This morning, Google scientists confirmed in a blog post that their quantum computer had needed just 200 seconds to solve a problem that they claim would take the world’s fastest supercomputer 10,000 years to complete.
This browser does not support the video element.
The team first ran the algorithm last spring using a 54-qubit processor called “Sycamore.” While the achievement is called quantum supremacy, it doesn’t mean that quantum computers are suddenly more capable than classical computers, since Google’s quantum computer only beat the competition at a single, highly contrived problem. Quantum computers with day-to-day applications may still be decades away, but this is an important scientific milestone when comparing quantum computers to their classical counterpart.
“For such large-scale endeavors it is good engineering practice to formulate decisive short-term goals that demonstrate whether the designs are going in the right direction,” Google’s John Martinis and Sergio Boxio, chief scientists of quantum hardware and quantum computing theory, wrote in the blog post. “So, we devised an experiment as an important milestone to help answer these questions.”
Quantum computers are a new kind of computing device that could one day be capable of solving problems that classical computers can’t. Instead of series of transistors linked together, representing two-choice bits like in classical computers, their base unit is the quantum bit, or qubit, a piece of hardware that mimics the behavior of a subatomic particle. Qubits communicate via the probability-driven theory of quantum mechanics instead of the regular rules of logic. They’re still two-choice systems that output binary code, but getting to the answer incorporates the quantum mathematical ideas of entanglement, superposition, and interference. This new architecture may one day excel at simulating the behavior of subatomic particles well enough to create new medicines and new materials. It might also be able to crack the code that modern-day encryption is based on.
Scientists must first find a physical system that assumes quantum properties. But quantum states are incredibly fragile—the slightest bump of heat or vibrational energy can make initialized qubits lose their quantumness and turn into regular bits. Google’s engineers built theirs from loops superconducting wire, controlled by quick, customized microwave pulses.
Google’s quantum supremacy experiment essentially sets up random circuits out of these qubits. Certain outputs become more common than others. It’s easy for Sycamore to find these outputted strings, but with each new qubit, it would take a regular supercomputer exponentially more time to come up with an answer. Google’s scientists ran the experiment repeatedly, incorporating a new qubit until the supercomputer simulating the quantum computer couldn’t keep up, according to the paper published in Nature.
The main application of such an experiment is that it can produce truly random numbers, something useful in various fields of science, cryptography, art, and of course, online gambling. But a hypothesis called the Church-Turing thesis claims that a theoretical computer called the Turing machine, which basically simplifies all computers to symbols on tape, is the most efficient way to solve computer problems. Google’s quantum computer provides evidence against this thesis.
Rumors and hype have surrounded the Google quantum supremacy announcement since the team published a paper describing how they’d achieve the milestone on the arXiv physics preprint server in 2016. Last month, the Financial Times reported that it had found the Google paper describing the completed quantum supremacy experiment on a NASA server, but Google would not confirm the veracity of the report until today.
Already, scientists are debating whether the quantum supremacy experiment actually demonstrates what it claims. On Monday, the IBM quantum team published a blog post arguing that a classical computer could more accurately run the Google problem in just 2.5 days. Simply put, it’s hard to prove a claim that a classical computer can’t do something.
Regardless, this experiment does not mean that quantum computers are suddenly going to appear in your iPhone; there’s a lot of work left. John Preskill, the CalTech physicist who coined the term quantum supremacy, told Gizmodo in March that while it’s certainly worth pursuing these supremacy experiments, “I think it’s more important to try and develop the tools that we need to scale up further: perfecting error-correction methods, improving the qubits, and addressing the system’s engineering issues that you need to control the platform with thousands or millions of qubits.”
It’s clear that we’ve entered a new era of quantum computing (it’s called the NISQ era), as companies now have noisy but functional devices that may actually be useful soon.
I’m at Google’s lab at Santa Barbara, California today to get a first look at the device. I’ll report back with more images of the computer and what it’s actually like.