The Internet and the Brain

Just thought I will share an article I read...

Having implants in your brain that let you turn on lights by thought alone probably sounds like science fiction. However, scientists are already performing experiments like “BrainGate,” which lets people “control electronics” with their minds. Such advances on the frontier between the brain and electronics are happening quickly. New inventions are changing how people view intelligence and providing fresh perspectives about the brain. Scientists are applying new discoveries about the brain to better understand and improve the internet’s functionality. These converging areas – studies of the brain and of the internet – are so close to merging that you could say “the Internet is evolving into a brain.” To understand what that means, and how viewing the internet this way will change it, consider the brain.

The brain – or, at least, its “outermost” layer – is more like a crumpled piece of paper with information encoded on it in Braille than it is like a grayish pink lump. The crumpling allows points on the paper (or neurons in the brain) to come close together, enabling fast communication. The brain conducts its activities with “distributed computing,” which uses “parallel processing,” not “serial computing,” so your brain can do several things at once.

The human brain hasn’t always been what it is today. “Six million years ago,” a human ancestor’s brain was the size of a chimp’s, but “2.5 million years ago” an evolutionary leap led to the development of the human cerebral cortex, which enables higher-level thinking. In fact, “there is nothing special about your brain. It is made up of the same sort of carbon molecules as the rest of your body. The brain breaks down to a physical network” of neurons, the memories in that network and the network’s active processing. The internet doesn’t yet have as many connections as the brain, but the internet is growing, advancing at a furious pace from resembling a brain to being a brain. It shares strengths and weaknesses with the brain and will become able to think. It is already “creating a collective consciousness.”

What Does It Mean to Build a Brain?

Computers process data extremely quickly. They can search their ideal, mechanized memories far faster than your brain’s 100 billion neurons can perform. It’s appealing to think that building faster machines with bigger memories will lead to intelligent, or even imaginative, computers. However, that’s not the case. Your brain doesn’t work in a computer’s focused way. Instead, your brain searches, then backs off, then ruminates, then pops up unrelated thoughts about your grocery list. That sort of recursive, leaping function – rather than simple, if speedy, calculations – seems to be integral to true intelligence. Creating an artificial intelligence would require a machine that uses a brainlike “loopy, iterative process.” This new intelligence would have to be able to describe its processes to itself and to think about them. It would need the capacity to question, self-examine, guess, “gaze out the window” and learn from repetition and mistakes.

“Pattern recognition,” a foundation of parallel processing, is central to human thought. As you sort data, you seek patterns, which lead to meaning and let you avoid extended, computer-like processing. Building computers that could recognize patterns would require assembling and networking “synthetic neurons” and then clustering them for greater ambiguity. Researchers working on such developments say that science could conceivably recreate the capacity of the human cortex (the locus of higher thinking). That would require a million common computers networked with a “terabyte of memory,” but it is possible. This kind of intelligence already exists, in a way, at Google’s “server farms,” which facilitate superior performance not with a single supercomputer but with “cloud computing”: many smaller computers working in unison.

The parallel between biology and technology matters because, rather than just being constructed, current technological advances are developing in an evolutionary way. Software advances parallel Richard Dawkins’s conception of evolution not as a process where species compete for survival but as a process where genes compete for dominance – each one selfishly pursuing its own genetic agenda over other genes. Dawkins also articulated the concept of a “meme,” or a “unit of culture” (any “thought or idea,” from a rock song to a way of using tools). Memes propagate from person to person (think of how jokes spread), mutating and changing their human hosts as viruses do. Such conceptual evolution applies to software, even “selfish software,” which spreads widely. In this model, technological evolution is simply an extension of biological evolution.

Intuition and Information

Your “brain is an excellent prediction machine.” While you probably aren’t intuitive about long-range forecasts or complex systems like weather fronts, all humans can extrapolate meaning and short-term future actions based on signs. This capacity, which evolution built into your brain over generations of hunting, is central to intuition, a defining trait of the brain. If you see a “coiled” shape on a path, you might jump away before you’re sure it’s really a snake. That’s intuition. Your brain combines the shape with its data on snakes and moves your body before your conscious mind can determine if the coil is a snake or a rope. It uses pattern recognition to decide quickly with limited data, a process that is core to human thought. The way intuition functions points to an apparent paradox.  It may seem as if the more data you have, the better decisions you’d make, and that you could make great decisions if you had “perfect information and perfect calculations.” In fact, access to fewer facts – such as having just “one good reason” – often leads to better, faster choices, while “too much information” can overwhelm your ability to decide.

Intuition is the positive side of a neural weakness: Because human beings recall things imperfectly and process data relatively slowly, the brain has had to become skilled at recognizing patterns with limited input. When you shift your perspective, your brain doesn’t have to acquire all possible data to define the new viewpoint. Instead, your brain quickly compares the new image to your “memory patterns” and recognizes it. Your brain does this with your life, too: You continually project forward into the future, modeling possibilities based on past experiences. This tendency is a major distinction between people and computers. Humans don’t operate with pure logic, and they never calculate all the possible choices before making a plan. Instead, by using “rules of thumb” or “heuristics” that recognize patterns, people quickly assemble a good-enough option based on a limited subset of possibilities. Such heuristic reasoning is central to some of the achievements in artificial intelligence, such as computers that win at chess.

Some commercial internet users want to turn the web into more of a predication machine. Netflix and Amazon are trying to improve the web’s ability to suggest products customers might like. The best algorithms sometimes make surprising suggestions that work like intuitive leaps, such as suggesting a book on getting rid of clutter for someone who requests a book on Zen. Making this connection requires uniting topics not through keywords or links but through conceptual implication: Both de-cluttering and Zen emphasize clarity and openness.

Memory and “Creative Destruction”

If you’ve ever forgotten a meeting or a phone number, you might wish for a perfect memory. However, that would backfire. If you could recall everything, “useless information” would overwhelm your brain. Even without that onslaught, your brain constantly disposes of “old thoughts” and replaces them with new ones. Memory works in a “constant ebb and flow”; new memories appear and others fade. This process, which is central to the brain’s workings, parallels economist Joseph Schumpeter’s principle of creative destruction, in which new firms, processes and products continually arise and displace old ones. Across the internet, sites bloom and die, “deprovisioning” as they yield to more useful replacements.

This principle is very visible in marketing. Campaigns once followed a set pattern: Figure out the customers, appeal to their desires and persuade them to buy a product. However, the web provides much more feedback a lot faster. You no longer have to figure out the best way to reach buyers. You can just try different tactics “and see what works.” You can create variations on every ad, test every tagline and run real-time experiments. These trends are happening on a larger scale with the web itself. “Web 2.0” brought an explosion of new “dynamic” activity: “collaboration, social networking, podcasting, wikis, blogs and collective consciousness.” Interactivity abounds. The evolution is ongoing. Tim Berners-Lee, the inventor of the World Wide Web, is working on a “semantic web” where the internet analyzes its own data and the relationships among its parts.

Neurons and Networks

Technically, the internet is a physical structure: “a network of computers and phone lines.” The web, by contrast, is a “system of interlinked hypertext documents.” This system – which allows the existence of the marvelous interactivity that defines the internet – functions much like your brain’s neurons. Just as neurons store memories and produce thoughts, websites store content and facilitate linkages. Each neuron “has about 7,000 connections,” so your brain has about “100 trillion connections.” The brain links emotions to your memories of the experiences that form these connections, so when you recall one element, you also summon associated elements.

Think of your memories as your brain’s software, the programming that allows it to function. Memories work like the internet’s websites. While the web doesn’t hold nearly as many connections as your brain, the web is still growing at a gallop, getting faster and more data dense. Like memes or organisms, sites fight to dominate their setting. They compete by being “novel and appealing,” filling their niches, linking usefully and fitting the way people process data. In the past, some people thought search engines would function, librarian-style, to find data on the internet. However, this tactic was an evolutionary dead end. Rather than using a clearly defined system like a library, the best search engines work like the brain – sorting information coded in language. As search engines advanced, they became capable of sorting the complex tangles of associations that define human language, especially those governing short words with multiple meanings. Now when search engines see “bat” on the same web page as “diamond,” they realize the page is about baseball, not flying mammals. Search engines sort websites by content, meaning, popularity, associations and the quality of their links.

According to an axiom called Metcalfe’s law, created by net pioneer Bob Metcalfe, networks benefit first movers substantially and grow continually, becoming exponentially more valuable as more users join. Think of the telephone. One phone is useless. With two, the users can call only each other, but a billion phones enable vastly abundant connections. Metcalfe’s law no longer holds when a network gets too big and strives for “equilibrium,” the limit that constrains any network’s workable growth. Like the brain, networks go through a period of intense expansion, a time of collapse or slowing growth and then equilibrium. If a network develops beyond equilibrium, it collapses. The growth of MySpace, with 100 million users, is slowing as its size becomes unwieldy. Facebook is less likely to stumble because it is essentially a “network of networks”: You don’t link to everyone, which would push the network beyond its functional limits. You connect to a cluster of people, that cluster connects to another, and so on.

The Internet, Business and the Future

An old saying maintains that people only use 10% of their brains. False. Humans use 100% of their brains nearly all the time; no vast, untapped capacity exists. Yet people can build tools that “expand intellect.” For now, the internet is the greatest such tool. As the net grows more sophisticated, it will get better at predicting your wishes. It will know your values and interests, anticipate your needs and search itself for your benefit, resulting in a “highly personalized” online experience, where the web presents information that matches what you want. This sounds very positive, but it also holds darker implications for the future.

The internet will track more of what you do online, correlating different elements of your life to pierce your privacy and reveal your secrets. Yet, fundamentally, the internet’s future resides in serving as a tool that multiplies the human mind’s power and efficacy, including by becoming intelligent itself. It is too soon to say what online intelligence will look like or what it will mean, but a kind of artificial intelligence will emerge, enabled by how the internet is evolving with a constant “flood of new websites.” New algorithms will let the web pull data from many sources and create fresh, meaningful patterns with more nuanced, less generic links. Over time, the internet will develop both intuition and common sense. “Expect certain systems on the internet to reach the level of consciousness we reserve only for the smartest of animals, including humans.”