The no-cloning theorem

You can't photocopy a qubit

Classical computers copy bits constantly — every backup, every cache, every Ctrl+C. Quantum mechanics forbids the equivalent: no machine, however clever, can take an unknown qubit state α|0⟩ + β|1⟩ and produce two independent copies of it.

The proof is one line of algebra. Quantum gates are linear: how they act on superpositions is fixed by how they act on |0⟩ and |1⟩. A copier would need to send (α|0⟩+β|1⟩) to (α|0⟩+β|1⟩)(α|0⟩+β|1⟩) — but multiplying that out gives cross-terms like αβ|01⟩, while linearity only delivers α|00⟩ + β|11⟩. The two answers disagree whenever α and β are both nonzero. No gate arrangement can square amplitudes, so no cloner exists.

Notice what linearity does produce: α|00⟩ + β|11⟩ is not two copies — it's entanglement! Try to copy a qubit with a CNOT and you entangle it with the 'copy' instead. Measuring either half now disturbs both.

Why teleportation destroys the original

No-cloning explains a rule you may have noticed in the teleportation lesson: the original state is always destroyed in transit. If teleportation left the original behind, you'd have two copies of an unknown state — exactly what the theorem forbids. Quantum information can move, but it can never fork.

It also explains why quantum algorithms can't simply 'save a checkpoint' of a superposition and retry the measurement. One state, one measurement, one outcome. Every protocol has to be designed around this scarcity.

Scarcity as a security feature

For cryptographers, no-cloning is a gift. An eavesdropper who intercepts a qubit in an unknown state cannot copy it, forward the original, and study the copy at leisure — classical wiretapping's standard move is physically impossible.

The best an eavesdropper can do is measure the qubit, which disturbs it in a detectable way. That trade-off — you can't look without leaving fingerprints — is the engine behind quantum key distribution, coming up in the quantum cryptography lesson.