Spotify’s Best Developers Haven’t Written Code in Two Months

“Our best developers have not written a single line of code since December.” That’s not from a think piece about the future of work—it’s from Spotify’s Q4 2025 earnings call.

The Core Insight

Spotify has built an internal system called “Honk” that lets engineers deploy code changes from their phones via Slack. Using Claude Code, an engineer on their morning commute can tell the AI to fix a bug or add a feature, receive a new app build pushed back to them, and merge to production—all before reaching the office.

The result? Over 50 new features shipped throughout 2025, with AI-powered Prompted Playlists, Page Match for audiobooks, and About This Song all launching in the past few weeks. Velocity has increased “tremendously.”

When a company with 751 million monthly active users tells shareholders that their top engineers have stopped writing code, we’ve reached a tipping point.

Why This Matters

This isn’t “AI assisted coding.” It’s AI-driven development with human oversight. The engineer’s job shifts from writing code to reviewing AI-generated changes and deciding whether to ship them.

The implications are staggering:

• Location independence: Why does it matter if you’re at your desk when you can ship from your phone?
• Speed multiplication: Time-to-production drops from hours to minutes
• Role transformation: Senior engineers become technical product managers, not coders
• Hiring criteria shift: What matters isn’t coding speed—it’s judgment about what to build and whether it’s correct

Spotify’s co-CEO Gustav Söderström called this “not the end of the line in terms of AI development, just the beginning.”

Key Takeaways

• “Honk” system enables mobile-first development: Bug fixes ship from Slack on your commute
• Claude Code is the engine: Not a custom model, but Anthropic’s coding assistant
• Data moat matters more than code: Spotify is building unique music preference datasets that LLMs can’t commoditize
• 50+ features shipped in 2025: AI velocity is showing up in actual product output
• 751M MAU reached: Q4 was their best quarter ever for user growth

The Data Play

The more interesting strategic angle is Spotify’s data advantage. When you ask “what’s workout music,” there’s no factual answer—Americans tend toward hip-hop, Europeans toward EDM, Scandinavians toward heavy metal. This preference data is unique.

“This is a dataset that we are building right now that no one else is really building,” Söderström noted. “It does not exist at this scale.”

While other companies worry about LLMs commoditizing their data, Spotify is accumulating preference graphs that can’t be scraped from the public internet.

Looking Ahead

If Spotify’s best engineers aren’t writing code, what are they doing? The answer seems to be: deciding what code should be written, reviewing what the AI produces, and thinking about product strategy.

This is the clearest signal yet that we’re moving from “AI as tool” to “AI as workforce.” The question for every software company is whether they can build their own “Honk”—or whether Spotify’s head start will compound into an insurmountable velocity advantage.

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