Google Quantum Supremacy: Here's What's Next

Google’s claim on Wednesday to have made a major breakthrough in computing called “quantum supremacy” is just an early step in the long road to perfecting its nascent technology.

The tech giant next hopes to expand on its announcement—that it had used a quantum computer to make a calculation at speeds far exceeding current technology—to achieve tougher milestones and, ultimately, create a new business.

For years, researchers at a number of companies have worked on quantum computers, which rely on strange quantum effects like “superposition” and “entanglement.” In achieving the long sought-after goal of quantum supremacy, Google has made a notable step forward that paves the way for advances in a range of areas from energy to health care to physics.

At a press conference Wednesday morning in a Santa Barbara laboratory—past wetsuits and sand-coated surfboards—Google’s team revealed some of its upcoming business and research plans. They discussed their next intentions, experiments, and potential ways the computers could be used for improving cybersecurity and making strides in other technical fields.

Cloud access granted

First, Google plans to let more people get their hands on its technology.

“I think the next right thing to do is to make [the computers] available to the world” through the so-called cloud, said Erik Lucero, a scientist on Google’s team. Housed in Google data centers, the machines will by next year be, he hopes, remotely accessible to the public over the Internet.

Lucero said Google would first open access to “close partners,” a short list whose few public names include Daimler, Volkswagen, and the U.S. Department of Energy.

Rival IBM, which contested Google’s supremacy claim on Monday, has been more vocal about its quantum computing collaborations, touting partnerships with ExxonMobil, JPMorgan Chase, Samsung, and others. The mainframe-maker has also offered public access to its quantum processors since 2016.

The two companies—plus others, such as Intel, Microsoft, Honeywell, startup Rigetti Computing, and more—are likely to clash more as they battle to recruit talent and persuade businesses to work with them.

Certifiable randomness

Google is working to commercialize at least one potentially near-term application that unexpectedly arose out of its supremacy experiment.

Wearing a puffy space-grey jacket and ripped jeans, Hartmut Neven, founder of Google’s quantum artificial intelligence lab, previewed the company’s plans for using the technology to create “certifiable random numbers” which can prove to skeptics their own trustworthiness. Such numbers could play an important role in computer security and digital privacy because they could lead to “better keys,” or tools for encrypting data.

“As you may know, randomness is an important resource in computer science,” Neven said. One can imagine Google offering this as a feature through Google Cloud, an enterprise business that has lagged behind rivals Amazon and Microsoft.

Since Google made the supremacy announcement, Neven said “many ideas have dropped on our lap. We’re in the process of sifting through them.”

One inbound proposal, Neven said, involved improving the performance of nuclear magnetic resonance imaging, an essential tool for scientists to understand the structure of molecules. Boosting such key analysis techniques could help advance efforts in biomedicine, chemistry, and materials science.

Google said it hopes to use its quantum computers to better simulate the interactions of nature’s building blocks, atoms and molecules, potentially expanding scientists” understanding of physics and accelerating their development of new medicines and batteries. Another further-out application could enhance machine learning algorithms, leading to more efficiency—and profits—for businesses of all types.

“If it’s a good idea we’re gonna want to run it on-chip right away,” said John Martinis, Google’s chief quantum hardware scientist.

Correcting the record

On the scientific front, Google is already readying new experiments.

Marissa Giustina, a Google quantum scientist who wore qubit-inspired earrings, drew with a marker on the conference room’s white-board walls a timeline indicating the team’s progress toward developing an “error correcting” quantum computer. It would be able to improve on the characteristically high failure rates that ruin many quantum computations.

“The big focus right now is on bringing down the error rates on the qubits,” said chief scientist Martinis, referring to the notoriously finicky quantum equivalent of classical computers’ bits. He is encouraged that, as Google’s team’s experiments have gotten more complex, error rates have maintained a consistency—not spiraling out of control, a much-feared potential obstacle.

Sergio Boixo, Google’s chief quantum theoretician, described in his raspy Spanish lilt how, in the spring of this year, the team feared it had hit a wall of insurmountable errors. The group eventually found a way over the hurdle by better calibrating their machines, he said.

Martinis mentioned one particularly thorny technical challenge that has occupied much of his time recently: Reducing the “noise,” or experimental interference, caused by a spaghetti tangle of wires that ferry signals to-and-from the team’s quantum processors.

“For me, the wiring problem is such a difficult problem that I have personally been working on advanced wiring solutions over the last three years,” Martinis said. “It’s boring. It’s not a very exciting thing, but it has to be done,” he said.

Martinis suggested that he has made a number of as yet unannounced breakthroughs in reducing the adverse impact of the wires. “We can do much much better than we’re doing right now” simply by making planned improvements, he said.

In the months ahead, if all goes well, expect error correction to be one of the next big headline-making experiments from Google following its supremacy announcement.

The team’s spirits are buoyed by their understanding of the error-induced pitfalls they have encountered to date. “That gives us confidence about building the next generation machine,” Giustina said.