Following on Cohen’s thoughts about searching as desire, and the ways in which our desire (or even personality) becomes amenable to data mining, I give you Crystal, the tool which will help you write “successful” emails to other people based on data mining their emails to you.
I think this might help unpack a few things in Cohen:
—Algorithms start modeling people. Or perhaps “people”: not people with deep subjectivity or psychology. Rather, data mining people’s online behaviors—searching, writing emails—at a population level lets you predict what people “want.”
—Presumably, using Crystal requires subjecting your own emails to such scrutiny: you can use the “knowledge” derived from data-mining only if you yourself subject your own email writing (read: searching) to data-mining.
—That also suggests that if other people start using Crystal on you (or on their emails to you, and also if Crystal works), that suggests that emails will now be less annoying, more productive, and so on.
—For a long time, we have been members of a “population” understood statistically at the level of, say, the United States Census, or polling during presidential elections, or testing new drugs. But in recent years, our membership in population-level statistical models has affected, in insidious and profound way, our intimate lives. Facebook predicts whose feeds we want to see, Google predicts what search results we want to see, and now Crystal promises to “improve” our email by giving our correspondents a statistical model of what we want in our emails.