Has AI-driven development reached a critical turning point? For Spotify, the answer appears to be a definitive yes. During the company’s fourth-quarter earnings call this week, co-CEO Gustav Söderström revealed that the platform’s top developers “have not written a single line of code since December.” This striking statement was part of a broader discussion on how Spotify is leveraging artificial intelligence to dramatically accelerate its engineering and product development cycles.

Throughout 2023, the company shipped over 50 new features and updates to its streaming application. More recently, it has launched a series of AI-powered additions, including Prompted Playlists, Page Match for audiobooks, and the About This Song feature—all released within the past few weeks.

A key driver behind this velocity is an internal system called “Honk,” which Spotify highlighted during the analyst call. This platform enables remote, real-time code deployment using generative AI, specifically Claude Code. Söderström offered a vivid example: “An engineer at Spotify on their morning commute from Slack on their cell phone can tell Claude to fix a bug or add a new feature to the iOS app. And once Claude finishes that work, the engineer then gets a new version of the app, pushed to them on Slack on their phone, so that he can then merge it to production, all before they even arrive at the office.”

Spotify credits Honk with speeding up coding and deployment “tremendously.” Looking ahead, Söderström emphasized that this is “not being the end of the line in terms of AI development, just the beginning.”

He also pointed to Spotify’s unique advantage in building a proprietary dataset that other large language models cannot easily replicate. Unlike factual resources such as Wikipedia, music-related queries often lack a single correct answer. For instance, definitions of “workout music” vary widely by individual and region: while many Americans might prefer hip-hop, millions opt for death metal, and while EDM is popular for workouts in parts of Europe, many Scandinavians favor heavy metal.

“This is a dataset that we are building right now that no one else is really building. It does not exist at this scale. And we see it improving every time we retrain our models,” Söderström noted.

When analysts inquired about Spotify’s stance on AI-generated music, the company clarified its current approach: it allows artists and labels to specify in a track’s metadata how a song was created, while continuing to monitor the platform for spam content.