FULL STORY

OpenAI Claims Landau-Siegel Proof, Math World Shaken

OpenAI claimed a proof excluding Landau-Siegel zeros and released 700 papers, hailed by Terence Tao as a 50-year breakthrough. Mathematicians then compressed the 8-page proof to five, as debate raged over AI's role in mathematics.

2026-10-17 ~ 2026-10-21 · 6 episodes · 34 posts

Episode 1 · 3Blue1Brown: There's Never Been a Better Time to Go Into Math (2026-10-17, 9 posts)

On October 17, 3Blue1Brown (Grant Sanderson) released "There's never been a better time to go into math," a systematic reflection on AI and mathematics; the next day Hesamation published a full 40-minute interview expanding on his views. The video starts from math students' anxieties, breaks down what mathematicians do, why math is valuable, and where funding comes from, then reasons about how each would change if AI capabilities keep growing.

Confirmed

  • The core claim: this is the best time ever to enter mathematics. Anyone can now access mathematical knowledge conversationally—exemplified by a commenter saying "I have a PhD in math; just tell me what to do"—and AI can take over tedious derivations.
  • In the Hesamation interview, Sanderson's rationale is that the math community as a whole recognizes it must change; many senior scholars are stepping back to seek directions and exemplars, giving newcomers a chance to shape that change.
  • The video unfolds through "how math operates → proof machines → conjecture machines → theory (construction)."
  • A 2.5K-like comment resonated widely: the line "AI does the math so you can do something else" dodges the core issue—what draws people into math and software is treating them as art and craft, achieving a kind of immortality by solving hard problems.
  • In the ensuing debate relayed by giffmana, the opposing view holds that people can code and do math purely for joy without maximizing productivity, rejecting the idea of stopping people from enjoying things just to preserve human usefulness.
  • joshwills added that PhD career pains make the issue more complicated than optimism.

Why it matters

  • Rather than taking sides on whether AI replaces mathematicians, this is a rare public exercise in itemizing the sources of value and funding in a craft-based knowledge profession under AI—relevant well beyond math, e.g., for software.
  • The comment's resonance (2.5K likes) and the follow-up discussion show that grief over losing one's craft and real PhD career anxieties coexist even among AI-tool believers, while the "learn for joy, not productivity" counterpoint offers an important corrective to the optimistic narrative.

Episode 2 · Debate Erupts Over Whether AI-Proven Theorems Count as Progress (2026-10-17, 3 posts)

OpenAI's AI-generated mathematical proofs have sparked fierce debate: critics like physicist Étienne Klein call AI solving major math problems meaningless, while defenders accuse elite academia of denying what would recently have been considered Fields Medal-level achievements.

Episode 3 · OpenAI Claims Quasi-Riemann Hypothesis Proof and 700 AI Papers; Tao Calls It Biggest Advance in 50 Years (2026-10-17, 6 posts)

On October 6, OpenAI claimed its AI model proved the quasi-Riemann Hypothesis (excluding Landau–Siegel zeros) with a 199-page proof, and simultaneously released 700+ AI-generated math papers reportedly resolving hundreds of open problems. The news, covered by Science, triggered anxious debate about the value of human mathematicians, though within days to two weeks human mathematicians solved several related problems.

Confirmed

  • Terence Tao called the result the biggest advance on the quasi-Riemann Hypothesis in 50 years, citing a huge new zero-free region and very clever new techniques, while noting it essentially combines three or four existing ideas in an unusual way.
  • OpenAI's claimed results include a solution to the long-open amenability problem for Thompson's group F; Justin Moore, who studied it for 19 years, wrote about his feelings, saying opening a bottle felt like a wake rather than a celebration.
  • The release of 700+ AI-generated papers was reported by Science.
  • Number theorist Hector Pasten and others praised the proof, and the math community briefly panicked about the value of human mathematicians.
  • Within days to two weeks afterward, human mathematicians solved multiple related problems.

Unconfirmed

  • Whether the 199-page proof has passed full peer review remains unclear in the source material.

Why it matters

  • If verified, it would be the most important advance in Riemann-adjacent research in decades and a signal of AI moving from auxiliary tool to primary force in frontier mathematics.
  • The episode also prompted reflection on the role of human mathematicians, whose rapid follow-up breakthroughs showed human research capability remains far from replaced.

Episode 4 · OpenAI's 370 Math Proofs Send the Mathematics World into Turmoil (2026-10-18, 3 posts)

After OpenAI released 370 math proofs on GitHub, Scott Aaronson said graduate students are terrified and mathematicians worldwide have paused their work to digest the results, while MIT students and researchers describe an existential crisis over the point of studying math.

Episode 5 · OpenAI's 8-page proof of no Landau-Siegel zeros shrinks to 5 (2026-10-19, 3 posts)

Mathematicians have rapidly compressed OpenAI's 8-page proof that Landau-Siegel zeros do not exist—Kerr and Green cut it to 6 and then 5 pages, and a new paper reduces it to an elementary exercise—sparking discussion that many open math problems may just be 'exercises for sufficiently determined readers.'

Episode 6 · 3Blue1Brown: There's Never Been a Better Time to Go Into Math (2026-10-19, 10 posts)

3Blue1Brown creator Grant Sanderson released the video "There's never been a better time to go into math," offering a systematic analysis of how rapid AI progress affects mathematics. His central argument: even if AI surpasses humans at mathematical proofs across the board, the value of mathematics won't die out—it may even enter its golden age. Society will still value human understanding of complex problems in the long run, mathematicians will remain a meaningful profession, and the upheaval creates opportunities for young researchers to redefine the discipline.

Confirmed

  • The video breaks mathematical research into components (proving theorems, posing conjectures/questions, building theories, applying theories, etc.) and assesses AI's impact and the remaining human advantages for each.
  • In follow-up debates with commenters, Sanderson added: if viewers conclude that "studying math is only for passion, with no economic value," then he failed to express himself well; he believes spending on fun and entertainment itself constitutes economic value, and math's other values won't be dissolved by AI either.
  • He also recommended a long-form conversation between Rutgers math professor Alex Kontorovich and University of Virginia professor Thomas Koberda, "The Dialectical Theorem," discussing the future of mathematicians' careers in the AI era.
  • Mathematician Alex Kontorovich said he finds it extremely hard to fully agree with anyone's views, but couldn't find a single flaw in this video, and would share it with every PhD student he knows.

Unconfirmed

  • User Markus Baumann raised a challenge: "learning for passion" sounds like consolation for when math becomes "the new humanities" and loses economic value and status—this dispute remains unresolved.
  • A pessimistic reading from X user Stocko holds that the video actually shows mathematicians' territory (proofs, questions, theory-building, applications) being conquered by AI one by one, leaving humans only "asking good questions"—but this is one interpretation, not the video's conclusion.

Why it matters

  • This is a systematic, frontal response to the anxiety over "will AI end math careers" from a leading math communicator, amplified by endorsement from working mathematicians, and it is shaping how young researchers think about career choices.
  • The debate over whether math is becoming "the new humanities" reflects a broader anxiety about the revaluation of fundamental disciplines under AI impact, and deserves continued attention.