3Blue1Brown — Channel Summaries
AI-powered summaries of 150 videos about 3Blue1Brown.
150 summaries
But what is a neural network? | Deep learning chapter 1
Handwritten-digit recognition becomes feasible once a neural network is treated as a layered math machine: each “neuron” computes a weighted sum of...
But what is a Fourier series? From heat flow to drawing with circles | DE4
Fourier series turn a messy, real-world initial condition—like a discontinuous step in temperature—into a controlled sum of simple, rotating...
But how does bitcoin actually work?
Bitcoin’s core trick is turning money into a shared, tamper-resistant ledger—so transfers don’t rely on a bank’s permission. The system works by...
The hardest problem on the hardest test
The probability that the center of a sphere lies inside the tetrahedron formed by four random points on its surface turns out to be exactly 1/4—a...
The most unexpected answer to a counting puzzle
A counting puzzle about two frictionless, perfectly elastic sliding blocks turns into an unexpected appearance of pi: when the incoming block’s mass...
But what is the Fourier Transform? A visual introduction.
Fourier analysis is built around one practical question: given a signal that’s messy in time—like the air-pressure trace from a sound—how can it be...
Solving Wordle using information theory
Wordle can be treated as a problem in information theory: each color pattern (green/yellow/gray) functions like a noisy “measurement” that reduces...
Vectors | Chapter 1, Essence of linear algebra
Linear algebra’s foundation is the vector—understood in three closely related ways—and the two operations that make vectors useful: adding vectors...
The essence of calculus
Calculus can be “invented” from a single geometric question: why a circle’s area equals πr². Starting with a circle of radius 3, the approach slices...
Transformers, the tech behind LLMs | Deep Learning Chapter 5
Transformer-based models—behind systems like ChatGPT—turn text into a stream of vectors, mix information across tokens with attention, and then...
Gradient descent, how neural networks learn | Deep Learning Chapter 2
Gradient descent is the engine behind neural-network learning: it repeatedly nudges thousands of adjustable weights and biases to reduce a single...
But why is a sphere's surface area four times its shadow?
A sphere’s surface area comes out to 4πR² for a reason that can be felt geometrically: when surface patches are “projected” onto a related flat...
Why do prime numbers make these spirals? | Dirichlet’s theorem and pi approximations
Plotting points (p, p) in polar coordinates—using radius r = p and angle θ = p radians—creates outward Archimedean spirals. When all integers are...
Why is pi here? And why is it squared? A geometric answer to the Basel problem
A classic infinite series—adding the reciprocals of the squares of integers—ends up equal to a multiple of π², and the surprising part is not just...
Linear combinations, span, and basis vectors | Chapter 2, Essence of linear algebra
Linear combinations turn two (or more) vectors into a whole geometric “shape” of reachable results—and that shape is the span. In the 2D coordinate...
Linear transformations and matrices | Chapter 3, Essence of linear algebra
Linear transformations in two dimensions are completely determined by where they send the two basis vectors, and matrices are just a compact way to...
Oh, wait, actually the best Wordle opener is not “crane”…
A subtle bug in the Wordle-simulation code changed which opening word comes out “optimal,” overturning the earlier claim that “crane” is the best...
How are holograms possible?
Holograms work because a flat recording can store the full “light field” around a scene—not just brightness from one viewpoint—by encoding both the...
Exponential growth and epidemics
Exponential growth in epidemics isn’t just a curve that looks steep—it’s a process where the number of new cases each day is proportional to the...
Eigenvectors and eigenvalues | Chapter 14, Essence of linear algebra
Eigenvectors are the vectors that stay on their own span under a linear transformation—meaning the transformation only stretches or squishes them,...
Backpropagation, intuitively | Deep Learning Chapter 3
Backpropagation is the mechanism that turns a network’s prediction error into specific, proportionate changes to every weight and bias—so the cost...
Large Language Models explained briefly
Large language models power chatbots by learning to predict the next word in a sequence—turning that prediction into fluent, context-aware responses....
Bayes theorem, the geometry of changing beliefs
Bayes’ theorem is presented as a disciplined way to update beliefs when new evidence arrives—without letting that evidence “decide” everything from...
The unexpectedly hard windmill question (2011 IMO, Q2)
A single, carefully chosen starting line can drive a “windmill” rotation that repeatedly uses every point of a finite planar set as the...
Differential equations, a tourist's guide | DE1
Differential equations are the language for describing change—when it’s easier to model how a system evolves than to pin down its exact state at...
Why colliding blocks compute pi
A pair of idealized, frictionless blocks can be tuned—by choosing a mass ratio—to produce a collision count whose digits match those of π, even...
Visualizing the 4d numbers Quaternions
Quaternions are a four-dimensional number system whose multiplication can be visualized as a pair of synchronized 90-degree rotations on a...
Simulating an epidemic
Epidemic control in these simulations hinges less on dramatic, late interventions and more on catching infectious people early and reliably. In an...
But what is the Riemann zeta function? Visualizing analytic continuation
The Riemann zeta function becomes understandable once its “analytic continuation” is treated as a rigid, geometry-driven extension: start with a...
Divergence and curl: The language of Maxwell's equations, fluid flow, and more
Divergence and curl turn messy vector fields into two crisp “local” diagnostics: divergence measures whether nearby flow behaves like a source or...
What's so special about Euler's number e? | Chapter 5, Essence of calculus
Exponentials are special in calculus because their derivatives are proportional to the functions themselves—and the constant of proportionality is...
Taylor series | Chapter 11, Essence of calculus
Taylor series turn local derivative information at a single point into accurate polynomial approximations nearby—often so accurate that, when enough...
Terence Tao on the cosmic distance ladder
Humanity’s first “cosmic distance ladder” wasn’t built with rockets or lasers—it was built with geometry, shadows, and timing. The central...
The determinant | Chapter 6, Essence of linear algebra
Determinants turn the messy question of “how much does a linear transformation stretch space?” into a single number: the factor by which areas (in...
Researchers thought this was a bug (Borwein integrals)
A family of integrals built from the “engineer sinc” function— \(\mathrm{sinc}(x)=\frac{\sin(\pi x)}{\pi x}\)—keeps landing exactly on \(\pi\) for a...
What does it feel like to invent math?
A geometric-looking “nonsense” identity—1 + 2 + 4 + 8 + … = −1—can be made meaningful once mathematicians redefine what “distance” and “infinite sum”...
Fractals are typically not self-similar
Fractals aren’t defined by perfect self-similarity. The more useful idea is that many rough shapes behave as if they have a non-integer “fractal...
But what is the Central Limit Theorem?
A single, chaotic process can be unpredictable ball-by-ball, yet the totals across many repetitions settle into a remarkably stable pattern: the bell...
The paradox of the derivative | Chapter 2, Essence of calculus
Calculus’s derivative isn’t a literal “instantaneous rate of change”—that phrase collapses under scrutiny because real change requires comparing two...
Matrix multiplication as composition | Chapter 4, Essence of linear algebra
Matrix multiplication isn’t just a computational trick—it’s a compact way to represent composing linear transformations. A linear transformation is...
How to lie using visual proofs
A run of “visual proofs” goes spectacularly wrong in three different ways—showing that convincing pictures can hide fatal assumptions about geometry,...
All possible pythagorean triples, visualized
Pythagorean triples—integer side lengths (a, b, c) satisfying a² + b² = c²—can be generated and visualized in a single, surprisingly structured way:...
Attention in transformers, step-by-step | Deep Learning Chapter 6
Attention in transformers is the mechanism that lets each token’s embedding absorb information from other tokens—turning context-free word vectors...
The other way to visualize derivatives | Chapter 12, Essence of calculus
Calculus intuition often gets trapped in graphs—slopes for derivatives, areas for integrals—but that graph-first mindset can make later topics feel...
Backpropagation calculus | Deep Learning Chapter 4
Backpropagation’s calculus boils down to one practical question: how much does the cost change when a single weight or bias nudges a network’s...
Inverse matrices, column space and null space | Chapter 7, Essence of linear algebra
Linear algebra’s payoff is practical: many real problems reduce to solving linear systems, and the geometry of a matrix determines whether solutions...
Group theory, abstraction, and the 196,883-dimensional monster
The monster group’s defining “size” is so specific—tied to a 196,883-dimensional structure—that it feels less like a random curiosity and more like a...
How secure is 256 bit security?
Breaking 256-bit cryptography boils down to an almost unimaginably unlikely guessing game: if an attacker must hit one specific 256-bit...
But what is a convolution?
Convolution is the mathematical “mixing” operation that turns two lists (or two functions) into a new list by multiplying aligned pairs and summing...
Derivative formulas through geometry | Chapter 3, Essence of calculus
Calculating derivatives stops being a memorization exercise when each rule is tied to a single geometric idea: a derivative measures how a quantity...
A tale of two problem solvers | Average cube shadow area
The average shadow of a cube—when light comes from directly above and the cube is tossed into every possible orientation—turns out to depend only on...
How (and why) to raise e to the power of a matrix | DE6
Matrix exponentiation—written as e^(At)—turns out to be a precise way to solve systems of differential equations where a state changes at a rate...
e^(iπ) in 3.14 minutes, using dynamics | DE5
The core insight is that the exponential function is uniquely characterized by the rule “rate of change equals the current value,” and swapping the...
Why “probability of 0” does not mean “impossible” | Probabilities of probabilities, part 2
Assigning a nonzero probability to every exact real value of an unknown parameter leads to a paradox: there are uncountably many candidate values, so...
Who cares about topology? (Old version)
The core breakthrough is a topological “collision” argument: for any closed loop in space, there must exist two distinct pairs of points that share...
Why this puzzle is impossible
The puzzle of connecting three utilities (gas, power, water) to three houses with nine non-crossing lines turns out to be impossible on a flat...
Dot products and duality | Chapter 9, Essence of linear algebra
Dot products don’t just measure “how much two vectors point together”—they secretly encode a linear transformation. That deeper link, revealed...
Thinking outside the 10-dimensional box
Higher-dimensional geometry stops behaving like “bigger 2D/3D,” and one of the clearest ways to see why is to track how the unit constraint on a...
But what is a partial differential equation? | DE2
The heat equation turns the everyday idea of heat flowing from warm to cool into a precise rule for how an entire temperature profile evolves over...
Newton’s fractal (which Newton knew nothing about)
Newton’s method turns a simple root-finding rule into an endlessly intricate fractal when it’s run over the complex plane. Starting from a seed...
How I animate 3Blue1Brown | A Manim demo with Ben Sparks
Manim—3Blue1Brown’s custom Python animation library—turns mathematical ideas into smooth, controllable visuals through a workflow that blends...
Integration and the fundamental theorem of calculus | Chapter 8, Essence of calculus
Integration is the inverse of differentiation in a precise sense: the accumulated area under a velocity curve produces a distance function whose...
But what are Hamming codes? The origin of error correction
Scratches, noise, and transmission glitches can flip 1s and 0s—yet many storage and communication systems still recover the original data exactly....
This pattern breaks, but for a good reason | Moser's circle problem
Moser’s circle problem starts with a tempting pattern: draw n points on a circle and connect every pair with a chord, then count how many regions the...
Pi hiding in prime regularities
A hidden arithmetic regularity—how primes split inside the Gaussian integers—turns a messy lattice-point counting problem into a clean alternating...
e to the pi i, a nontraditional take (old version)
The equation e^(πi) = −1 stops looking like black magic once exponentials are redefined as a bridge between two kinds of actions on numbers: sliding...
But why would light "slow down"? | Visualizing Feynman's lecture on the refractive index
Light bends in a prism because different colors drive different microscopic oscillations inside the glass, and those oscillations shift the wave’s...
Essence of linear algebra preview
Linear algebra often gets taught as a toolbox of computations—matrix multiplication, determinants, eigenvalues—without the geometric meaning that...
Euler's formula with introductory group theory
Euler’s formula, e^(πi) = −1, becomes far more than a numerical coincidence once exponentials are reinterpreted as a bridge between two kinds of...
The Hairy Ball Theorem
A continuous “comb-down” of a sphere’s directions is mathematically impossible: any continuous tangent vector field on a sphere must hit at least one...
Binomial distributions | Probabilities of probabilities, part 1
Online ratings tempt buyers to treat “% positive” as a direct measure of quality, but the number of reviews changes what that percentage really...
Three-dimensional linear transformations | Chapter 5, Essence of linear algebra
Linear transformations in three dimensions are fully determined by where they send the three standard basis vectors—so a 3D “grid-squishing” process...
Limits, L'Hôpital's rule, and epsilon delta definitions | Chapter 7, Essence of calculus
Limits sit at the center of calculus not as a new intuition, but as the rigorous language that makes “approach” precise—especially when derivatives...
Change of basis | Chapter 13, Essence of linear algebra
Coordinate systems aren’t just bookkeeping—they encode the geometry of space. In the standard setup, a vector like (3, 2) is interpreted as “3 units...
Hilbert's Curve: Is infinite math useful?
Hilbert’s curve earns its keep by solving a practical mapping problem: turning a 2D image grid into a 1D sequence of frequencies in a way that stays...
Why slicing a cone gives an ellipse (beautiful proof)
Slicing a cone at the right angle produces an ellipse—and the surprising part is that this “conic section” curve matches exactly the ellipse drawn by...
Cross products | Chapter 10, Essence of linear algebra
Cross products turn the geometry of a parallelogram into an algebraic object: in 2D, they produce a signed area, and in 3D, they produce a...
Olympiad level counting (Generating functions)
A counting problem about subsets whose element-sums are divisible by 5 turns into a clean formula once the subsets are encoded as coefficients of a...
The more general uncertainty principle, regarding Fourier transforms
Heisenberg’s uncertainty principle isn’t a one-off quantum oddity so much as a specific instance of a broader Fourier trade-off: signals that are...
Implicit differentiation, what's going on here? | Chapter 6, Essence of calculus
A calculus “weirdness” becomes manageable once tiny changes in two variables are given a geometric meaning: implicit differentiation is really about...
Visualizing the chain rule and product rule | Chapter 4, Essence of calculus
Derivatives of complicated expressions don’t come from memorizing formulas—they come from tracking how tiny input “nudges” propagate through three...
This open problem taught me what topology is
The core breakthrough is a topology-driven route to a classic geometric claim: every closed continuous loop in the plane contains a non-degenerate...
Why π is in the normal distribution (beyond integral tricks)
Pi’s appearance in the normal distribution isn’t a coincidence of algebra—it comes from geometry and from the way Gaussian shapes are forced by...
The impossible chessboard puzzle
A prisoner-style chessboard puzzle turns into a sharp impossibility result: if the board size (the number of squares) is not a power of two, no...
Some light quantum mechanics (with minutephysics)
Quantum mechanics’ most counterintuitive feature—probabilities replacing classical “splits” of energy—can be built from the ordinary wave physics of...
Complex number fundamentals | Ep. 3 Lockdown live math
Complex numbers become intuitive once they’re treated as a two-dimensional number system where multiplying by i performs a 90-degree rotation. That...
Trigonometry fundamentals | Ep. 2 Lockdown live math
Trigonometry’s “simple” graphs hide identities that are anything but obvious—especially once cosine is squared. By starting with nothing more than...
How might LLMs store facts | Deep Learning Chapter 7
Large language models don’t just “know” facts in a vague sense—those facts can be traced to specific internal computations, especially inside the...
Nonsquare matrices as transformations between dimensions | Chapter 8, Essence of linear algebra
Non-square matrices aren’t a special case—they’re the standard way to encode linear transformations between spaces of different dimensions. A...
How pi was almost 6.283185...
The commonly taught “pi” constant (3.1415…) became the default largely because of an 18th-century calculus textbook that spread a particular notation...
How wiggling charges give rise to light
Sugar water twists the polarization of linearly polarized light because its chiral molecules treat left- and right-handed circular polarization...
But how do AI images and videos actually work? | Guest video by Welch Labs
Text-to-image and text-to-video systems work because diffusion models can be understood as reversing a physics-like random process—then steering that...
Beyond the Mandelbrot set, an intro to holomorphic dynamics
Holomorphic dynamics turns the Mandelbrot set from a one-off curiosity into a recurring pattern: iterating complex-analytic functions produces stable...
What makes the natural log "natural"? | Ep. 7 Lockdown live math
Prime numbers turn out to be far less rare near a trillion than most people guess—and that “surprise frequency” is tightly linked to the natural...
Abstract vector spaces | Chapter 16, Essence of linear algebra
Linear algebra’s core move is to treat “vectors” as anything that supports two operations—addition and scaling—so long as they obey a fixed set of...
Five puzzles for thinking outside the box
A chain of geometry puzzles turns on one recurring insight: stepping into a higher dimension can make stubborn 2D questions tractable—and even when...
What is Euler's formula actually saying? | Ep. 4 Lockdown live math
Euler’s formula stops being a mysterious “imaginary exponent” once the exponential function is treated as a specific power series (exp), not as...
Music And Measure Theory
A ratio of musical frequencies can sound harmonious or cacophonous depending less on whether it is rational or irrational, and more on how well it...
What does area have to do with slope? | Chapter 9, Essence of calculus
Finding the average value of a continuous function turns out to be the same kind of calculation as measuring the slope of an antiderivative across an...
Solving the heat equation | DE3
The heat equation’s solutions aren’t determined by the differential equation alone: the temperature profile must satisfy the PDE in the rod’s...