In 1979, Giorgio Parisi succeeds in exactly solving one of the most resistant problems in the statistical physics of disordered matter: the spin glass model, a system in which tiny magnetic moments — spins — interact with each other through a random mixture of attractive and repulsive forces, generating a phenomenon known as frustration, in which it is impossible to simultaneously satisfy all the system's interactions. Unlike a conventional magnet, which in its minimum-energy state adopts a single ordered configuration, a spin glass becomes trapped in one of an astronomical number of states of similar energy, separated by energy barriers that grow with system size, so that the system never truly reaches thermodynamic equilibrium. Using the mathematical technique known as the "replica trick" — which involves analyzing identical, imaginary copies of the disordered system in order to average over the disorder — Parisi introduces a radically new order parameter: instead of a single numerical value as in conventional phase transitions, he proposes a continuous distribution of values, discovering a hidden hierarchical structure in the system's state space, today known as replica symmetry breaking. Rigorous mathematical validation of Parisi's solution — initially accepted for its extraordinary predictive power before being proved with full rigor — would still take several years to complete formally. What distinguishes this contribution from an isolated technical result in materials physics is its extraordinary capacity for conceptual export: the same mathematical structure Parisi discovered for spin glasses has turned out to be applicable, with adaptations, to the study of neural networks, proteins, combinatorial optimization — including the traveling salesman problem — theoretical ecology, and financial markets, making his solution one of the most transversal mathematical tools in complex-systems physics of the last half century.