Information measured in qubits
Classical information theory measures everything in bits — yes/no, 0/1, the smallest possible answer to a single question. Quantum information keeps that spirit but changes the carrier: instead of a bit, the unit is the qubit. A natural hope is that because a qubit lives in a superposition |psi> = alpha|0> + beta|1>, it must somehow hold *more* than one bit. Here is the honest answer, and it surprises almost everyone: when you read out a single qubit, you get exactly one classical bit. The amplitudes alpha and beta are real and they matter enormously while the qubit is being computed on, but measurement only ever hands you a plain 0 or 1, with probability |alpha|^2 or |beta|^2, and then the state collapses.
So where does the extra richness go? Not into one qubit, but into the relationships between qubits. Two qubits can be entangled — correlated more tightly than any pair of classical objects can be — and it is those correlations, not any single qubit's contents, that quantum information lets you do new things with. A result called Holevo's bound makes this precise: no matter how cleverly you encode, you cannot reliably extract more than one bit of ordinary information from one qubit you receive. The qubit is not a bigger bucket. It is a different *kind* of bucket, and the surprises come from how buckets can be linked.
Teleportation
'Quantum teleportation' is a terrible name — nothing material moves, nothing travels faster than light, and no person is ever beamed anywhere. What teleportation actually does is move the *quantum state* of one qubit to a distant qubit, without sending the qubit itself, by spending one shared entangled pair and two ordinary classical bits. That's the whole trick, and it's genuinely useful inside quantum networks and inside quantum computers.
- Ahead of time, Alice and Bob each take one half of an entangled pair. This is the resource they'll spend; it must be set up in advance.
- Alice has a third qubit in some unknown state |psi> she wants to send. She performs a joint measurement on her unknown qubit together with her half of the pair. This measurement destroys the original |psi> on her side — there is now no copy anywhere, exactly as no-cloning requires.
- Alice's measurement yields two classical bits (one of four outcomes). She sends those two bits to Bob over an ordinary channel — a phone line, fiber, anything classical. This step travels at normal speed, no faster than light.
- Using those two bits as instructions, Bob applies a matching correction (a Pauli gate) to his half of the pair. His qubit is now in the state |psi>. The state has been moved, not copied.
Read the bookkeeping closely, because it's where the honesty lives. The original is destroyed the moment Alice measures — so you never have two copies, and you never violate no-cloning. Bob's qubit stays useless noise until his two classical bits arrive, so no information outraces light. And those two classical bits are not optional flavor; they are mandatory. Entanglement alone moves nothing. You always pay the same toll: one entangled pair plus two classical bits, per qubit teleported.
Superdense coding
Superdense coding is teleportation's mirror image, and putting them side by side is the best way to feel what entanglement is worth. Here the goal flips: Alice wants to send Bob two classical bits, and she does it by physically transmitting just one qubit — provided the two of them already share an entangled pair.
- Ahead of time, Alice and Bob share one entangled pair — again, set up in advance, the resource they'll spend.
- Alice wants to send two classical bits (one of four messages: 00, 01, 10, 11). Depending on which, she applies one of four gates to her single half of the pair.
- Alice sends that one qubit to Bob.
- Bob now holds both halves. He measures them together and reads out all two classical bits Alice intended.
Notice this does not beat Holevo's bound. Alice transmitted one qubit *and* the pair carried a second qubit's worth of channel that was set up earlier — so two qubits were used in total to deliver two bits, which is perfectly ordinary. What's remarkable is the *timing*: the expensive part (distributing entanglement) can happen in advance, during quiet hours, and at send-time Alice moves only a single qubit to convey two bits. Superdense coding doesn't give you something for nothing; it lets you pay in advance with a resource called entanglement.
Toward a quantum internet
Teleportation is not a parlor trick; it's the proposed *backbone* of a future quantum internet — a network whose job is to distribute entanglement between distant nodes so they can then teleport quantum states to each other on demand. The headline application people often reach for is quantum key distribution (for example BB84): using quantum physics so two parties can share a secret key and detect any eavesdropper. Keep one distinction crisp — QKD is a physics-based key-exchange method, which is a different thing from post-quantum cryptography, the classical algorithms we run on ordinary computers to resist future quantum attacks. Both aim at security; only one needs special hardware.
Now the reality check this whole ladder is built on. A quantum internet is very early-stage. The hard part is that qubits can't be amplified or copied the way classical signals are repeated down a fiber — no-cloning forbids it — so you need quantum repeaters, which require storing and purifying entanglement, and those are still largely laboratory and short-link demonstrations, not deployed infrastructure. Entangled-photon links over tens to a few hundred kilometers and satellite experiments are real and impressive, but a general-purpose, always-on quantum internet does not exist yet. Treat anything promising one 'soon' with the same skepticism you'd bring to any frontier technology.
What entanglement buys you
Step back and the pattern is clean. In teleportation you spend one entangled pair plus two classical bits to move one unknown qubit's state. In superdense coding you spend one entangled pair plus one transmitted qubit to deliver two classical bits. The two protocols are duals: each converts between qubits, classical bits, and shared entanglement, and in both, [[entanglement|entanglement]] behaves like a currency — a consumable resource you set up ahead of time and then spend, one pair per use. This is why researchers literally keep ledgers of 'ebits' (units of entanglement) the way an accountant tracks dollars.
And it's why honesty matters more here than almost anywhere in computing. Quantum information does not mean a qubit secretly stores many answers, and it does not mean instantaneous communication. It means a genuinely new resource — entanglement — that, spent carefully and always paired with ordinary classical communication, lets you move and protect information in ways classical bits cannot. That's a smaller claim than the hype, and a far more interesting one, because it's actually true.