It has been almost exactly two months since Robert Kaye died.
Rob — known online as <mayhem>, a handle that was both a pose of defiance toward corporate interest and, on occasion, an accurate forecast of his management style — was the founder and Executive Director of the MetaBrainz Foundation.
He passed away on 20 February 2026, unexpectedly, at 56. I’ve been on the MetaBrainz board for more than fifteen years, and I’m still catching myself composing emails to him in my head. I miss his bullet-shaped head and wondering what colour his hair would be when I saw him next.
This post is three things: a rough eulogy, a reminder of what we’re losing to a handful of very big companies, and — eventually — a job advert. If you want to skip to the last part, scroll down. But I’d like you to read the rest first, because I believe more people need to understand what Rob built and why the world needs it more than ever.
The institution you’ve probably never heard of
MetaBrainz sits in a small, stubborn constellation of organisations that represent the best of the open internet: Wikipedia, the Internet Archive, IMDb (in its early, pre-Amazon form), Reddit (in its earlier, pre-everything form). Each of these is both an institution and a living community — a testament to a particular moment when the internet was still untamed and weird and largely built by people who thought it should belong to everyone.
MusicBrainz, the flagship MetaBrainz project, is an open music encyclopaedia: a painstakingly maintained, community-edited database of artists, releases, recordings, labels, and the relationships between them all. If you’ve ever had your music library tag itself correctly, or seen credits display accurately on a streaming service, or used a DJ application that knew what album your track came from — there’s a decent chance that data passed through MusicBrainz on its way to you. Rob’s life’s work is everywhere and his name is almost nowhere. Because that’s how infrastructure and maintenance works.
I went to university in Santa Cruz, which was also home — briefly — to IUMA, the Internet Underground Music Archive. IUMA didn’t survive. Its punk, outsider, “we’re-going-to-route-around-the-music-industry” ethos didn’t outlive the first dot-com bust. But the same spirit lived on in MusicBrainz, and that’s not an accident.
MusicBrainz was born angry
Here’s a piece of early-internet history that has mostly faded: the CDDB story. In the mid-1990s, when people started putting CDs into their computers, the discs themselves contained no useful metadata — just audio and track durations. So a community of music fans built a collective database, CDDB, where anyone who inserted a new CD could fill in the album details once, and every user from then on would benefit. Classic public-interest internet. Classic barn-raising.
And then — predictably, almost inevitably — the people running the server privatised it, rebranded it Gracenote, closed the APIs, and sold it to Sony. The volunteer contributors who had built the database got nothing. The data they had collectively assembled was now a commercial asset, locked up, licensed out, and no longer theirs.
Danny O’Brien wrote the definitive account of this in his 2021 EFF piece on CDDB. It’s short and worth reading. The punchline is that the enclosure was completed under the reasoning — offered, apparently in good faith, by the founders — that only a commercial structure could keep the project viable. “The focus and dedication required for CDDB to grow could not be found in a community effort,” one of them told Wired in 2006. I’m not going to name them here because their names don’t matter. What they built and then sold is gone.
MusicBrainz was, in part, Rob’s answer to that act of theft. CDDB has been atomised into features and sucked into other products and services. It didnt survive, MusicBrainz has.
Where Brewster Kahle founded the Internet Archive and Jimmy Wales founded Wikipedia with a certain noblesse oblige — earnest, civic-minded, grown-up — MusicBrainz was started in 2000 with a snarly sense of punk outrage. Fan-created data had been stolen from the community and monetised. Rob thought that was wrong. He thought it was going to keep happening unless someone built an alternative that couldn’t be captured. So he did. In 2004 he wrapped the collective in a California 501(c)(3), the MetaBrainz Foundation, with a simple constitutional principle: free for non-commercial use, supported (often very quietly) by the commercial entities that depend on the data. That model has funded the project on a shoestring for more than twenty years.
The EFF series that Danny wrote has a whole follow-up piece specifically about MusicBrainz as the counter-example — the project that didn’t get enclosed, didn’t get sold, didn’t stagnate, and didn’t die. That’s Rob’s monument. He built the thing that outlived the outrage.
The feral principle
You don’t get that outcome by being nice. You get it by being stubborn, by being willing to fight, and by being genuinely unafraid of the room.
Rob was all of those things. Cory Doctorow brought me onto the MetaBrainz board fifteen-plus years ago, when I was at Last.fm and clashing with CBS over where the company was heading (I ultimately left). I came in because I admired what Rob had built. I’ve stayed because of how he defended it.
He was, to be clear, feral about the things that mattered. He would not make strategic concessions on the data-licensing model. He believed that commercial partners shouldn’t extract value from the community without returning it. He was occasionally difficult to work with, and he was always, always principled. You always knew where you stood with Rob, and you always knew where MetaBrainz stood, because the two were hard to tell apart.
That’s hard to sustain in the long run — and Rob knew it, which is why he’s been building the board, the staff, and the community structures for years. But it’s the reason MetaBrainz exists at all. Somebody had to be unreasonable enough, for long enough, to keep the data open. Rob was that person.
Why it matters now more than it did in 2004
Here’s where this piece stops being an obituary and starts being an argument.
The digital music space has compressed. Three or four streaming services — Spotify, Apple, Amazon, YouTube — are now the dominant way most of the world experiences recorded music. For these platforms, music is a utility: a means of capturing attention, selling subscriptions, training recommendation models, and extracting more value from the artists and fans than they return.
This collapse mirrors the rest of the music industry. The rich label ecosystem that shaped twentieth-century music has compressed into three major holding companies that barely deserve to be called record labels anymore. Call them banks. Call them transglobal rights engines. Call them whatever you like. The actual fierce, weird, risk-taking label work is happening at the edges — Third Man, Top Dawg, Ninja Tune, Rough Trade, Beggars, Arts & Crafts, Domino, XL, Roc Nation, and dozens more. That’s where the culture is. The majors are increasingly upstream of it, monetising it.
Against that backdrop, the role of open music data is enormous and largely invisible. MusicBrainz data flows through the streaming services, through the distributors, through the rights administrators, through hundreds of established companies and thousands of startups. It underpins AI music recommenders, metadata-cleaning tools, catalogue analytics, artist-identity resolution, and — increasingly — training data for models that will shape how billions of people encounter music. Every one of those use cases needs a neutral, open, community-maintained reference layer to exist. That layer is MusicBrainz. That layer is Rob’s work.
Now look at what sustains it.
The Wikimedia Foundation runs on a $208 million annual budget and a substantial, professional organisation. The Internet Archive operates at around $37–40 million a year, largely from family foundations and individual supporters. These are the right comparables for MetaBrainz because they share the same structural role: critical infrastructure for the open internet, nonprofit, community-governed, commercially depended-upon.
MetaBrainz runs on a fraction of that. A small fraction. In an era when open data underwrites global AI training pipelines and the core plumbing of every music service in the Western world, this is not just modest — it’s precarious.
That’s the thing I want to be loudest about in this post: the gap between how important MetaBrainz is and the way it is resourced is one of the most striking mispricings in the current internet economy. Rob kept it going through sheer force of will, a lean operating model, and a licensing structure designed for exactly this problem. But the foundation needs to grow into the role the world has given it. That is a big ask.
MetaBrainz needs a new Executive Director. And you can read the job posting here.
We need someone who understands, in their bones, why open music data matters — to artists, to fans, to researchers, to the soul of the internet itself. Someone who can honour Rob’s ethos without trying to be Rob (an impossible assignment, and the wrong one). Someone with the operational skills to run a distributed open-source nonprofit, the political saavy to navigate relationships with the commercial partners that fund us, and the strategic imagination to push the organisation and the community into its next twenty-five years.
I believe we need to expand the mission, not just defend it. ListenBrainz — think of it as an open-source Last.fm — has extraordinary latent potential. Your listening history, your taste graph, the accumulated data of what you actually love: right now that sits inside Spotify, Apple, YouTube/Alphabet, and Meta, and it’s used to train models and target you with content that serves their interests, not yours. An open, portable, community-owned listening layer is one of the most important things the open internet could build in the next decade. The value to fans and artists is tremendous. MetaBrainz is already building it. And it is growing. But that garden needs tending and it needs more resources.
BookBrainz, Picard, the Cover Art Archive — there’s a whole portfolio of projects sitting under the MetaBrainz umbrella, each with its own community and its own moment of potential. The next ED gets to shape all of it.
Is it a hard job? Yes. You’d be taking over from a founder, from someone whose personality was load-bearing. You’d be fundraising for a nonprofit that’s chronically underfunded relative to its strategic importance. You’d be working with a global volunteer community that has strong opinions and long memories. You’d be communicating with some of the largest and most lawyered companies on the planet.
You’d also be running one of the most quietly important institutions on the open web, at a moment when the case for that kind of institution has never been stronger. You’d be continuing work that Rob started in a spare-time project in the late 1990s and turned into two decades of infrastructure that the entire music industry — including the parts that would probably prefer it didn’t exist — now depends on.
The ask
If you are that person, or if you know that person, get in touch.
You can reach me at matthew.hawn@metabrainz.org You can also read the MetaBrainz blog for the formal posting,
In the meantime, if you’ve never looked at MusicBrainz, ListenBrainz, Picard, or BookBrainz — do it. Edit an entry. Contribute a correction. Donate, if you can. Most of all: notice that it exists, and understand what it would mean if it didn’t.
Rob stubbornly built something that by all rights was unlikely to survive in the era of the corporate internet. But it did and has done for more than two decades, because he refused to let it be any other way. The rest of us get to decide what the next twenty-five years look like.