Two coins, one in your pocket and one carried to the Moon. Some pairs of quantum particles behave as if flipping one instantly fixes how the other lands, with no signal passing between them. That is entanglementA quantum link where two qubits' states become tied together, so acting on or measuring one affects the other. It is a key resource for quantum computing., and for thirty years the best physicists alive argued over whether it could possibly be real.
Measure one and you instantly know something about the other, however far apart they sit.
It is tempting to call this communication. It is not. You cannot use entanglement to send a message faster than light, because each particle on its own looks completely random, like a fair coin. The correlation only appears when the two sets of results are brought together and compared, and that comparison travels at ordinary, sub-light speed.1 Entanglement bends our sense of distance without ever breaking the speed limit.
Figure 1: A Bell test. One source sends two entangled particles to detectors measured far apart; their outcomes stay correlated more tightly than any classical link permits, and no signal passes between them. That violation is how we know the entanglement is real.
Einstein's objection, and Bell's test
Einstein hated this. In 1935 he and two colleagues argued that such spooky action at a distance had to mean quantum mechanics was incomplete, that each particle must secretly carry predetermined answers, what we now call local hidden variablesThe idea that each particle secretly carries predetermined answers before measurement, so nothing spooky is going on. Bell tests have ruled this out..2 For decades that looked like a question for philosophers, not something a laboratory could ever settle.
Then John Bell found the crack. He proved that if local hidden variables were real, the correlations between distant measurementsReading a quantum system, which forces it out of its blend of possibilities into a single definite 0 or 1 and ends its quantum behaviour. would have to stay under a strict numerical limit, a Bell inequalityA numerical limit on the correlations any 'local hidden variable' theory can produce. Quantum entanglement breaks it, which is how we know the link is real., that quantum mechanics openly predicts they break.3 The argument was suddenly decidable. When the experiments were finally done carefully enough to close every loophole, with the two particles measured 1.3 kilometres apart so nothing could quietly pass between them, the quantum prediction won, and the picture of particles pre-loaded with answers lost.4 Einstein was wrong about this one.
Fragile, and the engine of the machine
EntanglementA quantum link where two qubits' states become tied together, so acting on or measuring one affects the other. It is a key resource for quantum computing. is also delicate. The moment an entangled system brushes against its surroundings the link bleeds out through decoherenceThe loss of a qubit's quantum information as it interacts with the outside world. The main reason qubits are fragile., and it can disappear completely in a finite time, a collapseThe instant a measurement forces a qubit from a blend of possibilities into one definite classical value. abrupt enough that physicists call it the sudden death of entanglement.1 Holding particles entangled long enough to use them is one of the central engineering fights of the field.
Strip entanglement away, and a quantum computer is barely more than a classical one.
Horodecki, R., Horodecki, P., Horodecki, M. & Horodecki, K. “Quantum Entanglement.” Reviews of Modern Physics81, 865–942 (2009). DOI: 10.1103/RevModPhys.81.865. Preprint: arXiv:quant-ph/0702225.
Einstein, A., Podolsky, B. & Rosen, N. “Can Quantum-Mechanical Description of Physical Reality Be Considered Complete?” Physical Review47, 777–780 (1935). DOI: 10.1103/PhysRev.47.777.
Hensen, B. et al. “Loophole-Free Bell Inequality Violation Using Electron Spins Separated by 1.3 Kilometres.” Nature526, 682–686 (2015). DOI: 10.1038/nature15759. Preprint: arXiv:1508.05949.