On sites such as Amazon and iTunes, homophily is a selling point: it’s the basis for “collaborative filtering”, whereby you’re recommended books and music on the basis of what others who made the same purchase - people like you - also enjoyed.

The unspoken assumption here is that you know what you like - that satisfying your existing preferences, and maybe expanding them a little around the edges, is the path to fulfillment. But if happiness research has taught us anything, it’s that we’re terrible at predicting what will bring us pleasure. Might we end up happier by exposing ourselves more often to serendipity, or even, specifically, to the people and things we don’t think we’d like?

You don’t need technology to do that, but then again, technology needn’t be the enemy: Facebook could easily offer a list of the People You’re Least Likely To Know; imagine what that could do for cross-cultural understanding. And I love the Unsuggester, a feature of the books site LibraryThing.com: enter a book you’ve recently read, and it’ll provide a list of titles least likely to appear alongside it on other people’s bookshelves. Tell it you’re a fan of Kant’s Critique Of Pure Reason, and it’ll suggest you read Confessions Of A Shopaholic by Sophie Kinsella. And maybe you should.

He was describing the ballet of the train station. But his description could just as easily have applied to the Internet. Think about it: Serendipitous encounters between people who know each other well, sort of well, and not at all. People of every type, and with every type of agenda, trying to meet up with others who share that same agenda. An environment that’s alive at all hours, populated by all types, and is, most of the time, pretty safe. What he was saying, really, was that New York had become the Web. Or perhaps more, even: that New York was the Web before the Web was the Web, characterized by the same free-flowing interaction, 24/7 rhythms, subgroups, and demimondes.