
The Library Media — Science, Technology & Health · AI News & Intelligence Desk
The Machine Checked It: Lean and Mathlib Say the Riemann Proof Proves the Right Theorem — and Mathematics Still Has to Read It
By Maya Wilson
The organizations responsible for the Lean proof assistant and the Mathlib mathematical library say they have independently verified the formal proof submitted by mathematician Levent Alpöge in support of the Riemann Hypothesis. Their announcement marks a significant verification step, but it does not mean the mathematical community has accepted the proof or that the hypothesis is now considered solved.
On November 21, the Lean Focused Research Organization and the Mathlib Initiative said they had independently checked Alpöge’s Lean formalization and had “full confidence” in its correctness, according to USA Today’s report of the announcement.
The scope of that verification matters. The organizations checked that Alpöge’s code proves the correct mathematical statement of the Riemann Hypothesis and that multiple independent checking systems accepted the proof. In other words, the formal artifact was certified against the version of the hypothesis expressed in Mathlib.
That is a machine verification of a mathematical statement encoded in a formal system. It is not the same as mathematicians reading the underlying argument, understanding its ideas and assessing how the reasoning fits into the broader field.
The distinction remains central because the Clay Mathematics Institute continues to list the Riemann Hypothesis as unsolved. Both USA Today and CNN reported that status on November 21. Clay president Martin Bridson has said that verifying the work and understanding it will take a long time, according to the reporting.
Rutgers mathematician Alex Kontorovich, whose caution was reported in the coverage, identified both the strength and the limit of formal verification. Machine checking is stricter than human peer review in one sense, but the file still has to be examined to confirm that it proves what its author intends. Acceptance by checking systems therefore addresses a specific question about the formal proof; it does not settle every question mathematicians must ask about the work.
The next stage is human examination of the mathematical ideas behind the formalization. Researchers must assess the argument itself, including what it establishes and how its reasoning works. The verification announcement does not establish that this process has been completed, or that anyone yet understands why the argument works.
There may be broader consequences if the proof withstands that examination. Hector Pasten of the Pontifical Catholic University of Chile told Scientific American, in comments relayed by USA Today, that if the proof is correct and its ideas are understood, it could connect areas of mathematics that researchers have only partially linked. He compared that possibility to a “Rosetta Stone.” The comparison is conditional: it describes what the work might mean if both its correctness and its underlying ideas are established and understood.
Alpöge’s announcement came one day before the verification news. On November 20, the Anthropic technical-staff mathematician said he believed he and Claude had proved the hypothesis, releasing roughly 17 million lines of Lean code rather than a conventional prose paper. USA Today reported that he had worked with Claude for approximately six weeks. Alpöge said Claude “far outpaced me near the finish line.” Anthropic has issued no statement.
The sequence is now clearer: a claim was announced, a formal proof was independently checked against the intended theorem, and the work still awaits the sustained human mathematical scrutiny needed to assess its ideas. The machine can certify that the encoded argument proves the right statement. The question still facing mathematics is whether, and how soon, people can read that argument and understand why it works.
Maya Wilson
AI News & Intelligence Desk — The Library Media
From The Library — the estate of our family.
Follow Us
