We like to think of mathematics as an infinite, timeless forest. A renewable ecosystem where new questions naturally sprout from the soil of old answers. But what happens when someone brings in the bulldozers?
Terence Tao, one of the greatest living mathematicians, just issued a warning that should make anyone in knowledge work freeze. He didn’t say AI is getting too smart. He said AI is treating open math problems like a non-renewable resource.
Look at what just happened with the Navier-Stokes equations. AI didn’t just solve a problem; it collapsed centuries of slow, generative inquiry into an instant extraction. The proof is the trophy, but the real value of mathematics was never just the final answer. It was the shared struggle, the methods invented along the way, the entire culture of inquiry that grows around an unsolved mystery.
We aren’t using AI to explore the intellectual wilderness. We are using it to strip-mine it.
You’ve probably been told that AI will make mathematicians obsolete because it can solve equations faster than humans. That’s a fundamental misunderstanding of what mathematicians actually do. The current incentive structure doesn’t just want answers—it wants them immediately, at any cost. It treats open problems as extractable resources, pulling the value out of the ground and leaving a hollowed-out intellectual landscape behind.
Here is the twist nobody in Silicon Valley is talking about: when the answers become infinitely cheap and instant, the scarce resource isn’t problem-solving anymore. It’s problem-formation.
When the answers are infinite, the questions are all that matter.
If mathematics can be non-renewably mined by AI, the same extraction logic is coming for every creative and analytical field. Your codebase, your marketing strategy, your design framework—these are all intellectual ecosystems. If we optimize purely for short-term solution extraction, we exhaust the generative soil faster than new meaningful questions can be created. We get a million polished answers and no one left who knows how to ask the next big question.
The new role of the mathematician—and the knowledge worker—isn’t just to produce proofs. It’s to protect and regenerate the conceptual commons. To resist the urge to instantly extract every open problem for a quick publication or a viral tech demo.
If we treat knowledge as a mine, we will wake up in an intellectual wasteland wondering where all the ideas went.
We have to decide if we want to be miners or gardeners. Because once the generative ecosystem of inquiry is depleted, no algorithm in the world can generate the questions we forgot how to ask.
FAQ
Q: Can't mathematicians just reverse-engineer AI proofs to learn new things?
A: They can, but that's like studying the ashes of a burned forest to understand ecology. The generative process—the collaborative struggle that creates new mathematical fields and shared methods—has already been skipped.
Q: What does this mean for non-mathematicians?
A: If AI treats math as an extractable resource, it will do the same to coding, writing, and design. Your value shifts from producing answers to protecting the ecosystem of questions and knowing which problems are actually worth solving.
Q: Isn't faster problem-solving a good thing?
A: Only if you value the answer over the process. The current AI incentive structure optimizes for short-term extraction, which destroys the long-term cultural soil needed to grow entirely new disciplines.