The contest began with a deliberately provocative question: in an age of pervasive AI, is human creativity “a waste of time”? The New York Times assembled reporter and occasional songwriter Matt Richtel with four lifelong professional musicians for a weekend of writing and recording, then placed their work beside AI-generated songs for listeners to judge.¹

The paper framed the exercise as a “cage match: carbon life-forms versus silicon chips,” with the audience asked to choose between tracks including “Bury Me” and “Sweetgrass Clay.” But the point was not simply to crown a winner. The experiment set out to test whether people could distinguish human work from machine-made music, whether the humans could make something better, and whether the friction of collaboration was worth it at all.¹

Its account of the first day lingered on the unhurried rituals of composition — musicians standing around, listening and closing their eyes — a sharp contrast with AI’s near-instant output. That contrast is central to the Times’s case: human songwriting may be inefficient, argumentative and uncertain, but those qualities are part of the creative process rather than evidence against it.¹

A separate reaction was blunter. It called the matchup “not much of a competition,” arguing that the AI songs “suck” and that listeners can easily tell them from the human tracks when heard side by side. In that reading, the experiment does not reveal an existential dead heat. It reveals how little effort a machine needs to produce a song — and how much human judgment, experience and collaboration still matter when the goal is music worth hearing.²