DINNER WAS OVER at the Ferrucci household. It was a crisp evening in Yorktown Heights, a New York suburb ten miles north of the IBM labs. It was dark already, and the fireplace leapt with gas-fed flames. Ferrucci’s two daughters were heading upstairs to bed. In the living room, Ferrucci and his wife, Elizabeth, recounted a deeply frustrating medical journey—one that a retrained Jeopardy computer (Dr. Watson) could have made much easier.
Ferrucci had been envisioning a role for computers in doctors’ offices since his days as a pre-med student at Manhattan College. In graduate school, he went so far as to build a medical expert system to provide advice and answer questions about cardiac and respiratory ills. It worked well, he said, but its information was limited to what he taught it. A more valuable medical aid, he said, would scoop up information from anywhere and come up with ideas and connections that no one had even thought to consider. That was the kind of machine he himself had needed.
Early in the Jeopardy project, Ferrucci said, not long after the bake-off, he started to experience strange symptoms. The skin on one side of his face tingled. A couple of his fingers kept going to sleep. And then, one day, searing pain shot through his head. It lasted for about twenty seconds. Its apparent epicenter was a lower molar on the right side of his mouth. “It felt like someone was driving an ice pick in there,” he said.
When this pain returned, and then came back a third and fourth time, Ferrucci went to his dentist. Probing the tooth and placing ice on it, the dentist attempted to reproduce the same fearsome effects but failed. He could do a root canal, he said, but he had no evidence that the tooth was the problem. Ferrucci then went to a neurologist, who suggested a terrifying possibility. Perhaps, he said, Ferrucci was suffering from trigeminal neuralgia, more commonly known as the suicide disease. It was a nerve disorder so painful that it was believed to drive people to kill themselves. He recommended getting the root canal. It might do the trick, he said, and save them both from the trouble of ransacking the nervous system for answers.
Ferrucci got the root canal. It did no good. The attacks continued. He went to another neurologist, who prescribed anticonvulsion pills. When he read about the medicine’s side effects, Ferrucci said, “I didn’t know whether to take the pills or buy a gun.” He did neither and got an MRI. But putting on the helmet and placing his head in the cramped cylinder filled him with such anxiety that he had to go to another doctor for sedatives.
He had no idea what was wrong, but it wasn’t getting any better. As the Jeopardy project moved along, Ferrucci continued to make presentations to academics, industry groups, and IBM brass. But he started to take along his number two, Eric Brown, as a backup. “If I don’t make it through the talk,” he told Brown, “you just pick up where I leave off.”
In time, Ferrucci started to recognize a certain feeling that preceded the attacks. He sensed it one day, braced himself against a wall, lowered his head slightly, and awaited the pain. It didn’t come. He moved his head the same way the next time and again he avoided the pain. He asked his neurologist about a possible link between the movements of his neck and the facial pain. He was told there was no possible connection.
Months passed. The Ferruccis were considering every option. “Someone told us we should get a special mattress,” said Elizabeth. Then a friend suggested a specialist in craniofacial pain. The visit, Ferrucci learned, was not covered by his insurance plan and would cost $600 out of pocket. He decided to spend the money. A half hour into the meeting, the doctor worked his hands to a spot just below Ferrucci’s collarbone and pressed. The pain rocketed straight to the site of his deadened molar. The doctor had found the connection. A number of small contraction knots in the muscle, myofascial trigger points, created the pain, he said. The muscle was traumatized, probably due to stress. With massage, the symptoms disappeared. And Ferrucci kept them at bay by massaging his neck, chest, and shoulders with a two-dollar lacrosse ball.
He walked to a shelf by the fireplace and brought back a book, The Trigger Point Therapy Workbook, bristling with Post-it notes. On one page was an illustration of the sternocleidomastoid. It’s the biggest visible muscle in front of the neck and extends from the sternum to a bony prominence behind the ear. According to the book, trauma within this muscle could cause pain in the head. In a single paragraph on page 53 was a description of Ferrucci’s condition. It had references to toothaches in back molars and a “spillover of pain . . . which mimics trigeminal neuralgia.” Ferrucci could have written it himself.
If a computer like Watson, customized for medicine, had access to that trigger point workbook along with thousands of other books and articles related to pain, it could have saved Ferrucci a molar and months of pain and confusion. Such a machine likely would have been able to suggest, with at least some degree of confidence, the connection between Ferrucci’s symptoms and his sternocleidomastoid. This muscle was no more obscure than the Asian ornaments or Scandinavian kings that Watson routinely dug up for Jeopardy. Such a machine would not have to understand the connections it found. The strength of the diagnostic engine would not be its depth, but its range. That’s where humans were weak. Each expert that Ferrucci visited had mastered a limited domain. The dentist knew teeth, the neurologists nerves. But no one person, no matter how smart or dedicated, could stay on top of discoveries across every medical field. Only a machine could do that.
A few months earlier, on an August morning, about a hundred IBM employees filed into the auditorium at the Yorktown labs. They included researchers, writers, marketers, and consulting executives. Their goal was to brainstorm ideas for putting Watson to work outside the Jeopardy studio. The time for games was nearly over. Watson, like thousands of other gifted students around the world, had to start earning its keep. It needed a career.
This was an unusual situation for an IBM product, and it indicated that the company had broken one of the cardinal rules of technology development. Instead of focusing first on a market opportunity and then creating the technology for it, IBM was working backward: It built the machine first and was now wondering what in the world to do with it. Other tech companies were notorious for this type of cart-before-the-horse innovation. Motorola, in the 1990s, led the development of a $5 billion satellite phone system, Iridium, before learning that the market for remote communications was tiny and that most people were satisfied with normal cell phones. Within a year of its launch, Iridium went bankrupt. In 1981, Xerox built a new computer, the 6085 Star, featuring a number of startling innovations—a mouse, an ethernet connection, e-mail, and windows that opened and closed. All of this technology would lay the groundwork for personal computers and the networked world. But it would be other companies, notably Apple and Microsoft, that would take it to market. And in 1981, Xerox couldn’t find buyers for its $16,000 machines. Would Watson’s industry-savvy offspring lead to similar boondoggles?
In fairness to IBM, Grand Challenges, like Watson and the Deep Blue chess machine, boosted the company’s brand, even if it came up short in the marketplace. What’s more, the technology developed in the Jeopardy project, from algorithms that calculated confidence in candidate answers to wizardry in the English language, was likely to work its way into other offerings. But the machine’s question-answering potential seemed so compelling that IBM was convinced Watson could thrive in a host of new settings. It was just a question of finding them.
Ferrucci started the session by outlining Watson’s skills. The machine, he said, understood questions posed in natural language and could read millions of documents and scour databases at lightning speed. Then it could come up with responses. He cautioned his colleagues not to think of these as answers but hypotheses. Why the distinction? In every domain most of Watson’s candidate answers would be wrong. Just as in Jeopardy, it would come back with a list of possibilities. People looking to the machine for certainty would be disappointed and perhaps even view it as dumb. Hypotheses initiate a lengthier process. They open up paths of inquiry. If Watson came back from a hunt with ten hypotheses and three of them looked promising, it wouldn’t matter much if the other seven were idiotic. The person using the system would focus on the value. And this is where the vision of Watson in the workplace diverged from the game-playing model. In the workplace, Watson would not be on its own. Unlike the Jeopardy machine, the Watson Ferrucci was describing would be engineered to supplement the human brain, not supplant it.
The time looked ripe for word-savvy information machines like Watson, thanks to the global explosion of a new type of data. If you analyzed the flow of digital data in, say, 1980, only a smidgen of the world’s information had found its way into computers. Back then, the big mainframes and the new microcomputers housed business records, tax returns, real estate transactions, and mountains of scientific data. But much of the world’s information existed in the form of words—conversations at the coffee shop, phone calls, books, messages scrawled on Post-its, term papers, the play-by-play of the Super Bowl, the seven o’clock news. Far more than numbers, words spelled out what humans were up to, what they were thinking, what they knew, what they wanted, whom they loved. And most of those words, and the data they contained, vanished quickly. They faded in fallible human memories, they piled up in Dumpsters and moldered in damp basements. Most of these words never reached computers, much less networks.
That has all changed. In the last decade, as billions of people have migrated their work, mail, reading, phone calls, and webs of friendships to digital networks, a giant new species of data has arisen: unstructured data. It’s the growing heap of sounds and images that we produce, along with trillions of words. Chaotic by nature, it doesn’t fit neatly into an Excel spreadsheet. Yet it describes the minute-by-minute goings-on of much of the planet. This gold mine is doubling in size every year. Of all the data stored in the world’s computers and coursing through its networks, the vast majority is unstructured. Hewlett Packard, for example, the biggest computer company on earth, gets a hundred fifty million Web visits a month. That’s nearly thirty-five hundred customers and prospects per minute. Those visits produce data. So do notes from the company’s call centers, online chat rooms, blog entries, warranty claims, and user reviews. “Ninety percent of our data is unstructured,” said Prasanna Dhore, HP’s vice president of customer intelligence. “There’s always a gap between what you want to know about the customer and what is knowable.” Analysis of the pile of data helps reduce that gap, bringing the customer into sharper focus.
The potential value of this information is immense. It explains why Facebook, a company founded in 2004, could have a market value six years later of $50 billion. The company gathers data, most of it unstructured, from about half a billion people. Beyond social networks and search engines, an entire industry has sprung up to mine this data, to predict people’s behavior as shoppers, drivers, workers, voters, patients, even potential terrorists. As machines, including Watson, have begun to chomp on unstructured data, a fundamental shift is occurring. While people used to break down their information into symbols a computer could understand, computers are now doing that work by themselves. The machines are mastering human communication.
This has broad implications. Once computers can handle language, every person who can type or even speak becomes a potential programmer, a data miner, and an analyst. This is the march of technology. We used to have typists, clerks, legions of data entry experts. With the development of new tools, these jobs became obsolete. We typed (and spell-checked), laid out documents, kept digital records, and even developed our own pictures. Now, a new generation of computers can understand ordinary English, hunt down answers in vast archives of documents, analyze them, and come up with hypotheses. This has the potential to turn entire industries on their heads.
In the August meeting, Ferrucci told the audience the story of his recent medical odyssey and how a machine like Watson could have helped. Others suggested that Watson could man call centers, function as a brainy research assistant in pharmaceutical labs, or work as a whip-smart paralegal, with nearly instant recall of the precedents, both state and federal, for every case. They briefly explored the idea of Watson as a super question-answering Google. After all, it could carry out a much more detailed analysis of questions and piece together sophisticated responses. But this idea went nowhere. IBM had no experience in the commercial Web or with advertisers. Perhaps most important, Watson was engineered to handle one Jeopardy clue at a time. In those same three seconds, a search engine like Google’s or Microsoft’s Bing handled millions of queries. To even think about competing, the IBM team would have to build an entirely new and hugely expensive computing architecture. It was out of the question.
No, Watson’s future was as an IBM consulting tool and there were plenty of rich markets to explore. But before Watson could make a go of it, Big Blue would have to resolve serious questions. First, how much work and expense would it take to adapt Watson to another profession, to curate a new body of data and to educate the machine in each domain? No one could say until they tried. Second, and just as important, how much resistance would these new knowledge engines encounter? New machines, after all, are in the business of replacing people—not something that often generates a warm welcome. The third issue involved competition. Assuming that natural-language, data-snarfing, hypothesis-spouting machines made it into offices and laboratories, who was to say that they’d be the kin of a Jeopardy contraption? Other companies, from Google to Silicon Valley startups, were sure to be competing in the same market. The potential for these digital oracles was nearly limitless. But in each industry they faced obstacles, some of them considerable.
Medicine was one of the most promising areas but also among the toughest to crack. The natural job for Watson would be as a diagnostic aid, taking down the symptoms in cases like Ferrucci’s and producing lists of possible conditions, along with recommended treatments. Already, many doctors facing puzzling symptoms were consulting software tools known as medical decision trees, which guided them toward the most likely diagnoses and recommended treatments. Some were available as applications on smart phones. A medical Watson, though, would plunge into a much deeper pool of data, much of it unstructured. Conceivably, it would come up with hidden linkages. But even that job, according to Robert Wachter, the chief of hospital medicine at the University of California, San Francisco, was bound to raise serious questions. “Doctors like the idea of having information available,” he said. “Where things get more psychologically fraught is when a damned machine tells them what to do.” What’s more, once analysis is automated, he said, the recommendation algorithm is likely to include business analysis. In other words, the medical Watsons might come back not with the statistically most effective treatment but the most cost-effective one. Even if this didn’t happen, many would remain suspicious. And what if Watson had sky-high confidence in a certain diagnosis—say, 97 percent? Would doctors get in trouble if they turned a deaf ear to it? Would they face lawsuits if they ignored the advice and it later turned out the machine was right?
Then, of course, there was the possibility of disastrous mistakes resulting from a computer’s suggestions. Even if a bionic assistant scrupulously labeled all of its findings as hypotheses, some of them—just like Watson’s answers in Jeopardy—were bound to be nutty, generating ridicule and distrust. Others, perhaps more dangerous, would be wrong while appearing plausible. If a treatment recommended by a machine killed a patient, confidence in bionic assistants could plummet.
The other issue, sure to come up in many industries, boils down to a struggle for power, and even survival, in the workplace. “As every profession embraces systems that take humans out of it,” Wachter said, “the profession gets commoditized.” He noted the example of commercial aviation, where pilots who were once considered stars have ended up spending much of the time in flight simply backing up the machines that are actually flying the planes. The result? “Pilots’ pensions have been cut and they’re paid less, because they’re largely interchangeable,” he said. “Doctors don’t want to see that happening to them.”
For IBM, this very scenario promises growth. With more than $4 billion in annual revenue, the health care practice within IBM Global Services has the size of a Fortune 500 company. It runs large data centers for hospitals and insurance companies. It also helps them analyze the data, looking for patterns of symptoms, treatments, and diseases—as well as ways to cut costs. This is part of a trend toward statistical analysis in the industry and the rapid growth of so-called evidence-based medicine. But one of the most valuable streams of data—the doctor’s notes—rarely makes it into the picture, said Joseph Jasinski, who heads research for IBM’s health care division. This is where the doctor writes down what he or she sees and thinks. Sometimes it is stored in a computer, but only, Jasinski said, “as a blob of text.” In other words, it’s unstructured data, Watson’s forte. “There’s a strong belief in the community that if you could study clinical notes, you could analyze patient similarities,” he said. Neurologists’ notes—going back to Ferrucci’s case—could have pointed to common symptoms between patients with the suicide disease and others with knots in a muscle just below their shoulder blade. This analysis could expand, comparing symptoms and treatments, and later study the outcomes. What works? What falls flat? Which procedures appear to waste money?
Despite the growth of evidence-based medicine, many of these studies are hard to carry out, especially in the splintered $2.3 trillion American system. The doctor prescribes the treatment and the insurance company pays for it, but all too often neither of them gets the most vital feedback: how it worked. The feedback loop, in the language of statisticians, rarely closes. The most promising sites for this type of analysis, Jasinski said, are self-contained hospital networks, which keep voluminous records on patients and do extensive follow-up. They would include the Veterans’ Administration, the Mayo Clinic, and Kaiser Permanente in the United States. Many countries with national health care systems also have promising data. Denmark, where IBM has been running the data since 2006, could provide a national laboratory. There, a medical Watson could diagnose diseases, suggest treatments that have proven successful, and steer doctors away from those that have led to problems. Such analyses could save lives, Jasinski said. ”We kill a hundred thousand people a year from preventable medical errors.”
In fact, the potential for next-generation computers in medicine stretches much further. Within a decade, it should cost less than $100 to have an individual’s entire genome sequenced. Some people will volunteer to have this done. (Already, companies like 23andMe, a Silicon Valley startup, charge people $429 for a basic decoding.) Others, perhaps, will find themselves pressed, or even compelled, by governments or insurers, to submit their saliva samples. In either case, computers will be studying, correlating, and answering questions about growing collections of this biological information.
At the same time, we’re surrounding ourselves with sensors that provide streams of data about our activities. Coronary patients wear blood pressure monitors. Athletes in endurance sports cover themselves with electronics that produce torrents of personal data, reading everything from calorie burn to galvanic skin response, which is associated with stress. Meanwhile, companies are rushing into the market for personal monitoring. Zeo, a Rhode Island company, sells a bedside device that provides a detailed readout every morning of a person’s sleeping patterns, including rapid-eye movement, deep sleep, and even trips to the bathroom. Intel is outfitting the homes of elderly test subjects with sensors to measure practically every activity possible, from their nocturnal trips to the bathroom to the words they type on their computers. And each person who carries a cell phone unwittingly provides detailed information on his or her daily movements and routines—behavioral data that could prove priceless to medical researchers. Even if some of this data is shielded by privacy rules and withheld from the medical industry, much of it will be available. Machines like Watson will be awash in new and rising rivers of data.
But in the autumn of 2010, as Watson prepared for its culminating Jeopardy match, it had yet to land its first hospital job, and its medical abilities remained largely speculative. “We have to be cautious here,” Jasinski said. Though full of potential, Watson was still untested.
It may seem frivolous for the IBM team to have worked as hard as it did to cut down Watson’s response time from nearly two hours to three seconds. All of that engineering, and those thousands of processors were harnessed, just to be able to beat humans to a buzzer in a quiz show. Yet as Watson casts about for work, speed will be a crucial factor. Often it takes a company a day or two to make sense of the data it collects. It can seem remarkable, because the data provides a view of sales or operations that was unthinkable even a decade ago. But still, the delay means that today doesn’t come into focus until tomorrow or next week. The goal for many businesses now is to process and respond to data in real time—in the crucial seconds that a quick investment could net $10 million or the right treatment could save a patient’s life. Chris Bailey, the chief information officer at SAS, a major producer of analytics software, says the focus is on speed. “Our goal is to make the systems run a thousand or a million times faster,” he said. “That enables us to look at a million times more input.” With this speed, companies increasingly will be able to carry out research, and even run simulations, while the customer is paying for a prescription or withdrawing funds.
Computers with speed and natural language are poised to transform business processes, perhaps entire industries. Compared to what’s ahead, even today’s state of the art looks sluggish. Consider this snapshot of the data economy, circa 2011: A man walks into a pharmacy to renew his blood pressure medication. He picks up some toiletries while he’s there. He hands the cashier his customer loyalty card, which lowers his bill by a dollar or two, and then pays with his Visa card. This shopping data goes straight to Catalina Marketing in St. Petersburg, Florida, which follows the purchases of 190 million shoppers in America. Catalina scans the long list of items that this consumer has bought in the last three years and compares his patterns with those of millions of others. While he’s standing at the register, it calculates the items most likely to interest him. Bundled with the receipt the cashier hands him, he finds several coupons—maybe one for oatmeal, another for a toothpaste in a new upside-down dispenser. If and when he uses them, Catalina learns more about him and targets him with ever-greater precision.
That might sound like a highly sophisticated process. But take a look at how Catalina operates, and you’ll see it involves a painfully slow roundtrip, from words to numbers and then back again. “Let’s say Kraft Foods has a new mac and cheese with pimentos,” said Eric Williams, Catalina’s chief technology officer. The goal is to come up with a target group of potential macaroni eaters, perhaps a million or two, and develop the campaign most likely to appeal to them. The marketers cannot summon this intelligence from their computers. They hand the instructions—the idea—to Catalina’s team of seventy statisticians. For perhaps a week, these experts hunt for macaroni prospects in the data. Eventually they produce lists, clusters, and correlations within their target market. But their statistical report is not even close to a marketing campaign. For that, the marketers must translate the statisticians’ results back into words and ideas. “Trying to interpret what these people find into common language is quite a feat,” Williams said. Eventually, a campaign takes shape. Catalina concocts about six hundred to eight hundred of them a year. They’re effective, often doubling or tripling customer response rates. But each campaign, on average, gobbles up a month of a statistician’s work.
Williams’s fantasy is to have a new type of computer in his office. Instead of delivering Kraft’s order to his statisticians, he would simply explain the goals, in English, to the machine. It would pile through mountains of data in a matter of seconds and come back with details about potential macaroni buyers. The language-savvy machine wouldn’t limit its search to traditional data, the neatly organized numbers featuring purchases, dates, and product codes. It might read Twitter or scan social networks to see what people are writing about their appetites and dinner plans. After this analysis, Williams’s dream machine could return with a list of ten recent marketing campaigns that have proven the most effective with the target group. “If I don’t have to go to statisticians and wait while they run the data, that would be huge,” Williams said. Instead of eight hundred campaigns, Catalina might be able to handle eighty thousand, or even a million—and offer them at a fraction of today’s cost. “You’re talking about turning marketing on its head.”
His dream machine, of course, sounds like a version of Watson. Its great potential, in marketing and elsewhere, comes from its ability to automate analysis—to take people, with their time-consuming lunch breaks and vacations, their disagreements and discussions, and drive them right out of the business. The crucial advantage is that Watson—and machines like it—eliminate the detour into the world of numbers. They understand and analyze words. Machines like this—speedy language whizzes—will open many doors for business. The question for IBM is what Watson’s place will be in this drama, assuming it has one.
In the near term, Watson’s job prospects are likely to be in call centers. Enhanced with voice recognition software and trained in specific products and services, the computer could respond to phone calls and answer questions. But more challenging jobs, such as bionic marketing consulting, are further off. For each industry, researchers working with consulting teams will have to outfit Watson with an entirely new set of data and run through batteries of tests and training sets. They’ll have to fine-tune the machine’s judgment—the degree of confidence it generates for each response—and adapt hardware to the job. Will customers want access to mini-Watsons on site or perhaps gain access to a bigger one through a distant data center, a so-called cloud-based service? At this point, no one can say. Jurij Paraszczak, director of Industry Solutions and Emerging Geographies at IBM Research, sees versions of Watson eventually fitting into a number of industries. But such work is hardly around the corner. “Watson’s such a baby,” he said.
The history of innovation is littered with technologies that failed because of bad timing or rotten luck. If that $16,000 Xerox computer, with the e-mail and the mouse, had hit the market a decade later, in 1991, cheaper components would have lowered the cost by a factor of five or ten and a more informed public might have appreciated its features.
Technology breakthroughs can also consign even the most brilliant and ambitious projects to museum pieces or junk. In 1825, the first load of cargo floated from Buffalo to Albany on the new Erie Canal. This major engineering work, the most ambitious to date in the Americas, connected the farms and nascent industries of the Great Lakes to the Hudson River, and on to the Atlantic Ocean. It positioned New York State as a vital thoroughfare for commerce, and New York City as the nation’s premier port. The news could hardly have been worse for the business and government leaders in New York’s neighbor to the south, Pennsylvania, and its international port, Philadelphia. They’d been outmaneuvered. The only way to haul cargo across Pennsylvania was by Conestoga wagon, which often took up to three weeks. So that very year, the Pennsylvanians laid plans to build their own waterway connecting Philadelphia to the Great Lakes. They budgeted $10 million, an immense sum at the time and $3 million more than the cost of the Erie Canal.
The Pennsylvania Canal faced one imposing obstacle: the Alleghenies. These so-called mountains were smaller and rounder than the Rockies or Alps, but they posed a vertical challenge for canal designers. Somehow the boats would have to cross from one side to the other. So with engineering far more ambitious than anything the New Yorkers had attempted, the Pennsylvanians constructed a complex series of iron-railed ramps and tunnels. This horse-powered system would hoist the boats back and forth over the hills and, in a few places, through them. By 1834, boats crossing the state left the water for a thirty-six-mile cross-country trek, or portage, which lifted them up fourteen thousand feet and back down again. For the criss-crossing on ramps, the boats had to be coupled and uncoupled thirty-three times. For those of us not accustomed to Conestoga wagons, this would seem excruciatingly slow, even after stationary steam engines replaced the horses. The novelist Charles Dickens made the crossing in 1842. He set off on a boat and found himself, in short order, peering down a mountain cliff. “Occasionally the rails are laid upon the extreme verge of the giddy precipice and looking down from the carriage window,” he wrote, “the traveler gazes sheer down without a stone or scrap of fence between into the mountain depths below.” Dickens wouldn’t soon forget Pennsylvania’s $10 million innovation.
This canal project used current technology and heroic engineering to carry out work on a tight deadline. Would newer technology supplant it? It was always possible. After all, the Industrial Revolution was raging in Britain and making inroads into America. Things were changing quickly, which was precisely why Pennsylvania could not afford to wait. As it turned out, within a decade steam-powered trains rendered the canal obsolete. Yet the project laid out the vision and preliminary engineering for the next generation of transport. The train route over the Alleghenies, which followed much of the same path, was considered so vital to national security that Union soldiers kept guard over it during the Civil War. And in 1942, as part of the failed Operation Pastorius, Nazi agents targeted it for sabotage.
Will Watson be the platform for the next stage of computing or, like the Pennsylvania Canal, a bold and startling curiosity to be picked over, its pieces going into cheaper and more efficient technologies down the road? The answer may depend in part on IBM’s rivals.
Google is the natural competitor, but it comes at the problem from an entirely different angle. While IBM builds a machine to grapple with one question at a time, Google is serving much of networked humanity. Electricity is a major expense, and even a small increase of wattage per query could put a dent in its profits. This means that even as Google increasingly looks to respond to queries with concrete answers, it can devote to each one only one billionth the processing power of Watson, or less. Peter Norvig, the research director, said that Google’s big investments in natural language and machine translation would lead the company toward more sophisticated question-answering for its mass market. As its search engine improves its language skills, he said, it will be able to carry out smarter hunts and make better sense of its users’ queries. The danger for IBM isn’t head-to-head competition from Google and other search engines. But as searching comes closer to providing answers to queries in English, a number of tech startups and consultants will be able to jury-rig competing question-answering machines, much the way James Fan built his Basement Baseline at the dawn of the Jeopardy project.
As Google evolves, Norvig said, it will start to replicate some of Watson’s headier maneuvers, combining data from different sources. “If someone wants per capita income in a certain country, or in a list of countries, we might bring two tables together,” he said. For that, though, the company might require more detailed queries. “It might get to the point where we ask users to elaborate, and to write entire sentences,” he said. In effect, the computer will be demanding something closer to a Jeopardy clue—albeit with fewer puns and riddles.
This would represent a turnaround. For more than a decade, the world’s Web surfers have learned to hone their queries. In effect, they’ve used their human smarts to reverse-engineer Google’s algorithms—and to understand how a search engine “thinks.” Each word summons a universe of connections. Looking at each one like a circle in a Venn diagram, the goal is to organize three or four words—3.5 is the global average—whose circles have the smallest possible overlap. For many, this analysis has become almost reflexive. Yet as the computer gets smarter, these sophisticated users stand to get poorer results than those who type long sentences, even paragraphs, and treat the computer as if it were human.
And many of the computer systems showing up in our lives will have a far more human touch than Watson. In fact, some of the most brilliant minds in AI are focusing on engineering systems whose very purpose is to leech intelligence from people. Luis Von Ahn, a professor at Carnegie Mellon, is perhaps the world’s leader in this field. As he explains it, “For the first time in history, we can get one hundred or two hundred million people all working on a project together. If we can use their brains for even ten or fifteen seconds, we can create lots of value.” To this end, he has dreamed up online games to attract what he calls brain cycles. In one of them, the ESP game, two Web surfers who don’t know each other are shown an image. If they type in the same word to describe it, another image pops up. They race ahead, trying to match descriptions and finish fifteen images in two and a half minutes. While they play, they’re tagging photographs with metadata, a job that computers have not yet mastered. This dab of human intelligence enables search engines to find images. Von Ahn licensed the technology to Google in 2006. Another of his innovations, ReCaptcha, presents squiggly words to readers, who fill them in to enter Web sites or complete online purchases. By typing the distorted letters, they prove they’re human (and not spam engines). This is where the genius comes in. The ReCaptchas are drawn from the old books in libraries. By completing them, the humans are helping, word by crooked word, to digitize world literature, making it accessible to computers (and to Google, which bought the technology in 2009).
This type of blend is likely to become the rule as smarter computers spread into the marketplace. It makes sense. A computer like Watson, after all, is an exotic beast, one developed at great cost to play humans in a game. The segregated scene on the Jeopardy stage, the machine separated from the two men, is in fact a contrivance. The question-answering contraptions that march into the economy, Watson’s offspring and competitors alike, will be operating under an entirely different rubric: What works and at what cost? The winners, whether they’re hunting for diseases or puzzling out marketing campaigns, will master different blends. They’ll figure out how to turbocharge thinking machines with a touch of human smarts and, at the same time, to augment human reasoning with the speed and range of machines. Each side has towering strengths and glaring vulnerabilities. That’s what gives the Jeopardy match its appeal. But outside the Jeopardy studio, stand-alones make little sense.