No Better Time
Dedication
Dedicated to Matt, Sophie, & Claire
Contents
Dedication
Contents
Epigraph
Preface
Chapter 1 World Wide Wait
Chapter 2 Ascent to Israel
Chapter 3 Publish or Perish
Chapter 4 May Madness
Chapter 5 Secret Sauce
Chapter 6 Kings of Cache
Photo Section
Chapter 7 Science or Snake Oil?
Chapter 8 The Go-Go Days
Chapter 9 Overnight Zillionaires
Chapter 10 B’SIYATA DISHMAYA
Chapter 11 Gutted
Epilogue
Acknowledgments
Index
Copyright
Epigraph
“An algorithm must be seen to be believed.”
—DONALD KNUTH,
The Art of Computer Programming
Preface
In the spring of 2011, a friend asked if I was interested in a job co-producing an independent film tribute for the anniversary of the 9/11 attacks. The subject, he explained, was a passenger on the first plane to crash into the North Tower of the World Trade Center. From there, the story took on a life of its own.
It is the story of Daniel “Danny” Mark Lewin (1970–2001), who was almost certainly the first victim of the 9/11 attacks. It’s the story of an extraordinarily gifted young man who believed anything was possible and let nothing stand in his way. Of an all-American kid who moved to Israel against his will, ended up falling hopelessly in love with the country, and served as an officer in the most elite unit of the Israeli army. Of a young soldier who was trained to hunt and kill terrorists, and who—in a tragic twist of irony—later died at their hands. Of a loud, irreverent young computer science student who formed a beautiful friendship with a soft-spoken, reserved professor. Of a husband and father who spent years struggling to make ends meet and became a billionaire almost overnight. Of a theoretical mathematician who wrote a set of algorithms that would change the Internet forever.
Until now, it’s also a story that has remained largely untold. When Lewin was still alive, business journalists categorized him and his company, Akamai, as breakout stars of the first dot-com boom. After his death on 9/11, the mainstream media eulogized him as one of the 2,975 victims of the attacks but never fully investigated his actions on American Airlines Flight 11. His family and friends have remained largely silent. They are certain Lewin courageously tried to stop the terrorists, and that he was likely the first victim of the attacks, but they have been reluctant to share publicly why they believe this to be fact.
Some stories, I have learned, simply can’t be told without the passage of time. You can dig at them, push them, pound the pavement in search of ways to bring them to life, but you can’t be true to them until their keepers are ready to share a part of their lives long shrouded in privacy. It’s not until that point that they will reach willingly into the wells of their memory, perhaps even back to those moments hardest to relive, and begin to tell the story.
Chapter 1
World Wide Wait
“If people do not believe that mathematics is simple, it is only because they do not realize how complicated life is.”
—JOHN VON NEUMANN,
1947, first national meeting of the Association
for Computing Machinery
IN THE FALL OF 1996, a graduate student at the Massachusetts Institute of Technology came crashing into the quiet, cramped office of Professor Frank Thomson “Tom” Leighton with a tsunami-like enthusiasm. His name was Daniel “Danny” Lewin. Leighton distinctly remembered the favorable impression he formed of the energetic young student from Israel.
Leighton was not easy to impress—at least not in an academic setting. As a professor at MIT for more than a decade, he spent most of his days in the company of greatness. Leighton and his colleagues at MIT’s Laboratory for Computer Science (LCS){*} were legends in their field whose collective contributions read like a laundry list of computing and mathematical breakthroughs of the twentieth century. If you were one of the privileged handful of students accepted into the program each year, you quickly faced the fact that, at LCS, brilliance alone wasn’t rewarded; it was assumed.
Over the years, Leighton witnessed a steady host of students—most of them serious, industrious whiz kids—cycle through LCS. Leighton enjoyed working with them, excited by each new class and the possibility that, through his lectures or research projects, he might fan some spark in their nascent minds. But for Leighton, the most exciting possibility of all was this: the opportunity to work with a student smart and inspired enough to push Leighton himself beyond the boundaries of his own scientific pursuits. These kinds of students are statistical rarities, and Leighton was always on the lookout for them.
The semester of 1996 began like most at MIT, with thousands of students descending onto the university’s sprawling urban campus on the Charles River, directly across from Boston. They came from all over, the best and brightest minds drawn to MIT for its culture of cutting edge research and rigorous scientific inquiry. Churning out patent applications and spawning companies at an astounding rate, MIT has long been synonymous with the spirit of innovation. Founded in 1916 to serve the needs of a new industrial age, MIT quickly became one of the top research institutions in the world, and it still is today. And when it comes to computer science, MIT is almost unrivaled, consistently ranked as the first or second (to Stanford University or Carnegie Mellon) in the nation.
Lewin arrived in Cambridge late that summer on his own, leaving his wife, Anne, and two sons—Eitan and Itamar, a newborn—to remain in Israel until he settled into their new apartment. Lewin was American, born and raised in a suburb of Denver, Colorado. But he spent most of his formative years in Israel, moving there at age fourteen and attending a Hebrew high school. After graduating at age eighteen, he joined the Israel Defense Forces (IDF) and served almost four years, after which he attended college at the Israel Institute of Technology (the Technion) in Haifa. For Lewin, Israel was home, and moving back to America wasn’t easy. When he landed in Boston on a flight from Tel Aviv with little more than a few suitcases, Lewin experienced a bout of culture shock. The city was chilly, and he had nothing warmer than a fleece jacket in his wardrobe. His assigned apartment, located in a graduate housing complex near campus, was more modest than he expected, small for a family of four and their belongings. But as a former officer in one of the most elite units of the IDF, Lewin was quick to adapt. He set out to make Cambridge home and join the ranks of another elite cadre—LCS. He was immediately assigned to Professor Leighton (first as his teaching assistant, then as his research assistant), who was at that time head of the algorithms group at LCS.
Leighton soon found himself looking forward to Lewin’s visits, which punctuated his typically quiet, serious academic life with bursts of exuberance. And it wasn’t just the student’s gusto that intrigued Leighton. When he spoke about topics that energized him, which seemed to include almost everything, Lewin became so animated—arms gesticulating, eyes ablaze—that his enthusiasm was infectious. “He definitely wasn’t the shy, retiring type,” Leighton recalled. “What stood out to me was how engaging he was, almost like this live wire. When Danny was excited about something, you couldn’t help but get excited, too.” Like any incoming student eager to establish himself, Lewin not only dropped by Leighton’s office to pick up stacks of student papers to grade but also typically lingered long enough to strike up a conversation. “We started talking about research problems,” explained Leighton. “That’s when he began to really distinguish himself, because he had such smart things to say.”
In some ways, L
ewin’s appearance belied his intelligence. Lacking the physical traits of the stereotypical mathematician, he could have easily been mistaken for a high school athlete. Although he stood just five feet ten inches tall, Lewin was built like a bull—burly and broad-shouldered, sheer muscle from head to toe. He was boyishly handsome, with a soft, round face, blue eyes and brown hair that was prematurely receding, giving way to a long, smooth forehead. His smile was unyielding and almost impish, creating in those around him the urge to smile, too. In contrast to the quiet shuffle of most students across MIT’s campus, Lewin moved with a distinctive spring in his step, as if he were in a hurry to get somewhere.
In the cluttered, unkempt halls of LCS, over desks piled high with papers and textbooks, or crossing the campus quads, Lewin and Leighton spent hours absorbed in mind-bending conversations about math and computer science. “I felt like I was talking to an equal,” recalled Leighton. “He’d think of clever ways to take an idea in some new direction. That’s the best kind of research activity; it’s rewarding, enjoyable and just a lot of fun.” Lewin joined Leighton’s algorithms group, which was grappling with a challenging set of problems centered on this new mode of communication, the Internet, and some of the barriers to its growth.
At the time, neither Leighton nor Lewin could have predicted just how greatly the ground beneath them was beginning to shift with the stirrings of the dot-com boom.
While the Internet as we know it today was conceived as a collaboration between government, academic, and private sectors, its creation began in earnest at MIT in the 1960s. Over the course of the decade, a close-knit community of engineers and scientists worked in relative anonymity to develop the underpinnings of a vast computer network. These “fathers” of the Internet shared the belief that technology and humanity were inextricably linked, foreseeing a time when we would have the connective power of this electronic network at our fingertips. It was a revolutionary idea for a time when computers were expensive, colossal, calculating machines. One of the Internet’s earliest proponents was a mild-mannered Midwesterner, Joseph Carl Robnett Licklider. Known as J. C. R. or “Lick” to friends, Licklider—a preeminent Harvard psychologist-turned-MIT computer scientist—wrote a series of essays in the early 1960s documenting his vision for a globally interactive set of computers.
Around the same time, graduate student Leonard Kleinrock—a fast-talking kid from the Bronx on a full scholarship at MIT—was deep into his doctoral thesis on how to stabilize and increase the flow of information within this hypothetical network of computers. Kleinrock knew that the system of circuits used to transmit a telephone call, which has a centralized point of control, would not function well when moving large amounts of data between computer networks. His solution, which he outlined in his thesis, was a clever theory called “packet switching.”{†} Simply put, packet switching involves the breaking apart of data into separate, small blocks, each one tagged with an address. When the information blocks reach their destination, the network reassembles them. Packet switching later became one of the Internet’s fundamental networking technologies.{1}
This flurry of research at MIT gave rise to Project Mathematics and Computation (or Project MAC), which took off in the summer of 1963 with a $2 million grant from the Department of Defense for the development of a robust, fault-tolerant computer network. Under the leadership of Licklider, the Project MAC team immediately began building a large-scale network of computers called ARPANET, the precursor to the Internet.
The work on ARPANET was unusually collaborative; the Project MAC team joined forces with commercial computer companies such as Bolt, Beranek and Newman (BBN) to get the network off the ground. A computer at UCLA became the first “node” or point of connection for ARPANET. By December 1969, three more nodes had been added at the Stanford Research Institute (SRI), UC Santa Barbara, and the University of Utah, establishing a four-node network. As ARPANET grew in size and scope, so did the number of key players involved in its success. One of them was Vinton Cerf, who later became vice president and “chief Internet evangelist” of Google. With a PhD in computer science from UCLA, Cerf spent several years working directly with Robert Khan, an applied mathematician and computer scientist at MIT and BBN, to develop a “virtual handshake” (Transmission Control Protocol/Internet Protocol, or TCP/IP) that allowed computers in disparate places to talk to one another.{2}
By 1990 ARPANET had been decommissioned to make way for a much broader network supported by the National Science Foundation, and a growing number of commercial Internet Service Providers (ISPs). A year later, Al Gore—then a young senator from Tennessee—paved the way for the information superhighway by steering a bill through Congress that supported the creation of the Internet beyond academic and scientific institutions. The Internet had fully evolved from an academic experiment in shared intelligence into an explosive commercial enterprise.
Because of the time, there was even more of a buzz in the air than usual when Danny Lewin joined MIT’s LCS in the fall of 1996. By the mid-1990s, the lab was not only at the forefront of computing; it was also on the frontlines of the burgeoning dot-com boom. Under the leadership of Michael Dertouzos, formerly a computer scientist with Project MAC, LCS remained a hive of innovation. It inspired dozens of tech companies, including 3Com, a multibillion-dollar networking business co-founded by computer scientist Robert Metcalfe, and RSA Data Security, led by Ron Rivest, the inventor of the public-key encryption that allows us to securely enter information like credit card numbers online. LCS was also home to the World Wide Web Consortium (W3C), the Web’s international standards organization created by British computer scientist Tim Berners-Lee.
If Silicon Valley was the heart of the dot-com boom, Cambridge was its intellectual capital. The combined talent of the city’s crown jewels, MIT and Harvard, helped transform Cambridge from a biotech center into an incubator for the digital age. Clusters of startups with names like Frictionless and Viant rose out of its labs, offices, and dormitories. The greater Boston area was second only to Northern California in the number of computer companies that called it home. By the mid-1990s, 1,500 software companies had taken up shop in Cambridge and its surrounding areas.{3}
Ironically, the high-tech activity had little to no bearing on Lewin’s decision to attend MIT. Having been accepted to every one of the top ten graduate programs in computer science, he had had his pick of schools—Cal Tech, Stanford, and Carnegie Mellon, to name a few. Some even wooed him with generous scholarships. With the rigor of a mathematician, Lewin drew up a list, annotating it with the pros and cons of each school. The verdict wasn’t all his to determine, of course. He had a family to consider, and Anne was excited by the idea of moving to a sun-drenched campus in California. For Lewin, however, the decision had, in some ways, already been made.
Lewin had MIT in his sights as early as 1986, when he was just a teenager otherwise preoccupied with pretty girls. That September, he wrote a letter to his best friend, Marco Greenberg, in the U.S. In it, he told Greenberg he was trying to study for the bagrats (Israeli matriculation exams), but that flirtatious girls were distracting him. “The girls in school are trying the ‘FIRST STRIKE’ technique on me,” Lewin wrote, adding, “This is a very good opening to the school year!!!” It was a typical teenage letter, until page three, when Lewin casually noted: “If I graduate this year, I’ll go back for school. I don’t know where, but I would love any of the technical schools, [e.g.] MIT.” Lewin’s mention of MIT was one of several hints that tipped Greenberg off to his friend’s higher-than-average intelligence. A few years later, while the two were talking about college, Greenberg mentioned that, if Lewin wanted to go to an American school, he would have to complete the SAT (Standardized Aptitude Test). Much to Greenberg’s surprise, Lewin told him that he’d already taken the exam—for fun. When Greenberg pressed him about his scores, Lewin matter-of-factly answered: near perfect on the English, and perfect on the math.
Engineers built the hardware for com
puters and the Internet, computer scientists programmed the machines, and entrepreneurs forged new businesses on the vast new landscape. But nowhere in this developing ecosystem could one find a clearly defined role for the theoretical computer scientist.
In the 1990s, computer science was still a relatively new discipline. Naturally, the mathematical subset of this field, theoretical computer science (TCS), was even more novel and arcane. In part, its complexity was the cause, and it still is today. When you ask theoretical computer scientists to explain what, exactly, they do for a living, the answer is often too lofty for the average mind to grasp. Take, for example, Professor Leighton’s puzzling description of his early work at MIT: “I was interested in properties of networks and how to send messages through them. I wasn’t writing code, instead I was using math. So I’d say, for example, that if the network has N nodes, and every node needs to send a message to another node, there’s a way of routing all the messages so they don’t collide and they all get to their destinations at the same amount of time. If your problem is size N, where N is a variable, it will take some function of N, like 10 n.”
Leighton earned his PhD in math from MIT in 1981 and became a professor the next year. By the late 1980s, he was leading the theory group at LCS, the largest of its kind in the world, becoming somewhat of a celebrity in the rarefied field. TCS sits at the intersection between math and computation and relies heavily on algorithms, which are step-by-step procedures for calculations. Characterized by abstraction and complex ideas, TCS has long been relegated to a lesser status than other areas of computer science. If you’re a theoretician, then it’s often assumed you can’t also be an applied scientist. And if you’re not an applied scientist, then you’re not coming up with practical solutions. As Leighton writes in a paper examining the merits of his field: