What is quantum computing? It is computing that uses quantum states — superposition, entanglement, interference — to process information in ways classical bits cannot. The honest framing matters here, because the popular coverage wildly overstates it: a quantum computer is not a faster version of your laptop. It is a different kind of machine that is dramatically better at a narrow set of problems and worse at most others.
Classical Bits vs Qubits
| Classical bit | Qubit | |
|---|---|---|
| States | 0 or 1 | A blend of both until measured |
| Reading it | Gives you the value | Measurement collapses it to 0 or 1 |
| Copying | Trivial | Impossible to copy an unknown state (no-cloning theorem) |
| Scaling | Doubling bits doubles capacity | Doubling qubits squares the state space |
That scaling line is the source of both the power and the hype. Two qubits describe four states at once, three describe eight, and 300 qubits describe more states than there are atoms in the observable universe.
But — and this is the crucial caveat — you cannot simply read all of that. Measuring collapses it to a single outcome. The art of quantum algorithm design is arranging interference so wrong answers cancel and right answers reinforce, so that the final measurement gives you something useful with high probability.
What It Is Genuinely Good At
- Factoring large numbers — Shor’s algorithm breaks RSA encryption in polynomial time. This is the reason governments care, and the reason post-quantum cryptography is now being standardised.
- Simulating quantum systems — the original motivation, from Richard Feynman. Simulating molecules and materials on a classical computer is brutally hard because quantum state spaces explode. A quantum computer handles it naturally. This is the most likely near-term source of real value, especially in drug discovery and materials science.
- Unstructured search — Grover’s algorithm gives a quadratic speedup. Real but modest compared to the exponential one for factoring.
- Certain optimisation problems — an active research area with less certain payoff.
What It Will Never Be Good At
- Running your operating system or browser. There is no speedup for general sequential logic.
- Big data processing. Loading enormous classical datasets into a quantum computer is itself a bottleneck nobody has solved well.
- Training general AI models. No credible theoretical basis for a quantum advantage here.
- Anything requiring reading out large amounts of data. You get one measurement, and it is probabilistic.
The standard framing: quantum computers will not replace classical ones, they will be coprocessors called in for specific subroutines, the way GPUs are called for graphics and matrix work.
Why They Are So Hard to Build
The obstacle is decoherence. Qubits lose their quantum state through the slightest interaction with the environment — a vibration, a temperature fluctuation, a stray electromagnetic field. Once decoherence happens, the computation is garbage.
Consequences:
- Some designs operate at around 15 millikelvin — colder than deep space
- Error rates are far higher than in classical transistors, so enormous overhead goes into error correction
- Most current machines are NISQ devices: noisy, intermediate-scale, and unable to run the algorithms that would actually beat classical machines
The gap worth understanding: a machine with 1,000 physical qubits might yield only a handful of reliable logical qubits after error correction, and running Shor’s algorithm on meaningful numbers would need thousands of logical qubits. That gap is the whole engineering problem.
Where Things Actually Stand
Quantum computers exist and work — they do what the theory predicts at small scale. What has not been demonstrated is a practically useful quantum advantage on a real-world problem that classical machines cannot handle.
Various claims of “quantum supremacy” have been made, and several were subsequently matched or outperformed by improved classical algorithms on conventional hardware. That pattern has repeated more than once, and it is the main reason sober researchers are cautious about timelines.
Serious estimates for cryptographically relevant quantum computing range from a decade to several decades, with genuine disagreement.
What You Should Actually Do About It
That is what is quantum computing likely to mean for you in practice: not much, yet.
If you run systems that store sensitive data with a long secrecy lifetime — health records, state secrets, long-lived financial data — the relevant risk is “harvest now, decrypt later.” An adversary can capture encrypted traffic today and decrypt it once a capable quantum computer exists.
The practical response is to follow the post-quantum cryptography standardisation process and plan migration. For everything else, the honest advice is: this is not something to act on this year, and anyone selling you quantum urgency is usually selling something.
Related Reading
- Quantum entanglement explained — the phenomenon underneath
- What is an algorithm? — what Shor’s and Grover’s actually are
- How does machine learning work? — the thing quantum is not a shortcut for
