Google made a huge revelation on October 23, 2019, when it announced that it had reached something called “quantum supremacy.”
Via an article in the journalNature, Google said their quantum computer, called Sycamore, solved a particularly difficult problem in 200 seconds. For comparison, Google said the world’s current fastest classical computer — one called Summit owned by IBM that’s as big as two basketball courts — would take 10,000 years to solve that same problem.
This is what “quantum supremacy” means. It’s when a quantum computer — one that runs on the laws of quantum physics as opposed to the classical computers we’re familiar with (i.e. phones and laptops), which run on classical physics like Newton’s laws of motion — does something that no conventional computer could do in a reasonable amount of time.
The trouble here is that IBM responded to Google’s news to say that actually, Summit could solve the quantum computers’ problem in two and a half days — not 10,000 years as Google had suggested.
In this episode of Recode’sResetpodcast, host Arielle Duhaime-Ross and Kevin Hartnett, a senior writer for the math and physics magazineQuanta, break down exactly what quantum computing is and why Google dunking on IBM both was and wasn’t a huge deal.
“It’s insane because if we build a working quantum computer, it demonstrates that we have achieved a kind of physical mastery over matter in the universe. At the most fundamental level, we are controlling it. We’re manipulating it to our own ends and we’re performing calculations with it and we’re performing calculations with it. That’s kind of stunning. And the fact that engineers are now actually pulling this off is kind of amazing,” Hartnett pointed out.
Later in the episode, the two also dive into what applications a working quantum computer would have in the real world from pharmaceutical drug discovery and financial modeling to breaking the internet by undoing a common form of encryption called RSA encryption.
”Right now, that’s how we keep a lot of information on the internet safe. But with a very powerful quantum computer … the time it takes to break a large number down becomes really short. And that means the key you use to encrypt stuff is easy to figure out and the internet is kinda screwed. But that is still mostly theoretical at this point,” Duhaime-Ross explained.
Listen to their entire discussion here. We’ve also shared a lightly edited transcript of Harnett’s conversation with Duhaime-Ross below.
You can subscribe toReseton Apple Podcasts, Stitcher, or Spotify.
The term “quantum supremacy” sounds frankly ridiculous. It always makes me feels like electrons are going to take over the world. But people have been waiting for this to happen for a long time. Because what Google announced means that a special type of computer, called a quantum computer, is finally doing things. Instead of just feeling like …
… this almost mythical dream for decades. It’s felt like for years we’ve been waiting for an announcement of what people call quantum supremacy. And that’s what Google announced this week.
So shit’s getting real [with] quantum computing and two of the biggest tech companies, Google and IBM, are obsessed with it. But first I gotta make you talk about physics. Don’t worry. Cause I got you. I am here for you. And we’re all gonna do this together.
So Kevin — what’s a quantum computer?
A quantum computer is a machine that performs calculations using the laws of quantum physics as opposed to, I guess, your classical computer like your phone or your desktop computer, that behaves according to the laws of classical physics. So a classical computer computes using bits and bits can be either 1 or 0.
So like the foundation of code basically.
Exactly. That’s what code is. Now, a quantum computer doesn’t use bits, it uses quantum bits or qubits. And these qubits are made out of quantum material and a cubit can be in a position of 1 or 0. Or it can be described by the probability that it’s in one or the probability that’s in the position zero.
The advantage of that system is you have these many different possible states. Your quantum computer could be in when you kind of are working through your calculation and each of those states has a probability assigned to it.
ADR: That’s the key. A regular computer (or what computing experts would call a “classical computer”) like your laptop or your phone — those computers can only try one possible pathway at a time to get to an answer. But a quantum computer can get to the right answer more efficiently.
That’s not super intuitive, so here’s a thought experiment: imagine a ball and a hill. Your task is to get that ball to land on a specific spot at the bottom of the hill.
Imagine that you’re at the beginning of the comp, your computation. You are at the top of the hill and you’re going to roll a ball down the hill. And it could kind of go any number of ways down the hill and end up at any number of different spots at the bottom. But there’s only one spot at the bottom that’s the correct answer. So if you have a classical computer, the best you can do is kind of roll that ball down the hill and see where it lands and see if it’s the correct answer or not.
So you’re constantly running the ball up the hill and then you get to drop it once and then maybe you get the correct answer if you don’t. You got to roll it back up the hill and try all over again.
It’s a lot of effort, and it takes a lot of time. [But with a quantum computer] you have the ability to kind of structure your cubits so that the path the ball takes down the hill is in some ways the most likely path it takes down the hill is the one that corresponds to the correct answer at the bottom of the hill.
So you can kind of tip the hill, tip the scales in your favor in order to get to the correct answer a whole lot faster. A quantum computer should theoretically be able to get you an answer a heck of a lot faster than a classical computer. But when it comes to speed, sometimes the newest, flashiest mode of transportation isn’t the most reliable.
So, you have to ask yourself: when does performing a calculation with a quantum computer become undeniably more efficient than doing the same thing with a classical computer?
That, in a nutshell, is the quest for quantum supremacy.
Quantum supremacy is a benchmark that was set about seven years ago in 2012 by a physicist named John Preskill at the California Institute of Technology, Caltech. And what quantum supremacy means is a quantum computer can do something that no ordinary classical computer can match.
There is this idea that quantum computers should be able to perform calculations much faster than regular computers. But a lot of people doubted whether it was possible to actually achieve a calculation like that in practice. So quantum supremacy is the moment a quantum computer actually performs calculations that a classical computer simply can’t keep up with.
Because if you give a classical computer enough time, maybe even more time than we have left in the universe, it can do anything a quantum computer can do. Supremacy means a quantum computer can do something in a categorically faster way than a classical computer can do.
The world’s most powerful classical computer is called Summit. It’s owned by IBM, and it’s as big as two basketball courts.
Google’s quantum computer, on the other hand, probably fits in your bedroom. It’s called Sycamore.
So what Google did was give Sycamore a very specific problem to solve, called a random circuit sampling problem.
What they showed is that their computer can solve this random circuit sampling problem in about two hundred seconds. And the most powerful classical computer in the world, which is called Summit. Well, Google estimated it would take that computer 10,000 years to solve the same problem that their computer solved in 200 seconds.
So what Google announced is that it basically dunked really hard on IBM. Most powerful computer, the most powerful computer, classical computer in the world.
It dunked hard in the sense of this particular problem, and you could say it killed it.
It is a big deal. It’s a big deal in a practical sense, right? There are lots of things we can do if they only take a couple minutes to carry out the calculation that we can’t do if we need years or tens of thousands of years.
Except that if you ask IBM, Google didn’t dunk as hard as it says it did.
IBM is certainly skeptical. They issued a paper saying, Not so fast, they haven’t actually done it because Google estimated in their paper that it would take the most powerful, ordinary computer 10,000 years to carry out this calculation. And IBM said, actually, that most powerful computer which we designed can do this in two and a half days and probably even faster than that if we had time to fine-tune how we did it.
I should also say that IBM themselves, they are perhaps Google’s biggest competitor in the effort to build a quantum computer.
Is IBM just a sore loser or is this a very good point that they’re making?
Well, the point they’re making is good, and it does point to something important about supremacy, which is you’re trying to prove that a quantum computer can do something faster than any classical computer can. So proving supremacy, truly proving supremacy would involve proving there’s no way a classical computer can do this kind of problem as fast as a quantum computer. And we don’t have that kind of proof here.
It took the most powerful classical computer in the history of the world, a computer that occupies an area the size of two basketball courts. This computer can solve a problem in 200 seconds that takes the world’s second-fastest computer a minimum of two and a half days to do this calculation. If Google’s quantum computer gets even a little bigger — if instead of 53 cubits, they have 70 cubits — now, a classical computer that fills two basketball courts won’t be able to keep up at all. You need a classical computer the size of a city to simulate that same calculation.
And that’s the sense in which the quantum computer is doing things categorically faster than a classical computer. And I think that difference is really what people have in mind when they talk about quantum supremacy. And that’s why most people are pretty comfortable saying that what Google did, if it fully checks out, is in fact a demonstration of quantum supremacy.
So is IBM also going to demonstrate supremacy anytime soon?
IBM has been on the record for a while now that they are not chasing the goal of quantum supremacy. They view it as an artificial benchmark where you could cross it and still not be doing anything useful. So they, at least in their public statements, have been much more practically minded.
This is a difference between Google being the kind of basketball player that wants to straight up flip in the air before dunking. And IBM that just wants to do a lay-up because they know that it works. And it’s the same amount of points.
I’d say so. And it seems like Google has been more animated by the basic science of it and the sense of possibility and exploration.
My man Kevin with the perfect segue! We’ve established that Google’s quantum computer is significantly faster than the most powerful classical computer in the world.
But what exactly are the possibilities with this? Are personal quantum computers in our future?
To find out, listen to the full conversation and subscribe toReseton Apple Podcasts, Stitcher, Spotify, or wherever you listen to podcasts.