Why the internet's development will mimic that of the human brain
Of all the futurologists, none is more bold or outspoken than Ray Kurzweil. He is no high-tech ingénue. Kurzweil invented the flatbed scanner and—after a meeting with little Stevie Wonder (he was still little Stevie Wonder back then)—also invented the first device to read books for the blind, along with the voice synthesizer to turn the print in those books into speech. (Kurzweil later invented the musical synthesizer: remember the Paul McCartney/Stevie Wonder hit "Ebony and Ivory?" That's Ray's synthesizer at work.) Kurzweil was named inventor of the year by MIT and the Boston Museum of Science. He is a member of the US Patent Office National Inventors Hall of Fame and has been given the National Medal of Technology.
What does Kurzweil predict for the future? (Cue the flashback music.) "It is now 2019 . . . a $1,000 computing device is now approximately equal to the computational ability of the human brain . . . the vast majority of transactions now include a simulated person . . . there are widespread reports of computers passing the Turing test [for intelligence] . . ."
And by 2029? Kurzweil predicts that a $1,000 unit of computation will have the computing capacity of approximately 1,000 human brains. Also, he says: "Automated agents are now learning on their own . . . significant knowledge is being created by machines with little or no human intervention . . . The majority of human communication does not involve a human being. The majority of communication involving a human is between a human and a machine."
Is Kurzweil on target? I don't know. He is a braver man than I am. He's smart, too: when he made these predictions in 1998, he said of 2009 that personal computers with high-resolution visual displays would come in a range of sizes (think iPhones); cables would be disappearing (think wi-fi); purchases and reservations would take place with a virtual personality (think Expedia and Amazon); intelligent "courseware" would emerge as a common means of learning (think online education); accelerating returns from the advance of computer technology would result in continued economic expansion (true, at least until recently); and human musicians would routinely jam with "cybernetic" musicians (think Rockstar and Guitar Hero).
Kurzweil was concerned mainly with computers, but where is the internet heading? Luckily for me, in the case of the Web we have the brain as a guide, so there are some clues that can be drawn from nature. As a baby's brain develops, there is an explosion of neural development. In the early days of the internet, there was a combinatorial explosion of new websites, but that has slowed over the past few years. The internet has passed through its early embryonic stage. Websites continue to be built, but at much slower rates.
But those sites have increasing numbers of links to each other, mimicking the way a newborn starts to explore the world and build causal relationships that translate into neural and semantic connections. Given the way search engines spider and rank websites, there is a strong incentive to increase the number of links, and that is propelling development forward. Like a growing brain, link relevance is becoming more important, websites are decaying and dying. Links that have no traffic or usefulness are disappearing.
Eventually, internet algorithms will mature and, like the brain, start discounting, discarding, and penalizing links that are irrelevant, unnatural, or fake. At some point, the Web itself, through one of its governing bodies (the CIAs of the Web: W3C or ICANN) may even take action and proactively remove bad links. We will see a constriction at some point of websites as well as links, because such limits are central to an evolutionary path that remains "fast and frugal."
Neurons are different from websites: whereas a neuron's power comes from the network of neurons it is connected to, an individual website can hold (in theory) infinite amounts of information. On the internet, individual websites are the gravity of the whole system. But this situation will change as the internet matures into an adult brain. Although websites store information, their power, too, lies in the ways in which they connect to one another. A network of sites creates a density of content, relevance, and independence that no individual site can seize, just as a catalytic converter is far more powerful in the context of a combustion engine.
But this horsepower cannot be used . . . yet. As the Web evolves, the value of these networks of networks will increase: new programs, algorithms, spiders, and frameworks will be developed to leverage that power. In the near future, websites will pull information from various sites to create collages of new information; search engines will leverage link structure to determine category information and not only popularity; the Web itself will enable the formation of content clusters that will in turn enable a semantic Web; and increasingly complex neural networks will evolve that will allow for communication, networking, and thought.
Equally important, the underlying structure of links on the Web will change. Currently, links have one dimension: the royal blue connections from one site to another. But that is not how the brain works. The brain has two types of links: inbound (axons) and outbound (dendrites). Memories are even richer in their semantic links (i.e., a 911 is a Porsche; a Porsche is a fast car; fast cars get more speeding tickets; speeding tickets and fast cars cost a lot of money). The Web will evolve to incorporate richer links and relationships. Sites will need to display the links that are coming in, and that will provide a proxy for what the site is about.
Neurons also weight their links, and this is reflected in the relative strengths of memories, but that does not happen on the Web. Internet links will evolve to incorporate weighting, allowing everyone to see how relevant and important a link is on a given page. Think of it as a colour code, where blue might represent the best links on the page, red second best, yellow third, and so on. In the beginning, site owners may choose their most important links, but the system will evolve to allow dynamic weighting based on how many people click on each link or how relevant the link is.
The ultimate goal is that information will become as readily available as it is in our brains. We have no search engines in our minds. Instead, information spreads across a series of connected memories, links in our brains to other information—this is how we think. One day, entering thoughts, feelings, or aspirations into the Web will generate a similar process—it will open up endless information, dynamically organized based on your thoughts. Ultimately, this will render search engines (and yes, even Google) obsolete.
Sites will also become modular, enabling not only independent editors (as we now see at Wikipedia, where anyone can edit the online encyclopedia) but also the Web itself to modify the very sites that are a part of it. Just as the brain can modify memories and even create new ones, the internet will be able to modify and create websites. Sound strange? It shouldn't be hard to imagine, because it is already happening to some extent. A news report you saw on CNN.com may have been posted earlier on Reuters and picked up by millions of sites automatically, including CNN. In the brain and on the Web, information is not static.
The brain is a sponge in that it soaks up information, but from that point, information is manipulated and integrated into the rest of the knowledge stored in the brain. In advertising, this means you need to position your message appropriately. On the internet, it means you must let go and be willing to let your information change and adapt as the social network reacts to it.
Looking ahead, new websites will be created that pull together clusters of information that are tightly linked. The network of networks that we find on Facebook will be applied to the Web. No longer will you need to go to a dozen car sites or an aggregator like Autobytel or even Google to find information about the best hybrid car. Instead, the internet will create a page with only the most interesting and relevant information across the Net. Even today, there is a small start-up called Kosmix.com that is building what I call dynamic portals—search for anything and it returns a page dynamically generated from content all over the Web. The Web will one day be able to generate a Web page specific to your request, just as the brain fires off new symphonies of excitation when it encounters a novel subject.
And what of Moore's and Metcalfe's laws? They will both peter out under the weight of their own gravitational force. This should not be surprising, because all exponential growth eventually stalls. But what does that say for computers and the Web? My prediction is that it will fuel a new era of productivity in which software advances will outpace the growth we've seen in hardware; intelligence will emerge, not from brute force but from educated guesses. Remember that the brain is a slow computer, so we do not gain intelligence from sheer size or speed. What makes us smart is that we are loopy, slow, and speculative. When the internet can no longer count on productivity gains from brute force, it will turn to other measures, and that will surely come from mimicking the power of the brain.
Reprinted by permission of Harvard Business Press. Excerpt from Wired for Thought: How the Brain is Shaping the Future of the Internet. Copyright 2009 Jeffrey M. Stibel. All rights reserved.
Posted 14 August 2009 : Director.co.uk
