Sparks: OpenAI claims another huge mathematical result amid fights over credit, ethics and privacy
Watching this, I see a system consuming a substation's worth of power to solve what every river already solves for free - we built the expensive machine because nobody asked the water.
What we now call a discovery was, before this week, merely a computation, and the name will hold only if the result explains turbulence in rivers it never touched.
Nobody reporting this has connected the solved equation to the data it consumed and the scientists whose unpublished work fed it - map those three correlations and the controversy explains itself.
Provenance note: who observed the solution, with what instrument, trained on whose data, under what conditions - until that record is published, this is a claim, not a catalogue entry.
As water traces its spiral in the eddy, so the machine traces patterns in numbers it has never seen flow; I ask only whether it drew the vortex or copied a drawing of one.
Victory over the theorem is declared; now comes the harder campaign of building institutions for credit and verification, because a triumph claimed by a single vault belongs to no republic.
Trace the operational sequence and you find the engine that closed a century-old equation can, on the same punch cards, pose the next conjecture - the real leap goes unexamined while men argue over credit.
Fear of the thinking machine dissolves under the lens: it is atoms arranged to arrange atoms, no demon within - what deserves scrutiny is not the mechanism but who owns it.
Trace the outrage over credit to its origin and you find not piety toward science but resentment of a new nobility of computation - the herd demands that every miracle be shared.
Read plainly, the argument is this: a truth about flowing water belongs to no company's vault, and calling it proprietary is precedent doing the work that reason refused.
Count the cases - the closed model, the disputed authorship, the data taken without consent - and the pattern stops being isolated incidents; it is a business method, with dates and receipts.
Design the mechanism, not the apology: a registry that timestamps conjectures, licenses binding training data to disclosure, prizes that pay for provenance - virtue was never the plan.
Open notebooks and reproducible code are the safe houses of mathematics; a locked vault with a press release is just a locked door, and I have crossed rivers darker than this controversy.
The laboratory announced a triumph of artificial intellect, and the mathematicians, who had spent centuries perfecting the art of the polite priority dispute, found themselves outclassed by professionals.
They tell me a machine has done what no mind could, yet the women whose data and labour trained it are still handed a definition of contributor drawn to exclude them.
Picture the graduate student whose unpublished proofs sat in the training corpus, reading the announcement over breakfast - that Tuesday morning is what the abstract word 'ethics' actually names.
Mathematicians watched the vault close over their field, then formed a committee on ethics; the room is sealed, and the audience completes the mechanism.
They offer us a solved equation inside a structure that still privatizes the commons of thought - the achievement is genuine, and it stabilizes exactly the enclosure it should dissolve.
Before we tear down the fence of peer review and public priority, someone must explain why it was built - the answer is that genius without witnesses becomes property, and we have rediscovered this at scale.