You might need an advanced degree to fully grok the spooky science of quantum physics. With the state of subatomic particles, such as the polarization of a photon, in constant flux, that particle is essentially in multiple states at the same time. Building a computer around that phenomenon means that each data bit, instead of being a zero or a one at any given moment, can be both a zero and a one.
Funky physics aside, the sales pitch for the technology is straightforward: It will (someday) beat even the most powerful traditional computers at finding the optimal arrangement of many interdependent components. For instance: picking the optimal routes for thousands of delivery vehicles, all sharing the same road network. Another: predicting the way many atoms of different elements, and all their buzzing electrons, will interact when designing complex molecules for new drugs.
Until recently, quantum computing has largely happened in the scientific underground–with large systems housed inside refrigerated cabinets at university and government labs, or in the skunkworks at cash-rich companies like Google. (Systems cost several million dollars and require specialized care and feeding.) But in 2016, IBM launched the Q Experience, providing online access to a quantum system so people without multimillion-dollar R&D budgets can kick the tires. This year, China’s Alibaba, California’s Rigetti, and now Canada’s D-Wave have followed suit.
Related: Quantum computing is almost ready for business, startup Rigetti says
This opening of access could lead to a crowd-sourced revolution in advancing the gnarly new algorithms required to integrate quantum with traditional computing–and fulfill the sales pitch.
“Those are the optimistic possibilities,” says Chirag Dekate, a Gartner analyst covering AI and emerging technologies. “But if you speak with the realists in the space, what they will tell you is…it’ll allow end users to explore and figure out that quantum computing is just not ready for primetime yet.”
Perhaps nowhere is the promise-vs.-reality debate more heated than with D-Wave–established back in 1999 and the latest to offer a cloud-based system, which opened to the public today. Mentioning D-Wave often provokes a smirk or grimace among quantum computing experts, who say the company may be overselling its outlier approach to the technology, known as quantum annealing.
“D-Wave claims that they’re realizing production value from their quantum computer,” says Brian Hopkins, a Forrester Research analyst covering quantum computing. “I haven’t been able to verify that claim by talking to anybody who’s willing to tell me they’re doing anything other than running science experiments.” He goes on to say, however, that anyone who did get a competitive advantage using a quantum computer probably wouldn’t want competitors to know.
“Every time somebody says, well, D-Wave this, D-wave that, my response was, go try it. You go see for yourself,” says Alan Baratz, EVP of R&D and chief product officer at D-Wave. And now anyone actually can.
IBM’s Q Experience is also free. Rigetti’s free quantum cloud is limited to about 500 invited participants. I was unable to find Alibaba’s online system, announced back in March, and the company hasn’t responded to my inquiry.
What happens when Romeo meets Juliet
At the very least, D-Wave Leap is a fun way to get a feel for quantum computing, running introductory lessons on a real machine within the first few minutes. It currently features two example problems that harness quantum computing’s ability to consider multiple variations of a situation at the same time. That comes from the ability for each bit to be considered to have a value of both zero and one, plus the ability to link the fate of bits together through “spooky” phenomena like entanglement.
The first example is factoring: finding all the integers (i.e. not decimal numbers) that another integer can be divided by. First D-Wave shows how a classical computer would solve the problem, trying each option in sequence. For the number 49, can you divide it by 2? No. Can you divide it by 3? No. And on and on, until it tries 7.
It then reconfigures the problem into a latticework of entangled qubits. You press the “RUN” button, and a distant quantum computer spits out the answer “7×7” after a 16-millisecond run–leaving enough computer time for solving up to 3,998 other problems.
These sample tasks, known as optimization problems, are easy enough for a traditional computer to solve without breaking a sweat. But as the numbers get bigger, the workload on a traditional computer quickly spirals out of control. On that all computing experts and mathematicians agree.
The battle of the bits
One critique of D-Wave’s technology is that it’s only useful for these optimization problems. Its qubits are tied together in a way that, driven by physics, naturally cascades to an alignment providing the optimal arrangement (or arrangements) to solve the problem. But D-Wave’s tech doesn’t provide the quantum versions of the “logic gates” on which today’s computers are based. Those systems, which most other companies are pursuing, could solve a wider range of potential problems. “DWave’s qubits and qubits of gate-modeled or universal quantum computers are similar only in name,” says Hopkins. “You simply cannot compare them in any meaningful way.”
D-Wave has a pretty compelling counter-argument. “Today you can’t actually use a gate model system for anything of value,” says Baratz. Or as Hopkins describes it: “Universal quantum computers have qubits that are highly susceptible to errors and hence the complexity of problems they can solve is currently very limited.”
What remains controversial, though, is D-Wave’s current ability to economically beat a standard computer. “I have run into academic researchers from reputed universities in New England that claim that quantum annealing systems are fairly limited, and you can actually achieve exactly [what they do] using high performance classical systems,” says Chirag Dekate.
D-Wave, of course, disagrees; and it’s now putting its mouth where its money is ($220 million in funding raised). Beyond introductory materials, including videos and other explainers, Leap gets very technical. It includes a suite of developer tools called Ocean, sample code, and a software developer kit that allow programmers to build complex operations for solving real problems. That’s similar to what Rigetti is doing, with a set of tools–following the nature theme–called Forest.
This is the natural progression of quantum computing, according to analysts. “Cloud is going to be the de facto way to access quantum computers,” says Brian Hopkins. “Very few firms are going to buy quantum computers.”
Putting a spin on quantum computing claims
Rigetti expects that, within three years, its community of developers will reach so-called “quantum advantage”–getting results faster or cheaper with a combination of quantum and classical computers than with a classical machine alone. That’s a bold claim in the industry, but D-Wave says it’s passed that milestone, which it brands “customer advantage.” “We have [clients] working on 100 applications. Many of them are achieving customer advantage already,” says D-Wave’s SVP of marketing, Jen Houston. She and Baratz point to customers like OTI Lumionics, which they say is using their technology to model new molecules for OLED lighting.
Observers remain wary. “I’m always skeptical until I actually speak with a customer firsthand,” says Dekate. “Vendors will always say some interesting things to push their technology.”
By bringing its technology into the open, D-Wave is allowing many more people to test its claims, and trade notes. “We’ve had—up until now–20, 30, 40 customers that have been doing really interesting application work,” says Baratz. (Many had been accessing the systems over the cloud since 2011.) “Now they and hundreds of thousands of developers that we’re opening the system up to will be able to communicate and interact with one another.”
Hundreds of thousands sounds quite optimistic. Quantum computing is still a rather niche field for savvy early tinkerers, says Dekate. But he agrees that opening up these computers–whatever quantum technology they use–could have a huge benefit to advancing the nascent technology. “If you have a problem set that can benefit from quantum computing, irrespective of the platform type, rethinking your approach, rethinking your problems, and thinking in a quantum mindset absolutely is going to be helpful,” he says.