✘ The Web, have we lost it? Can it become semantic? Is music metadata a solution?
And: The algorithm failed music; Fairness in music recommender systems; Salvation for music criticism; alive internet theory; New Taylorism
Dead Internet Theory. Alive Internet Theory. Have bots taken over? Do people still exist on the Web? If your music gets streamed on Spotify, does anyone hear it? These are all significant and big questions. It’s ontology. It’s hermeneutics. It’s also politics and economics. During the recent hype of blockchain, crypto, NFTs, and DAOs there were many, me included, who liked to talk about Web3. Now, this wasn’t a new concept. Web3, in the spirit of Tim Berners-Lee, was supposed to be the Semantic Web. A fully interoperable and agentic Web where people retain their digital sovereignty while what we call the Internet works for us.
Web. Internet. internet.
We’ve scaled down our expectations of the World Wide Web. Initially, people agreed on certain standards so that we could use the beauty of things like hypertext to get websites to interact with each other. This became Web1. Then, we entered the platform era of the Internet. Instead of people building websites, they created profiles on websites. This became Web2. While this happened, actually when this was in its early stages, Tim Berners-Lee - together with James Hendler and Ora Lassila - wrote about The Semantic Web.
“Most of the Web’s content today is designed for humans to read, not for computer programs to manipulate meaningfully. Computers can adeptly parse Web pages for layout and routine processing—here a header, there a link to another page—but in general, computers have no reliable way to process the semantics … The Semantic Web will bring structure to the meaningful content of Web pages, creating an environment where software agents roaming from page to page can readily carry out sophisticated tasks for users.”
This doesn’t sound too different from the PR around AI agents coming from major tech companies and start ups alike. There is, however, one major difference. The Semantic Web was still a ‘web’ and, more importantly, decentralized.
The original premise of Web1 centred on its decentralized nature, encapsulated in shared protocols and programming languages. With Web2, this decentralized aspect got pulled to the background in favour of platforms and apps. Instead of building websites linked through hypertext, the interoperability ended on the platform basis. Web2 wasn’t really a Web anymore. In this period, the language used for the World Wide Web shifted to the Internet, with a capital I. As we retreated further into apps and platforms, the capital I disappeared as well. It became simply the internet, a tool under the surface of all the apps we use. The Semantic Web promised something very different.
Another piece of evidence that the Semantic Web failed to break through is where it did seem to work. Google Search used semantic tooling to give summaries. Facebook developed OpenGraph, but really that was to make sure even less people clicked away from Facebook and back onto the open Web as Facebook got able to simply display all the information of a website inside its app.
It’s all about metadata
Just as with music, the Web runs on metadata. Simply put, this is data about data. For music, this is songwriter, genre, bpm, etc. For a website it’s word count, summary, keywords, etc. Developers were, perhaps, simply not incentivized to provide good metadata to make the Web more interoperable and semantic. Cory Doctorow immediately caught on to the metadata fallacy. In response to the Semantic Web article by Berners-Lee and his colleagues he wrote about Metacrap. Cory provides seven reasons why the Semantic Web won’t work:
People lie
People are lazy
People are stupid
It’s impossible to know yourself
Schemas aren’t neutral
Metrics influence results
There’s more than one way to describe something
All of the these still hold and we see them in music, too. The first four get compounded by Fake Artists or money-making schemes that involve tacking on a famous (real) featured artist for playlisting purposes. Schemas are indeed not neutral. While DDEX is accepted around the world as the general music metadata standard, it lacks information around, for example, gender. Metrics aren’t neutral either, whoever determines them determines what matters - and, by extension, what gets pushed to the background.
Even if the ideal world of the Semantic Web would work, or would have worked, it still would have been determined by those who set the terms - as open as they were intended.
Reclaim the Web
Last week, I wrote about community and community works best when it doesn’t take place on a centralized platform. Instead, the community should have data sovereignty. Music is historically difficult when it comes to data. From the 1950s onwards, labels and publishers controlled the data. Now, it’s the platforms. In between, we find pockets of artists and other creators that are trying to reclaim the Web. These are the artists who build their own release plans, their own minigames, their own AI models, etc. Most importantly, they claim ownership of their fanbase. This can be messy and involve many different chats, Discords, groups, and more. If we can make the Semantic Web work for one thing, it would be this: artist communities, scenes, collectives, and more all sharing a programming language that is interoperable. People are always trying to build this, from Bandcamp to Nina and from Metalabel to Substack. Most of these end up being another platform. We don’t need this:
Instead, we need to get diligent about our metadata and focus on building stuff we own instead of building on rented land.
LINKS
🪫 The algorithm failed music (Terrence O’Brien)
“More companies will probably start offering off-ramps as algorithm fatigue grows. But, eventually, companies will figure out how to create the illusion of serendipitous discovery. They will serve up algorithmic recommendations, but package them in a way that feels more natural.”
✘ A sort of companion read to the piece above. I love the little takedown of any kind of optimism in this highlighted quote.
🪟 Fairness & Transparency in Music Recommender Systems: 2025 Review (Dmitry Pastukhov)
“Working at the crossroads of research, industry, and artist practice has given us a close-up view of how the conversation around fairness in music recommendations has evolved. What once felt like an abstract, technical topic is now part of a much bigger discussion — one that spans research circles, label boardrooms, and artist communities alike. In the past few years, we’ve seen fairness move from theory to practice, becoming a shared concern about how music is discovered and consumed in an increasingly AI-driven world.”
✘ This is about algorithms and there’s not many people in the world better positioned to comment on this than Dmitry. Music Tomorrow has been a crucial cog in understanding how platforms platform music and then how to work with that instead of against it. Eventually, the conclusion here is also to work towards more collaboration - otherwise known as interoperability.
✒️ As music journalism marches towards oblivion, a plea for salvation (Sean Adams)
“Disclaimer: Sorry for the clickbait title, but this could be an academic essay entitled something dry like “Music Journalism As Infrastructure” but that lacks the intrigue or friction these platforms demand if I want to write to you, my fellow music fan.”
✘ Writing about music is a skill. Interviewing is a skill. It’s necessary infrastructure in a healthy music ecosystem. Overall, I wish we would share more music with each other.
🧟 alive internet theory (Spencer Chang)
“This is why I keep finding myself coming back to playhtml. I’ve been able to imagine and create new ways of being with people online, sharing experiences I’ve never had before. I’m in wacky territory now, where the ideas are messing with fundamental assumptions about how websites and the internet work. I’m not sure if that’s what will lead to this vision, but it feels like it’s in the right direction.”
✘ Since I mentioned it at the top of my piece today, I wanted to point out that if you read MUSIC x, we’re both humans. All we need to do is reach out to each other and boom, we’re two humans connecting on the Internet. You can also just look at this alive internet website that will serendipitously drop you into the human-made Web.
🛡️ Protecting the Family: A few more notes on “the new Taylorism” (Toby Bennett)
“When all is said and done, then, this is not an album about the triumph of arts and crafts over business interests but of their ever closer entwinement. The Life of a Showgirl purports to pull back the curtain on the off-stage workings of the music industry. What it actually reveals is that the inside story of the battle for Taylor Swift, into which all of us who listen to this album have been enrolled, is that of one capitalist pitted against another. The victim may have become the victor but it is a victory for asset management, not for the visions of a popular cultural wealth for all embodied by the playwrights, poets, visual artists and folklorists that she routinely calls on to bolster her creative persona. The logic of the god-like founder, with the singular personal authority to exclude and control, prevails.”
✘ I don’t have a lot more to say, this is both a great piece of analysis and fun and witty writing. Taylor is a unicorn and she is the rod along which we measure the health of our industry.
MUSIC
Hilary Woods tries to re-enter her body. She does so through her new songs on a record called Night CRIÚ. It’s a beautiful collection of songs which focus on her vocals more than her instrumentation. Listening to it, I feel transported into her layered being and the many different Hilary’s that she has to contend with. We all face this struggle, of course. It’s just that we don’t all have this same method of dealing with it. You could say it’s cathartic.