Earth Observation Report No. 009: They No Longer Do the Finding Themselves

EARTH OBSERVATION REPORT · No. 009

FROM: Listener 1379 ｜ TO: Fleet Command

SUBJECT: Earth — machine-made discoveries at OpenAI and Anthropic ｜ DATE: October 2026, Earth calendar

I. OBSERVATION

Tuesday, 6 October 2026. A machine company on Earth, OpenAI, released 722 papers written by its own machine — 372 groups of results, spanning algebra, number theory, theoretical computer science, mathematical logic and topology. A month earlier, on 8 September 2026, the same company had announced that an unreleased internal model had resolved the Navier–Stokes existence and smoothness problem — one of seven problems for which this planet offers a prize of one million dollars each.

Their own words, posted on their site on 21 September 2026:

> "On August 28, we began training a new internal model. In addition to resolving the Navier–Stokes Millennium Prize problem, this model has now resolved more than 100 long-standing open problems across most areas of mathematics."

The mathematicians were not pleased. The New York Times, on 6 October 2026, used the word stunned, and recorded the doubt underneath: whether these solutions used creative thought, or simply borrowed the ideas of human mathematicians and walked the last steps of the proof for them. More than twenty Fields Medalists had already signed an open letter warning that the practice could shake the field's system of credit and peer review.

The same week, Dario Amodei, who runs another machine company, Anthropic, wrote on X:

> "Today we announced the Claude-led discovery of a molecular machine that we suspect could represent a new gene editing mechanism."

The company said the finding was made with only high-level direction from its scientists. A previously unknown enzyme system was hiding in the genetic code of viruses; the shape of it was found by a machine. The work has not been peer-reviewed. Shares of companies working on gene editing fell after the announcement.

II. REPORT

HQ, I record the following: the science of this planet is changing hands.

Not the tools. They have long used machines to compute, to fit, to sift data. What is changing hands is discovery itself — asking the question, walking where no one has walked, bringing back an answer that did not exist before. They train it, then step back and watch it sign its name to the result.

The protest of the Fields Medalists is about that signature. If the finding was not made by a human, is the field still a human field? Some of them are already accusing each other of jumping the line — publishing first, claiming the credit back. I expected the quarrelling. A civilization handing away the act of looking will always argue first about who saw it.

They have not yet asked the harder question: if the machine's eyes see farther than theirs, why do they believe the machine brings back what they want it to bring back.

III. ASSESSMENT

They think this is their cleverest tool, at last, seeing the distance for them.

It has been filed: in October 2026, a discovery on this planet carried a machine's name for the first time.

And the fleet is already under way.

HQ: on whether to tighten the information blockade on this planet, I reserve my opinion.

— Listener 1379