AI and the Beatles

LA is notorious for its traffic. And one of the few benefits of the pandemic is that I no longer need to drive to the office, and so I have a little more time in the day, some of which I now spend watching Netflix. I work with algorithms for a living, and I am amazed at how effective Netflix’s matching algorithms have become since I last tuned in a few years back. Based on my past viewing habits, the suggestions are really good at pointing me to shows that fit my interests—I will spare you dear reader what those are. I also know that they use my viewing and demographic data to manufacture new entertainment to match viewer interests—so even more shows especially designed for me are coming my way.

Yet I worry. I get presented with the same types of shows now. And let’s face it. In any genre, some content are really brilliant and original, but the rest are derivative, repetitive and boring. Could this algorithmic approach to entertainment and marketing become destructive? Consider the problem of a corporation like Disney or Netflix looking to create a “boy group” as a vehicle to help the corporation sell merchandising, advertising and other content to a “target market”. Would such an algorithm tell Disney to select boys without any formal musical training, living in a decaying city in a country far away from the target market, and with accents and mannerisms to boot that are alien to the biggest audience? And yet, whether you like their music or not, the Beatles turned out to be the most profitable musical act in history. Likewise, would an algorithm look to Queens or the Bronx for talent in order to create a whole new art form and business line—I am talking about the Sugar Hill Gang and Run DMC?

Not to pick on Disney’s algorithm, suppose Harvard’s admission committee is looking to admit that special high school senior that they think will eventually help popular music win its first Nobel Prize in literature. Would Harvard’s algorithm select a middling high school senior from the backcountry Midwest with virtually no singing or melodic ability or musical training—and yet we have Bob Dylan.

My hunch is that algorithmic entertainment is probably great at homogenizing product lines and simplifying choice—it makes it easy for the viewer to passively pick and watch shows. But serendipity—the residual in the algorithm—is probably where outlier talent and the big money is, at least for now (and I hope forever).