Episode 34 - Quantum at Work

Tomlin Hawthorne

You have a phone within reach. Inside it is a chip, about the size of a fingernail, that does the thinking.

Tomlin Hawthorne

That chip is made of switches. Tiny electrical switches, each of which can be on or off.

Tomlin Hawthorne

How many?

Tomlin Hawthorne

A thousand? A million?

Tomlin Hawthorne

If it's a recent phone, somewhere between fifteen and twenty billion.

Tomlin Hawthorne

That's more than twice the number of people on Earth, on a fingernail. Every one of them was placed deliberately, and every one of them works. Each is so small that a few thousand would fit across the width of a hair. It's the most intricate object human beings have ever made, by a very wide margin, and we make them by the hundred million and give them to children.

Tomlin Hawthorne

Not one of those switches would work if quantum mechanics were false. For the last seven episodes I've been telling you how strange this theory is. Today I'd like to show you that it's also the most practical thing in the series.

Tomlin Hawthorne

Last time you looked down the left-hand column of the periodic table and predicted what caesium does in water. If you then found the film, you saw a small lump of metal dropped into a bath and the bath come apart. You read that off a position on a chart. That's what Mendeleev did, and it's what Pauli explained: electrons filling the atom row by row, and one lonely electron on the outside wanting out.

Tomlin Hawthorne

This is the last episode of Part Six. The question's in two halves. Where is quantum mechanics in everyday life? And what is this second quantum revolution we keep hearing about?

Tomlin Hawthorne

Murray Hill, New Jersey. The sixteenth of December, 1947. It's the main research laboratory of the American telephone company, Bell Labs. We've been here before. It's where Shannon rode his unicycle, where Davisson and Germer baked their nickel, and where, seventeen years after this, two engineers will find a hiss in a radio antenna that turns out to be the afterglow of the creation of the universe. I don't think any other building has a record like it.

Tomlin Hawthorne

Jon Gertner wrote a fine history of the place, called The Idea Factory. And if you're interested in how organisations produce new things, it's worth your time. Bell Labs was the research arm of a legal monopoly. That meant patient money, and no need to show a profit this quarter. The man who ran it believed in putting theorists, experimenters and engineers in the same corridors and making them eat lunch together. He designed the building with very long hallways, so that you couldn't get to the canteen without walking past a dozen open doors.

Tomlin Hawthorne

The telephone company had a problem. Its whole network ran on two components. Mechanical switches, which wore out. And vacuum tubes, which are like light bulbs: they're hot, they're fragile, they burn out, and they use a great deal of power. The company wanted something better. Something solid, with no moving parts and no glowing filament.

Tomlin Hawthorne

On that December afternoon, two physicists, Walter Brattain and John Bardeen, had a contraption on the bench. A small slab of a grey element called germanium. Pressed on to it, a plastic wedge with a strip of gold foil wrapped round the tip, slit with a razor blade to make two contacts a hair's breadth apart. The whole thing held down with a bent paper clip.

Tomlin Hawthorne

They fed a small electrical signal into one contact. Out of the other came the same signal, a hundred times stronger.

Tomlin Hawthorne

It was the first transistor. Brattain told his car pool that evening that he'd done the most important experiment of his life. He was right. It may be the most important invention of the century.

Tomlin Hawthorne

There's a coda about the human side. Their group leader, William Shockley, wasn't in the room. He was brilliant, and he was furious to have been left out. He shut himself in a hotel room and, in a few weeks, designed a better version, the one that was actually manufactured. All three shared the Nobel Prize. Shockley then moved to California to start his own company, near his mother's home in Palo Alto. He was such a terrible manager that within a year, eight of his best young engineers walked out together and founded their own firm. Then their people left and founded more. That's why the computer industry is in a valley south of San Francisco. It's there because nobody could stand working for William Shockley.

Tomlin Hawthorne

Now, why does a transistor need quantum mechanics?

Tomlin Hawthorne

Remember episode twenty-eight. An electron in an atom can only have certain energies. Rungs on a ladder.

Tomlin Hawthorne

Now bring a huge number of atoms together into a solid crystal. Each atom's rungs are affected by its neighbours. And what happens is that the sharp rungs smear out into broad bands. Ranges of energy that electrons in the solid are allowed to have, with gaps in between that are forbidden.

Tomlin Hawthorne

And then Pauli's rule, from last time, tells you how the bands fill. From the bottom up, one electron per state.

Tomlin Hawthorne

Now it all depends on where the filling stops.

Tomlin Hawthorne

If the top band is only partly full, there are empty states right next to the occupied ones. Electrons can shift into them easily. They can move. That's a metal. That's why copper conducts.

Tomlin Hawthorne

If the top band is completely full, and there's a wide gap before the next one, the electrons are stuck. Every seat is taken and the next row is a long way up. Nothing can move. That's an insulator. Glass. Rubber.

Tomlin Hawthorne

And there's a third case. The band is full, but the gap to the next one is small. In that case the electrons are stuck, but only just. A little nudge, a small voltage, a touch of warmth, a few atoms of impurity added on purpose, will lift some across the gap, and then they can flow. That's a semiconductor. Silicon. Germanium.

Tomlin Hawthorne

A semiconductor is a material balanced on the edge between conducting and not. And that makes it controllable. With a small electrical signal, you can switch it from one to the other. On, off. That's a transistor. A switch with no moving parts, flipped by electricity.

Tomlin Hawthorne

You can't get there without quantum mechanics. Before it, nobody could even say why copper conducts and glass doesn't. The difference between them is a difference in how standing waves stack up.

Tomlin Hawthorne

From that paper clip to twenty billion on a fingernail took about seventy-five years. One of the men at the bench, John Bardeen, a quiet, golf-playing Midwesterner, went on to explain how superconductors work and won a second Nobel Prize for it. He's the only person ever to win the physics prize twice. Almost nobody knows his name.

Tomlin Hawthorne

That's one. Here are some more, quickly, because I'd like you to see how much of your day is quantum.

Tomlin Hawthorne

The laser. Remember the two tribes from last time. Photons are the gregarious ones. In 1917, Einstein, in the middle of the war, noticed something in the theory. If an atom is in an excited state, ready to drop down a rung, and a photon of exactly the right colour passes by, the atom is nudged into dropping, and gives out a second photon. And the second one is an exact copy of the first. Same colour, same direction, perfectly in step.

Tomlin Hawthorne

Now you have two. They pass two more excited atoms. Four. Then eight. Put mirrors at each end so the light runs back and forth through the material, and you get an avalanche of identical photons, all marching together.

Tomlin Hawthorne

The first working laser was built in 1960, by Theodore Maiman, with a rod of synthetic ruby and a photographer's flash lamp. His paper announcing it was turned down by the leading physics journal. And when it was finally demonstrated, nobody knew what it was for. One of Maiman's own colleagues called it a solution looking for a problem.

Tomlin Hawthorne

It's now how the supermarket reads a barcode. It's how the signals carrying this podcast went down a glass fibre. It's how surgeons reattach a retina, and how those gravitational waves were detected. Flexner's essay again, from episode twenty: the usefulness of useless knowledge.

Tomlin Hawthorne

Next: tunnelling. This one's a pure piece of wave behaviour. Remember that an electron is described by a wave. When a wave meets a barrier, it doesn't stop dead. It fades out, a little way inside. And if the barrier is thin enough, a faint bit of wave is still there on the far side.

Tomlin Hawthorne

And where there's wave, there's some chance of finding the particle. So now and then, an electron that hasn't the energy to get over a wall simply turns up on the other side. Not over it. Not round it. It's as if you rolled a ball gently at a hill, not hard enough to reach the top, and one time in a million it appeared in the next valley.

Tomlin Hawthorne

The memory in your phone works that way. Each bit is a tiny island of metal, completely surrounded by insulator. To write to it, electrons are made to tunnel through the insulating wall. Once they're in, they're trapped. That's why your photographs are still there when the battery's flat.

Tomlin Hawthorne

And here's a larger example. The Sun.

Tomlin Hawthorne

The Sun shines by fusing hydrogen nuclei together. But those nuclei are both positively charged, and they repel each other fiercely. To get them close enough to fuse, you'd have to slam them together very hard, which means very hot. And the centre of the Sun, at fifteen million degrees, isn't hot enough. Not by a wide margin. By classical physics, the Sun shouldn't shine.

Tomlin Hawthorne

It shines because of tunnelling. Every so often, two nuclei that haven't the energy to get over the barrier between them tunnel through it. It's rare. For any given proton, it happens about once in ten billion years. But there are a great many protons.

Tomlin Hawthorne

So every sunrise is a quantum effect. So is every leaf, and everyone who ever ate one.

Tomlin Hawthorne

And briefly: the screen you may be looking at is made of light-emitting diodes, which are semiconductors run in reverse, so that electrons dropping across the gap give out photons. An MRI scanner in a hospital is reading a quantum property of the hydrogen nuclei in your body. And satellite navigation depends on atomic clocks, which keep time by counting the oscillations of the light given off when an atom drops between two particular rungs. The second, the unit of time, is now defined that way.

Tomlin Hawthorne

All of that is what people call the first quantum revolution. It used quantum mechanics to understand how matter behaves in bulk. But notice that in a transistor, or a laser, the truly strange things, a single particle in a blend of two states, two particles entangled across a room, aren't being used directly. They're averaged over billions of atoms.

Tomlin Hawthorne

The second revolution is the attempt to use the strangeness itself. To take single atoms, or single photons, hold them, and control their superposition and their entanglement deliberately. And the headline act is the quantum computer.

Tomlin Hawthorne

I need to be careful here, because there's more nonsense written about this than about anything else in physics. So let me say what one is, what it isn't, and where things stand.

Tomlin Hawthorne

An ordinary computer works with bits. A bit is a switch. It's a nought or a one.

Tomlin Hawthorne

A quantum computer works with quantum bits. Qubits. A qubit is some small quantum system, a single atom, say, or a tiny loop of superconducting metal, that has two states, which we label nought and one. But being quantum, it can be in a superposition. A blend of nought and one, like the electron going through both slits.

Tomlin Hawthorne

And several qubits can be entangled, so that they share one joint state that can't be broken down into separate states for each.

Tomlin Hawthorne

Now, here's the description you'll have read. A quantum computer tries every possible answer at once. Because each qubit is nought and one at the same time, three hundred of them can represent more combinations than there are atoms in the universe, and the machine explores them all in parallel.

Tomlin Hawthorne

That's wrong. Or it's so incomplete as to be wrong. Here's the catch. When you measure the qubits at the end, to read off the answer, the superposition collapses. You get one result. At random. If the machine really were just holding all the answers at once, then reading it would give you one of them by chance, and you'd be no better off than if you'd guessed.

Tomlin Hawthorne

So what's the trick? It's the double slit.

Tomlin Hawthorne

Remember the stripes. The wave goes by two routes, and where the routes agree, they reinforce, and where they disagree, they cancel. A quantum algorithm is a way of choreographing the computation so that the paths leading to wrong answers cancel each other out, and the paths leading to the right answer reinforce. Then when you measure, you get the right answer with high probability.

Tomlin Hawthorne

It's interference. It's not brute force. It's more like tuning an instrument so that only one note sounds.

Tomlin Hawthorne

And that's hard to arrange. It only works for problems that happen to have the right kind of structure. For most of what you do with a computer, writing, browsing, spreadsheets, video, a quantum computer would be no faster at all. It's not a faster computer. It's a different kind of machine, good at a few particular things.

Tomlin Hawthorne

What things? Two matter most.

Tomlin Hawthorne

One is simulating quantum systems themselves. Molecules, materials, chemical reactions. An ordinary computer chokes on these, for the reason I gave at the end of the last episode: the equations are too hard. A quantum computer speaks the language natively. That could matter enormously for designing drugs, or catalysts, or batteries.

Tomlin Hawthorne

The other is more alarming. In 1994, a mathematician at Bell Labs, Peter Shor, found a quantum method for breaking a large number into the smaller numbers that multiply to make it. That sounds like a parlour game. But the difficulty of exactly that problem is what protects most of the encryption on the internet. Your bank, your messages, a great many state secrets. A large enough quantum computer would open them.

Tomlin Hawthorne

So how close is it? This is the honest part.

Tomlin Hawthorne

The obstacle is the one from episode thirty. Decoherence. A superposition is destroyed the moment anything in the surroundings records it. Schrödinger's cat can't stay in a blend because the air is constantly taking notes. A qubit has the same problem. A stray vibration, a passing photon, a little warmth, and the computation is ruined. So these machines are run in vacuum, shielded from everything, at temperatures a hundred times colder than outer space.

Tomlin Hawthorne

And even then, the qubits make errors. A lot. The remedy is to use many physical qubits working together to make one reliable one, constantly checking each other. But that only helps if each qubit is already good enough. Otherwise, adding more just adds more errors. For twenty-five years, that was the wall.

Tomlin Hawthorne

In December 2024, Google reported a chip on which, for the first time, making the group bigger made the error rate go down. That threshold had been the goal since the 1990s. It's a real milestone.

Tomlin Hawthorne

And the 2025 Nobel Prize in Physics went to three men, John Clarke, Michel Devoret and John Martinis, for experiments they did in California in the 1980s, showing that an electrical circuit you can hold in your hand, made of billions of atoms, can behave as a single quantum object. It can tunnel. It has rungs. That was the proof that you could build a qubit out of wire, and it's what those chips are made of.

Tomlin Hawthorne

But. The machines that exist have, at most, on the order of a thousand qubits, and noisy ones. To break real encryption, the usual estimates call for something like a million good ones. Nobody serious will give you a date. The estimates run from five years to thirty. The sensible people are, in the meantime, replacing the vulnerable encryption with kinds that a quantum computer can't break. That's happening now, and you won't notice it.

Tomlin Hawthorne

By the time you hear this, those numbers may have moved. It's a fast field. The principle won't have.

Tomlin Hawthorne

The myth of the week is a single word used as an adjective. Quantum. Quantum healing. Quantum wellness. Quantum skincare. A quantum leadership retreat.

Tomlin Hawthorne

It means nothing in those phrases. It's there because it sounds deep and can't be checked. You now know what the word refers to: a theory in which energy comes in lumps, particles are described by waves, and Pauli's rule decides where electrons sit. If someone's selling you a quantum pillow, you're entitled to ask where the superposition is.

Tomlin Hawthorne

And the question to take away. It's a practical one, and I suspect it bears on your working life.

Tomlin Hawthorne

How should an intelligent person who isn't an expert judge a claim about a technology whose principles they can't personally check? You can't audit a quantum computer. Neither can I, really. But decisions get made, money gets invested, policies get written, by people in exactly that position.

Tomlin Hawthorne

So what questions would you ask now, that you wouldn't have asked eight episodes ago? Here are three that I'd offer. What exactly can it do that an ordinary machine can't, and for which problems? How many good qubits does that need, and how many do you have? And who, with no stake in it, has checked?

Tomlin Hawthorne

I'd say that's a decent template for hard claims in general.

Tomlin Hawthorne

Your challenge is one of my favourites, and it's a real quantum effect you can do with a phone and a pair of sunglasses. They have to be polarised sunglasses, the kind sold for driving or fishing.

Tomlin Hawthorne

Turn your phone screen up bright, with something white on it. Hold one lens of the sunglasses in front of it, and look through the lens at the screen. Now slowly rotate the sunglasses, like turning a steering wheel.

Tomlin Hawthorne

At one angle, the screen goes black. The light from your phone is polarised, which means its waves are all lined up one way, and the lens only lets through light lined up its own way. Crossed, nothing gets through.

Tomlin Hawthorne

Now the good part. If you can get a second pair, or a loose polarising filter, hold the first lens at the angle where the screen is black. And slide the second lens in between the first one and the screen, tilted at forty-five degrees, halfway.

Tomlin Hawthorne

The screen comes back.

Tomlin Hawthorne

You've added an obstacle, and more light gets through. Two filters block everything. Three let some pass. That's not something a simple sieve can do. The middle filter is making a measurement on each photon, and the measurement changes the photon's state. It's about the cheapest piece of real quantum mechanics there is.

Tomlin Hawthorne

That's the end of Part Six. We have a theory in which things are waves until they're found, in which the world isn't made of gloves, in which nobody agrees what's real, and which runs your telephone.

Tomlin Hawthorne

Next time we begin Part Seven, and go further down. Inside the atom is the nucleus, the fly in the cathedral. What's inside the fly? We start in Paris, in the Bibliothèque nationale, where there's a set of notebooks you can only read after signing a waiver. They belonged to Marie Curie, and they're still radioactive.