Can you actually teach your music app what you like?

YouTube and Spotify are introducing features that let you move beyond passive listening or watching. YouTube now allows you to create custom feeds by typing in a description of what you want, like a mix for a specific commute or videos that match a relaxing mood. Spotify is launching a feature called Taste Profile, which lets you see how the app interprets your preferences and gives you a way to request changes. You can tell the app to prioritize certain genres, exclude specific sounds like nighttime white noise, or help you branch out into new areas. Both services are moving toward a model where the user provides the roadmap instead of just letting the platform guess.

Every media platform uses a recommendation system to keep you engaged. Think of this as a super-fast librarian that tracks everything you have ever interacted with. It looks at your history and compares it to millions of other users to guess what might keep you listening longer. The new tools rely on artificial intelligence models—the technology behind tools like ChatGPT—to act as an interpreter. When you give a prompt, such as asking for music that inspired a specific song, the AI breaks your request into concepts, searches a database of hundreds of millions of items, and builds a new playlist or feed that matches your specific criteria. It is effectively overriding the standard guesses the app usually makes by forcing it to prioritize your current, explicit instructions over your past behavior.

These tools acknowledge a fundamental frustration: the one-size-fits-all algorithm is rarely perfect. By letting users step in and tweak the settings, companies are moving toward a more personal experience. However, this also shifts the responsibility back to you. The more you interact with these tools, the more power you have to shape your digital environment, but it requires you to be an active participant. Instead of just accepting what is served, we are entering an era where we must curate our own digital spaces. The real question isn't just whether these features work, but whether we actually want to spend our time managing our preferences, or if we would rather just let the mystery of the algorithm keep playing.