You’re not out of new music, you’re just using the same five doors to look for it.
Limit how much you use the ‘safe’ stuff as a starting point. If you only ever hit Discover from the same three artists, the algorithm just keeps circling the same drain. Start discovery from different places: a label you like, a city, a playlist made by a stranger, a live set, a radio show. When something outside your usual lane clicks, follow that artist, label, or scene for a bit instead of running back to the comfort genre. You’re not abandoning your taste, you’re adding more doors into it.
Use both, but for different jobs. Algorithms are great for volume and quick ‘more like this’ suggestions when you’re already in a lane. Humans—friends, DJs, writers, curators—are better at giving you context, emotional shortcuts, and big surprise leaps you’d never click on yourself. Let the algorithm do the grunt work of serving candidates, then lean on human-made stuff when you feel stale or want to jump scenes entirely. The trick is to keep curating what you save so your algorithms have good raw material.
If you’re constantly digging but never replaying, the music doesn’t really sink in, and everything starts feeling disposable. A simple balance is to set specific discovery windows—maybe 2–3 short sessions a week—and let the rest of your listening be guilt-free repeats. In those sessions, be ruthless: if a track doesn’t grab you in a minute or two, move on; if it does, save it somewhere you’ll actually revisit. That way your library deepens over time instead of turning into an endless pile of half-heard songs.