Gravity is the curvature of spacetime

Einstein's great insight was not merely that gravity could be described by geometry — it was that gravity is geometry. Spacetime is a 4-dimensional pseudo-Riemannian manifold \((M, g_{\mu\nu})\). Mass-energy curves the manifold according to the Einstein field equations:

Here \(R_{\mu\nu}\) is the Ricci curvature tensor, \(R\) is the scalar curvature, \(g_{\mu\nu}\) is the metric tensor, \(\Lambda\) is the cosmological constant, and \(T_{\mu\nu}\) is the stress-energy tensor. Every term on the left is pure differential geometry (Layer 4). The right side encodes the physical content — how matter and energy source curvature.

Why the mapping works: Physical spacetime has the structure of a smooth manifold. Gravity, uniquely among the forces, is universal — it affects all objects equally (the equivalence principle). This universality means gravity cannot be a force in spacetime; it must be a property of spacetime itself. Differential geometry is the mathematics of curved spaces, so the mapping is not a convenience but a necessity.

Predictive power: The geometric framework predicted gravitational lensing (confirmed 1919), gravitational time dilation (confirmed by GPS satellites, which correct for a \(\sim 38\;\mu\text{s/day}\) drift), frame dragging (confirmed by Gravity Probe B, 2011), and gravitational waves (confirmed by LIGO, 2015).

Quantum theory is a probabilistic theory set in Hilbert space

The mathematical framework of quantum mechanics consists of three pillars, each drawn from a different mathematical layer:

1. States are vectors \(|\psi\rangle\) in a complex Hilbert space \(\mathcal{H}\) (linear algebra, Layer 3).
2. Observables are self-adjoint (Hermitian) operators \(\hat{A}\) on \(\mathcal{H}\), with eigenvalues as possible measurement outcomes (algebra + analysis).
3. Measurement probabilities follow the Born rule: the probability of obtaining eigenvalue \(a\) when measuring \(\hat{A}\) in state \(|\psi\rangle\) is \(P(a) = |\langle a | \psi \rangle|^2\) (probability, Layer 6).

The time evolution of a closed system is governed by a unitary operator:

This is a one-parameter group of unitaries — pure linear algebra and operator theory. The entire framework of quantum mechanics is, in this sense, the theory of linear operators on Hilbert spaces augmented with a probabilistic interpretation.

Why the mapping works: Quantum superposition demands linearity (if \(|\psi_1\rangle\) and \(|\psi_2\rangle\) are valid states, so is \(\alpha|\psi_1\rangle + \beta|\psi_2\rangle\)). The probabilistic nature of measurement demands a framework for assigning probabilities to outcomes. Hilbert space — the natural habitat of both linear algebra and \(L^2\) analysis — provides exactly this.

The security of the internet depends on pure mathematics

Modern public-key cryptography rests on the computational asymmetry of certain mathematical problems:

RSA (1977). Key generation: choose large primes \(p, q\); compute \(n = pq\) and \(\varphi(n) = (p-1)(q-1)\). Choose \(e\) coprime to \(\varphi(n)\); compute \(d \equiv e^{-1} \pmod{\varphi(n)}\). Encryption: \(c \equiv m^e \pmod{n}\). Decryption: \(m \equiv c^d \pmod{n}\). Security relies on the difficulty of factoring \(n\) — a number-theoretic problem with no known polynomial-time classical algorithm.

Elliptic Curve Cryptography (ECC). An elliptic curve over a finite field \(\mathbb{F}_p\) forms an abelian group under a geometric chord-and-tangent addition law. The discrete logarithm problem on \(E(\mathbb{F}_p)\) — given points \(P\) and \(Q = nP\), find \(n\) — is believed to be harder than integer factorization for equivalent key sizes. ECC achieves the same security as RSA with dramatically smaller keys (256-bit ECC \(\approx\) 3072-bit RSA).

Why the mapping works: Cryptography needs one-way functions — operations easy to compute but hard to invert. Number theory and algebra provide a rich supply of such functions because the algebraic structure that makes computation easy (modular arithmetic, group laws) does not make inversion easy. The security guarantee is ultimately a statement about the computational complexity of algebraic problems.

The quantum threat. Shor's algorithm solves integer factorization and the discrete logarithm problem in polynomial time on a quantum computer, threatening both RSA and ECC. Post-quantum cryptography has shifted to problems from lattice theory (algebraic geometry) and coding theory (discrete math) — mathematics again providing the security foundation.

Neural networks are compositions of linear algebra and nonlinear analysis, trained by statistical estimation

A feedforward neural network computes a function \(f: \mathbb{R}^n \to \mathbb{R}^m\) as a composition of layers:

where each \(W_k\) is a weight matrix (linear algebra), each \(\sigma_k\) is a nonlinear activation function (analysis), and each \(b_k\) is a bias vector. Training minimizes a loss function \(\mathcal{L}(\theta)\) via gradient descent:

This is multivariable calculus (Analysis, Layer 5). Backpropagation computes \(\nabla_\theta \mathcal{L}\) via the chain rule — the most consequential application of differential calculus in modern technology.

The layers involved:

Why the mapping works: Learning from data is fundamentally a problem of statistical estimation in high-dimensional space. The function approximation capabilities of neural networks (the universal approximation theorem) come from analysis. The efficient parameterization comes from linear algebra. The training procedure comes from optimization theory. Machine learning is not applied to one branch of mathematics — it sits at the intersection of at least four.