(First paragraphs of my next book. I like them.)
In 2001, the American philosopher Hubert Dreyfus published On the Internet. The book continued the kind of work that Dreyfus had been doing throughout his career: debunking, from a pragmatic and phenomenological perspective, the hype surrounding emergent forms of information technology (IT). In this case, Dreyfus’s target was the chaotic organization, or rather disorganization, of what we then called the World Wide Web. It could be pretty frustrating, at the turn of the twenty-first century. As long as you knew what specific www you wanted, it worked fine (connection speeds aside). But if you only knew what general kind of site you wanted, the millenial internet was almost useless. An online search for information about tortoises, to use Dreyfus’s bookish but accurate example, might lead you to sites on pre-Socratic metaphysics (because a tortoise features prominently in Zeno’s famous Paradox). For Dreyfus, the failure of online search was predictable, rooted in the fundamental inability of computer networks, being merely syntactic (formal) engines, to capture the semantics (meanings) of the symbols they processed. A “tortoise” was a “tortoise” was a “tortoise” was a “tortoise” (…) as far as the web was concerned. What was worse, the amount of information available online was growing exponentially—into a wilderness of tortoises, within which it could only become harder and harder to find the specimen you wanted. The ineffectiveness of then-available search engines was an empirical matter, pegged even by their designers at about 30% precision. Dreyfus’s theoretical argument was to the effect that we had absolutely no reason to think that figure would improve. “One thing is sure,” Dreyfus concluded, grimly: “As the Web grows, Net users who leave their bodies behind and become dependent on syntactic Web crawlers and search engines will have to be resigned to picking through heaps of junk in the hope of sometimes finding the information they desire.”
In 2009, Dreyfus brought out a second edition of On the Internet. The revised first chapter draws a big red line through the original. For as Dreyfus freely and frankly acknowledges, his critique of less than a decade before was being undermined—in something like the etymological, fatal, sapper’s sense of that word—even as he was writing it. There was this computer scientist at Stanford called Terry Winograd, Dreyfus recalls; and he had a couple of bright graduate students (Larry and Sergei) working on the search problem. The young men realized that the meaningfulness of a given website for a specific search term could be quantified by clicks under that term. So, for example, if searches for “tortoise” yielded many clicks on a given zoological website (among those on the initial hitlist), that data counted as “votes” from the searching page for the clicked-on page. A page that received many such votes was “important” for the search term; the relative “importance” of the searching page could also be factored in; and the whole process would become stronger, in recursive (or self-reinforcing) fashion, every single time it was run. Out of these insights, the grad students built an algorithm. Around the algorithm, a program. Around the program, a search engine. And the rest, as they say, is now. The Dreyfus of 2009 makes no attempt to fold Google into any ongoing critique, reaching back to 2001. He just notes, baldly and a little glumly, that “pessimism has turned to optimism” in this area of computer science. He even gives Page and Brin, in the revised first chapter, the last word, to the effect that there is, after all, “a bright future for search.” The first chapter of On the Internet, 2nd edition, presents a rare instance of a brilliant, distinguished, and highly polemical scholar looking back over his own previous work—noting its falsification by the very forces he had attempted to critique—and finding that all he can say is: “Oh.”
Libber leeber 