Google’s Bristlecone Quantum Computing Chip Achieves 72 Qubits

The Reality of Quantum Supremacy and A Look at Future Quantum Applications

When a quantum computer is finally able to calculate operations faster than the most powerful supercomputers currently in existence, it will achieve what is known as “quantum supremacy”. Google Quantum AI Lab revealed a new gate-based superconducting quantum computing chip called Bristlecone last week with a square array of 72 qubits (a portmanteau for quantum bits). They are going for quantum supremacy, but they may be a few qubits short.

Say hello to Bristlecone, Google’s latest quantum processor (left). In the graphic of the device, each “X” shape represents one qubit, with “nearest neighbor connectivity”. (Image courtesy of Google.)

Say hello to Bristlecone, Google’s latest quantum processor (left). In the graphic of the device, each “X” shape represents one qubit, with “nearest neighbor connectivity”. (Image courtesy of Google.)

Classical Bits Versus Qubits

Qubits are analogous to classical binary bits but are primarily differentiated by their ability to be in a superposition of both 0 and 1 states at the same time, whereas classical bits are either in a 0 state or a 1 state. For classical bits, the 0 and 1 represent a mutable dichotomy (on and off, yes or no, true or false), however a processed bit is known for its dichotomy of two different DC voltage levels (x volts for the “0” state, y volts for the “1” state). These switch through what’s known as “the forbidden zone”, which represents the rate of transference between the two DC voltage levels that represent the two bits in either the “0” state or the “1” state.

Though the qubit has a similar dichotomy of states to a classical bit, meaning it can be 0 or 1, the superposition of a qubit is represented by two basis vectors, |0) and |1), which represent [0 1] and [1 0], or “ket 0” and “ket 1”. These basis vectors, or base states, can then be combined into pairs and multiple qubits.a

A key difference between classical bits and qubits is that qubits can hold two bits of information using superdense coding, while a classical bit has a limit of one bit. Using superdense coding, two bits or information are encoding in a state of a single qubit under the assumption that each are sharing an entangled state. With this assumption, the limit of classical bits that can be encoded in a qubit in two, thus the term “superdense”.

This doubling of efficiency per qubit compared to classical bits is why adding qubits to a quantum computer increases its computing power exponentially versus adding classical bits to a classical computer, which does not.

Google’s Bristlecone and The Reality of Quantum Supremacy

The first quantum computer had 2 qubits in 1998, and the previous record for highest qubit count in a quantum computer was 50 qubits by IBM in 2017. Google’s John Martinis, who is leading their quantum computing efforts believes that achieving quantum supremacy is reachable milestone in the next few months or by the year’s end.

Research Scientist Marissa Giustina installs a Google Bristlecone chip at the Quantum AI Lab in Santa Barbara, CA. (Image courtesy of Google.)

Research Scientist Marissa Giustina installs a Google Bristlecone chip at the Quantum AI Lab in Santa Barbara, CA. (Image courtesy of Google.)

Well, don’t get your hopes up, because there definitely won’t be a commercial version available to try out at home anytime soon.

What happens when quantum supremacy is achieved?

First, creating qubits are insanely complicated because their circuits must be made of superconductive material, which needs an environment of ultra-low (in Bristlecone’s case less than -450 degrees Fahrenheit) temperatures to keep electrical resistance to an absolute minimum.

It’s not just temperature that needs to be controlled, it’s any level of vibration. The tiniest vibration can cause qubits to lose their quantum state. When this happens, the rate of erroneous calculations begins to slink upward.

Another factor that affects the ability of a quantum computer to produce errors is the number of qubits. The more qubits there are, the more error-prone the computer becomes. More qubits mean more complexity and more power dedicated towards creating the perfect conditions to stabilize the fragile quantum state. The computational stability is perpetually caught in this extremely fragile house of cards, though enough strides have been made in a few supporting areas including supercooling technology to keep the dream of quantum supremacy within site.

Also, the speed of classical supercomputers isn’t standing still, and the two most powerful among them sit inside China’s borders, so it might be impossible to test against them. IBM in 2017 believed that a 49-qubit computer would do the trick, but simulations performed on a classical computer indicated that a 100-qubit computer would need to be achieved before quantum supremacy could be achieved, and any useful quantum applications could be developed. IBM Q is an example of an initiative to build commercially available universal quantum computing systems based on a quantum processor prototype.

Google, whose majority income is based on advertising revenue, via AdWords, is pushing research in a few different bleeding edge technologies like AI, might beat IBM to achieve “quantum supremacy” over China’s classical supercomputers with Bristlecone.

How to Benchmark A Quantum Computer

How do you create a problem that can’t be solved on classical computers to test a quantum computer? 

The task should be impossible to perform on classical computers, but Martinis and researchers at Google are going as far as they can, attempting to solve an algorithm that ranks extremely close to the limits of known classical supercomputers. Then, if they add just one more qubit, they may just pass the fastest supercomputers.

Even if Google achieves quantum supremacy, what then? There are virtually zero quantum computing applications that have any usefulness, so besides discovery, on a practical level, what can quantum computers do?

This chart conceptualizes the relationship between error rate and number of qubits. The research direction is in red, where researchers at the Quantum AI Lab hope to create and access near-term applications en route to the construction of an “error corrected quantum computer”. (Image courtesy of Google.)

This chart conceptualizes the relationship between error rate and number of qubits. The research direction is in red, where researchers at the Quantum AI Lab hope to create and access near-term applications en route to the construction of an “error corrected quantum computer”. (Image courtesy of Google.)

They are difficult to manage, costly to maintain, incredibly fragile and are ridiculous even as hypothetical replacements for less cumbersome classical computers. Classical computers will also continue to improve, and over-hyped technologies can easily fall into yet another layer of technological obscurity through too much media exposure.

Most of the legwork that needs to occur to keep a possible 100-qubit quantum stable requires the most focus now, before any quantum applications that are beyond the capabilities of classical supercomputers can be made viable.

Potential Quantum Applications

What are the problems that scale better for quantum computers to solve versus classical computers? One category might turn out to be machine learning. Algorithms are being developed by companies like D-Wave Systems to leverage quantum-computing capabilities to perform machine learning tasks better than classical computers, which are already doing extremely sophisticated things like beating the best human Go players. Quantum computing could help with optimization problems, which generally revolve around trying to find the best answer out of a complicated and large set of alternatives. Designing molecules might be an example of a near-future quantum application. NASA is investigating the usefulness of quantum computing to solve problems like locating exoplanets and tackling enormous logistical issues.

Bottom Line

The problems of superconductivity and tiny vibrations make quantum computing in its current state an unbelievably fragile enterprise, suited only for extremely wealthy companies like Google and prestigious legacy companies and organizations that pioneered computing applications, like IBM and NASA.

It’s hard to get a grasp even from experts, who disagree on the viability of quantum computing at almost every level. But Google’s Martinis is determined to break through to achieve quantum supremacy. After that, he wants to lead the way with proving quantum computing’s usefulness by creating some killer quantum apps.