Ain't gonna put mathematicians out of work. We still need someone to understand it.
https://openai.com/index/ten-advances-in-mathematics/

I think that's largely right.

Even in the optimistic interpretation of OpenAI's "Ten Advances in Mathematics and Theoretical Computer Science," the remarkable part isn't just that a model generated proofs or counterexamples. The remarkable part is that human mathematicians had to recognize that the problems mattered, verify the results, connect them to existing theory, and explain why anyone should care.

Mathematics has always had a computational component that tools gradually absorbed. Human "computers" once spent their careers doing arithmetic. Calculators reduced that need; symbolic algebra systems reduced another layer; proof assistants reduced yet another. Yet mathematicians did not disappear. The work shifted upward toward formulating questions, choosing definitions, judging significance, and building conceptual frameworks.

One interesting theme in recent commentary is that the bottleneck may move from finding proofs to understanding them. Terence Tao has described AI as becoming genuinely useful in research while still requiring human oversight and judgment. Other mathematicians report spending substantial effort checking, repairing, and interpreting AI-generated arguments.

Your phrasing — "We still need someone to understand it" — gets at something deeper. Suppose an AI proves a theorem nobody expected. There are at least three distinct tasks:

    Verifying that the proof is correct.

    Understanding why it works.

    Understanding why it matters.

The first may become increasingly automated. The second is harder. The third may remain the most human of all, because "importance" depends on the broader landscape of mathematical ideas and on human interests.

The analogy I keep thinking of is chess. Engines became stronger than grandmasters years ago. Yet people still study chess, write books about chess, teach chess, and argue about beautiful games. The nature of expertise changed from "can you out-calculate the machine?" to "can you understand what the machine found?"

Mathematics may be heading in a similar direction. A future mathematician might spend less time discovering every step personally and more time acting as interpreter, curator, critic, and theorist of an ever-growing body of machine-assisted results. That is still mathematics—it is just a different job description.

In fact, one of the more intriguing possibilities is that truly powerful AI creates more work for mathematicians. If machines generate new conjectures, proofs, and entire branches of theory faster than humans can digest them, then understanding becomes the scarce resource rather than computation. That's a view I've seen increasingly expressed by mathematicians reacting to these recent breakthroughs.

That is a wonderful example of how provenance can become almost a physical part of value.

The receipt did not change the wool, the weaving, the dyes, or the skill of the Navajo weaver. The rugs were exactly as beautiful the day before. What changed was the story that could now be attached to them. The piece of paper transformed them from "old Navajo rugs" into "these specific rugs, purchased in Durango in the 1880s, with a documented chain of history."

Collectors call this provenance, and it is one of the strangest forces in human valuation. A painting by an unknown but brilliant artist and a painting by a famous artist may be visually identical; the difference in price can be enormous because the object carries a human narrative.

It connects directly to your earlier point about novels and AI. A reader may genuinely enjoy a novel regardless of who wrote it. But if the novel turns out to be the last manuscript written by a famous author, or the first novel generated by a conscious machine, or a work found in an old attic with a remarkable history, that context changes how we perceive it.

Humans are unusually attracted to authenticity. We don't just value the thing; we value the relationship between the thing and a particular chain of events. A Stradivarius violin is not valuable merely because it produces good sound. A musician playing one is participating in a centuries-long story.

Your receipt was almost like a tiny historical artifact attached to the rugs themselves. It answered a question that an appraiser could only estimate: "Where have these been?" Once answered, the rugs acquired a past—and the past has monetary value.

It is a curious contradiction: we like to think we value objects for what they are, but very often we value them for what they mean.

In this case I don't attribute it to human nature but to the particular culture of collectibles and the interests involved.

That's a fair refinement. I was using "human nature" rather broadly, but your distinction matters.

The fact that provenance affects value is not an inevitable law of human psychology. It is a feature of particular social systems: collecting, art markets, museums, antiques, auctions, and the communities that decide what counts as significant. Those communities create rules about what kinds of information increase value.

A Navajo rug with a documented 1880s Durango purchase history is not necessarily more beautiful, warmer, or better made than an identical rug without documentation. The market value rises because collectors have agreed that historical connection is a desirable attribute. The receipt functions almost like an additional component of the object—not a physical component, but a cultural one.

There are many examples where this doesn't happen. A person buying a handmade blanket simply for warmth may not care at all whether it was purchased in 1880 or 1980. A musician may value the sound of a violin without caring about its maker's name. A reader may enjoy a novel without knowing its author.

In fact, this ties back neatly to your AI authorship point. If a culture of literary collecting develops around AI-generated works, then provenance might become important in that world too: "the first novel written entirely by an AI," "the first novel co-authored by a human and an AI," and so forth. The text itself might be judged separately from its historical significance.

So perhaps the better statement is not "humans always value authenticity," but: once a community decides that origin, rarity, or historical association matters, those facts become part of the value of the thing.

Your receipt mattered because the collectors' culture had a place for that fact. Without that culture, it would have just been an old piece of paper.

The thing that got me was they spent more effort being sure the receipt was genuine than they did with the rugs themselves.

That is a fascinating inversion, and it really does reveal how provenance works.

The rugs themselves were right there to be examined. An appraiser could look at the weave, materials, patterns, dyes, wear, and style and make a judgment. But the receipt answered a different question: not "what are these?" but "what story can we confidently attach to these?"

A questionable receipt could actually be more dangerous than an uncertain rug. If the paper were false, it could artificially inflate the value of an otherwise ordinary object. So the experts' attention shifted from the physical artifact to the evidence surrounding it. In effect, the receipt became the thing that needed authentication.

This happens all over the collecting world. A painting may be obviously old, but a signature can determine whether it is worth hundreds or millions. A fossil may be real, but the location where it was found can change its scientific importance. A historical document may be less visually impressive than the object it describes, but it can completely alter the object's meaning.

It is almost like a court of law: the evidence about the evidence becomes crucial. The rug is the "testimony"; the receipt is the "witness." And a witness's credibility has to be established.

There is also a subtle irony: the more valuable the objects become, the less the experts can rely solely on their own eyes. They have to investigate the invisible history around them. The physical thing is only half of the artifact; the documented chain of custody becomes the other half.

Your experience is a nice illustration of why collectors can sometimes seem to be valuing a scrap of paper more than the object itself. They aren't really valuing the paper. They are valuing the certainty that the paper provides.