Terence Tao: Did AI Just Solve the $1,000,000 Navier-Stokes Problem?

Look closely at the world around you, whether you're observing the majestic spiral arms of distant galaxies or watching the milk swirl into your morning coffee, you are witnessing the same fundamental phenomenon, fluid flow. It's beautiful, it's mesmerizing, and it seems to be consistent until we look deeper. We believe that a single set of rules governs all of this motion, from the largest cosmic scales down to the smallest eddies in a stream. But there is a haunting question that keeps physicists and mathematicians like Terence Tao awake at night. What if those rules can break? What if the mathematical code that describes this motion contains a hidden flaw? Today we're embarking on a journey into the hidden heart of complexity, where we ask if our current understanding of the physical universe is complete, or if there's a glitch waiting to be found in the very nature of how fluid and other particles move. To understand the glitch, we first have to understand the rules. For that, we need to travel back 200 years to the 1800s. Two brilliant minds, Claude-Louis Navier and George Stokes, penned what we now call the master equations for fluid flow. Starting with Isaac Newton's deceptively famous and simple second law, force equals mass times acceleration, which works perfectly for solid blocks sliding down a pramsay, but they adapted it for a continuous substance. This was a massive leap. They created a mathematical recipe to describe substances that don't hold their shape, like water or air. These equations, later known as the Navier-Stokes equations, are the bedrock of modern physics. They give us instructions for how fluids move, how fluid flows, how ocean waves swirl, and describe a lot of video game and AI generated physics demonstrations, and they provide the basis for modern simulations, including their incarnations in video game design. They're the very reason we can design airplanes and predict weather. But writing down the equations is one thing. Solving them is something totally different. This mathematical sentence is the source code for classical physics regarding fluids. It looks intimidating, but it actually tells you something dynamical. It represents a precise balance of forces. On the left side, we have the acceleration of the fluid, how its motion is changing from one moment to the next. On the right, we have the forces pushing and pulling on that fluid, F equals ma. It's an equation that updates our reality and our understanding millisecond by millisecond. It tells the water in a pipe or the wind over a wing exactly what to do next, but there's a catch. While the equation looks compact, it hides infinite complexity. It's what's called a differential equation, and such equations must hold true at every single point in 3D space all at once, and hidden within these symbols is a specific mechanism that makes the math incredibly difficult to entangle. Let's break the equation down into its characters to see where the drama comes from. We have the velocity, which tracks changes over time. We have the pressure gradient, which dictates how fluid moves from high pressure to low pressure. We have the viscosity term, which represents internal friction. Think of it as the stickiness of honey that tries to smooth things out. But then there's the troublemaker, the term in the middle, u dot del times u. This represents non-linearity, u times u, u squared. In simple terms, this is the fluid carrying itself around. It means that the flow of energy depends on the flow itself. It seems like self-dealing, but this self-referential loop is where chaos is born. It turns simple inputs into wildly unpredictable and complex outputs. Whenever you see non-linearity, these types of behaviors arise. When the other terms try to keep things orderly and smooth, this non-linear term is constantly trying to twist the system into unpredictable knots, whirls, eddies, vortices. Because of that troublemaker term, these equations govern two completely different worlds. On one side, we have laminar flow, the predictable world. This is smooth, gentle, and soluble. We can calculate exactly what a slow-moving river will do. But turn up the speed and you enter the second world, turbulence. This is the realm of chaos. In turbulence, tiny changes at the beginning magnify. They don't fade away. They get amplified. A butterfly flapping its wings causing a tornado is not just a metaphor, it's a mathematical possibility here. We use these equations every day to engineer technology, to design airplanes and wind turbines, and even to mimic reality in video games so you can get that extra millisecond of frag speed. But the frightening truth is that for turbulent flow, we can only approximate the chaos. We cannot solve it exactly. We literally do not know if the math holds together or whether it breaks down when the water gets rough. This uncertainty is so profound that the Clay Mathematics Institute has named one of the seven Millennium Prize problems. They have placed a $1 million bounty on the answer. The challenge is not to calculate a specific flow, but to prove something fundamental about the equations themselves in three dimensions. The challenge asks for a proof of one of the two following scenarios. Scenario A, do solutions always exist and stay smooth and well-behaved forever? Or Scenario B, can you show a specific example where a smooth start leads to a disaster? The Millennium Problems ask for a mathematical guarantee. Either the universe is safe from runaway chaos and the equations will work forever, or there's a trapdoor in reality where the physics literally falls apart. What does it mean for physics to fall apart though? Is it just a catchphrase that mathematicians use to make something sound more breathlessly exciting than it actually is? No, physicists and mathematicians call something a singularity or a blow-up whenever a following phenomenon takes place. Imagine a fluid flowing smoothly, then purely through its own internal mechanics it begins to twist and tighten. It forms a knot of energy that gets smaller and smaller, spinning faster and faster. A blow-up or singularity is the theoretical moment where the velocity or pressure at that point becomes infinite. And the terrifying part is that the math suggests this could happen in a finite amount of time. It's like a camera zooming in on a whirlpool, and before you can stop it, the speed hits infinity and the laws of physics break. If this can happen, it means the equations describe a universe that will eventually destroy itself, at least the fluid mechanical part of itself. You may ask, why should I care about a math glitch when real water doesn't explode like this? And you'd be right. And that's what led Terence Town on his adventure to try to solve this problem. Real water does not possess infinite velocity. But that is exactly why this matters. If the Navier-Stokes equations predict a blow-up, but nature never actually demonstrates one, it means our equations, our fundamental source code for a classical world, are flawed or perhaps incomplete. It could be a crisis for physics if this were true. It would imply that the tool we have used for two centuries to build our civilization, or at least the technology in it, is wrong at a very, very deep and fundamental level. It tells us that our map does not match the territory. Solving this isn't just about the money, obviously. It's about verifying that we actually understand how the physical world was constructed. For decades, no one could prove it either way. Then, in 2014, Terry Tao got on the scene. He couldn't initially prove a blow-up in our real equations, so he tried something clever. He built a toy universe. He created a mathematical model with equations that were slightly modified. Very similar to ours, but distinct. In this parallel mathematical universe, Terry proved that a finite-time blow-up or singularity is possible. He showed that fluid in this toy model could indeed crush itself into a singularity. This was a warning shot. It wasn't real life. It wasn't our real universe. But it showed us that the real Navier-Stokes equations are likely sitting right on the cusp of a singularity, of a blow-up. We are balancing on the knife edge between stability and chaos. To solve the real problem, Tao developed a roadmap focused on what's called the energy cascade phenomenon. In turbulence, energy moves. It transfers from large scales like big swirls in a river down to smaller scales and then even smaller scales. Eventually, we end up in tiny, almost microscopic eddies. Usually, the viscosity kills the energy at the bottom. But the danger zone is if this cascade happens too fast. If the energy concentrates into an infinitely small point faster than friction can dissipate it, you get a blow-up, a singularity. Tao's program is essentially an attempt to track the energy, to bookkeep it, to make sure that we can prove mathematically that the energy can never pile up fast enough to break the system. He needed to tame the cascade. The strategy to tame the cascade relies on finding invariants. These are things that, due to symmetry, never change. Think of conservation of energy. In the chaotic mess of a storm, are there quantities that refuse to change? Are there symmetries we can exploit? These are called critical norms. If you can find a quantity that remains invariant under scaling, meaning it doesn't get bigger even as the swirls get smaller and smaller, you have a leash to tame the chaos. The mathematical goal is to prove that these critical quantities stay bounded, finite. If they have a limit, then the velocity can never go to infinity and the system can never really blow up. You're hunting for the mathematical seatbelts that keep the universe from falling out of the car, but finding them has proven quite difficult for human intuition alone. This is where the story enters the modern age. The complexity of 3D turbulence might be beyond what human brains can visualize on their own. So a new partner has joined the hunt. AI. Artificial Intelligence. We aren't just using computers to run simulations and make better video games. We're using deep learning and AI reasoning models to think alongside mathematicians like Terry. These AI models can navigate high dimensional spaces that leave human brains baffled, even brains as bright as Terry Tao's. They serve as a partner for mathematicians and look for the breakage points and blowups and singularities in the math. It's a fusion of rigorous human logic and the unmatched pattern recognition prowess of machines. How does this partnership actually work though? The AI follows Terry's roadmap in two specific ways. First, it acts as a pattern hunter. It analyzes massive data sets of turbulent simulations to find those hidden invariants, the statistical norms, the laws that humans simply can't see in the bulk amidst the noise. Secondly, the AI that they have developed helps build better toy models, more accurate to help us reverse the engineering data to propose new simplified equations for mathematicians and physicists to study. In this way, AI is helping our intuition by bridging the gap between the unsolvable real world and the solvable but incomplete toy model. It's checking the math, probing for weaknesses, and guiding the human researchers towards the final proof. We're now standing at the precipice. The Navier-Stokes equation solutions unsolved for nearly two centuries in their glitz and glory, in their full glory, is finally within grasp. It's coming from a perfect storm, a pairing of Terence Tao's brilliant human roadmap guiding the immense computational power of AI. Together, Terry and his collaborators are doing more than just hunting to solve a 1 million dollar prize. They are decoding the fundamental rules of our universe. Whether they prove that the universe is safe and smooth or that it contains a hidden singularity, the answer will change physics forever. We're about to learn if the code of Stay tuned for more episodes about these vexing problems and what they mean for both physics, mathematics, and even video games. And click here to watch my two interviews with Terence Tao. One walk and talk on the future of science and the other a sit-down interview in his office at UCLA about his work, about all of his explorations in mathematics. Don't forget to like and comment and subscribe.