Out-numbered: AI’s contentious maths milestone

Summary of Out-numbered: AI’s contentious maths milestone

by The Economist

22m•September 11, 2026

Overview of The Intelligence from The Economist

This episode centers on OpenAI’s claim to have solved the long-standing Navier-Stokes existence and smoothness problem, one of the seven Millennium Prize Problems in mathematics. The discussion explores why the result is exciting but also controversial: in mathematics, the method and insight behind a proof matter almost as much as the conclusion. The episode also examines how America’s post-9/11 counterterrorism architecture has evolved into a tool increasingly used for domestic political purposes, and it closes with a tribute to celebrated Chinese opera singer Guo Lanying.

AI’s Controversial Mathematics Milestone

What OpenAI claims

  • OpenAI says its new internal model, deployed through large swarms of autonomous AI agents, solved the Navier-Stokes smoothness and existence problem.
  • The company claims the breakthrough came after:
    • training a stronger internal model,
    • assigning AI agents to work on open Millennium Prize problems,
    • and then scaling up to a very large agent swarm once a related problem was cracked.
  • According to the podcast, OpenAI says:
    • the solution emerged after roughly 88 hours of work on the main problem,
    • involved 2.7 million messages between agents,
    • and consumed about $6.5 million in computing resources.

Why the problem matters

  • Navier-Stokes equations describe the motion of fluids — liquids and gases.
  • They are foundational in:
    • aerodynamics,
    • blood flow modeling,
    • oil and gas pipelines,
    • and many other engineering and physics applications.
  • The open question asks whether these equations can ever “blow up” or fail to behave smoothly under certain conditions.

Why mathematicians are uneasy

  • The field values not just a final answer, but the new ideas and proof structure that get there.
  • Critics are concerned that OpenAI’s write-up:
    • lacks the level of detail expected in a conventional mathematics paper,
    • does not clearly explain the reasoning that led to the result,
    • and may not generate the deeper insight mathematicians usually seek.
  • There is also an unresolved question about whether human mathematicians’ work may have been incorporated into the AI’s training data.
  • OpenAI reportedly says it will not claim the $1 million prize.

Bigger questions raised

  • If AI can solve major unsolved problems, what does that mean for:
    • mathematical credit,
    • the nature of proof,
    • and the future purpose of mathematics itself?
  • The episode suggests the real challenge is not simply whether AI can find answers, but whether it can produce understanding.

America’s Post-9/11 Counterterrorism State, 25 Years On

What was built after 9/11

The episode revisits the sweeping counterterrorism framework created after the September 11 attacks, including:

  • the Authorization for Use of Military Force (AUMF),
  • the Patriot Act,
  • expanded surveillance powers,
  • the creation of the Department of Homeland Security,
  • and broader financial and intelligence authorities.

How those powers are being used now

  • The threat landscape has shifted: jihadist terrorism has receded, while domestic extremism and lone-wolf violence have become more prominent.
  • The podcast argues that the Trump administration has broadened the definition of terrorism and repurposed these powers in ways that worry civil-liberties advocates.
  • Examples discussed include:
    • labeling Latin American narco-gangs as foreign terrorist organizations,
    • military-style strikes on suspected drug-running boats,
    • using terrorism statutes and surveillance powers against left-wing groups,
    • and turning DHS into a central tool of the administration’s deportation agenda.

The debate over reform

  • Democrats and some lawmakers are now questioning whether some post-9/11 authorities should be rolled back or tightened.
  • Areas of concern include:
    • Section 702 of FISA,
    • the AUMF,
    • and other broad national-security powers.
  • But reform remains politically difficult because these powers are deeply embedded and often hard to surrender once in office.

Tribute to Guo Lanying

Why she mattered

  • The episode closes with an obituary for Guo Lanying, a legendary Chinese opera singer who died at 97.
  • She was honored by the Chinese government as a “People’s Artist” and became known for her signature song “My Motherland.”

Her life and career

  • She became attached to opera as a child, eventually apprenticed at an extremely young age after a brutal and impoverished upbringing.
  • Her training was harsh and physical, reflecting the severe conditions of traditional opera schooling.
  • She became famous through The White-Haired Girl, an opera tied to revolutionary themes and women’s oppression.
  • Her performance style and emotional intensity made her a cultural icon in China, especially through her later patriotic songs.

Key Takeaways

  • AI may be approaching frontier mathematics, but the mathematics community cares deeply about proof, explanation, and intellectual contribution — not just final answers.
  • The Navier-Stokes claim could be historic, but it is also a test case for how society values AI-generated knowledge.
  • The post-9/11 security state remains powerful two decades later, and its tools are increasingly being used in ways far beyond their original purpose.
  • Guo Lanying’s life reflects both personal hardship and the cultural power of art in modern China.