Concepts explained
The Quantum Computer — A Switch That Is Neither On Nor Off
This series opened with 'a computer knows two things, on and off' and closed by saying that never changed. A machine that breaks that sentence is now being built. What it means for a bit to move by probability, and what that machine can and cannot do.
A coin spinning in mid-air
Toss a coin. Until it lands and stops in your palm, is the spinning coin heads or tails? Neither. It is not yet decided. There is only some chance of heads, some chance of tails.
A qubit is a switch that can hold this 'mid-air' state — tilted, say, 70% toward on and 30% toward off. This is called .
So put it this way. A bit stores an answer. A qubit stores the possibilities of an answer.
Possibilities that cancel each other
If that were all, a qubit would be a switch with odds attached — no different from a die. The truly strange part comes next.
Flip a coin as often as you like; it stays fifty-fifty. But take a qubit at off, 'half-turn' it to fifty-fifty, then half-turn it once more — it returns to off. Every time. Impossible for a coin.
A quantum computer is a machine that puts this cancelling to work. Spread the candidate answers across qubits as possibilities, choreograph the turns so that the possibilities leading to wrong answers cancel each other and only the right one survives, then look once at the end.
The usual explanation — 'it computes every answer at once, hence fast' — is half right. They are spread out at once, but reading gives you only one, and a naive read gives you any old answer. The real trick is not the spreading but the cancelling, and it is fast only on problems shaped for cancelling.
Tool The Qubit Board — A Switch Neither On Nor Off Four boards in order: tilt a qubit, half-turn it, bind two together, and finally find the treasure among four boxes the quantum way, in a single look. Every press explains what you just saw.Binding two together
Two qubits can be bound in a special way. Once bound, looking at one settles the other — across a room or across a continent. This is . Einstein called it spooky and refused to believe it; half a century later, experiment ruled it real.
Piece two said each extra switch doubles what you can hold. Qubits do the same, with a twist: eight bits hold one of 256 values, while eight bound qubits hold all 256 possibilities at once in a single state. Three hundred of them hold more possibilities than there are atoms in the observable universe. This is where ordinary computers stop being able to imitate.
Where the idea came from
In 1981 a physicist put it this way: nature runs on quantum rules, and an ordinary computer strains to imitate them. To imitate nature you need a machine that runs on nature's own rules. The starting point of the quantum computer.
For a decade or more it remained a thought experiment; nobody could show what it was for. In 1994 a mathematician proved on paper that such a machine could split large numbers into prime factors overwhelmingly fast — and that problem happened to be the pillar of internet cryptography.
A little further in
What it will do, and what it will not
A quantum computer will not replace the ordinary kind. Email, documents, games, video — no gain there. Just as piece six's GPU took only the picture-shaped work, the quantum computer will be a third worker taking only work shaped for cancelling.
drugs, fertiliser catalysts, battery materials — nature is quantum, so quantum imitation fits
it can open today's locks — which is why the world is already changing them
finding the best among vast options — logistics, design, some AI calculations
email, documents, games, video — not shaped for cancelling, so no gain
Why not yet
Qubits are far too delicate. A little heat, a faint vibration, one passing radio wave, and the mid-air state collapses into plain on or off. So qubits live in refrigerators colder than deep space — and still last a fraction of a second.
Piece five's check numbers are needed here too, at a far higher price: keeping one qubit trustworthy is reckoned to cost hundreds to thousands of qubits. Today's machines are roughly the vacuum-tube computers of piece one — filling a room, breaking somewhere almost daily, and yet already running on the real principle.
Back to the series' first sentence. A computer knows two things, on and off. A qubit knows the between. And the moment we look, it returns to one of the two. The knowing is still ours — the ones who read it.
The question that remainsA machine whose answer is many until someone looks — when does its answer become an answer?