A Map of Mathematics · from Euclid to today
🗺️ A map between Euclid and now

Everything between Euclid
and now

A guided tour of mathematics from Euclid to this decade's breakthroughs — thirteen chapters, each with things you can drag, break, and prove. The map below is the table of contents: every field links into its chapter.

Start reading · Chapter 1 →

IAncient & Classical

From Euclid's Elements to Newton's calculus — about two thousand years where geometry came first and "rigor" meant a careful Greek proof.

Classical Number Theory

~500 BCE · classical
Pythagoras · Diophantus · Fermat

What it is. The study of whole numbers and their patterns — which numbers are prime, which divide which, and which equations have a whole-number answer.

What it unlocked. Foreshadows modern cryptography, coding theory, and modular arithmetic. Fermat's "infinite descent" introduced a proof technique still used in computational theory today.

Euclidean Geometry

~300 BCE · classical
Euclid — Elements

What it is. A system of shapes built from points, lines, and angles, where every fact follows logically from just five starting rules.

What it unlocked. All of classical physics and engineering — and the template for proof itself. The way mathematics has been taught for two thousand years.

Trigonometry

~140 BCE → 1600 CE
Hipparchus · Aryabhata · al-Tusi

What it is. The math of angles and ratios in triangles — sine, cosine, tangent — that turns an angle into a length.

What it unlocked. Navigation, astronomy, and everything periodic. The bridge from static geometry to oscillation and waves — and eventually to Fourier analysis.

Classical Algebra

~820 → 1600
al-Khwarizmi · Cardano · François Viète

What it is. The art of manipulating equations as symbols instead of as pictures, with step-by-step procedures for solving them.

What it unlocked. Reasoning about unknowns as objects. Viète's symbolic notation became the language of every later branch of mathematics — and of programming.

Analytic Geometry

1637
René Descartes · Pierre de Fermat

What it is. The moment shapes became equations — every curve lives on an x/y grid and can be written down algebraically.

What it unlocked. Functions and their plots; graphical computing; the calculations Newton and Leibniz needed to invent calculus.

Probability Theory

1654 → 1713
Blaise Pascal · Pierre de Fermat · Jacob Bernoulli

What it is. The math of chance — how to put a number on uncertainty and reason about outcomes when you can't know which one will happen.

What it unlocked. Statistics, decision theory, and (eventually) machine learning. Bernoulli's law of large numbers gave empiricism a mathematical backbone.

Calculus

1670s–1690s
Isaac Newton · Gottfried Leibniz

What it is. The math of change — how fast a quantity is moving at one instant, and how to add up infinitely many tiny pieces to recover the whole.

What it unlocked. All of modern physics (motion, forces, fields), optimization, and integration. Leibniz's notation became the universal language; nothing in engineering works without it.

IIThe 19th Century

The century when math turned inward. Calculus got rigorous foundations, geometry split into many geometries, and a few quiet papers (Boole, Cantor, Galois) planted the seeds of everything that follows.

Real Analysis

1821 · 1821 → 1860
Augustin-Louis Cauchy · Karl Weierstrass

What it is. The careful, line-by-line rebuilding of calculus — pinning down what "limit", "continuous", and "infinite sum" really mean instead of relying on intuition.

What it unlocked. Proof that calculus actually works. Weierstrass's epsilon-delta style became the standard register for all higher mathematics and made numerical computing possible.

Complex Analysis

1825 · 1814 → 1850
Augustin-Louis Cauchy · Bernhard Riemann

What it is. Calculus stretched onto the plane — letting i be a real coordinate instead of an apology, and asking what functions on that plane behave like.

What it unlocked. The math of waves, fields, and signals; conformal mapping; everything in engineering that uses e^(iθ). Some of the deepest results in number theory rest here too.

Differential Geometry

1827 · 1827 → 1854
Carl Friedrich Gauss · Bernhard Riemann

What it is. The geometry of curved surfaces measured from the inside — how to tell a sphere from a saddle without ever stepping off it.

What it unlocked. General relativity. Riemann's generalization to arbitrary dimensions gave Einstein the language for spacetime; today it underlies graphics, robotics, and shape analysis.

Non-Euclidean Geometry

1832 · 1830 → 1870
János Bolyai · Nikolai Lobachevsky · Bernhard Riemann

What it is. What happens when you drop Euclid's parallel-lines rule — entire consistent geometries appear where lines curve, triangles add up wrong, and "straight" needs redefining.

What it unlocked. The realization that geometry is a choice, not a universal truth. Made Einstein's curved spacetime conceivable and opened the door to studying many possible "spaces" at once.

Group Theory

1832 · 1830 → 1880
Évariste Galois · Arthur Cayley

What it is. The math of symmetry — what stays the same when you turn, flip, or shuffle a thing, and how those operations combine.

What it unlocked. All of modern abstract algebra; the answer to why quintic equations have no general formula; the language for gauge theory in physics and for permutation patterns in computer science.

Boolean Algebra & Logic

1854 · 1847 → 1854
George Boole · Augustus De Morgan

What it is. Logic re-cast as algebra — AND, OR, NOT acting on the two values true and false, with all the rules of arithmetic but cleaner.

What it unlocked. Every digital computer. Boolean operations are the hardware-level language of computing; De Morgan's laws remain a daily tool of circuit design and program optimization.

Set Theory

1874 · 1870s–1880s
Georg Cantor · Richard Dedekind

What it is. A careful language for collections — what's inside, what's outside, how big, and how to compare sizes (including infinite ones).

What it unlocked. The common ground beneath every modern branch of math. Cantor showed there are different sizes of infinity; the language is now everywhere from type systems to database semantics.

Topology

1895 · 1736 → 1895
Leonhard Euler · Henri Poincaré

What it is. The study of shape properties that survive any amount of stretching and bending — what counts as "connected", what counts as a "hole", with no rulers allowed.

What it unlocked. Algebraic topology, manifolds, and a whole new way to classify spaces. Poincaré's Analysis Situs set the stage for everything that comes next on this map.

III20th–21st Century

Mathematics becomes its own ecosystem — abstract, layered, and braided with computer science. Most of what powers modern cryptography, machine learning, and physics was invented in this stretch.

Category Theory

1945 · 20th century
Samuel Eilenberg · Saunders Mac Lane

What it is. The math of structure-preserving composition — the same rules work in different settings (algebra, topology, logic, programming), so once you learn the patterns you see them everywhere.

What it unlocked. Functors, natural transformations, and monads — the abstractions that power modern functional programming (Scala, Haskell). Modern algebraic geometry, type theory, and proof assistants all speak categorical language.

Abstract Algebra

1921 · 1920s–30s
Emmy Noether · Bartel van der Waerden

What it is. A move away from "find x" — instead, study the structures (groups, rings, fields) by the rules they obey, regardless of what their elements are.

What it unlocked. Modern ring theory, representation theory, and homological algebra — and almost every typeclass and functor pattern in Scala / Haskell traces back here.

Statistical Inference

1925 · 1920s–30s
Ronald Fisher · Jerzy Neyman · Egon Pearson

What it is. A toolbox for telling signal from noise — hypothesis tests, p-values, and confidence intervals that turn data into defensible claims.

What it unlocked. Experimental design across science, medicine, and A/B testing. Modern machine learning and causal inference build on (and argue with) Fisher's legacy.

Mathematical Logic & Gödel

1931 · 1900 → 1931
David Hilbert · Kurt Gödel · Alfred Tarski

What it is. The proof that any formal system rich enough to do arithmetic must contain true statements it can't prove — and can never prove its own consistency.

What it unlocked. Computability theory and the limits of mechanizable reasoning. Modern proof assistants (Coq, Agda, Lean) inherit the formal-systems machinery this set up.

Functional Analysis

1932 · 1900 → 1930s
David Hilbert · Stefan Banach · Frigyes Riesz

What it is. Linear algebra cranked to infinity — treating whole functions as if they were vectors in a space, with notions of length, angle, and convergence.

What it unlocked. The native language of quantum mechanics, PDEs, signal processing, and modern optimization. Convergence theorems here back every iterative solver you'll meet in numerical code.

Measure Theory & Probability

1933 · 1900 → 1933
Henri Lebesgue · Andrey Kolmogorov

What it is. A rigorous way to assign "size" to weird subsets of the line — and the foundation of probability that came out of doing it.

What it unlocked. Every modern probabilistic algorithm, Monte Carlo simulator, stochastic process, and machine-learning loss function rests on Kolmogorov's 1933 axioms.

Computability

1936 · 1936
Alan Turing · Alonzo Church

What it is. What "computable" actually means — pinned down with mathematical precision a decade before anyone built a digital computer.

What it unlocked. All of theoretical computer science. The lambda calculus that Church proposed is the soul of every functional language you've written (Haskell, Scala, ML, Lisp).

Algebraic Topology

1942 · 1895 → 1940s
Henri Poincaré · Emmy Noether · Solomon Lefschetz

What it is. Using algebra (groups, in particular) to count the holes and connectivity of shapes — turning topology into something you can compute with.

What it unlocked. Homology and cohomology became the workhorse invariants of geometry. Homotopy type theory marries these ideas with computation and gives proof assistants new powers.

Game Theory

1944 · 1944 → 1950s
John von Neumann · Oskar Morgenstern · John Nash

What it is. The math of strategic interaction — what rational players should do when their outcomes depend on each other's choices.

What it unlocked. Modern economics, auction design, distributed-system consensus, and AI alignment. Nash equilibria show up in mechanism design and blockchain protocols.

Information Theory

1948 · 1948
Claude Shannon

What it is. A way to put a number on how much information a message carries — and on how much can be sent through a noisy channel.

What it unlocked. Every ZIP file, every WiFi protocol, every error-correcting code, and every neural-network loss function traces back to Shannon's entropy.

Modern Algebraic Geometry

1960 · 1950s–70s
Alexander Grothendieck · Jean Dieudonné

What it is. A re-foundation of "the geometry of polynomial equations" using schemes and sheaves — geometry rebuilt as deeply abstract algebra.

What it unlocked. Wiles' proof of Fermat's Last Theorem leans on this machinery. Modern parts of it appear in topological data analysis and the highest layers of category theory.

Dynamical Systems & Chaos

1963 · 1890s → 1960s
Henri Poincaré · Edward Lorenz

What it is. The study of how systems evolve in time — and the discovery that perfectly deterministic equations can produce wildly unpredictable behaviour.

What it unlocked. Chaos theory, the limits of weather prediction, nonlinear control. The same ideas explain turbulence, neural-network training dynamics, and the butterfly effect.

Type Theory & HoTT

1972 · 1908 → today
Bertrand Russell · Per Martin-Löf · Vladimir Voevodsky

What it is. A way to organize math (and programs) so paradoxes can't form and proofs can be checked by a machine — types depend on values, values prove propositions.

What it unlocked. Proof assistants (Agda, Idris, Lean) and the entire tradition of typed functional programming. Homotopy type theory marries types with continuous spaces.

Number Theory & Cryptography

1976 · 1976 → 1995
Diffie · Hellman · Rivest · Shamir · Adleman · Wiles

What it is. Encryption built on mathematically hard problems (factoring, discrete logs, elliptic curves) — anyone can lock the box, only the owner can open it.

What it unlocked. Every HTTPS connection, every digital signature, every blockchain wallet. Wiles' proof of Fermat's Last Theorem opened the same theoretical doors that elliptic-curve crypto walks through.

Fractal Geometry

1982 · 1967 → 1982
Benoit Mandelbrot

What it is. Geometry of shapes that repeat themselves at every scale — coastlines, clouds, turbulence — measured by a fractal dimension that needn't be a whole number.

What it unlocked. A language for the rough parts of nature; image compression; procedural graphics; surprising appearances in network topology and finance.

IVThe Frontier

Mathematics since the personal computer: amateurs settle 50-year-old problems, proofs get checked by type-checkers, and the most famous shape of 2023 was found by a hobbyist playing with paper cutouts. This part of the map is still being drawn.

Surreal Numbers & Games

1974 → 1982 · logic
John Conway · Berlekamp · Guy

What it is. Numbers that grow out of two-player games — every position has a value, and the value system holds infinities, infinitesimals, and everything between.

What it unlocked. Combinatorial game theory; for a programmer the whole construction is one recursive datatype — a game is two lists of games.

Modern Knot Theory

1984 → 2020 · geometry
Vaughan Jones · Lisa Piccirillo

What it is. Telling tangled loops apart with computable fingerprints — polynomials and colorings — instead of squinting.

What it unlocked. Invariants tying knots to quantum physics; in 2020 a grad student settled the 50-year Conway-knot question in a week.

Sandpiles & Self-Organized Criticality

1987 · applied
Bak · Tang · Wiesenfeld

What it is. Drop sand grain by grain; tall cells topple onto neighbors, and avalanches of every size organize themselves — no tuning knob.

What it unlocked. A model for earthquakes, blackouts, cascades; the final pile never depends on toppling order — confluence, if you speak rewrite systems.

Expander Graphs

1980s → 2025 · applied
Lubotzky · Phillips · Sarnak

What it is. Sparse networks that behave dense — few wires, no bottlenecks, random walks mix almost instantly.

What it unlocked. Error-correcting codes, derandomization, robust networks; a decades-old bet about them settled in 2025.

Topological Data Analysis

~2000 → today · geometry
Edelsbrunner · Carlsson

What it is. Grow balloons around data points and record when blobs merge and loops are born and die — a barcode of the data's shape.

What it unlocked. Topology as a data-science tool; long bars are signal, short bars are noise.

Compressed Sensing

2004 → 2006 · applied
Candès · Tao · Donoho

What it is. Rebuild a signal from far fewer samples than the textbook minimum — provided it's mostly zeros in some basis.

What it unlocked. Faster MRI, single-pixel cameras; sparsity plus randomness beats the official sampling rate.

Optimal Transport

1781 → 2010s · analysis
Monge · Kantorovich · Villani

What it is. The cheapest way to move a pile of dirt into a set of holes — grown into a geometry where the points are entire probability distributions.

What it unlocked. The Wasserstein distance inside GANs and diffusion models; two Fields Medals and a Nobel en route.

Random Matrix Universality

1955 → 2010s · analysis
Wigner · Tao · Vu

What it is. Fill a matrix with coin flips and its eigenvalues space themselves in one universal pattern — the same as heavy nuclei, Riemann zeta zeros, and the bus schedule in Cuernavaca.

What it unlocked. Universality — wildly different systems, one statistical fingerprint.

Prime Gaps

2013 → 2014 · algebra
Yitang Zhang · James Maynard · Polymath8

What it is. An unknown lecturer proved primes arrive in bounded pairs forever; a public crowd project drove the bound from 70,000,000 to 246.

What it unlocked. First finite bound on gaps; twin primes within shouting distance; proof that breakthroughs still come from nowhere.

Sphere Packing in High Dimensions

2016 · geometry
Maryna Viazovska · Thomas Hales

What it is. The densest way to pack spheres — solved exactly in dimensions 8 and 24, via a "magic function" nobody expected.

What it unlocked. The E8 and Leech lattices; the same geometry runs error-correcting codes; Fields Medal 2022.

Formalized Mathematics

1976 → today · logic
Voevodsky · Buzzard · Tao

What it is. Proofs written as programs a computer type-checks — when it compiles, the theorem is true.

What it unlocked. Lean's mathlib; a Tao paper formalized by a crowd in three weeks; Curry–Howard cashed in at industrial scale.

Chromatic Number of the Plane

1950 → 2018 · geometry
Hadwiger · Nelson · Aubrey de Grey

What it is. Color every point of the plane so any two points exactly distance 1 apart differ — 4, 5, 6, or 7 colors? Nobody knows.

What it unlocked. In 2018 an anti-aging biologist showed 4 isn't enough; the gap is now 5–7 and anyone can play.

ML as Mathematical Collaborator

2021 → today · applied
DeepMind · He · Williamson

What it is. Neural networks as lab partners — trained on math data, they flag patterns humans then prove: a new knot-theory theorem, "murmurations" in elliptic curves.

What it unlocked. Machine-suggested conjectures, IMO-level proof bots, a 56-year-old matrix-multiplication record broken.

Random Structures & Thresholds

1960 → 2022 · applied
Erdős · Kahn · Kalai · Park · Pham

What it is. Grow a random network edge by edge and properties don't creep in — they snap, like water freezing.

What it unlocked. Kahn–Kalai located every snap point; proved in six pages in 2022; phase transitions across CS and physics.

Aperiodic Monotiles

2023 · geometry
David Smith · Kaplan · Myers · Goodman-Strauss

What it is. One 13-sided shape that tiles the entire plane — yet no matter how you lay it, the pattern never repeats.

What it unlocked. A 50-year problem ended by a hobbyist with paper cutouts; "the hat", then the mirror-free "Spectre".

Busy Beaver

1962 → 2024 · logic
Tibor Radó · the bbchallenge collective

What it is. Of all 5-state machines that eventually stop, which runs longest? Proved in 2024 by online amateurs: 47,176,870 steps.

What it unlocked. Undecidability you can touch — BB(6) already outgrows ordinary math; the proof is machine-checked in Coq.

Kakeya Sets

1917 → 2025 · analysis
Besicovitch · Hong Wang · Joshua Zahl

What it is. How little area does a needle need to turn all the way around? Bizarrely — almost none.

What it unlocked. The 3-D version fell in 2025 after a century ("once in a century" proof); load-bearing for half of harmonic analysis.