Wired For Thought by Jeffrey M Stibel
Harvard University Press, 2009.
Hardcover versions available online from Barnes and Noble and Amazon.
Kindle version also available from Amazon.com
Reviewed by Georgy S Thomas
In his preface to Wired for Thought, internet pioneer and brain scientist Jeffrey M Stibel offers the following explanation for his undertaking:
‘‘You can take away any phenomenon and study its parts for years, but until you step back far enough to see it in its entirety, you will not understand how it works and where it may go.” Stibel’s method of stepping back to view the internet is by comparing it with the human brain.
In his view, many of the biggest internet enterprises are successful because they understand the similarities between the brain and the internet. He points out the presence of brain scientists amongst decision-makers in several of the leading internet enterprises.
In the introduction, Stibel first lists out his big ideas.
The internet is a brain because it manifests intelligence, rather than merely reflecting it.
Humanlike thinking will emerge from the internet because of its network approach and mimicking of human weaknesses.
The internet is an evolution of the human brain.
The brain as a prediction machine is different from the way computers work, but is similar to the internet.
Creative destruction is another shared trait between the brain and the internet.
Language, considered uniquely human, is at the heart of the most important internet tool: search.
The internet will crash, but will get bigger and stronger with each collapse. Again a trait similar to the brain.
He then takes them up in detail chapter by chapter. Along the way, we get acquainted with concepts and terms like memes, intuition, forecasting, heuristics, fuzzy logic, polysemy, synset, spreading activation, encephalization, etc. We also encounter a few fascinating characters like Dan Dennett, Jim Anderson (both mentors of Stibel), Robert Metcalfe (whom Stibel rather shamelessly uses as a straw man) and Ray Kurzweil.
The book is peppered with nuggets of information. For instance, did you know that many of the biggest internet properties hand code particularly important web pages for easy retrieval? I didn’t until I read Stibel. On searching for confirmation on the net, I ran into a Q&A session by New York Times design director Khoi Vinh where he admitted to the practice at the Times. The shared concern seems to be that usage of web development applications like Dreamweaver and FrontPage (now Expression Web) as well as other WYSIWYG editors spoils the uniformity of the page’s look and feel across various browsers.
Also when Stibel, a pioneer and one of the thought leaders of the internet revolution, lists for us the characteristic features of the best web sites, we better make a note of that.
Of the big ideas discussed, the one with which Stibel appears to struggle a bit seems to be the argument that the internet is the evolution of the brain. In the beginning part of the book, while tracing the origins of cloud computing, Stibel discusses Richard Dawkins’ twin ideations of the selfish genes and the selfish memes. This allows Stibel to wonder, ”Could it be, then, that the selfish gene became frustrated with the slow evolution of the human brain, and so leapt the fence from the organic world to the inorganic? Could selfish genes have created selfish memes to do their work? And is that what has led us to selfish software? Is that why humankind, imprisoned as we are in carbon molecules, is driven to invent machines made of sand and metal?”
He then goes on to answer these questions with an ”I am not completely convinced…but… the perspective is exhilarating” approach.
But a few pages further, while discussing evolution, Stibel seems to have shed his diffidence. He says: ”The human brain evolved as a hardwired device until about 150,000 years ago, when….the mind began to outrace evolution. Rather than experiencing a radical improvement in the brain itself, humankind began to develop software _ cultural software _ that has improved over time.” A few sentences later comes the emphatic statement, though still hanging onto the coattails of Richard Dawkins and Dan Dennett: ”…the evolution of memes is not merely analogous to genetic evolution; it’s an extension of it.” Looks like the author’s views have also evolved during the course of writing the book!
In a similar vein, the book is also not completely free from factual inconsistencies. In page 74, we read the statement: ”Each neuron in the brain has about 7,000 connections, for a total of some 100 trillion connections.” By the time we reach page 124, we read thus: ”The average neuron has roughly 10,000 connections to other neurons in the brain.” It’s evolution at work again!
The Centrality of Language for Search
The chapter on language being central to both the brain and the internet offers a fascinating insight into how search engines derive meaning from language. Stibel is able to speak from experience because Simpli.com, a search engine which he founded, used the expanded version of a program called WordNet to do just that. He first states the problem: Choosing the right meaning of words having many possible meanings comes effortlessly for people, but is downright impossible for computers.
Then he goes on to explain how WordNet overcame the limitation by first forming a hierarchy of words, and then because words have multiple meanings, building sets of synonyms or synsets which operate together for a single meaning. WordNet then uses a process called spreading activation, or the wiring and firing together of synsets, to build context into language. Google, which acquired the knowhow through an acquisition, now uses it to power AdSense.
The Long Tail Doesn’t Apply
Stibel observes how Google and other search engines tie word meanings to frequency of use for ordering and ranking results. At work here is Zipf’s law, named after linguist George Kingsley Zipf. The law predicts that the more frequent words are used a great deal more than less frequent words. What this means is that as one goes down the list, there’s a very rapid drop in frequency of appearance.
And it would seem that the law also applies to the degree of ambiguity of words. ”The more frequently a word appears in text, the more different meanings it has.” Thus the most common words are also the most ambiguous. Now you know why academicians have an easy time writing technical articles for journals and a harrowing time writing an easy-to-read general interest book, Stibel tells us. The author, mercifully, doesn’t suffer from this weakness.
The Zipf law’s direct relevance to internet search is in the accessibility of web pages: The most common pages are accessed very frequently, while the bulk of the web pages are almost never seen. Based on this, Stibel goes on to state that the long tail does not generally apply to websites, internet, or search. Chris Anderson, are you listening?
The Laws of Networking
In the course of developing his big idea that the internet will shrink like the brain, Stibel lists out for us the laws of networking, seen through its three stages. Because he cannot as yet prove his contention that the internet will one day collapse, he goes on to take up case studies of networks within the internet which have experienced the three stages of growth. The lessons he draws up from this analysis are invaluable to all web entrepreneurs. Want to know why MySpace shouldn’t fight its slowing growth? And what are the prospects for Facebook? Well, you’ve the answers in these pages.
I particularly liked Stibel’s observation that ”the value of a network does not increase with size when the size of the network makes it impossible to derive value from it”. How true. Stibel ends the book by sticking his neck out and charting out a possible path through which the internet could evolve. Does he see Google still presiding over as the lord of the worldwide web manor? I’m not letting the cat out of the bag here. Better read the book and find out on your own.
The author is a Bangalore-based tech entrepreneur.