Chapter 1The Dentist's Son

Foreword

There is a shape to the founder story that anyone who has covered technology for a living learns to recognize, because the same shape arrives, with small variations, several times a year. A young person builds a thing in a room. The thing works. Money notices. And then the second act begins, and the second act is almost always the same: the grown-ups arrive. The venture capitalists arrive with their term sheets and their board seats. The recruited chief executive arrives, hired away from some larger firm to bring what the industry calls, without embarrassment, adult supervision. And the young person who built the thing is eased, gently or otherwise, out of the center of it, retitled as a founder or a chairman or a visionary, given an office and taken off the wiring diagram. I have written that handoff more times than I can count. It is the load-bearing story of the whole industry — the story of how the kid's company becomes the grown-ups' institution.

Mark Zuckerberg is the one where the handoff never came.

That is the fact that pulled me toward this book, and it is a stranger fact than it sounds. The company he began in a Harvard dormitory in the winter of his sophomore year is now a structure that more than three billion human beings pass through every day, and the share arrangement that governs it has not been meaningfully altered since he set it up as a twenty-year-old. The board cannot remove him. The shareholders cannot outvote him. No investor was ever able to install the adult who, in every other version of this story, walks in around year three and takes the wheel. The wheel was never available. It has been held, without interruption, by the same hands that first reached for it — and those hands belonged, and in some structural sense still belong, to a nineteen-year-old who had concluded that the most important thing a person could refuse to give up was control.

So I went into the reporting expecting to find the seam. I expected, somewhere in the twenty-two years, the moment the sophomore became the CEO — the crisis, the humbling, the meeting after which he was a different kind of man running a different kind of company. I read for it. I worked from what the record makes available: the published memoirs of the people who were in the rooms, the depositions and filings from the lawsuits that have followed him since the first spring, the two decades of contemporaneous press that tracked every product decision and every apology. I was looking for the before-and-after. I kept not finding it.

What I found instead was continuity, and continuity is the harder thing to write, because it does not come with a scene. The instinct that built a messaging program to connect a dentist's office to the family kitchen is the same instinct that built the dorm-room directory, is the same instinct that turned down a million dollars from Microsoft at seventeen rather than surrender a piece of software, is the same instinct visible at forty-one in the man who could redirect the largest communication layer ever assembled and answer to no one while he did it. The company did not grow the boy up. The company grew up around him, a vast professional apparatus of lawyers and lobbyists and executives arranged in concentric rings around a fixed point, and the fixed point did not move.

This is why I think he is worth a book, and worth this one. We keep waiting for the institution to correct the founder, the way institutions are supposed to, and it has not, and it is not built to. To write his life is to sit with a discomforting proposition: that the deepest operating logic of one of the most consequential companies in the history of commerce is not the product of a committee, or a market, or a mature corporate mind, but of a set of convictions a teenager formed about people and networks and control before he was old enough to rent a car. I did not want that to be true. It made the reporting harder and the conclusions less comfortable. It is, as far as I can tell, true anyway.

I came to this subject to explain how the sophomore became the man in charge. I leave it having concluded that the question was backward — that the man in charge has, in the way that counts, never stopped being the sophomore, and that the whole world now lives downstream of that fact.

— Sterling Aldridge, Atlanta

Chapter 1 · The Dentist's Son

I.

Dobbs Ferry is a village on the eastern bank of the Hudson River, fifteen miles north of the George Washington Bridge. The commuter trains run along the water. In the mornings the fathers board them and ride to Grand Central, and in the evenings the same trains bring them back to the platform at the foot of Main Street, and they walk up the hill to houses with porches and lawns that slope down to the road. It is the kind of place that exists in the American mind as a stage set for a certain kind of childhood, and it is also a real place, with real property taxes and a real school district, and in the 1980s it was a place where professionals from New York City moved when they wanted their children to grow up somewhere other than the city but close enough that the city's gravitational pull still reached them. It is forty minutes by train to Midtown. On a clear night you can see the haze of Manhattan's light rising above the trees to the south. Mark Zuckerberg was born in the hospital at White Plains, fifteen miles east, on May 14, 1984. The family lived in Dobbs Ferry.

The house at 775 Broadway is a three-story white clapboard on a corner lot. It has a wraparound porch. There is a driveway that curls to the side and a dentist's office at the rear of the property, attached to the house through a door in the kitchen, so that the man who lived there could walk from his breakfast table to his patients' mouths in less than thirty seconds. His name was Edward Zuckerberg. He was known to his patients as Painless Dr. Z, a brand he had chosen deliberately and printed on his business cards and on the signs outside his practice. He was one of the first dentists in the Hudson Valley to advertise on the internet. He had built the practice's first website himself, in 1995, on a dial-up connection, and he kept the office computers running on a small network he had wired through the walls.

His wife was Karen Kempner. She had trained as a psychiatrist, completed her residency, and then, after the children began to arrive, stopped practicing medicine and became the office manager of Painless Dr. Z. She handled the scheduling. She handled the insurance. She handled, in the early years, the four children: Randi, the oldest, born in 1982; Mark, born May 14, 1984; Donna, born in 1987; Arielle, in 1988. Three girls and a boy. The boy in the middle.

The siblings were exceptional in ways that would each, in time, become public. Randi would become Facebook's Director of Market Development, then leave in 2011 to found Zuckerberg Media, and later write a bestselling book and perform on Broadway. Donna would earn a PhD in classics from Princeton and become editor-in-chief of Eidolon, an online journal of classical studies, and publish Not All Dead White Men, a scholarly examination of the alt-right's appropriation of classical texts. Arielle would move into Silicon Valley venture capital. The Zuckerberg household was not merely comfortable. It was intellectually competitive in a way that produced, in each of the four children, a particular kind of ferocity — the ferocity of people who had been rewarded, since childhood, for being the smartest person in the room, and who had never found a room they could not eventually master.

The house had a name, among the children, though the children did not think of it as a name. It was just the house. But later, when the boy in the middle became famous and reporters came to Dobbs Ferry to interview the neighbors and the old teachers, the house would acquire a kind of weight in the story, because it was in this house, in a room on the second floor, that a pattern of activity began in 1990 or 1991 that would, thirty years later, reshape the communication of the human species.

II.

To understand what happened in that room, it is necessary first to understand Edward Zuckerberg. He was not an ordinary suburban dentist. He had grown up in Brooklyn, the son of a mailman, and he had worked his way through Brooklyn College and then through NYU dental school, and he had built his practice from nothing in a town where the older dentists already had their patient lists. He was a careful man. He was also, by the standards of his profession in the 1980s, a technologist. He subscribed to computer magazines. He bought an Atari 800 in 1982, the year his first child was born. He bought an IBM PC/AT in 1985, when his second child was a year old. In the home office at the back of the house, among the insurance forms and the X-ray films, there was always a computer, and the computer was always being upgraded, and the upgrades were always being performed by Edward himself, because he liked doing it and because the children, watching, liked watching it.

Mark, the middle child and only son, sat on his father's lap during some of those upgrades. He was four years old, then five, then six. He was a small boy with wide-set eyes and curly brown hair and a stare that his sisters found unsettling and that his father found promising. He did not speak much. When he did speak, the sentences came out already complete, with subordinate clauses, as if he had been working them out silently before opening his mouth. Randi, two years older, would try to engage him in the games that older sisters play on younger brothers, and he would either ignore her or respond with a question so literal that it stopped the game. What are the rules, exactly. Who decided the rules. What happens if someone does not follow the rules.

When Mark was still a boy, Edward began teaching him to program. He started him on Atari BASIC, the language built into the family's Atari, partly because he had decided he wanted to automate parts of the dental practice's billing and thought he might write the software himself. The boy took to it at once. He did not approach it the way a child approaches a game. He approached it the way a graduate student approaches a textbook — methodically, with a pencil.

Edward noticed. Edward had the kind of instincts about his children that his wife's training as a psychiatrist had taught her to be cautious about — the instinct to identify a talent and then to pour resources into it. But with Mark, Edward did not hesitate. By the time Mark was seven, Edward was sitting with him in the evenings after dinner, working through simple programs on the IBM. By the time Mark was eight, Edward was consulting him on upgrades. By the time Mark was ten, Edward had concluded that the boy had outgrown what he himself could teach, and he did something that was unusual for a suburban father in 1994: he hired his son a tutor.

The tutor's name was David Newman. He was a software engineer and a graduate student in computer science, and he came to the house at 775 Broadway once a week and sat at a desk in the upstairs room and worked with Mark on programming. Newman would later tell reporters that Mark was a prodigy, that it was impossible to keep ahead of him, that he had to prepare lessons the way he would prepare a graduate seminar and still the boy would finish the material in half the allotted time and ask questions Newman could not answer. "It was tough to stay ahead of him," Newman said years later. "He had a hunger for it that I had never seen in a student of any age." Around the same time, Edward enrolled Mark in a graduate-level software design course at nearby Mercy College. The professor, on the first day, saw a small child walk into his classroom and assumed the child was lost, there to be collected rather than to enroll. Edward, by the account later given, explained that his son was there for the course.

The boy was eleven. The Quantex 486 DX running Windows 3.0 that Edward had given him the previous year — at a time when a machine like that cost around $1,000 and required a knowledge of DOS to operate — was already too slow for what Mark was trying to do. By age twelve he had exhausted it and persuaded his father to upgrade. The upgrades would not stop for the rest of his life.

III.

The program that Mark built that year, in the upstairs room at 775 Broadway, was called ZuckNet. It was a messaging system. It ran on the small network of computers that Edward had wired through the house, connecting the home computer to the computers in the dental office at the rear of the property. The purpose of ZuckNet was simple and revealing. In the dental practice, when a patient arrived for an appointment, the receptionist at the front desk had to walk back through the door to the main house and tell Edward, because the office and the house were connected but the communication between them was not. ZuckNet fixed this. The receptionist could type a message on her computer and it would appear on Edward's computer — or, if Edward was upstairs in the home office, on the computer there. The family computers all ran ZuckNet too. Randi could message Mark. Mark could message his mother. The messages were short. They were instant. They replaced the shouting up and down the stairs that had been the previous communication protocol of the Zuckerberg household.

It is worth pausing over this. When Mark built ZuckNet, around 1996, internet-based instant messaging was only beginning to reach the general public. ICQ, the first widely used internet messaging program, developed by the Israeli company Mirabilis, was released in November 1996; AOL Instant Messenger did not arrive until 1997. The twelve-year-old boy in the upstairs room in Dobbs Ferry had built, for the use of his family, a functional instant messaging system of his own at the very moment such tools were first reaching the wider world. He had done this not because he had read about messaging systems or been inspired by a concept he had encountered elsewhere, but because he had identified a specific, concrete need — the dentist's office and the house could not communicate in real time — and had built the simplest possible solution to it. This was the engineering instinct at its purest. Not curiosity about what was possible in the abstract, but attention to what was broken in the particular, and the desire to fix it with code. He had not invented it from first principles — he had built it on existing networking primitives, the way any programmer does — but he had identified a need, designed a solution, and shipped it, in the full engineering sense of the word shipped. The software worked. It was used every day. It changed, in a small but real way, the communication patterns of the people around him.

What it reveals is not just technical precocity. Lots of twelve-year-olds could, with enough tutoring, write a basic messaging program. What it reveals is a specific instinct: the instinct to look at a human communication problem — Edward needing to know a patient had arrived — and to solve it by inserting software between the people involved. The software was not a tool the family picked up and put down. It became the layer through which their communication passed. Once ZuckNet was installed, the Zuckerbergs did not go back to shouting up the stairs. The layer, once added, was not removed.

This is the pattern. It will repeat.

IV.

He was not, despite what the later mythology would suggest, a friendless child. Randi, his older sister, was his closest collaborator and his fiercest rival, and the two of them ran a kind of intellectual arms race through the Dobbs Ferry years, each trying to out-read and out-argue the other. Donna and Arielle, the younger sisters, were too small at first to be competition, but by the time the family moved into the second half of the decade the four of them constituted a small, competitive, verbal, argumentative unit that their parents had to referee. Dinner at the Zuckerberg house was not quiet. The children were encouraged to argue. Edward, in particular, believed that arguing was how children learned to think, and he would stage debates at the dinner table on topics that ranged from the plausibility of a particular piece of science fiction to the ethics of a particular political decision. Karen was the calmer influence. She watched. She intervened when the arguments got personal. She was, one neighbor remembered, the emotional regulator of the household, and she regulated mostly by not overreacting, which was itself a form of regulation because it prevented the children from learning to escalate for attention.

Mark went to Ardsley High School, the public school a few miles away, for ninth and tenth grade. He was an excellent student. He was also, by every account, socially peculiar. He was small for his age. He made eye contact too intensely or not at all. He spoke with a precision that other fifteen-year-olds found either hilarious or alienating. He was on the fencing team. He was bad at fencing at first and then, through the same combination of analytical focus and sheer repetition that he brought to programming, he became very good, and he captained the junior varsity team. The fencing was his primary extracurricular activity during the Ardsley years, and it produced something besides athletic competence: it produced a physical discipline for dealing with an opponent who was trying to defeat you. In fencing, as in the coding competitions that occupied his evenings, the objective was not to be clever. The objective was to be faster and more precise than the person across from you. The sport and the discipline reinforced each other in ways that would, twenty years later, lead him to take up mixed martial arts with the same methodical seriousness.

He would build, across his high-school years, several more pieces of software. He had built a Risk clone that could play the Roman Empire against a human opponent and defeat it. And in his senior year at Phillips Exeter he built a music recommendation engine called Synapse Media Player — with his classmate Adam D'Angelo, who would later co-found Quora — that learned a user's listening habits and suggested songs based on a primitive collaborative-filtering algorithm. Synapse was good enough that AOL and Microsoft both tried to buy it. PC Magazine gave the software a 3-out-of-5 review. Microsoft offered, according to accounts later given, close to a million dollars for the software and a job for Mark when he was older. He was seventeen. He turned it down. He released Synapse as a free download.

He turned it down. This is the first of several moments in the story when Mark Zuckerberg will be offered an enormous sum of money for a piece of software he has built and will refuse it. It is tempting to read these moments as a consistent philosophical position — the founder who refuses to sell because he has bigger ambitions. But at seventeen, in a dormitory at Exeter, the refusal looks more specific than that. He did not yet know what he wanted Synapse to become. What he knew was that selling it would mean losing control of it. Losing control of the layer. Handing over to someone else the ability to decide what the software did and who it did it to.

Control. The word will matter.

The Microsoft offer had come through a recruiter who had reached out to Edward after a description of Synapse appeared on Slashdot, the technology news site that was, in 2002, one of the primary venues through which software developers shared and discovered new projects. The Slashdot post had generated thousands of visitors to the Synapse download page in a single day — a traffic event that the web hosting service struggled to handle. The sudden visibility was a preview, in miniature, of what would happen in February 2004 when Thefacebook's link began spreading through Harvard's email lists. The mechanics were the same: a piece of software, passed from person to person through digital networks, reaching more people than the person who built it could have reached by any other means. The scale would differ by many orders of magnitude. The mechanism was identical.

The AOL offer had been similar in structure to Microsoft's, though smaller: acquisition of the software, relocation to California for the programmer. Mark had turned down both. He had released Synapse on the web for free and watched as the downloads accrued — thousands, then tens of thousands, from users across the country who had found it through Slashdot and through the small community of music software enthusiasts who traded recommendations on early forums. The downloads were not revenue. They were evidence of reach. The evidence that a piece of software, given freely to the public, could spread further than a product sold through a corporation's distribution channels — this was the lesson that would, three years later, make Thefacebook free to its users and generate revenue from advertising against their attention rather than from charging them for access.

V.

In the spring of 2000, when Mark was in the tenth grade at Ardsley, Edward and Karen made a decision that would reshape their son's life. They decided that Ardsley, the perfectly respectable suburban high school, was not going to be sufficient for a boy who had already outgrown everything it could teach him about computers. They began investigating boarding schools. The one they chose, after much research, was Phillips Exeter Academy, in Exeter, New Hampshire — one of the oldest and most elite boarding schools in the United States. It had been founded in 1781. Its alumni included presidents and novelists and a large fraction of the East Coast's professional class. It was the kind of school where the children of the American elite were sorted and groomed. It was also, in 2000, just beginning to acquire a reputation in a new area: computer science.

Mark applied. Mark was accepted. In the fall of 2000, at the age of sixteen, he boarded a train with his parents and rode north, past New Haven and Hartford and Boston, to the small town in New Hampshire where, for the next two years, he would live in a dormitory and eat in a dining hall and study Latin and ancient Greek and, in his free hours, teach himself everything he could about how human beings connected to each other through networks.

The house at 775 Broadway receded. The family he had built ZuckNet for was now at the other end of a long train ride. He did not write home often. When he did, the letters were short.

He was already looking at the next layer.

What the two years at Ardsley had given him, beyond his programming skills, was a first encounter with institutional limits. Ardsley was a perfectly adequate school. It was not equipped for a student who had already built a functional instant messaging network as a boy and had been taking graduate-level computer-science coursework since middle school. His teachers did not quite know what to do with him. When a student outpaces his curriculum at age fifteen, the school faces a choice: accelerate him or contain him. Ardsley, like most public schools, was not organized around acceleration. Edward and Karen, watching, made the decision.