further development. Paul Kocher is among the handful of people who are making significant contributions. Among Kocher’s accomplishments is the formulation
of a method to defeat RSA encryption, one of the most widely used forms of securing transactions. (Breaking existing codes is part of a cryptographer’s job description.) He is also one of the designers of the Secure Sockets Layer (SSL) 3.0 protocol, which many Web sites use to encrypt and authenticate credit information.
Kocher’s career shows how innovators are able to cross disciplinary lines: Although his bachelor’s degree from Stanford is in biology, he taught himself everything he’s needed to know about computers. He’s also an entrepreneur: founder, president and chief scientist of Cryptography Research as well as cofounder and chief scientist of ValiCert. His lack of academic credentials in computing hasn’t stymied his career. Says professor emeritus Martin Hellman at Stanford: “He knew more than most people who had completed PhDs in the area. When people call me for consulting, Paul is my first recommendation.”
Getting complex machines to work well is tough enough; getting machines to work together intelligently is much more difficult. Steven Leeb is attacking this problem in several areas of a field known as “mechatronics”: a combination of mechanical engineering, electronics and intelligent computer control. Leeb’s forays into mechatronics could ultimately pay off in a remarkable range of fields: from artificial muscles to drug delivery to control of electricity and lighting in buildings.
On the biomechanical side, Leeb and his colleagues have developed gel polymers in which ferromagnetic materials are embedded. The polymers contract when exposed to magnetic fields, rendering them useful for artificial muscles and drug-delivery systems; related gels might be used to make braille and 3-D displays. Leeb has also developed a technique called nonintrusive load monitoring, a way of determining the major electrical loads in a building from measurements made only where the current enters the building. This is not only a simpler way to collect such data, but one that opens the way for intelligent power controllers and quality monitors. “I am enjoying just watching all of the neat stuff that comes out of Steve’s lab,” says colleague Jim Kirtley. “Steve Leeb has already had a major impact and promises to be very influential in the future.”
Consider two facts: 1) Hideo Mabuchi is the only Caltech graduate student in physics in the last decade to be offered a professorship before he received his doctorate; 2) Nobelist David Baltimore, president of Caltech, says: “Make no mistake: Professor Mabuchi’s sights are set at nothing less than changing the world.”
Mabuchi’s plan to change the world leads through the quantum computer. A small but growing group of physicists hope to manipulate the quantum effects that govern the world of atoms to build powerful computers and communications devices. One hurdle to building even simple quantum devices is that experimentalists have been forced to observe quantum effects in tightly controlled systems, counting themselves lucky if they can catch even a fleeting glimpse of these phenomena. Mabuchi is working to change that.
He has conceived and performed experiments that measure quantum effects in real time in an open system, creating what he calls “realtime movies” of the interaction of an atom and photon. In his next experiments, he will try to control quantum interactions using this real-time feedback. That kind of control could be a first step toward making quantum devices a reality.
If robots are to play a larger role in our lives, we will need to make them more dexterous and responsive. Akhil Madhani has been a leader in this pursuit since his days as a graduate student in MIT’s Artificial Intelligence Lab. As his doctoral project, Madhani created “Black Falcon,” a robot that performed surgery under the command of a surgeon at a remote location.
Although the military had been working on “telesurgery” for some time, Madhani’s system represented a significant advance because it was able to do minimally invasive surgery (MIS)–a widely used technique in which the entire operation is carried out through a small incision. What’s more, Black Falcon improves on conventional MIS, because it can reach into areas of the body that would be difficult for a surgeon alone to reach. The obvious potential of this system hasn’t been lost on the private sector: Intuitive Surgical has licensed it for commercial applications. In the meantime, however, in one of the boundary-crossing transformations that characterize contemporary innovation, Madhani has moved on to Walt Disney Imagineering, where he is working on what he will describe only as “next-generation, interactive robotic characters.”
In March, Shoko Manako was named a “Young Researcher of the Year” by the Japan Society of Applied Physics. The purpose of the award is to encourage young Japanese scientists in their work. Then again, given the success Manako has achieved, she may not need much encouragement.
After graduating from Toho University in 1989, Manako went to work on new synthesis methods for high-temperature superconductors. She presented a paper based on that work to a 1989 meeting of the Society of Applied Physics and her career took off. Since then, Manako has spearheaded research at NEC to further develop next-generation methods using electron beams (rather than light) to etch features onto computer chips.
In particular, Manako has fabricated patterns in a polymer resist (the material used to pattern semiconductor chips) as small as 7 nanometers wide. “This width is the most narrow resist pattern that has ever been obtained in the world, and gives a way to a new device such as a quantum-effect device,” says Kiminari Shinagawa of Toho University.
The human capacity for working together in groups is deeply rooted in our evolutionary history and we often take it for granted. Robots are starting to get some of that same capacity thanks to Maja Mataric, director of USC’s Robotics Research Laboratory. Mataric and her students have been developing techniques to enable groups of robots to communicate with one another, coordinate activities and even learn from one another. Examples: Two robots using algorithms her group developed to cooperate in an effort to move a box, groups of robots working together to move in formation, groups that learn “social rules” such as sharing information and yielding to one another. “I cannot think of a single paper dealing with group robotics which does not refer to Dr.Mataric’s work,” says George Bekey, founder of USC’s robotics program.