Yandex Music Updated Its Recommendation Algorithm to Prevent Musical Tastes from Stagnating

Yandex Music has updated its recommendation algorithm to help users avoid stagnation of their musical tastes. The updated version of 'My Wave' now more actively suggests new artists and tracks from different genres, helping musical preferences develop with age.

Why Musical Tastes Become Rigid

A 2018 study revealed a funny yet sad fact: people on average stop listening to new music around age 33. After this age, musical preferences freeze, and new genres and artists remain on the periphery of attention. This happens to almost everyone — with age, taste develops, favorite artists emerge, and venturing beyond their boundaries seems strange or unnecessary.

This statistic is easy to understand: when you've found music you love, and the choice of millions of tracks becomes tiresome. But this also leads to missing out on a lot of great music. The problem is acute in the age of streaming: there's too much choice, but without a good algorithm, people get lost in the catalog and listen to the same things over and over.

Algorithm Update in 2026

In early 2026, Yandex Music introduced a significant update to the 'My Wave' algorithm. Technically, the system has become more powerful: it better analyzes listening context, accounts for preference dynamics, and predicts when a user is ready to discover something completely new. The results are impressive.

Over four years of development, the algorithm has learned to generate approximately half of all songs in users' favorite playlists. This doesn't mean people have stopped adding tracks on their own — it's simply that recommendations have become so relevant that users save them as their own discoveries. For comparison: on YouTube Music or Apple Music, most users curate their playlists independently.

On Yandex Music, the ratio is almost equal: fifty-fifty between independent additions and system recommendations.

How the Algorithm Expands Horizons

The updated system works in several directions simultaneously:

• Suggests new artists similar to favorites but not identical to them
• Finds tracks that fit the current mood and expand genre boundaries
• Balances between stability (familiar genres) and novelty (unexpected discoveries)
• Considers context: music for work differs from music for sports or leisure
• Adapts to seasonal changes (summer playlist differs from winter)

Such systems work because they don't try to guess taste once and for all. They constantly learn, seeing which recommendations users save, which they skip, and how their mood and behavior change.

What This Means

The update shows that recommendation algorithms are becoming a full-fledged tool for developing taste. Instead of becoming stuck with the same artists at age 33, people can now discover new music at 40, 50, and beyond. And this is thanks to AI that has learned to predict not what's popular and trivial, but what's truly interesting for each individual.