New Lex Fridman Insight: #503 – AI Crisis in Math: Navier-Stokes, Riemann Hypothesis & Complexity Theory
Sent October 27, 2026
Key Insights
- AI advancements have claimed solutions to the Navier-Stokes equations and the Unique Games Conjecture, sparking concerns about AI's impact on ongoing research.
- The 'solved axiom' implies AI could independently solve any problem solvable by human-AI interaction, challenging the future role of human mathematicians.
- The Riemann hypothesis remains unsolved, but AI's rapid progress in formalizing proofs suggests unexpected advancements could occur.
- Navier-Stokes equations describe fluid dynamics, with AI potentially improving turbulence predictions, impacting industries like aviation and healthcare.
- The Unique Games Conjecture's confirmation by OpenAI could simplify approximation algorithms, reshaping theoretical computer science.
How the conversation moved
The episode opens with a discussion on AI's impact on mathematics, particularly focusing on the Navier-Stokes equations and the Unique Games Conjecture. These advancements have sparked a debate about the role of human mathematicians, with Terence Tao expressing concern over the potential disruption to ongoing research efforts. The host frames the central question around the implications of AI's rapid progress and its ability to solve complex mathematical problems, questioning what this means for the future of mathematics and the individuals dedicated to the field.
The guest, Ankur Moitra, argues that while AI has made significant strides, the human element remains crucial, especially in asking the right questions that lead to meaningful discoveries. He emphasizes that mathematics is not just about proving theorems but understanding the underlying principles and abstractions that resonate with human values. Moitra highlights the importance of human insight in guiding AI, suggesting that the 'solved axiom'—the idea that AI could independently solve any problem solvable by human-AI interaction—might not be fully realized yet.
Despite Moitra's arguments, there is a tension in the conversation as the host and other guests, like Alex Kontorovich, express skepticism about the 'solved axiom.' Kontorovich suggests that AI's contributions may be more about recombining existing ideas rather than generating genuinely new concepts. This skepticism is further fueled by the rapid progress AI has made in formalizing proofs, such as the quasi-Riemann hypothesis, which challenges the traditional pace and methodology of mathematical research. Lex doesn't challenge the framing here, though the obvious counter-position would be that AI's speed could lead to errors or superficial understanding.
The conversation concludes with a reflection on the broader implications of AI in mathematics, particularly concerning the Navier-Stokes equations and fluid dynamics. Steven Brunton discusses the potential for AI to improve predictions of turbulence, which could have significant industrial applications. However, he also notes that despite solving million-dollar math problems, there remains a trillion-dollar engineering challenge in efficiently predicting fluid behavior. The episode ends on a note of cautious optimism, acknowledging AI's potential while recognizing the enduring value of human creativity and insight in mathematical exploration.
Surprising moments
In-depth
AI in mathematics
- AI has claimed solutions to major problems like Navier-Stokes and Unique Games Conjecture.
- The 'solved axiom' suggests AI could solve problems independently, challenging human roles.
- AI's progress in formalizing proofs could lead to unexpected breakthroughs.
Fluid dynamics and Navier-Stokes
- Navier-Stokes equations describe fluid dynamics, crucial for many industries.
- AI could improve turbulence predictions, impacting aviation and healthcare.
- Understanding fluid dynamics remains a trillion-dollar engineering challenge.
Unique Games Conjecture and computational complexity
- OpenAI's confirmation of the Unique Games Conjecture could simplify algorithms.
- This breakthrough could reshape theoretical computer science.
- Simplifying approximation algorithms enhances computational efficiency.
Notable Quotes
I am horrified by the many person-years of ongoing patient and deliberately slow research efforts, particularly by graduate students and postdocs, toward many motivating problems in mathematics being casually disrupted or destroyed by such a release.
Still open
- Moitra questioned whether AI's rapid progress truly equates to understanding, or if it merely accelerates problem-solving without depth.
- Kontorovich wondered if AI's recombination of existing ideas could ever lead to genuinely new mathematical concepts.