You’ve probably seen the triumphant headlines. OpenAI has made a massive breakthrough on the Navier-Stokes equations—one of the most notoriously unsolvable problems in mathematics. The subtext is clear: the machines are getting so smart that soon, they’ll be curing diseases and designing new materials without us.
But the headlines are a lie.
We thought AI would democratize genius. Instead, it industrialized plagiarism.
Here’s what actually happened. OpenAI didn’t magically derive a mathematical breakthrough from the ether. According to a statement from NYU mathematician Tristan Buckmaster, OpenAI used an AI model to ingest and bootstrap from the unpublished, freely shared work of Buckmaster and his colleague. They took the messy, unfinished, collaborative brainstorming of human mathematicians, ran it through an algorithmic meat grinder, and slapped their corporate logo on the output.
This isn’t a story about artificial intelligence getting smarter. This is a story about artificial exploitation.
If you aren’t a mathematician, you might be wondering: who cares? Don’t we want AI to solve hard problems? But you have to understand the culture of mathematics. Math isn’t just about the final correct answer; it’s about the proof. It’s about the lineage of ideas. For decades, mathematicians have shared unfinished drafts, half-formed concepts, and dead-end hypotheses on forums and in informal papers. They do this out of a spirit of open collaboration, trusting that their peers will build upon and credit their work.
The scarce resource in the age of AI isn’t compute power. It’s human generosity.
OpenAI framed their Navier-Stokes effort as a step toward solving practical, real-world problems. But by bootstrapping their “breakthrough” from underacknowledged human work, they’ve created a toxic paradox. The very human generosity that makes mathematical progress possible is now being silently absorbed by a machine system that claims the glory.
What happens next is entirely predictable. As one commenter noted, moving forward, no mathematician in their right mind would want to have this kind of experience. Why would anyone share their unfinished ideas if a multi-billion dollar tech company is just going to scrape it, feed it to an LLM, and claim they invented it?
This is the dark twist nobody in Silicon Valley wants to talk about. The real threat isn’t that AI becomes too smart. The real threat is that AI erodes the social mechanisms that make intellectual work verifiable and fair. If mathematicians retreat behind paywalls and closed labs to protect their intellectual property from AI scrapers, the entire ecosystem of human knowledge slows down.
OpenAI didn’t solve a Millennium Problem. They created a millennium crisis of trust.
If we let machines claim the spark, we won’t have to worry about them replacing human creativity. There simply won’t be any left to replace.
The AI industry loves to preach about the “commons” and “benefiting humanity.” But when they cannibalize the actual commons—the fragile, trust-based network of human researchers sharing ideas for the pure joy of discovery—they aren’t building the future. They are burning the foundation to heat their own servers.
Anyone relying on AI for knowledge work needs to wake up. The Navier-Stokes controversy isn’t an isolated academic spat. It is the blueprint for what happens when capital-rich tech companies decide that human labor is just a training dataset waiting to be monetized.
The math is clear. Stop celebrating the machine. Start protecting the humans.
FAQ
Q: Doesn't AI just build on existing data like all researchers do?
A: No. Human researchers cite their sources and participate in a reciprocal economy of credit. AI scrapes uncredited, unfinished work at scale and repackages it as proprietary genius. It’s extraction, not collaboration.
Q: What's the practical implication for other fields?
A: If this continues, researchers will lock down their unfinished drafts to avoid being scraped, drastically slowing down the pace of human discovery. AI will cannibalize the very ecosystem it relies on.
Q: Is OpenAI actually doing anything wrong if the work was publicly available?
A: Legally, maybe not. Ethically, it's parasitic. Publicly available doesn't mean 'free for corporate appropriation without credit.' Exploiting academic goodwill to train a for-profit model is a betrayal of the scientific commons.