Search Engines lead us into new information frontiers. But will we ever find what we’re really looking for?
When my son was young, he trembled in fear and joy. Like all of us, he arrived incomplete, and the neurons in the limbic system of his toddler’s brain were still growing. The limbic system is the seat of human emotion. When enough neural pathways functioned, he developed the sensations of anger, delight, sorrow, excitement.
Those feelings enlarged his world, but they shook him. He stood transfixed on the kitchen floor, newly arrived in something that he would eventually call love, perhaps, or anguish. Each irrevocable change was jarring, though, as it might be if at once our eyes commanded the ultraviolet spectrum, or if we heard with the clarity of owls.
We envy children the intensity of feeling things like pure joy, or even raw sorrow. They are untempered by history or boredom: Everything is new. If we could recall that dismaying onrush of new feelings, we would better grasp what technology is doing now. Soon enough, thanks to an ever-denser Internet and increasingly effective search technologies, we should possess enormously augmented brains, relentlessly searching new landscapes for knowledge and connection.
The neurons may be electronic, but that may not matter much. Our two primal urges—to know the world and have the world know us—will flower as never before, encoding us in a continual, global bazaar of talk, products, and sensation. We will have virtual selves on the Internet at all hours. They may earn us real money, or involve us with real friends and political movements, or help us find parts of ourselves we now see only dimly, if at all.
For certain, we will access and filter an unimaginable amount of information. According to Hal Varian of the School of Information Management and Systems, in 2002 we filmed, wrote down, or otherwise encoded some five exabytes of information. That is roughly the same amount of data contained in all preceding human conversations—ever. In 2005, Varian figures, we made twice as much. And he sees that three-year doubling cycle continuing for the foreseeable future.
Ninety-five percent of this material is in digital form and is easily uploaded to private or public networks, in particular the all-encompassing Internet, where it is filtered, snipped, remixed, and otherwise given new life. Above all, it is searched for meaning and links to other information, then served to our hungry minds.
If our minds were not so hardened in their habits, we might tremble like my toddler at the ways computing and communications, guided by search, now urge us toward an enormous collective awareness. The booming, buzzing mass of networked computers want data, from hard facts to family photos. The more data there is on the Internet, the better chance search engines have of finding what we want, and the more possibilities of what we want are created. The more people contribute, the more people there are to contact, to see what they have to offer.
Many of us already seem to be in constant contact—bees in a massive intelligence hive—with social-networking software for pheromones, and weblogs for bee dances in which we come and go with a single click. Soon, however, today’s 1 billion Internet users will be able to meet the planet’s 5 billion other people, connected by cheap cell phones and $100 laptops. Watch the bees dance then. Amplify that mix with 10 times more sensors than people on the Internet. The sensors will aggregate data and, aided by pattern-searching software, will see now-unobserved movements in nature and society. That search may create cures for the common cold by watching the flow of air through a city, or change agriculture by watching the jungle grow at night. It may create new commerce or communities by watching the time it takes me to trust someone with whom I’ve just connected online.
Along with the evanescent hive, we may gain a kind of solitary immortality, with all of our personal memories encoded and stored online forever, underscored with software that mimics our own consciousness.
Most likely, we will end up in some yet-unimagined combination of the immortal self and the fleeting node. Like toddlers feeling love before they can talk about it, we have as yet no language to frame this. We have to stretch our minds to understand what technology will make us, much the way we struggle to imagine the state of mind in ancient Egypt. Perhaps, most chillingly, we will be oblivious to the larger mind, to a conscience. We’ll be just possessed and straining bees, or neurons.
In Cal classrooms, students stare at their wireless laptops during a lecture. But look closer: They have created a temporary community on an Internet chat channel, sharing websites that amplify the talk, or exchanging notes on their prof’s performance. On another side of some screens, browsers seek the latest hometown news from Tulsa or Shanghai. At such times it is hard to say where, exactly, the students’ minds are.
To hear its builders talk, technology will remake the human species, changing the ways we think, even what we desire. “There is a kind of beauty in having a surrogate memory. The fact is, I’ve delegated a lot of my remembering,” says Gordon Bell, a distinguished senior research fellow at Microsoft and five-decade veteran of the industry. One of his projects today is “MyLifeBits,” a collection of his lifetime’s letters, cards, photos (25,000 of those alone), phone calls, and anything else he can put online. Looking ahead, he sees a world where billions of sensors act as proxies for our own perceptions and “everything will be maintained … there will be intelligence everywhere.” After all, each of those little electronic eyes, those minuscule ears, feeds back to a very big brain.
One of the fine ironies of this future is that much of our new selves will be built on today’s desires. The quality of the results on search engines like Google or Yahoo, the core of competition among them, relies on such disciplines as artificial intelligence, machine learning, and parallel processing—problems whose resolution will empower the Web. Thanks to the surprising success of search-based advertising, those companies also have the revenues to develop technology like never before. Getting the best brains onto building better electronic brains is (forgive me) a no-brainer.
Google and Microsoft recently joined Sun Microsystems in funding a new center at Cal that aims to build computers smart enough to serve millions of Internet users at once. Computer science professor David Patterson, who earlier worked on the computer technology behind the Inktomi search engine, figures the center will start to crack the problem of creating true artificial intelligence, through software capable of optimizing the patterns of a million different clickstreams, or pathways users take as they navigate websites. This gathered intelligence, combined with pattern recognition, is the key to creating virtual intuitive intelligence, something that could revolutionize marketing as well as strategic business or international political negotiations.
But building out an eBay or bidding on a mansion, electro-fragging a boring professor while cruising Amazon.com, or even building monster computers with high-end machine learning in a nearby Cal lab are just the tiny tip of this big iceberg.
The last comparable revolution in communications—the creation of the printing press—quickly created an industry of classical translation into the peoples’ tongues, expanding the Renaissance and spurring the Reformation. Hearing people expound on today’s search revolution makes it feel like a time of that magnitude. “We want to organize all the world’s information, and make it universally accessible” says Eric Schmidt, the Berkeley Ph.D. in computer science who heads Google. The service also offers free instant messaging and e-mail and has released Internet telephone software, so the information can move around quickly. Even so, Schmidt figures organizing “all” the information, everything that can be known about the planet and about ourselves, could take 300 years. As always, success may be uncertain, but one thing is clear: According to Schmidt, “It is important that we not be stopped.”
Schmidt is an engaging, circumspect billionaire, the kind of person you might want involved in a major change in human consciousness. He came to Google after running computer-networking company Novell, and before that he was chief scientist at Sun Microsystems. He moved to Google in 2001, when it was a reasonably profitable small company with 230 employees. Today, it has 5,000 employees and a stock price that valued the company at $140 billion this winter.
That sounds like a lot, until you consider that the creation and manipulation of American consumer desire through advertising, product information, or sales is worth $2 trillion, out of our $11 trillion national economy. Google, which promotes search as a better way to sell, proposes to upend much of that.
“People come to us in weakness,” Google cofounder Sergey Brin told me in 2003. “They are admitting that they don’t know something. When we tell them what they want to know, it creates a strong bond.” As any salesperson will tell you, if you know the need and you have the bond, you’re most of the way to closing.
The Google crew saw the social dimension of the Internet. That bond that Brin talked about came from looking at the way one person linked his website to others. Those links between pages were like votes on how valuable one site was relative to similar sites. The better ones got more links. In effect, they saw that the Web is a continual flux between individuals and collectives.
Today, this type of collective intelligence is harnessed as “tagging.” Millions of users share favorite photos, bands, or websites, highlighting or tagging the ones they like best, adding a little information about them, then moving them along to others. When the subway bombs went off in London last summer, people notified their loved ones that they were unhurt through a website called Flickr, one of the tagging pioneers.
This incremental, collective work is also the ruling genius of Wikipedia (see “Open source“), an Internet encyclopedia written by thousands and edited by no one. Its thousands of authors/editors inform on more than 100,000 subjects, with no central authority guiding the process. Someone writes on a topic in which they feel expert and posts it on Wikipedia. Someone else is free to amend that article, even erase it and write his or her own. In time, the most authoritative work is expected to arise. At first it sounds lunatic, yet Wikipedia’s collective capabilities put it ahead of Encyclopedia Britannica in entries and only slightly ahead of it in error rates.
There is the somewhat more self-conscious collective intelligence of open-source software. In the freedom of open source, a multitude of engineers—strangers to each other in language or custom—construct machine instructions of staggering complexity with no road maps or deadlines. They connect over the Internet even as they build out the beast, and the prime rule is that everyone shares their creation.
The massive computer powering Google (it’s really a collection of smaller electronic brains joined as one by clever software) uses Linux. Now, the open-source insight to which many builders freely contribute to create high-quality, cost-free product is drifting from software into the physical world. “Opencola soda,” a homebrewed version of Coke, lists its ingredients and manufacturing steps on the Web, with the same open-source license agreement as Linux and much other software. There is another site for beer. San Francisco’s Institute for OneWorld Health is applying the principles to drug development. Blogging, with its many voices on each topic, may evolve into open-source media.
Schmidt also applied the idea of collective intelligence to running Google. Every Monday, each employee notes what he or she did the previous week for an internal website that other employees access. The idea is to share as much information about Google itself as widely as possible within the company. Thousands of engineers share product ideas via e-mail, forming ad hoc teams to build software on the fly. It is posted on the Google site, and within a few hours management knows if it has a hit on its hands. Orkut, Google’s social-networking site for strangers to meet and share interests, was launched this way. Though not popular in the United States, Orkut has gained millions of users in Brazil and Iran.
Few at Google are hired for specific jobs; people come into the fold and get horse-traded based on their skills and the moment’s need. Like nodes in a network that are assembled and taken apart for different tasks, it is assumed that people will be used for one function now, another later. Twenty percent of most employees’ time is spent on Google projects apart from their main job, communicating with people from other departments—the tasks arrived at by a kind of email consensus.
Google’s halls are as crowded as the cells of a hive, only this time the buzz comes from coffee urns marked “regular” and “strong.” People work three to a room, not just because the place is growing fast but to speed conversation. Managers park physicists, say, next to mathematicians. A problem in one specialty, it seems, is often solved by examining it through the eyes of another specialist in another realm. The point is to break down the categories and unnecessary distinctions, get the best possible data, and move it through the system at the highest possible velocity.
“All the employees are in constant conversation,” says Schmidt. The contact and shared knowledge fueling the mission, he says, hold people long after they have made their millions from stock options.
Google is hardly the only company to realize the rise of the Internet’s collective brain. Yahoo, which began life logging by hand which Internet sites were particularly useful, has during the past year spent more than $200 million on hot companies specializing in tagging, like Flickr, and something called “mash-ups.” These involve borrowing from several pieces of software around the Web to build yet more software. An instant message with a party address, say, might connect to a mapping site and a directory of phone numbers, so you can tell a cabbie just where you need to go.
In conjunction with Cal’s School of Information Management and Systems, Yahoo has a research arm on University Avenue in Berkeley dedicated to the emerging social Web.
At Yahoo’s Berkeley lab, the engineers wonder about the coming impact of 1 billion camera-equipped cell phones, transmitting pictures and location data to a big central database. Search will then monitor much of the physical world, and record the visual marks of both our treasured and monotonous days. They ask what it says about us that Flickr has grown not just 70 million images but up to 80 tags per picture. They look at a Taiwan-born, Korea-incubated community service for answering any imaginable question. Ask a question of the group (What does Mars weigh?), and some anonymous altruist provides the answer (6.42 x 1023 kilograms). The accuracy is remarkably good, discounting such unanswerable questions as “Should I sleep with my boyfriend?”
Few people involved in creating the internet, from software writers at Google to the guys hauling cables around connecting office computers, talk about the rise of a collective mind. They are technicians, ostensibly with faith in scientific laws. In computing, the holiest of those is Moore’s Law. It is based on an observation Intel founder Gordon Moore made 40 years ago, that the density of transistors on an integrated circuit doubles about every 24 months. Now, following the curve of Moore’s Law, they forecast computers in 2020 will have the processing power of a human brain. Thirty years later, by the same thinking, they expect to welcome a net computational power of all the brains on earth.
As Schmidt points out, it is “reasonable” to assume that machines will be indistinguishable from humans in many behaviors. Many biological systems, by this logic, will be simulated and duplicated. Whether this means computer implants in humans, keyboards operated by our brain waves, or DNA-based computing, it certainly means that more categories will break, more distinctions fall. Implanted thermostats, hormone regulators, antioxidants, and antivirals unleashed when called for by automatic sensors are logical extensions of search and feedback mechanisms.
Another tech rule, Metcalfe’s Law, states that the value of a network is equal to the square of the number of its users. Because value is good, in effect the network wants to be big, maximizing its use and utility.
There is also Gordon Bell’s Law, which talks about regularly timed decade-long reigns of dominant computing platforms: The mainframe begat the microcomputer, which begat the PC, which begat the Internet, which begat the multipurpose cell phone computer. Bell’s Law implies a march of electronic intelligence expanding the self. In the future, Bell sees the world dominated by sensors, extending our intelligence into remote locations like space, or the ultraviolet spectrum.
When intelligence gets on a roll—with all three laws operating—it wants to grow, and it wants to connect. The engineers, with their belief in the power of the collective mind, turn out to be accidental mystics. What they know for certain is that the need for computing and communication is virtually insatiable.
Search advertisers respond with ever more powerful searches for customers. In the old days of 2000, advertisers bought the right to put their ads on search results pages that depended on a single word. For example, a health food store would pitch when someone searched for “granola.” Nowadays, multinationals mine data of ever-bigger searches, using powerful software that calculates millions of probable word combinations. Then they purchase rights to 300,000 or more words and phrases at a time, updating their deals every hour. The algorithms used in the world’s bond markets are now also used to indicate that people typing in search terms such as “biting” or “Jeep Sahara Limited XT” are likely new parents, maybe interested in purchasing a stroller.
My son, now almost a teen, assumes the new technology as a given, and grows with it. His peers also like being part of the collective mind. On MySpace.com, millions of kids memorialize themselves and create communities around bands and art they would never otherwise know, in constant contact with one another through instant messaging.
On fanfiction.net, people write new chapters to hundreds of beloved books, bringing Willy Wonka scores of additional authors. At the online game site Second Life, people build their own characters for a digital world, and make up their own adventures. Linden Labs, the owners of Second Life, originally tried to control the game, but the players rebelled and took control. Now they access Linden’s software and build each other’s bodies, clothes, weapons, and buildings. Then they sell them for real-world money. A Chinese woman based in Germany develops Second Life property to the tune of $150,000 a year. Linden contents itself with controlling the currency, releasing additional land, and watching the adventures roll along.
Across the Web, people build networks and communities at maximum velocity. E-mail is too slow, so communication is IM—in pairs, in groups—with multiple conversations on one screen. Couples IM each other on one end of the computer screen, then cut and paste the dialog on another message channel, where their friends, say, critique a boyfriend’s sincerity or a girlfriend’s promises. Love gets coached, in real time—a kind of Wikipedia for emotional life. My friend in Virginia eats dinner with his wife, son, and son’s laptop, its Internet camera bringing the boy’s girlfriend to the kitchen table every night.
We are just a few years from multiple cameras with sensors, and artificial intelligence algorithms that come out of Yahoo—or Professor Patterson’s Berkeley lab—learning which ones we like best, encoding our tastes, and giving us more of what we want: contact.
It all improves as it grows, and look where it is growing now: In China, IM provider Tencent claims 450 million registered users, 18 million of them chatting online at any given moment. Their new social-networking site, a kind of MySpace for China, though still in its testing phase, quickly filled an experimental limit of 20 million users.
In 2004 a group of language-translation specialists came out of academia and into Google, and with their first big funding and big computational resources, they won the National Institute of Science and Technology’s test in translation from Chinese. Geography has already surrendered to the global brain, and language is the next barrier to fall. What the printers did by publishing translations of a few classical texts in the 16th century has its corollary in creating a common worldwide conversation in the 21st. And it is one with mighty pattern-finding software and open-source logic that sweeps the system for whatever we then seek—illumination, music, sex, revenge, surprise—beckoning and bringing it without stop.
What, ultimately, can we say about a future like this? We should begin with the most basic things: We will still be human. We will yet propel ourselves to connect. We will want change and transformation as much as ever. We will want to change others. We will dream of residing in someone else’s memory. Algorithms will do this to and for us on an unimaginable scale. Search will take us irrevocably into a fevered collective intelligence. This will seem normal—to both my son and me.
Quentin Hardy lectures at the School of Information Management and Systems. He is also the Silicon Valley bureau chief of Forbes magazine. Abelardo Morell’s photographs appear courtesy of Bonni Benrubi Gallery, New York, NY.
From the March April 2006 Can We Know Everything issue of California.