WTF is quantum computing?
We haven’t got to where we are today because of our muscles. If you compare us to the rest of the animal world, our physicality is pretty hopeless. Linguist Dereck Bickerton speaks of the haplessness of early humans: “It is more than likely that some of our ancestors suffered the ignominious fate of being eaten by weasels. Some of them, it’s pretty certain, were eaten by birds.” I mean come on, that’s pretty sad. Getting eaten by a lion has at least some degree of gravitas, dying in the jaws of a weasel, that’s just embarrassing. So no, we aren’t where we are today because of our muscles. We’re here because of our brains.
Evolution did great work. Our brains got bigger and bigger and we became more and more like we are today. But our brains could only ever be so good and so big. After all, our mothers are bipedal and consequently endowed with narrow hips. Humans are not very well designed for large headed babies. So for years we bumbled along with our pretty big brains, mostly content with our abilities as a species. Until along came computers, an innovation that revolutionised the horizon of what we could achieve. Computers extended our brains beyond our own biology and into silicon chips.
The kind of computers that we use today are known as classical systems. They have been and continue to be fantastic. They allow us to send rockets to space, communicate instantaneously with relatives across the world, and take belfies — A self-taken photograph of one’s own buttocks.
Classical computers have been improving at an astonishing rate for the last half-century. Every two years we have been able to roughly double the number of transistors that we can fit onto a microchip. Transistors are the fundamental building blocks of computing, the switches that control the flow of electrical input, that allow for the binary on/off, zero/one system that is fundamental.
So every two years, we have been doubling our computing power — something known as Moore’s law. But Moore’s law is slowing down. We’re reaching the physical limits of what is possible. To cram more transistors onto a chip, we have of course been making those transistors smaller and smaller. Nowadays we can get 4.3 billion onto a chip the size of a fingertip. The world’s smallest transistors are now just 1nm long. That’s 500 times smaller than the diameter of a red blood cell. The thing is that when you get that small, things start to get real freaky. Quantum mechanics pops up and starts messing around. Basically the transistors cease to be reliable, due to something called quantum tunnelling.
What is quantum tunnelling you ask? Magic, young humanoid. Bloody magic. In a nutshell, when a transistor wants to stop the flow of current and become a zero, or adopt an off state, it puts a wall in the way. Walls are good at stopping things. Donald Trump wanted to build one to stop his people running away to Mexico. But Donald Trump’s people aren’t quantum particles. You see quantum particles are magic. They can basically teleport through walls — that’s what quantum tunnelling is. Therefore we can’t just keep making transistors smaller, because smaller transistors won’t work.
So great computer scientists had a bright idea. Instead of fighting quantum mechanics and it’s insanity, why not just build a quantum computer that takes advantage of the nuttiness. In fact quantum computing promises to be amazing. To understand why let’s think again for a minute about classical computing.
There are some problems that classical computers will never be able to solve. Take coffee. Inside of coffee is caffeine. Caffeine is a molecule. It’s more complex than water but less complex than a molecule of DNA or protein. To accurately model a caffeine molecule using today’s computers is impossible. If you took every single atom in our galaxy and used them to build a classical computer, that computer would still be incapable of simulating a caffeine molecule. That’s insane, isn’t it? Quantum computing will make this possible.
So why will quantum computers be so powerful? It boils down to two technicalities.
Quantum particles possess something called superposition. This means that they can be in all of their different possible states at the same time. That is, until we observe them, at which point they are forced to pick a state. In a classical computer we represent data as bits. Each bit can either be a one or a zero. In a quantum computer we represent data as qubits. Each qubit can be both zero and one at the same time due to the wonders of superposition. So in a classical system of 4 bits, there are 16 possible variations of 0 and 1, and the bits can only represent one variation at a time. In a quantum system 4 qubits can be in all 16 of these possible variations at the same time. Take that to just 20 qubits and your system could simultaneously hold over 1 million values. This is truly parallel computing and totally revolutionises the way that we can solve problems.
And then there’s a second technicality. Quantum particles can be entangled. Something Einstein called “spooky action at a distance”. Once you have entangled two quantum particles they will behave in the sum opposite way to their entangled partner. That means that when we observe one of the particles it will pick a state and the other particle will instantaneously choose the opposite state. So if one spins in a downwards direction, the other will spin in an upwards direction. And this behaviour isn’t limited by distance. You could entangle two quantum particles and then separate them by a distance of light years. The crazy thing is that when you observe or change one, the other will still instantaneously react. We don’t fully understand why or how, but it’s just the way the quantum world works. This entanglement means that read/write operations will only have to be done on one of the two qubits that you want to know about or change.
These superpowers will allow quantum computers to reach into the bag of previously unsolvable problems and solve a whole host of them with ease. The implications for progress in fields such as medicine, cryptography, chemistry, and artificial intelligence are immense. That’s right, you thought AI was scary, well wait until those AI have quantum brains.
So how far away from quantum computers are we? Last year IBM released a 15 qubit system. The things that you can do with 15 qubits are relatively trivial. It is thought that we can start tackling meaningful problems somewhere in the region of 20–100 qubits. IBM plan on releasing a 50 qubit quantum computer this year. And Google, well they already have a 72 qubit system.
The space is developing extremely fast, but don’t go cracking open the celebratory KitKat just yet, because… well, qubits are crazy. They’re like your aunt after a glass of Prosecco — extremely error prone. Somewhere just under 10% of the time, they just say ‘bugger it’ and fail to do their jobs. We’re going to have to continue wrangling this error percentage down, but they’ll always be a bit nuts, it’s just in their nature. Expert humans think that to have a truly fault-tolerant logical 100 qubit system, the quantum computer will actually have to contain 1 million qubits. That means that 10,000 qubits are needed to distil a singular “logical” qubit from the sum of their madness. We’ve got a long way to go.
But don’t despair, that celebratory KitKat will be making its way through your gut in the not too distant future. Experts said in San Francisco a couple of weeks ago that quantum computers will capable of important real-world applications within 5 years.
Around 3.85 billion years ago life was born on earth. Life that was simple, and took many years of Darwinian evolution to become intelligent. Then we popped up, having crawled out of the sea some 400 million years ago and into the trees, and then down onto the ground to stand on two legs. Our brains stewarded us to where we are today, they’ve been good to us. We no longer get eaten by weasels. But there are problems that only computers can solve. And there are problems that only quantum computers can solve. I’ll leave you with a quote from the father of quantum computing, Richard Feynman: “Nature isn’t classical, dammit, and if you want to make a simulation of nature, you’d better make it quantum mechanical, and by golly it’s a wonderful problem, because it doesn’t look so easy.”
This piece was transcribed from The Zip Files — an irreverent weekly 20–25 minute podcast that I produce to help the busy millennial catch up with all of the week’s most important tech news. Here’s the episode in which this piece was featured:
On Apple Podcasts:
The Zip Files by Evan Thomas on Apple Podcasts
Download past episodes or subscribe to future episodes of The Zip Files by Evan Thomas for free.