What a qubit is, and what it is not

Quantum computers are often described as machines that try every answer at once. That picture is wrong, and the real one is more useful.

Ask most people what makes a quantum computer powerful and you will hear that it “tries every answer at the same time”. It is a memorable line, and it is the main reason quantum computing is so often misunderstood. If a quantum computer simply checked every answer in parallel, it would solve almost any problem quickly. It does not, and the reason why is the best place to start.

Bits and qubits

A classical bit is either 0 or 1. A qubit can be in a superposition of 0 and 1, described by two numbers called amplitudes. Amplitudes are not probabilities: they can be negative, or even complex, and that is what matters. When we measure a qubit we get a plain 0 or 1, with probabilities that come from those amplitudes, and the superposition is gone.

So a register of qubits can hold a superposition over a very large number of states, but a measurement returns only one of them. If you prepare every possible answer and then measure, you get a random answer. That is no better than guessing.

Where the advantage comes from

Because amplitudes can be positive or negative, they can interfere, just as waves do. A quantum algorithm is a carefully designed sequence of operations that makes the amplitudes of wrong answers cancel out and the amplitudes of right answers reinforce each other. By the time we measure, the right answer is the likely one.

Designing that interference is hard, and it works only for problems with the right kind of structure. That is why the list of problems where quantum computers are expected to give a large speed-up is fairly short: simulating molecules and materials, factoring large numbers with Shor’s algorithm, and certain search and optimisation tasks where the gain is smaller.

Entanglement

Entanglement is a correlation between qubits that has no classical equivalent. Measuring one entangled qubit tells you something about the others, in a way that cannot be explained by each qubit carrying its own hidden value. Entanglement is a resource that quantum algorithms use, but it does not allow faster-than-light communication, and on its own it does not make a computation fast.

Three things worth remembering

A quantum computer does not try every answer and pick the best. It uses interference to make good answers likely.
Quantum speed-ups apply to specific kinds of problems, not to computing in general. Your laptop will not be replaced.
Measurement destroys superposition, so getting useful information out of a quantum computer is part of the design problem, not an afterthought.

Seen this way, quantum computing is less magic and more engineering: a new kind of hardware with its own strengths, limits and costs. That is a better foundation for deciding where it will genuinely help.