Orlowitz, a 12-year WMF veteran who built the Wikipedia Library, argues the Foundation's response to the Wikimedia Workers Union is the textbook consultant-grade union-busting script: captive-audience meetings, 'we're a family not a company' framing, neutral-sounding HR emails timed to organizing announcements, slow-walked bargaining, and selective policy enforcement. His point is that none of this is improvised — it's move-for-move what Amazon and Starbucks do, and an institution whose brand eq
By submitting the piece under the framing 'Big Tech's Anti-Labor Playbook Has Come for Wikipedia,' the submitter explicitly endorses the thesis that WMF's behavior is indistinguishable from the standard corporate anti-union response. The 457-point score and front-page slot indicate the developer audience finds the parallel credible and uncomfortable.
The editorial pushes back on the dismissive 'every employer does this' reaction by arguing that WMF is not a normal employer: it is the steward of the fifth-most-visited site on the internet, the single largest training corpus for every frontier LLM, and the closest thing the web has to a public utility for facts. When that specific institution adopts the labor practices of the companies Wikipedia exists as an alternative to, the stakes extend beyond one workplace dispute.
Jake Orlowitz, who spent 12 years at the Wikimedia Foundation building the Wikipedia Library and other community programs, published a long post on Medium arguing that WMF leadership is now running the same anti-union playbook he'd expect from Amazon or Starbucks. The piece, titled *Wikipedia Is Doing the Capitalist Thing*, hit 457 points on Hacker News with the front-page slot devs reserve for stories that make them physically uncomfortable.
The specifics he names are not novel. They are, in fact, the point. Captive-audience meetings where managers explain why a union would be bad for you. 'We're a mission, not a company, we're a family' framing. Carefully neutral-sounding HR emails that arrive the same week the organizing committee goes public. Slow-walked bargaining. Selective enforcement of communication policies. Orlowitz's argument is that none of this is improvised — it's the standard consultant-grade response, and the Wikimedia Foundation, whose entire brand equity rests on *not* being a normal tech company, is running it move-for-move.
The Wikimedia Workers Union went public earlier this year. WMF's public response has been the usual mix of "we respect our employees' rights" boilerplate and operational friction. Orlowitz's piece is the first detailed inside account from someone with a decade-plus tenure and no obvious axe to grind beyond the one he names: that the institution he gave his career to is behaving like the institutions Wikipedia exists to be an alternative to.
For most readers, the immediate reaction is "so what, every employer does this." That reaction is exactly why the story matters. The Wikimedia Foundation is not most employers. It is the steward of the fifth-most-visited site on the internet, the single largest training corpus for every frontier LLM you've ever used, and the closest thing the web has to a public utility for facts. When that institution adopts the labor practices of the companies whose extractive behavior its existence was supposed to push back against, the asymmetry that justified its special status starts to collapse.
The playbook Orlowitz describes is well-documented. Amazon's 2022 Bessemer campaign included mandatory anti-union meetings and texts to workers' personal phones. Starbucks closed unionizing stores under cost-cutting pretexts that the NLRB has since ruled illegal in multiple cases. Google's response to the Alphabet Workers Union was quieter but structurally identical — reorgs that conveniently dispersed organizing cores, contractor-vs-FTE wedges, and a steady drumbeat of "we already give you everything a union would." The throughline is not malice. It's that once an organization crosses a certain size and budget, the consultant class that gets hired to handle "employee relations issues" is a small world, and they all sell the same deck.
What's different about WMF is the donor base. Wikipedia raises roughly $180M a year from small-dollar donors who give specifically because they believe they are funding something that is not a normal corporation. If those donors start reading Orlowitz-style accounts in their feed every quarter, the fundraising banner that pops up every November stops working. The financial model and the moral model are the same model. You cannot crack one without cracking the other.
There's also a more uncomfortable systems-level point. The open-knowledge infrastructure devs depend on — Wikipedia, OpenStreetMap, the Internet Archive, the major Linux foundations — has been quietly professionalizing for a decade. Salaried staff replaced volunteer coordinators. Foundations grew legal and HR departments. The governance ratchets in one direction: toward looking like the companies they were founded to be different from. The labor question is where that drift becomes visible, because labor is the one place where the interests of the institution and the people doing the work openly diverge.
If you depend on Wikipedia data — and if you ship anything that touches search, retrieval, summarization, or any LLM-backed feature, you do — this is a supply-chain story. The humans who run the bots, fix the schema, handle the abuse queue, and keep the dump pipeline from breaking are the same humans currently in a labor dispute with their employer. Burnout, attrition, and the slow rot of institutional knowledge are the failure modes here, not a dramatic outage. You won't get a status page. You'll get a slow decline in dump quality, slower response on category restructures that break your scrapers, and a thinning of the editorial defense against the LLM-generated slop that's already arriving in volume.
Practically: if your RAG stack pulls from Wikipedia, now is the time to pin specific dump revisions and stop assuming the live API is a stable contract. If you have a budget line for "open data," point some of it at the Wikimedia Endowment with a note about staff conditions — donor pressure is the only governance lever external to the foundation that has ever worked. And if you maintain bots or tools in the Wikipedia ecosystem, the people you collaborate with on-wiki are likely the same people organizing. Treating that as relevant context is just professional courtesy.
The broader takeaway for anyone building on open infrastructure: the foundations are not neutral. They have CFOs, board politics, and the same incentive gradient any organization with $100M+ on the balance sheet has. Plan accordingly.
The Wikimedia Workers Union will either win recognition or it won't, and either outcome will set a template for the rest of the open-knowledge nonprofit sector — Mozilla, Apache, the Linux Foundation, OSI — that is watching closely. The optimistic read is that a successful union at WMF locks in the labor conditions that keep the institution producing the public good devs rely on for the next twenty years. The pessimistic read is that the playbook works, organizing collapses, and the slow professionalization-into-normal-corporation completes. Either way, the days when you could donate to Wikipedia and tell yourself it was categorically different from donating to a tech company are ending. The next time the orange banner shows up, read it with that in mind.
Some English Wikipedia (enwiki) editors are striking. They are predominantly non-technical that are forced to maintain their own shadow IT-style infrastructure that Wikimedia (nonprofit owners of Wikipedia) doesn't provide. It is very difficult to be a productive editor without custom tooling a
I spent ~2 years actively editing Wikipedia for multiple hours every day. I remember taking my laptop out at airports for 20 minutes between transfers, just to tweak an article or improve a source. While I originally started because I found some articles lackluster, I quickly realized how vigorous t
17 months of operating expenses are actually not a lot for a foundation. Especially one whose goal is to preserve something for a long horizon.Unions exist to combat the monopsony power of corporations. Corporations and unions can exist in constant tension with each other because ultimately both are
> The Wikimedia Foundation closed last fiscal year with $208.6 million in revenue. It holds $296.6 million in reserves, 17.1 months of operating expenses.The actual physical cost of hosting Wikipedia is < $5 million per year.
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.
To give context, it seems like what happened is WMF did two separate things:- Fired one of the original developers of MediaWiki (the open source project that powers wikipedia) - Brooke. This person was at one point in contention to basically be BDFL of MediaWiki. She is somewhat less publicly promin