The textbook definition, and why it’s incomplete
You’ve probably heard that entropy is “a measure of disorder,” and that the second law of thermodynamics says entropy always increases — a shuffled deck of cards doesn’t spontaneously sort itself, a broken glass doesn’t reassemble. That’s true, but as Quanta Magazine’s Zack Savitsky lays out, two centuries after the idea was introduced, physicists have arrived at a deeper — and stranger — way of thinking about what entropy actually is.
From steam engines to information
Entropy was introduced roughly 200 years ago by a French engineer trying to understand the limits of steam engines: no matter how well-built an engine is, some energy always ends up as useless, spread-out heat rather than useful work. That “spreading out” is entropy increasing.
The modern reframing
The article explains a shift in how many physicists now think about entropy: not as some fixed, objective property that a physical system simply has, but as a reflection of how much an observer doesn’t know about a system’s exact microscopic state. A box of gas has enormous entropy not because it’s “disordered” in some absolute sense, but because there are astronomically many possible arrangements of its individual molecules that would all look identical to us from the outside — we simply can’t track every molecule.
This reframing connects entropy directly to information theory: entropy becomes a measurement of missing information, not just physical messiness.
Why this matters
This isn’t just a philosophical distinction. It’s actively driving new research into “information engines” (devices that can convert information itself into usable work) and quantum thermodynamics, where the old, purely mechanical picture of entropy stops giving clear answers. Understanding entropy as ignorance rather than just disorder turns out to be the more useful — and more accurate — way to think about one of physics’ most fundamental concepts.