THE VOICE IN THE MACHINE

What the people of New Jersey knew about Bell Labs that some of us had to learn from the inside.

Why am I writing this? I am aware of the trap. I look back at the place where I was a long time ago and remember a golden age that, examined closely, turns out to have been ordinary, or at least less fantastic than memory paints. I have worked in and visited many laboratories over the course of my career, most of them excellent, with amazing people, some of whom are still among my most loved friends. But as honest and objective as I can be, none of those places ever gave me the feeling of magic that Bell Labs did. Perhaps that was because I was much younger. Perhaps because I had just moved to a new country, where everything was big and strange to my eyes. Perhaps there really was magic in that place, and I was lucky enough to arrive while the magic was still there, or as we say for beauty, also magic is in the eye of the beholder. I have never quite answered this question, and I am not sure I want to.

Let’s rewind the reel to one late afternoon in February 1988, when I landed at JFK in the middle of one of those North-East Coast snowstorms that make you wonder, while you are still in the air, whether the plane will ever find a landing strip in the milky fog and snow, and manage to land. I had seven suitcases holding everything my family thought we would need for a year, and probably a little more. My wife and our son, two and a half years old, would join me, from Italy, the place where I was born, grew up, and studied, a week later. I had barely seven days to find somewhere to live, buy a car, and take care of a hundred other things I did not yet know existed.

Aaron Rosenberg, the Bell Labs researcher who had invited me to spend a year as a visiting scientist, came to collect me at the airport. Americans, in general, are not so solicitous. America is a country of pioneers, and the unspoken understanding is that if you can manage on your own, good, and if you cannot, you would have done better to stay home. But Aaron was a man in his fifties, kind and understated, erudite across a remarkable range of fields, music, history, biology, and of course his own, voice recognition and verification, and with what I came to think of as a European soul. He had braved the snowstorm to pick me up. “This is going to be the worst part of your trip,” he told me as he loaded the suitcases into his classic wood-paneled station wagon, and we set off while the snow was still falling in a white curtain lit by oncoming headlights. He drove me to dinner at his house in New Jersey before dropping me at the hotel in Murray Hill, a short walk from the laboratories.

We crossed the Verrazzano Bridge, which pulled an involuntary wow out of me, then Staten Island, the Goethals Bridge, and suddenly we were in the suburban world of New Jersey, when the snow stopped and the sky opened, clear and full of stars. It was the first time I had seen suburban America. On earlier trips I had visited New York, Boston, Pittsburgh, but never the suburbs, and my first impression, that crisp winter evening, was that I had landed in an enormous campground: low wooden houses scattered among the trees, their windows glowing faintly in the dark, and a sense of quietness after the crazy turnpike traffic. Aaron took me home, where his wife Judith, motherly and warm, had made dinner for me, and then to the hotel where I would stay for a few days, or at least until I would find a permanent accommodation. Everything would begin the next morning.

There is a small detail I have never quite lived down. The day after I arrived with my seven suitcases, I did the count again and found only six. I had left one at the airport. But by then I had so much else to think about that I decided the missing suitcase could wait.

During that first week, while I was looking for somewhere to live, I heard for the first time something I would go on hearing for years. When landlords asked where I would be working and I answered Bell Labs, they would smile and say, without fail, “Ah, Bell Labs. Strange people, but very smart.” The woman who would become my landlady said it too. Miss Keller was a retired chemist who lived alone in Springfield. The apartment she rented me was just above hers, on the second floor, in a condominium of small buildings spread across a green. She was missing several fingers on each hand, and on the right she wore a kind of polished steel pincer, connected by cables to her forearm, that opened and closed as she moved it. It frightened my son a little, but she was cheerful and kind, and always enjoyed a good glass of wine, which she drank through a straw, the pincer being ill-suited to lifting a glass. We would learn later that she had managed to blow some of the fingers off both hands with a solution gone wrong. When I told her where I worked she gave the same verdict as everyone else, “strange people… but very smart,” and then added a question of her own: “Is that fellow who walks backwards still there?”

Life inside the myth

Bell Laboratories, the Labs to anyone who lived nearby, were a myth. A place that has since vanished, the way so many beautiful things do, leaving the myth intact and the reality behind it gone. What I want to describe here is the place as it was in those first years, when it still felt like the center of the technological world.

Aaron came to pick me up at the hotel, as promised, the next morning, a day of sun with the snow still white on the ground, and drove me to the legendary 600 Mountain Avenue. AT&T’s entire R&D arm was called Bell Laboratories, and it counted more than twenty thousand people spread mostly across the northeastern United States, with laboratories as far away as Colorado. Its headquarters, Murray Hill, tucked among the unexpectedly green rolling hills of a New Jersey that outsiders knew mainly for the ugly refineries near Newark, was considered one of the ivory towers of American research, and not only the industrial kind. The research group I would join for a year, though I did not yet know it would be for many more years after that, belonged to area 11, also called core research, the heart of the heart of Bell Labs, where the most forward-looking work of the whole organization was done. There was even a small museum in the lobby, open to the public and mentioned in tourist guides, and a bust of Alexander Graham Bell inscribed with his famous sentence: “Leave the beaten track occasionally and dive into the woods. You will be certain to find something that you have never seen before.” That inscription turned out to be true, for me and for many others who joined the Labs, because we did, without exception, find something we had never seen before.

What made area 11 different was not the equipment, the location, or the salaries. It was freedom. This was one of the very few places in the world of industrial research where, if you had the fortune to work there, you were free to follow your own curiosity, your own inventiveness, your own desire to build something new. Everyone else at AT&T did development, the D of R&D, with goals defined by the company, however advanced. We decided what our research was about, and we defined the goals ourselves, individually. My group, Information Principles Research, worked on the algorithms and techniques for handling voice and images, and its director was Jim Flanagan, a pioneer of digital speech processing and a gentleman from another era, when researchers still wore a jacket and tie, in his case, always coordinated with the handkerchief in his breast pocket.

Why does a place like that acquire a reputation for strangeness? I think the reason is simple. Tell talented people that no one will ever ask them to justify the why and the how of their ideas, and most of them will pour that freedom into the work, growing a little unworldly, a little eccentric, absorbed to the point of forgetting the ordinary rules everyone else obeys. A few will go further and become genuinely strange. To the outsiders the distinction did not matter; brilliant or genuinely strange, we all came back simply as strange. From the inside I could see the gradations the outside missed.

The tour of the Labs that every newcomer received made the point better than any speech. Larry Rabiner, who was already regarded as one of the godfathers of speech recognition and ran the speech research lab under Flanagan, came to my office at a run, as he always did, with the forms I needed to sign under his arm, and took me straight to one of the great attractions of Murray Hill: the stockroom. It was a wonder, and the pride of every resident. Pens, notebooks, and the famous invention notebooks bound in blue cloth with numbered pages, where you recorded your ideas as dated proof of their originality, were all self-service. But you could also walk up to the counter and ask for a voltmeter, a centrifuge, an oscilloscope, a microscope, perhaps a portable mass spectrometer, and carry it away with nothing more than a signature in the register.

The cafeteria was another point of pride, far better than your usual company canteen, with a menu that changed daily and the famous Frank’s Eggs Benedict on Wednesdays. You lined up at the grill in front of Frank, a gruff but likeable man with enormous eyebrows and sideburns and a baritone voice who spoke an incomprehensible slang, whatdoyawonjunman, and who poached the eggs with the patience of a Carthusian monk. Next to the cafeteria there was a restaurant, the Oak Room, where if you had time, you could sit and be served a full three-course meal, including the unmissable sundae. That was the place we went once a week, especially when we had guests, either from outside, or from other Bell Labs locations, like Holmdel. Holmdel was a location some 40 miles south of Murray Hill, where other members of the Information Principles Research department, whom we used to meet regularly, were located, including people like Yann LeCun, today universally known for his contributions to modern AI, and, a few years later, Vladimir Vapnik, one of the fathers of statistical machine learning. 

The offices were assigned by strict hierarchy: a corner office with four windows for a director, two windows for department heads and senior researchers, one window for the younger ones, and for newcomers like me a shared office, often with no window at all. Mine had no window.

I had barely settled in when a distinguished man with white hair, gold-rimmed glasses, and a bow tie appeared and, hearing I was Italian, began speaking to me in fluent, almost accentless Italian. This was Bela Julesz, the inventor, among other things, of the random-dot stereogram, and he invited me into his office, sat me in front of one of those clouds of meaningless dots, and told me to stare until something changed. After a few minutes three-dimensional shapes began to rise out of the noise, and he told me, with evident satisfaction, that my brain had just grown new synapses. On my way back to my own office I passed a man with a grey beard who walked along hugging the walls, touching them with one hand, his eyes never leaving the floor, as if afraid to cross open ground. A few days later I learned that this was Dennis Ritchie, the inventor of the C language and one of the fathers of Unix.

The strange ones

The scientists I have described were eccentric in the ordinary way of absorbed, brilliant people. The ones I mean now were something else, strange in a way you did not forget.

Miss Keller’s question, whether the man who walked backward was still there, had stayed in my mind, so I asked around, and one day I saw him myself. He was a statistician of some fame, gaunt and past sixty, with long thin unkempt hair. I watched him walk backwards down one of the long, windy corridors of Murray Hill — the wind in the Bell Labs corridors was real, though it deserves a story of its own — reach his office door, open it from behind with a contortionist’s maneuver, and disappear inside, still moving in reverse. The reason, I discovered when I later came to know him, was not eccentricity but tragedy. A stroke some years earlier had destroyed the part of his brain that lets us walk forward without thinking, the way we all do, unconsciously, while leaving intact the part for walking backward. He could still walk forward if he wished, but only by consciously controlling every muscle, checking his balance, moving one slow deliberate step at a time. Walking backward was simply easier, because he did not have to think about it. He remained a brilliant statistician, collaborating actively with members of my group. Some years later I saw him in a wheelchair, and then he was gone.

Another figure in the folklore of the Labs was a man whose name I never learned, who went about every single day of the year dressed as an explorer: khaki shorts, a safari jacket, mountain boots, and the classic cork pith helmet. Sometimes I would pass him on my way to work as he pedaled in on his bicycle. Even when the snow was coming down thick and the temperature was well below freezing, he kept to the shorts and the bicycle. He turned out to be a kind of modern Luddite who refused every form of modernity at home, apparently owning no electrical device of any kind, no television, no washing machine, no oven, imposing a pioneer’s life on his family while working for one of the most advanced technology research centers in the world.

Then there were the night creatures. Those who work intimately with computers, today they would be called hackers, and in those days I was one. They prefer the hours of the night, and the Labs had a particular charm in the hours when ordinary people go to dinner and to bed, a charm that inspired the wildest and most creative ideas. Under the yellow light of the flickering fluorescent lamps you would meet Jon, who lived in Manhattan, spoke in bursts at a hundred syllables a minute, arrived in the late afternoon and left after midnight; or Paul, who spent every night at the Labs and, when he took a few days off, went parachuting onto the wild snowfields of northern Canada, hoping to make it down the mountain alive. I would run into them in the department kitchen, warming a slice of pizza that had been sitting on a desk long enough to be a perfect candidate for carbon-dating, or continuously filling their mugs with black coffee, sometimes grumbling that they did not have enough computer cycles to finish their experiments.

I should explain those late nights, because they were not characterized by laziness. I was usually hunting a bug, an error hidden somewhere in the lines of code, and debugging is an activity that takes hold of your mind completely. You cannot think of anything else, you forget to eat and sleep, it is detective work, and you are not satisfied until you find the culprit. Working from home was not really an option then. At home I had a terminal connected by a modem sending signals on a telephone line, but at those times the digital connections were hundreds of thousands of times slower than what we take for granted today, good only for checking results, not for real work. For that I had to be at the Labs.

One night, deep in one of those hunts, I heard a sound from the corridor, like a sudden rush of water, then a second, then another, and another. I went out to look, and the scene was surreal. The old wing where I worked had chemistry laboratories, so the corridors were fitted at regular intervals with decontamination showers, each with a large yellow head and a heavy handle you could pull in case of a spill of chemicals on your body. In the yellowish light I saw a shadowy figure in a black waxed raincoat with the hood up, a small bucket in hand, moving from shower to shower at a measured pace, pulling each handle and collecting some of the liquid in the bucket. When I asked the next morning who it might have been, no one could tell me a thing. To this day I do not know whether I truly saw it or whether it was produced by a higher than usual dose of American coffee.

Fun for profit

The head of the Labs, the VP of Research at AT&T, in those years was Arno Penzias, who had shared the 1978 Nobel Prize with Robert Wilson for the discovery of the cosmic background radiation, the faint three-degree-Kelvin hum that arrives from every direction in the sky, the fossil echo of the Big Bang itself. Arno was approachable. You would see him wandering the laboratories, asking researchers about their latest results, giving good talks in the great auditorium. It was he who said that the purpose of Bell Labs was “Fun for Profit,” and that phrase captured the whole spirit of the place. The researchers were children set loose in a magnificent toy shop, free to do as they pleased, and it was the parent company, the telecommunications giant AT&T, whose job was to turn the results into profit. Something, with hindsight, that AT&T never managed to do very efficiently, even though most of the innovations changed our lives. But that is another story.

At the end of each year Arno collected the best papers and results from every director and rewarded them with a symbolic prize, a red T-shirt he had printed for the occasion. Across the front were the letters 3K, for the three degrees Kelvin of his radiation, scattered in every size and direction, and in the center his own optimistic revision of a line from T. S. Eliot. Eliot had written that the world would end not with a bang, but with a whimper. Arno had turned it around: the world began not with a whimper, but with a BANG. The three-degree-Kelvin shirt was a coveted prize, though not one you could wear on the street. How many people in suburban New Jersey would know that 3K meant a measure of radiation, and not other less inspiring things? So I kept mine at home.

The foreigners

There was another kind of strangeness at the Labs, gentler and closer to my own experience, and it belonged to those of us who had come from somewhere else. After a couple of months I was fully integrated, with friends of roughly my age, immigrants like me, and at lunch we would compare our countries with America and always find something to be puzzled by. The Germans complained about the imprecision of the trains, the disorderly lines at the post office, though they should have seen the ones in Italy, and the general absence, or excess, of clear bureaucratic rules. The Swiss complained even more. We Italians, the French, and the Spaniards complained about the food, and above all about the coffee. In those days Starbucks was nothing more than the first mate of the Pequod, for those who had read Moby-Dick, and for everyone else a name that meant nothing at all, with no connection to the baristas and cappuccinos of today. Most of us made our peace with the black liquid that filled the American cups, huge by an espresso drinker’s standard, which in Italy would have been cups for caffellatte. The more adventurous, and I was one of them, bought tins with vaguely Italian names like Caffè Medaglia d’Oro, or South American ones like Café Bustelo. We had brought our little moka stovetop pots from home, and we invited one another over to drink what no one in Italy would ever have accepted as coffee.

I have not written here about the mystery of the hidden piano room, the one Mark, the janitor I befriended, finally helped me find, where a mysterious pianist was playing Bach’s Goldberg Variations behind a locked door in the deep of the night, and then vanished through an emergency exit without setting off the alarm. No one really knew who he or she was. That story deserves its own telling, and I will come back to it in a future post. What I want to leave you with instead is the feeling of those first years, and a thought about what they meant.

What the strangeness was for

The reputation was not wrong, only imprecise. Most of the people at the Labs were not strange at all, only brilliant and a little unworldly, absorbed by their work to the point where convention simply fell away. A few crossed over into something stranger, the man who walked backward because it was easier for him, the explorer on his winter bicycle, the shadow collecting decontamination fluid at two in the morning, the unseen pianist and his (or her?) Goldberg Variations in the middle of the night. But eccentric or strange, they drew on the same permission. When you let people follow their curiosity without asking them to justify it in advance, you get all of it at once: the transistor, the C language, Unix, the proof sitting in a red T-shirt that the universe began with a bang, and also the men who walked the corridors in their own private worlds. Claude Shannon, William Shockley, Dennis Ritchie were never asked to explain the why and the how of their revolutionary ideas, because they already knew, and no one felt the need to ask.

That is the part I find myself returning to, now that the Labs I knew are gone, sold off in the breakup of AT&T and slowly worn down by the demand for profit at any cost, until being a scientist meant justifying every idea before you were allowed to have it. We have not stopped producing clever people or useful things. If anything the world has changed for the better in countless ways, and we have far more today than we had then, the Web, smartphones, an incomprehensible AI that does not stop surprising, and sometimes scaring, us. What we have made scarcer is the particular kind of place that tolerates strangeness long enough for brilliance to grow inside it, that trusts curiosity before it can show its return. Would something like Bell Labs, or Xerox PARC, or the old IBM Research be possible today? Probably not, at least not in that form. One of the answers to that question was brilliantly developed by Andrew Odlyzko in his 1995 essay “The Decline of Unfettered Research,” which I discuss in a separate essay, part of the From Spark to System series on this blog.

Epilogue

Am I telling you all this to mourn a world that has ended? I hope not, because that is not how it feels from the inside. Bell Labs was an extraordinary adventure, one of the great pieces of luck in my life, the one I walked into one snowy February afternoon with seven suitcases, six of which actually arrived. But it was the beginning of something rather than the whole of it. The years since have carried me through other laboratories, other countries, other interesting people and technologies I could not have imagined while standing in that windowless office in Murray Hill, and each of them, in its own way, asked me again to leave the beaten track and dive into the woods. The adventure did not end when the Labs did. It is still going on today, and I intend to keep following it wherever it leads, now and for as long as there are woods left to walk into.