Spotify crossed the 100 million track mark in 2024. Apple Music is somewhere north of that. The sheer volume of recorded music available to anyone with a phone and a broadband connection is historically unprecedented, and it should feel like a golden age. Instead, it mostly feels like noise. The paradox of choice is real, and the streaming platforms have responded to it with algorithmic curation that promises to know you better than you know yourself. But if you have spent any meaningful time with services like NTS Radio, Rinse FM, Lot Radio, or Dublab, you already know the truth: a human being with taste and conviction will always outperform a recommendation engine built on engagement metrics.
This is not nostalgia. This is a practical argument about how music discovery works, what gets lost when you hand it to machines, and why volunteer-driven internet radio may be the most important infrastructure electronic music has right now.
The Algorithmic Discovery Problem
The way Spotify's Discover Weekly or Apple Music's personalized playlists work is well documented. Collaborative filtering compares your listening habits with millions of other users, identifies patterns, and surfaces tracks that people with similar profiles have enjoyed. Audio feature analysis examines tempo, key, energy, and spectral characteristics. Engagement metrics track skips, saves, repeats, and playlist additions. All of this feeds a model that optimizes for one thing: keeping you listening.
The problem is that "keeping you listening" and "expanding your taste" are not the same goal. An algorithm that notices you skip tracks with harsh textures or unusual time signatures will stop showing you those things. It learns your comfort zone and reinforces it, nudging you toward the center of your existing preferences. Over time, your Discover Weekly starts to sound like a shuffled version of what you already know.
This is not a bug. It is the intended behavior of a system designed to minimize friction. But friction is where musical growth happens. The first time you heard a Drexciya record, or an early Plastikman track, or a 160 BPM juke edit, it probably did not feel comfortable. It felt alien, confusing, maybe unpleasant. And then something clicked. That moment of disorientation followed by recognition is the core experience of discovering music that actually changes you, and algorithms are structurally incapable of delivering it.
The DJ as Editor
What stations like NTS and Unchained Radio understand is that curation is an editorial act. When a DJ builds a two-hour show, they are not assembling a playlist. They are constructing an argument. There is a thesis in the sequencing: this track follows that one because the DJ hears a conversation between them that you might not have noticed. The transition from a deep house groove into an Italo disco rarity into a broken beat experiment is not random. It is a perspective.
Charlie Bones did this for years on NTS with his Breakfast Show, treating the morning slot like a freeform radio journal that might move from Ethiopian jazz to UK garage to library music in a single hour. Throwing Shade built shows around the intersection of Middle Eastern music and club electronics. Shanti Celeste's Rinse FM residency wove together disco, house, and ambient in ways that no algorithm would ever attempt because the connections were too personal, too specific, too rooted in an individual sensibility.
This is what "taste" means in practice. It is not a preference profile. It is a point of view that has been developed through years of listening, collecting, playing out, and paying attention. When you tune into a curated radio show, you are borrowing someone else's ears for an hour. You are trusting their judgment. And that trust is what enables the surprise, the discomfort, the unexpected left turn that sends you down a rabbit hole you never would have found on your own.
The Value of Passive Listening
There is something fundamentally different about passive listening to a curated stream versus actively searching for music on a streaming platform. When you search, you are in control. You have expectations. You are looking for something specific, or at least something adjacent to what you already know. The experience is shaped by your existing taste, which means the ceiling for discovery is low.
When you tune into a radio station and let it play, you surrender control. You do not know what is coming next. You cannot skip ahead. You are in someone else's hands, and that vulnerability is precisely what makes the experience valuable. The best listening sessions work because the listener has agreed to be taken somewhere unfamiliar.
This is why internet radio works so well as a background companion for work, commuting, cooking, or just existing in your apartment. It provides a sonic environment that is shaped by human intelligence but does not demand your active participation. When something catches your ear, you look it up. When it does not, you let it wash over you. The ratio of passive to active engagement is the opposite of what Spotify encourages, and it leads to a qualitatively different relationship with music.
Red Light Radio and What We Lost
In 2020, Red Light Radio in Amsterdam closed its doors after eleven years of broadcasting from a former prostitution information center in the De Wallen district. The station had become one of the most respected platforms in electronic music, hosting everyone from Antal to Palms Trax to local Amsterdam selectors who never played outside their neighborhood. Its archive, thousands of hours of shows, remains one of the most valuable documents of 2010s dance music culture.
The closure hit hard because it illustrated something important: these stations are fragile. They depend on cheap or donated space, volunteer labor, small grants, and a community that shows up. They do not have Spotify's billions in venture capital or Apple's hardware revenue to fall back on. When the rent goes up or a pandemic shuts down the city, they can disappear overnight.
But the response to Red Light Radio's closure was also instructive. The community mourned, and then it built more stations. Niche internet radio did not contract after 2020. It grew. New stations launched on Mixcloud Live, Twitch, and standalone platforms. The impulse to broadcast, to share music through a curated lens, proved more durable than any single institution.
Why This Still Matters for Electronic Music
Electronic music has always depended on informal distribution networks. Before Spotify, before Beatport, before even widespread MP3 sharing, the way most people encountered new dance music was through a DJ playing it in a club or on the radio. The music was contextual. You heard it in a mix, surrounded by other tracks, shaped by someone's selection. The context was inseparable from the content.
Streaming platforms strip that context away. A track on Spotify is an isolated object, sitting in a database alongside a hundred million others, waiting to be surfaced by an algorithm that knows nothing about its history, its scene, or the specific 3 AM moment it was designed for. Internet radio puts the context back. When you hear a track on a house or electro session, you hear it as part of a story someone is telling.
That story is what makes electronic music worth caring about. Not the individual tracks in isolation, but the culture of selection, sequencing, and sharing that gives those tracks meaning. Algorithms cannot tell stories. They can only optimize for engagement. And engagement is not the same as love, or understanding, or the slow accumulation of knowledge that turns a casual listener into someone who genuinely hears the difference between a Detroit electro record and a Rotterdam one.
Curated internet radio is not a relic. It is not a hipster affectation. It is the mechanism through which electronic music renews itself, the way new sounds get tested, new artists get heard, and new connections get made between genres and generations. Spotify has 100 million tracks. That is a warehouse. NTS has taste. That is a curatorial vision. The difference is everything.