Spotify says its best developers haven’t written a line of code since December, thanks to AI

Has the rise of AI coding marked a pivotal shift in software development? For Spotify, this seems to be the case. During its recent fourth-quarter earnings report, co-CEO Gustav Söderström revealed that the company’s top developers have not written any code since December, a testament to the transformative power of artificial intelligence in their operations. Spotify has notably leveraged AI to enhance its development processes, successfully deploying over 50 new features and updates to its streaming app throughout 2025. Among the latest innovations are AI-driven features like Prompted Playlists, Page Match for audiobooks, and About This Song, all of which have been introduced in recent weeks. Central to this rapid development is an internal tool named "Honk," which enables engineers to expedite coding and product rollout. This system allows for real-time, remote code adjustments using generative AI, specifically utilizing Claude Code. Söderström provided a vivid example: an engineer can instruct Claude to rectify a bug or introduce a new feature for the iOS app while commuting via Slack on their mobile device. Once Claude completes the task, the engineer receives an updated version of the app through Slack, allowing them to merge it into production before even arriving at the office. Spotify credits this AI integration with significantly hastening its coding and deployment processes. Söderström expressed optimism regarding future developments, indicating that this progress is merely the beginning of AI's potential in the company’s operations. Additionally, Söderström highlighted Spotify's unique dataset, which he claims cannot be replicated by other large language models (LLMs) that rely on publicly available resources like Wikipedia. He noted the subjective nature of music preferences, which can vary widely based on location. For example, while hip-hop may be favored by many Americans, others might prefer genres like death metal or EDM. "This is a dataset we are actively building at an unprecedented scale, and we see its quality improving with each model retraining," he pointed out. In response to inquiries about AI-generated music, Spotify clarified that it allows artists and labels to denote how a track was created in its metadata while actively monitoring the platform to combat spam.