Daylila
Quantum computing, without the hype

Lesson 2 of 9

A qubit is not a bit that is both at once

Explain what a qubit's state actually is, and why "both at once" misleads.

01 · Learn · the idea

Two sealed boxes. Both give a clean fifty-fifty, forever. They are opposite states, and one gate tells them apart every time.

How it works

  1. A classical bit is a value: 0 or 1
  2. A qubit's state is a DIRECTION on a globe north = 0, south = 1
  3. Sideways is the case people call “both at once”
  4. But sideways has a WAY ROUND, and opposite ways are different states
  5. A Hadamard swings those opposite points up to the poles now they read 0 and 1 with certainty

In numbers

How often 0 comes up in 100 measurements

Two Hadamards return you to certainty. Real randomness cannot be undone — so the randomness was never in the qubit.

Side by side

A sideways qubit and an actual coin

Qubit on the equator A coin being flipped Measure it once fifty-fifty fifty-fiftyIs there a hidden answer already? no — there is a direction no — there is nothingRepeat the same operation returns to certainty stays random foreverCan two of them be opposites? yes, and one gate proves it no — all fair coins are alike

The common misread

That a qubit in superposition is secretly already 0 or 1 and you just don’t know which. If that were true, applying the same gate twice could not bring back certainty — and it does, every time.

The words

gate
one operation applied to a qubit, like an AND gate in an ordinary chip
Hadamard (H)
the gate that swings a pole to the equator and an equator point back to a pole
superposition
a state pointing somewhere other than the two poles — sideways, with a definite way round

Carry this

The randomness is not in the qubit. It appears when you look, and the whole machine works in the region before you look.

Full lesson
Two boxes you cannot tell apartWhat the state actually isThe move that reveals itWhy "both at once" has to goWhat that leaves you holdingSitting with it

Two boxes you cannot tell apart

Someone puts two sealed boxes in front of you. Each one produces a qubit when you press it. You measure the qubit from box A: you get 0. Press again: 1. Again: 0, 0, 1, 0, 1, 1. After a thousand presses it is a clean fifty-fifty.

Box B does exactly the same. Fifty-fifty, no pattern, no bias.

Every measurement you can make says these boxes are identical. They are not. And the way you tell them apart is the beginning of everything a quantum computer does.

What the state actually is

A classical bit is 0 or 1. There is nothing else to say about it.

A qubit’s state is not a value. It is a direction. Picture a globe. Straight up at the north pole means “definitely 0”. Straight down at the south pole means “definitely 1”. Every other direction on that globe is also a legitimate state of the qubit — and there are infinitely many of them.

A qubit pointing at the equator is the case people describe as “both at once”. That description is doing real damage, because it throws away the only thing that distinguishes one equator point from another: which way round the equator it is pointing.

Box A produces a qubit pointing at one spot on the equator. Box B produces one pointing at the opposite spot. Both are “fifty-fifty” if you measure them top-to-bottom. They are as different as two states can be.

The move that reveals it

There is an operation — a gate, in the same sense as an AND gate in an ordinary chip — called a Hadamard. Written H. What it does is rotate the globe so that the two opposite equator points swing up to the poles.

Apply H to box A’s qubit and then measure. You get 0. Not usually 0. Always 0. Run it ten thousand times and you get 0 ten thousand times.

Apply H to box B’s qubit and measure. You get 1. Always.

The information was there the whole time. Measuring top-to-bottom was simply the wrong question, and asking it a million times would never have helped.

Why “both at once” has to go

Follow the consequence, because it is sharp.

Start with a qubit that is definitely 0. Apply H. It moves to the equator, and if you measure now, it is fifty-fifty. So far the popular description survives.

Now don’t measure. Apply H a second time instead.

If the qubit were genuinely “randomly 0 or 1” after the first H, then a second H would leave it just as random — a coin flipped twice is still a coin. What actually happens is that the qubit lands exactly back on 0, with certainty. Every single time.

  • Definite 0, measured: 0 comes up 100 times in 100
  • After one H, measured: 0 comes up about 50 in 100
  • After two H, measured: 0 comes up 100 in 100

The randomness was never in the qubit. It appeared only when you looked, and it went away again when you stopped looking and kept computing. Something that can be undone was not random.

What that leaves you holding

So the honest picture is this. A qubit carries a direction. Measuring it in one particular way collapses that direction into a single 0 or 1, and if the direction was sideways, which one you get is genuinely unpredictable. But between operations, before you look, there is no coin and no dice. There is a state, it is completely definite, and it moves in ways you can predict exactly.

Everything a quantum computer does happens in that “before you look” region. The whole art is arranging the directions so that when you finally do look, the answer you want is the one pointing at the pole.

Sitting with it

It is worth noticing what went wrong with “both at once”. It was not too simple. It was simple in the wrong direction — it kept the part that sounds strange and threw away the part that does the work.

That failure has a shape you will meet far outside physics. The popular version of a hard idea usually keeps whatever is most quotable, and quietly loses whatever was load-bearing. Being able to notice the difference is most of what it means to understand something rather than to have heard of it.

02 · Try · the lab

03 · Check · quick quiz

1. Box A and box B both give a clean fifty-fifty when measured. How can you tell them apart?

  • Apply a Hadamard gate first, then measure — A gives 0 every time and B gives 1 every time
  • Measure many more times; the difference shows up in the long run
  • You cannot — fifty-fifty is fifty-fifty
  • Measure faster, before the state has time to settle
Answer

Apply a Hadamard gate first, then measure — A gives 0 every time and B gives 1 every time — The two states point at opposite spots on the equator. Measuring top-to-bottom asks the wrong question however often you ask it. The Hadamard swings those spots up to the poles, where the difference is total.

2. A qubit starts as definitely 0. You apply a Hadamard twice and then measure. What do you get?

  • 0, every single time
  • 0 or 1, about fifty-fifty
  • 1, every single time
  • 0 about 75 times in 100
Answer

0, every single time — The second Hadamard rotates the state exactly back to the north pole. If the first one had really made the qubit random, nothing could undo it — a coin flipped twice is still a coin. This is the clearest proof that it did not.

3. What is wrong with saying a qubit in superposition is "0 and 1 at the same time"?

  • It throws away the direction, which is the only thing that distinguishes two fifty-fifty states
  • Nothing — it is a fair simplification of what is really happening
  • It understates it; the qubit is actually in many more than two states
  • It is only wrong once the qubit has been measured
Answer

It throws away the direction, which is the only thing that distinguishes two fifty-fifty states — Two states that both read fifty-fifty can be opposites, and behave completely differently one step later. "Both at once" keeps the part that sounds strange and loses the part that does the work.