In 1865, the German physicist Rudolf Clausius coined a new word — entropy — and used it to formulate one of the most consequential statements in science: the entropy of an isolated system never decreases. Heat flows from hot to cold; ordered states decay into disordered ones; the universe, taken as a whole, is running down toward uniform tepid equilibrium. The Second Law of Thermodynamics was the first physical law to single out a direction of time, and it has resisted every attempt to wriggle out of its implications.
The microscopic explanation came from Boltzmann in the 1870s. Entropy is the logarithm of the number of microstates compatible with a given macrostate — the more ways a system can be arranged while looking the same from outside, the higher its entropy. Order is statistically rare; disorder is statistically generic. The Second Law says systems drift from rare configurations to common ones, and the drift is overwhelmingly probable rather than logically necessary — but for systems with billions of particles, overwhelmingly probable is effectively certain. Life is a local entropy decrease, paid for by a much larger entropy increase elsewhere (the Sun pours out radiation; we organize a small fraction of it into ourselves). The heat death of the universe — Clausius's term — is the eventual end state in which all energy gradients have been spent and nothing more can happen. Whether this is the actual cosmological future is a question modern physics has not closed: dark energy, accelerating expansion, and unresolved questions about quantum gravity all complicate the picture.
Information theory (Shannon, 1948) showed that entropy is also a measure of information — the deep connection between thermodynamic disorder and information uncertainty is one of the surprising convergences of twentieth-century science. The Second Law is invoked in arguments about biological complexity, the arrow of time, the limits of computation (Landauer's principle), and climate change (the planet's energy balance is a thermodynamic system). It is one of the few physical laws that no working physicist expects to be revised.