Aliens Would Count. They Probably Wouldn't Write e^x.

Lex Fridman opens with a question that is really about the whole conversation: if intelligent life exists elsewhere, would its mathematics look like ours? Grant Sanderson, the educator behind the YouTube channel 3Blue1Brown, does not hesitate. He thinks it would probably be very different. Notation would obviously differ, and he suspects notation does more than dress the math—it can steer what the math itself becomes.

Some pieces, he says, really do look universal. Other minds would count. Notions like one, two, three—the natural numbers—are "almost why we put that name to it." Once you can count, you have repetition: count by two, two times, and addition and multiplication arrive almost for free. The contested choices start later, when you extend those integers into a continuum. He points to the surreal numbers, a different system that still captures continuity. You could, in principle, model planetary motion with that as the backend and keep a similar frontend of physical laws.

What fascinates him is how little of the menu we have even seen. The alternative number systems mathematicians have already invented may be only a scratch on the surface of what a completely different mode of thought, or a different way of interacting with the universe, would produce. Notation is the difference you would notice first. The deeper divergence, he thinks, would be the mode of thought itself—though he still insists notation carries more cognitive weight than we usually admit.

Asked for a least-favorite piece of notation, Sanderson names the habit of writing the exponential function as e to the x. The symbols imply that you start with a special constant of nature and repeatedly multiply it by itself. That story is fine for integer powers, strained for square roots, and actively misleading once you write e to the power of πi. At that point, he says, you are taking the idea of repeated multiplication, twisting its arm behind its back, and ordering it to apply to complex numbers. He does not think that is pedagogically helpful.

The exponential is better introduced as the function that solves a simple differential equation: rate of change depending linearly on the current value. Population growth and compound interest fit that mold, so repeated multiplication feels natural there. Circular motion is a cousin of the same law, except the operation relating velocity to position is a 90-degree rotation rather than a scaling. Parse e to the πi honestly, he says, and it tells you to pick an origin, always move perpendicular to the vector from that origin, and walk a distance of π times the radius—you end up halfway around. Trying to force that geometry into "multiply a constant by itself" has, in his view, wasted an uncountable number of intelligent hours among scientists, electrical engineers, and students.

So is Euler's identity the most beautiful equation in mathematics? Sanderson thinks the identity is pretty, not mystical, and that the mystery is mostly notational. Historically, e is the constant of self-reinforcing growth; π is geometric; i is the so-called imaginary unit. They get stuffed into one bed even though they would rather sleep apart. A cleaner picture, for him, is a single central function, exp: plug in 1 and you meet e; move in the imaginary direction and the function is periodic with period τ. Those two constants come from orthogonal directions of the same function. The famous equation exaggerates how intimate they are. There is also a small cognitive joke he likes: e is treated as the numerical mascot of calculus, the study of change, which is an odd job for a constant.

The alien question is a doorway to the older one: is math discovered or invented? Sanderson's answer is a cycle. You discover things about the universe that tell you which mathematics will be useful. The mathematics itself is, in a sense, invented—but of all the possible maths you could have invented, discoveries about the world do the filtering. It is not that every civilization would land on the same principles. It is that the world nominates a subset.

The Pythagorean theorem is his working example. Geometric proofs with rearranging squares feel like discoveries, probably because they were built from physical objects. Then look at how we formalize the plane as R², the set of real pairs, and how we define a metric on it. Distance is defined so that the theorem comes out true, and suddenly the sacred feeling thins. Other metrics on the same set of pairs would be mathematically consistent. They just would not describe the physical world we actually study, because they are not the ones in which Pythagoras holds.

So the sequence, as he tells it, is this. A genuine discovery informs the invention of an abstract representation—R² with a particular distance. Studying that abstraction on its own then produces new ideas, inventions, and mysteries. Those mysteries can yield further discoveries, which hint at what else is worth inventing. Math switches roles. At different times it is playing a different part, and the either/or question never quite fits.

Fridman then borrows a Feynman-flavored question: what is the difference between physics and math? Sanderson's first pass is tidy. Mathematics studies abstractions over patterns and pure logic; physics is grounded in a desire to understand the world we live in. He immediately undercuts the tidiness. Talk to different mathematicians and you will hear very different answers, because mathematicians are not one species.

He sketches three temperaments. Some people are turned on by pure puzzles—combinatorics, a toolkit of problem-solving moves aimed at patterns for their own sake. Some are physically motivated: they hunt for new math in veins they already know will matter to physics or computer science. Chaos theory is his example, a body of statements that are purely mathematical and heavily motivated by applications. A third type loves abstraction itself, the feeling of power that comes from a system that can describe as many things as possible. Those are the people who invented topology, later fell into category theory, and can talk for a long time about infinite categories.

Each camp will describe the math–physics border differently. He cites Vladimir Arnold, author of admired books on differential equations, who liked to say that mathematics is a branch of physics—an unsurprising slogan from someone whose subject is PDEs. Category theorists who are not thinking about physics at all will instead celebrate generality, and treat usefulness in the world as a happy coincidence. That coincidence is the surprise hiding under Fridman's next question: why should puzzles and abstraction also describe quarks?

Why, Fridman asks, are the fundamentals of quarks and the rest of reality compressible into equations that are cleaner than they had any right to be—much simpler than they could have been? Sanderson says he does not know, then offers guesses. One is a selection effect. The systems physicists find interesting are the ones simple enough to describe mathematically. Cross a threshold of complexity and the subject gets reclassified as biology or something else. The universe may not be cooperating; the disciplines may just be sorting.

From an information-theory angle, you can always wash away the unimportant parts of whatever you are studying and be left with a low-information core. He also admits he has trouble even conceiving what intrinsically complicated fundamental laws would mean: a tangle of equations you cannot decouple, or a world whose rules are different everywhere, a deeply distributed chaos that no closed formula could capture. We take for granted that the laws are the same in every place. That assumption is doing a lot of work.

An anthropic twist follows. Would a world with incompressible laws even allow beings who ask these questions? Maybe the universe is not so compressible, and evolved brains simply perceive the compressible slice—Chomsky's reminder that we are the descendants of apes, limited biological computers. Sanderson's reply is practical: we can fly spaceships. Over-simplified or not, the models let us twist the world and get the response we predicted. "The fact that we can fly is pretty good." That is a reality check the armchair argument does not get to skip.

Is the universe a simulation? Sanderson finds the idea conceivable. What he does not buy is the popular probability argument: if simulations are possible, simulated worlds will eventually run their own, layers will stack, and a dart thrown at the stack almost surely hits a simulated layer. That story, he says, quietly assumes that information processing has no ceiling.

Physics already suggests a ceiling. The resources required to simulate a universe like ours are enormous just in the number of bits at play, so nested layers cannot proliferate without bound. The real question becomes which is more plausible: that the highest-level universe has unbounded processing capacity, or that capacity is finite and the number of available layers is therefore small. He has no idea how to put a probability distribution on that capacity. People, he thinks, often assign something like a uniform prior over an infinite stack of meta-layers, which is a Pascal's-wager move: the mere existence of that infinite set is not being given a low enough prior.

Fridman enjoys the hypothesis as a thought experiment about computation and AI, a way to ask whether scaling Turing machines far enough reproduces the interesting features of the universe. Sanderson's lasting interest is more specific. General relativity and quantum mechanics together limit how many bits you can store in a given region before it collapses into a black hole. Naive technological optimism says we can always get cleverer and cram more into less space. Physics says no. Information is physical; it has a finiteness. That fact alone, he says, is thought-provoking, whether or not we live on a simulated layer.

Fridman admits he is psychologically uncomfortable with infinity; Sanderson is fairly at ease. How many possible words exist as strings of letters? You do not have to list them. The operative property is that however many you have written, you can throw an "a" on the end and obtain another conceivable word. That "always add one more" is what infiniteness consists in. You never need to hold the whole set in your head at once—and if you tried, Sanderson jokes, your brain would collapse into a black hole.

He redefines abstraction while he is at it. Abstraction is not woo, and it is not ineffable. Every time Fridman moves his face, the analogue of pixels on Sanderson's retina changes completely, yet a coherent notion of "this is Lex" survives. That unification of many different images under one compressed, resilient label is abstraction. Infinity, used that way, is the same kind of object: a description that applies to many situations sharing a common attribute. It should be no more worrying than Lex-ness.

If the word *abstract* is what makes people nervous, he suggests swapping it for *general*. The danger Fridman is circling is then just over-generalization: a concept so general it applies to nothing in particular in a useful way. Sanderson is not claiming we can visualize an infinite bag. Written "1, 2, 3, 4, …" only insinuates the add-one-more property. Concretize infinity into an actual completed collection, and he is as uncomfortable as anyone.

A lot of the motivation for 3Blue1Brown, Sanderson says, is the feeling he gets reading mathematics: this is so abstract I cannot wrap my head around it; I just want something concrete. You can spend an hour on a couple of pages and come out with definitions and a sense of self-defeat. He does not even think of himself as particularly good at math in that sense. Visualization is his way of insisting on a lower rung of the ladder.

Understanding, in his account, does not start from the top. A textbook will define a vector space, list axioms, and promise that everything else follows. That is not how understanding works. You start as low as you can: not a vector space, but a concrete list of numbers, or an arrow on the page with an x-distance and a y-distance. Video (and any textbook that forces itself to include many images) makes that choice unavoidable. The arrow on screen is an actual arrow. It is not a quotation-marked vector that could apply to anything. In the code you have to declare a direction and a magnitude. Every such choice is a concrete example, and every concrete example shortens someone's path.

He catches the same vice in his scripts. The first instinct is always to say the abstract, general, powerful thing and fill in examples later. Every time, the reverse is more compelling. Show a bunch of examples and let the viewer's brain do the pattern recognition; the similarity will already be felt. When the definition or formula finally arrives, it names something already in the head, instead of dumping a high-cognitive-load string of symbols that the audience is supposed to populate later—assuming they are still with you.

What is the most beautiful idea he has come across in mathematics? Often, he says, beauty lives in the things you understand a little and not entirely. Residual mystery is part of what makes them beautiful—almost a personal discovery, a small leap. As a child he had a series of tiny volumes, Wooden Books, with a short description on the left and a picture on the right. On the last page of one of them, in a corner, sat a formula: a sum over all natural numbers of one over n to the s, set equal to a product involving all the primes.

Any young math enthusiast has been told how chaotic the primes are, and they are. On one side of that equation were the counting numbers, as understandable as mathematics gets; on the other, every prime. The feeling was: wait, they are related *like this*? He later learned it was the Euler product for the zeta function, and that the identity essentially encodes the fundamental theorem of arithmetic: every number factors uniquely into primes. Deeper still, it lets you talk about primes not only as themselves but through another set of numbers—the non-trivial zeros of the zeta function. He remains, by his own account, in over his head. The more he learns, the prettier it gets, and enough mystery remains that it stays beautiful.

The path he describes is ordinary and specific: reading, for example the analytic-number-theory article in *The Princeton Companion to Mathematics*, then chasing its tree of references; visualizing whenever possible. His Riemann-zeta video started because he was programming, tried to see what a complex function looked like, and found the picture elucidating. Recurring cousins of the Fourier transform in those texts were a signal to build Fourier intuition even when the books did not name it. The beauty, pressed, is that the patterns are unpredictable and intricate, yet it never feels as if humans made an arbitrary choice in studying this. You feel as if you are speaking to the patterns themselves, or to nature. Arbitrary objects struggle to feel beautiful; that is what the word *contrived* is for. Non-arbitrary means a clean abstraction can get a grip. That is also why he thinks another civilization would be on the zeros of the zeta function, or some clearly equivalent cousin, eventually.

His favorite video to make is *Who Cares About Topology?* It opens on the inscribed-square problem: draw any loop; must four points on it form a square? The question is useless in the physical sense and interesting because we still cannot answer it. Weaken it to rectangles and the problem becomes hard but solvable, by a path that runs through a torus and a Möbius strip. As a child, those surfaces felt like construction-paper toys. Who cares that a twisted band has one side? He thought math was about solving problems. The argument in a paper by a mathematician named Vaughan lets those shapes arise with a job to do, in a way that does not feel contrived.

The topology is abstract and non-constructive, so putting it on screen meant extra work. He used Grapher, built into macOS, and could not anticipate the surface. Once drawn, it looked like the Sydney Opera House, or a Frank Gehry design: it meets the plane curve, has that curve as boundary, sits above the plane—and still the actual shape was a surprise. The forcing function of needing an image added something the original paper did not show. That, for him, is the satisfaction of the medium.

The painful part of the job is a topic he wants that will not yield a script. One such idea had been stuck for two or three months. The right move, he thinks, is to set it aside. You do not want to overwork a narrative, because the more you have thought about it, the less you can empathize with the student who does not yet understand. The judge saying "this sucks" is his past self—the only person he really knows. That method has a cost. His quaternions video answered *his* question: what are they doing in four dimensions, can we see it? Robotics engineers and graphics programmers usually want 3D rotations instead. He would not recommend that video to them, unless they later ask in what sense the objects act in four dimensions. Neural networks, similarly, can be filmed from physics, neuroscience, robotics, or statistics; he has already made three videos on Euler's formula and plans at least one more. By now he understands Euler's identity well enough to see the version that just woke up, hair messy, bags under its eyes. The zeta function is still in the high-dopamine dating stage.

Would he live forever if he could? He thinks so. Fridman argues that finiteness is what makes life meaningful: this room closes in forty minutes, and that is why he loves it more. Sanderson is not convinced that mortality is doing that work in his own days. Only a small percentage of his time is spent with death salient. Most of the time he is still motivated to make things, to be with people, to experience love, and that motivation shows up while death is not in his mind at all.

Maybe that is youth. Maybe the thought is subconscious. Maybe, as Fridman needles, he has built an illusion that lets him escape mortality by enjoying the moment. Gun to his head, Sanderson does not think that is it. Fridman keeps pushing, in the melancholy register he blames on a bit of Russian background, and suggests math and science are a beautiful escape into something seemingly infinite. Sanderson's dry reply is the line he jokes would make a good Twitter bio: "Just because you're trolling me doesn't mean I'm wrong." On a more mathematical note, he had already said that as long as you do not live forever, there will always be enough mystery. Questions appear faster than answers. Even immortality might not drain the subject—though Fridman would still want an ending for the plot.

Knowing when a video is finished is, for Sanderson, the easy part. He writes a script around an aha moment, puts visuals to each sentence, narrates, and edits. Once the script exists, the endpoint is clear, even if wording and occasional structure change in animation. Knowing when the *script* is done is harder. Problem-solving videos end when someone who did not understand the solution now could. Neural networks and similar subjects have too many angles; the standard becomes: this asks a meaningful question and answers it.

For beginners, and for people mid-career who want math back, his advice is blunt. Your temptation will be to spend more time watching lectures or reading. Force yourself to do more problems than you naturally would. Focused time should go to specific problems, preferably from well-curated lists. Flip to the exercises at the end of a chapter before you read it. Most will not make sense yet; the few that do are the best ones to think about. Glance, read a little, glance again, and do not consider the chapter finished until you have actually worked through a couple of exercises. He admits the advice is hypocritical. His own videos almost never come with exercises. He sees himself as a different part of the ecosystem, which is also an admission that watching is only a partial education.

Absolute beginners, he says, can do worse than Khan Academy, which has a large question bank from basic algebra and calculus up. Another path is to learn to program and let mathematical need pull you forward. He knows many people who disliked math until programming turned them on to it; living in the Bay Area may bias the sample, but he is willing to bet the path generalizes. Teaching, even more so. He recites a probably apocryphal ladder—remember about 10 percent of what you read, 20 percent of what you hear, 70 percent of what you actively interact with, 90 percent of what you teach—and says the numbers may be meaningless while still ringing true. He is willing to claim he learns nine times better by teaching than by reading, and that might be a lowball. Do something that explains the idea. That is how knowledge consolidates.

Asked for a happy moment outside family, he remembers music, which used to occupy more of his life. After a gig at a ski resort, he, his brother, and a friend jammed on the gondola ride down; the car came over a ridge, city lights appeared, and the feeling was simply happy—not transformative, not even mainly about the people, just the doing. Fridman, also a musician, wants music back in Sanderson's creative work. Sanderson half-promises a music video by the end of 2020, violin as his main instrument, guitar and piano as side vices. Fridman closes with Paul Lockhart's line from *A Mathematician's Lament*: the first thing to understand is that mathematics is an art, and the difference from music or painting is that our culture does not recognize it as such. He thanks Sanderson for showing that art to millions. Sanderson says thank you for saying it.