The way engineers work at Spotify It's changing from top to bottom. In a matter of months, the company's most valued technical profiles have gone from spending hours in front of the code editor to managing artificial intelligence systems that do much of the heavy lifting for them.
According to the platform's executives streamingThe star developers haven't written any code manually since December. This isn't just hype; it's the result of a firm commitment to integration. Generative AI throughout the entire software development flow, something that is already noticeable both in the inner workings of the app and in the new features that reach users in Spain and the rest of Europe.
“Spotify’s best programmers haven’t written a single line of code since December.”
During the presentation of the results of fourth quarter of 2025Spotify's co-CEO, Gustav Söderström, uttered a phrase that has resonated throughout the tech sector: their best engineers "They haven't written a single line of code since December."In other words, top-performing profiles no longer program in the traditional way.
In that same speech, Söderström explained that these developers have replaced the keyboard with a role much closer to that of architects and supervisors of automated systems. Instead of typing instruction by instruction, they define what needs to be built, how the system should behave, and validate whether the code proposed by the AI meets the requirements.
The executive insisted that this shift is not an isolated experiment, but rather the starting point of a new stage for the company. The idea is clear: engineers stop being “builders” who write code and start conducting the orchestra, while AI models take care of repetitive and high-volume tasks.
This discourse fits within a context in which the Artificial Intelligence It is already surpassing the average human in linguistic creativity, according to studies such as the one from the University of Montreal. Many professionals fear that this same ability will be completely transferred to programming, relegating traditional human work to a secondary role.
In the case of Spotify, however, the official message is that the engineers still have the last wordAI proposes and generates the code, but it is people who review it, correct it, and decide what goes into production.
Honk and Claude Code: This is how the automation brain works at Spotify
The heart of this change is HonkIt's an internal system developed by Spotify to integrate generative AI into the daily work of its teams. It's not just a one-off assistant, but an infrastructure that connects advanced models with the engineers' standard work tools.
Within Honk, the use of Claude CodeA specialized programming model that can generate new features, fix bugs, or rewrite entire modules. What's interesting is how it integrates into... Slack and in automation systems (ChatOps), so that developers can work with AI from their mobile phone or laptop without leaving their usual communication environment.
Söderström gave a very graphic example to illustrate this: a Spotify engineer, on his morning commute to the office, opens Slack on his mobile phone and asks Claude to Fix a bug or add a new feature to the iOS appWhile the person is still on the subway or bus, the model generates the change and builds an updated version of the application.
Once the process is complete, that new build is sent back to Slack so the engineer can review it. If everything is correct, it can be integrated into the production line without even turning on the office computer. According to the company, this cycle significantly reduces “tremendous” deployment times and brings the changes to the user almost in real time.
This scheme fits with an increasingly visible trend in the European technology ecosystem: AI is ceasing to be an accessory tool and is becoming a centerpiece of the development pipeline, deeply embedded in processes, reviews and launches.
More than 50 new features in one year powered by AI
The commitment to automation isn't just talk. Spotify has highlighted that, by 2025 alone, it had deployed over 50 relevant functions and settings on its platform. A pace that the company directly links to the adoption of Honk and Claude Code in its internal workflows.
Among the most visible new features for users in Spain and the rest of Europe are the following: text-generated playlistsThese are also known as Prompt Playlists. They work simply: the user types in natural language what they want to listen to—for example, “relaxing rock to study on a rainy afternoon”—and the system automatically creates a music selection tailored to that request.
In the field of audiobooks, the platform has introduced Page MatchThis tool allows you to scan a page of a physical book and jump to the same point in the audio version. This functionality combines text recognition with AI models to synchronize both reading experiences.
Another addition is About This SongThis option offers contextual information about the song playing: production details, interesting facts, or the story behind the track. This extra layer is already integrated into the listening experience of many European users, adding depth to the platform beyond simple playback.
Furthermore, the push for automation has gone hand in hand with other recent functionalities, such as AI-generated playlists, improvements to discovery tools and continuous adjustments to the interface, which benefit from both language models and the ability to deploy changes very quickly.
New Spotify features: chats, ChatGPT integration, and audio improvements
Alongside the change in programming methods, Spotify is rolling out a range of features that directly affect the daily use of the app, many of them developed and tested in that environment. AI assisted programming.
One of the most striking is the incorporation of individual and group chats within the applicationThanks to this feature, users can comment on songs or audiobooks in native chat spaces, without having to jump to WhatsApp or other platforms to share recommendations or discuss a collaborative list.
The system is designed with certain limitations: You can only start chats with people with whom you have previously shared content.For example, through collaborative playlists, Jams, or Mixes. With this, Spotify attempts to maintain some control over who can contact whom within the app's environment.
Another relevant novelty is the Spotify integration in ChatGPTWhen mentioning the service in a conversation with the chatbotIt is possible to request lists tailored to different moods or situations — from training to concentrating at work — and receive personalized recommendations almost instantly.
The option to exclude specific songs from the Taste ProfileThis is especially useful for those who use the platform for white noise, children's music, or specific sounds that they don't want to interfere with their recommendations. In this way, the music suggestions are better tailored to the listener's actual preferences.
Lossless Audio, Mix and the new Weekly Wrapped
In the realm of audio alone, Spotify has strengthened its offering for users of Spotify Premium with the arrival of lossless audioAvailable in FLAC quality up to 24-bit/44,1 kHz in all markets where the service operates. This improvement is aimed at those seeking higher fidelity playback, something especially valued in European markets with a strong audiophile culture.
In addition to this, the company has enhanced the function MixThis feature allows you to create smooth transitions between songs, adjust the equalizer, and regulate the volume for more seamless playback. The goal is to make the transition between tracks as smooth as possible, something that's noticeable both in workout sessions and when listening to playlists designed for concentration.
Another recent bet is the version Wrapped weeklyIt's a kind of miniature version of the classic year-end summary. Each week, users receive personalized statistics about their listening habits, with data on their most frequently played artists, songs, and genres.
This weekly Wrapped includes a Image ready to share on social media and the ability to send these statistics directly from the app to friends within Spotify or through external platforms like Instagram or WhatsApp. With this, the company reinforces the social dimension of listening, especially popular among young users in Spain and other European countries.
According to the platform itself, many of these features have been able to be deployed so quickly thanks to the intensive use of AI in the code and in data analysis, closing the circle between development, experimentation and final product.
From the developer who types code to the architect who directs AI
The change taking place at Spotify reflects a broader transformation in the world of software engineering: the classic programmer role It begins to fall short in describing what the most senior profiles do in the big European tech companies.
With systems like Honk, writing code line by line is no longer the core activity. The most experienced engineers take over. to identify problems, design solutions, and monitor quality of what the models generate. The mechanical work is reduced, but the weight of technical criteria and responsibility for the result increases.
Spotify also insists on a hybrid approach: while senior profiles rely heavily on AI, junior developers continue learning. “in the old-fashioned way”, writing code manually to consolidate basic knowledge of structures, algorithms and good practices in languages such as Java.
The company argues that this hybrid model prevents new generations from blindly relying on AI without understanding what lies beneath, while freeing veteran professionals from repetitive tasks that a machine can reasonably reliably handle.
Outside of Spotify, other developers consulted by specialized media outlets such as Ars Technica acknowledge that, in just a few months, AI tools have gone from offering minor assistance to being capable of solve complete tasks on your own. From fix tests Many who fail to implement entire functions describe productivity increases ranging from several times to, in some cases, ten times faster than doing everything by hand.
Productivity, doubts and fear of job replacement
Not everyone is applauding this shift. Although many professionals see AI as a an ally for getting rid of tedious tasksDoubts are also growing about the reliability of the generated code and, above all, about its impact on employment in the medium term.
Developers interviewed by various media outlets admit that the quality of current tools is very high for programming and debuggingAlthough they are not yet ready for complex creative work in other fields, such as literary writing, in the software sector, many agree that "it is already changing everything."
The main fear revolves around what will happen to jobs: first, routine coding is handed over to AI, then part of the architecture design, and later, perhaps even the product managementSome argue that those who fail to adapt to working hand in hand with these tools will have increasingly fewer opportunities in the market.
At Spotify, for now, the official line is that AI hasn't replaced engineers, but rather changed their daily work. More experienced developers focus on higher value-added tasks, while junior professionals continue to build a solid foundation of technical knowledge.
This debate is very much present in Europe, where EU authorities are working on regulatory frameworks on AI which will affect both the internal use of these technologies and their impact on entire sectors, including the software industry.
A unique music dataset as an advantage for Spotify's AI
Beyond the way it programs, the Swedish company sees artificial intelligence as a key resource to exploit its greatest asset: music consumption data of millions of people worldwide, including users in Spain and the rest of the continent.
During the results presentation, Söderström emphasized that Spotify is building a proprietary and difficult to replicate datasetBased on listening patterns, tastes, and cultural behaviors. Unlike open resources such as Wikipedia, in music there is usually no single correct answer to many questions.
Examples like “what music works best for working out?” illustrate this diversity: in the United States, a significant portion of the public favors hip-hop, while millions prefer much more extreme styles, such as death metal. In Europe, the picture is even more complex, with the EDM, heavy metal, or Latin pop competing for the soundtrack of gyms and nightlife.
This entire mosaic feeds into a dataset that, according to Spotify, is enriched every time They retrain their modelsThe more users listen, save songs, or skip tracks, the more the AI learns and the better the personalization becomes. In turn, a more tailored experience keeps people on the app longer, creating a virtuous cycle that strengthens the platform's position against competitors.
From an engineering perspective, this volume of information also serves to quickly test and validate new features, which fits perfectly with the automated development model that drives Honk.
AI, creativity and the future of programming
Spotify's move comes at a time when the Generative AI is advancing at great speed in fields traditionally considered creative. The University of Montreal study, which suggests that certain models outperform the average human in linguistic creativity, has raised concerns about the future of professions related to writing, design, and programming.
Even media figures like Elon Musk They have reignited the debate by suggesting that programming as a profession could be seriously affected before the end of the decade, due to AI's ability to write and maintain complex code. Large tech companies like Google, Anthropic, and Microsoft are already using these systems to automate some of their internal processes.
In this context, Spotify presents itself as a case study of what the near future might look like: engineers don't disappear, but the weight of their work shifts towards problem definition, supervision and high-level decision-making, while AI takes over day-to-day execution.
The question is how the labor market in Europe and other developed markets will be reconfigured if this model spreads. New specialized profiles are likely to emerge in manage and audit AI systemswhile some of the tasks currently performed by junior programmers are being fully automated.
Meanwhile, users are already noticing the effects of this transition: more features, faster launches, and an app that changes more frequently, all supported by a technology that largely writes the code in the background.
What's happening at Spotify illustrates the extent to which Software development is entering a new stageThe company's best programmers have practically stopped typing code by hand, but their role is more crucial than ever in guiding, controlling and leveraging the artificial intelligence that powers the platform, both in Spain and in the rest of Europe.
