#472 – Terence Tao: Hardest Problems in Mathematics, Physics & the Future of AI

Terence Tao, a renowned mathematician and recipient of the Fields Medal, discusses the evolution of human memory and technology, noting how modern tools like Google have shifted the burden of memorization from individuals to machines. He expresses concern that this shift may diminish inspiration and discovery. The conversation explores the future of automation and AI, emphasizing the need to guide technological advancements for human flourishing, particularly in business management where human intuition remains crucial.

Tao reflects on his early research, particularly the Kakeya problem, which involves determining the least area required to rotate a needle in a plane. He connects this problem to broader mathematical concepts, including partial differential equations and geometry, and discusses its implications for wave behavior and singularities in mathematical physics. The Navier-Stokes regularity problem is also addressed, questioning the conditions under which singularities may form. Tao emphasizes the need for absolute certainty in mathematical proofs, contrasting this with other fields where near-certainty may suffice.

The complexities of fluid dynamics governed by the Navier-Stokes equations are discussed, including their significance for weather prediction. Tao likens the challenge of proving generalities about these equations to Maxwell's Demon, illustrating the difficulties in establishing certain mathematical phenomena. He notes that while fluids can become turbulent, they do not transfer all energy from larger eddies to smaller ones, which is essential for understanding potential blow-up scenarios.

Tao's approach to proving global regularity for Navier-Stokes involved averaging the equations and selectively modifying interactions to force energy to blow up in finite time. He highlights the importance of identifying ineffective techniques in tackling complex mathematical problems, particularly in the context of supercriticality, where competing forces create unpredictability. He discusses the distinction between supercritical, critical, and subcritical equations, emphasizing the role of nonlinearity in fluid dynamics and weather modeling.

The conversation introduces a theoretical machine that functions like a Rube Goldberg device, using water to perform computations akin to modern computers. This concept involves representing bits with water pulses, where collisions create logic gates to construct a Turing machine entirely from water. While this idea offers a roadmap for addressing the Navier-Stokes problem, it remains theoretical, facing practical challenges.

Tao explores the dichotomy in mathematics between structure and randomness, discussing the challenges of proving the existence of patterns in sequences, such as the digits of pi. He references the infinite monkey theorem and emphasizes the need for careful reasoning when dealing with limits. The conversation highlights the limitations of non-human intelligence in creating complex narratives, contrasting it with the unique capabilities of human mathematicians.

The discussion touches on the interplay between mathematics, physics, and engineering, emphasizing the reliance on unrealistic assumptions to simplify complex realities. Tao notes the rise of experimental mathematics, facilitated by computers, which allows for the exploration of extensive data sets. He discusses the challenges of combinatorial explosion in mathematics and the role of AI in enhancing the experimental aspect of mathematics.

The allegory of Plato's Cave is referenced to illustrate the limitations of human perception, questioning the possibility of accessing true reality. Effective modeling is discussed, emphasizing that successful models should have fewer parameters than data points to avoid overfitting. Tao highlights the universality in mathematics, noting that complex systems can emerge from simple interactions.

The conversation contrasts two styles of mathematicians: "hedgehogs," who possess deep expertise in a specific area, and "foxes," who have a broader understanding across multiple fields. Tao identifies as a fox, valuing the exploration of connections between different mathematical domains. He emphasizes the importance of analogies and narratives in problem-solving.

Tao discusses the role of notation in mathematics and the evolution of physics concepts, particularly in quantum mechanics. He highlights the ongoing challenge of unifying quantum mechanics and general relativity, expressing a theoretical belief in a "theory of everything." The historical collaboration between physicists and mathematicians is noted, with string theory mentioned as a leading candidate for unification.

The conversation addresses the unique contributions of humans in mathematics, such as inventing new theories and abstractions, which AI may struggle to replicate. Tao reflects on the complexities of transforming mathematical problems and the potential for AI to enhance problem-solving capabilities. He acknowledges the challenges of formalization in proofs and the increasing role of AI in mathematics, particularly through tools like Lean.

The discussion highlights the importance of collaboration in mathematics, particularly during initial brainstorming phases. Tao contrasts traditional collaboration with lean programming, which allows for a divide-and-conquer approach. He emphasizes the potential for distributed contributions to mathematical problems, especially as platforms like Lean and GitHub could significantly scale experimental mathematics.

Tao addresses the challenges of coding in mathematics and the implications of large-scale projects like the Equational Theories Project. He critiques current academic metrics and discusses the importance of recognizing contributions in collaborative efforts. The conversation shifts to the integration of AI in mathematics, particularly with DeepMind's AlphaProof, and the challenges of mapping natural language to formal language.

The discussion touches on the Poincaré conjecture and Grigori Perlman's solitary work in solving it. Tao reflects on the emotional investment in mathematical problems and the value of making mistakes in research. He acknowledges enduring mathematical problems like the twin-prime conjecture and the Riemann hypothesis, discussing their complexities and the challenges of proving them.

The conversation concludes with reflections on the evolving nature of education, the importance of adaptability in learning, and the potential for new tools to engage the public in math research. Tao expresses hope in the creativity of the younger generation and the need for a healthy infrastructure to foster a more intelligent and rational community.