Hot ice cream, sleepless cats and affordable quantum computers

Some things are not impossible, but are, still, implausible – in the absence of creativity, consistency of efforts and bold thinking. Enterprise-easy quantum computing is one such oxymoron. Let’s find out why voodoo-factors and plateaus matter here

Exactly a decade back, in 2012, the Nobel Prize in Physics went in the lap of Serge Haroche and David Wineland. It was for their work in measuring and manipulation of individual quantum systems.

As we sit in 2022, we see this year’s Nobel Prize in Physics going again into Quantum’s kitty with Alain Aspect, John Clauser and Anton Zeilinger for their work on quantum entanglement and the strange quantum nature of reality.

From Richard Feynman’s quantum simulator in 1982, to David Deutsch’s quantum Turing machine in 1985, to Peter Shor’s iconic algorithm in 1994—a lot has happened in the quantum world.

“We do not expect a ‘transistor moment’ yet, where a particular approach to quantum computing would break away. Superconducting, Trapped ion/atom, Photonics, Electron spin, Topological, etc. and possibly combinations of them will continue to receive significant R&D investment.”

– Shahin Khan, Founding Partner and Analyst, OrionX

The strides being made in this field are well, pun intended, quite quantum. And yet, we are yet to

see quantum computers in the grocery carts of enterprises. Affordability and enterprise-readiness are two big walls that hold these big and beautiful

machines back.

Shall we ask why and dig into the ‘why not’?

But before that, a recap of what makes quantum what it is.

Under The Big Hood

Unlike a classical computer that works on 1s and 0s, a quantum computer works on quantum states and phenomena. It works on not bits but qubits. A qubit can be realized via the polarization of a single photon, or electron spin, or currents in a superconductor, experts tell. But what makes these qubits unique is that while they are being measured, they can exist as 0 and 1 simultaneously. This leads to the ability of a quantum computer to perform massive calculations, to tap quantum entanglement, to use quantum superimposition and many such quantum capabilities.

These muscles promise a lot of disruption and roof-breaking scales in areas where classical computers always used to hit a glass ceiling. Like drug discovery, climate research, genome work, space exploration, cryptography and what not. No wonder, top guns like Google, IBM, Microsoft-PsiQuantum etc. are vying for quantum supremacy.

However, it will take some time before quantum computers can become mainstream. And there are lot of reasons for that—the chief one being affordability.

Quantum is Heavy – on the pocket, too

Shahin Khan, Founding Partner and Analyst, OrionX reminds us of the saying in quantum computing—that it is simultaneously overhyped and underestimated. “While there is significant progress across the board, the technology is still at its infancy.” He picks out four main challenges to consider here. “Quantum hardware systems with sufficient scale and fidelity are not yet available. Programming quantum computers is difficult in general and too specific to each technology approach/modality. New software abstraction layers are becoming available and can separate coding from the ‘voodoo factor’ of the underlying quantum science.”

And we still need to get past the hype to the hyphens that matter.

CChirag Dekate, VP Analyst at Gartner recalls the latest Hype Cycle for Compute 2022 from Gartner and opines that quantum computing continues to be a peak hype technology, where hype exceeds capability today. “Gartner’s current estimate is that quantum computing will likely take more than 10 years before it achieves ‘plateau of productivity’. In his perspective, the three primary inhibitors for practical adoption are as follows. The scale of quantum systems—which continues to be in the 10’s or 100’s of qubits. Then there is lack of virtuous business value cycles, given limited understanding of where quantum technologies can be applied and lack of business value from implementations. And then there is also limited enterprise maturity around quantum, which often requires CXO sponsorship, and commitment to organizational, skills and long-term investment cycles.”

“we have installed IBM Quantum System One systems in Germany, and Japan, with installations in progress in Korea, Canada, and in the U.S. at Cleveland Clinic.”

– L Venkata Subramaniam, Sr. Manager – AI Science and IBM Quantum Ambassador, IBM Research India.

Khan also points out how programmers want to focus on what happens vs. how it happens. “Applications in materials research and optimization, as well as early examples of Machine Learning, look promising but we expect getting data into and out of quantum computers will remain a challenge and will inhibit adoption.”

Also the rationale to switch to quantum computers can get weak when quantum-inspired algorithms often perform so well in a classical computer as to reduce the need for a quantum computer. “Quantum sensing (measurements) and quantum communication will likely become available before quantum computing,” Khan quips.

Apart from these issues, quantum computers have other wrinkles to iron out too. The power in qubit calculations increases exponentially unlike the 1:1 ratio of transistors. Also, quantum computers have high error rates and have to be kept in cold environments. Even the slightest disturbance in temperatures can shake up the qubits and ruin the ‘coherence’. Using error correction codes, further, necessitates a high scale of qubits, leading to another affordability set-back. They need isolated facilities for hitting the desired levels of efficiency. Maintaining quantum mechanical state is a delicate thing and quite challenging from a commercialization angle.

One of the major barriers to quantum computing reaching enterprise readiness and affordability is to keep the equipment extremely cool and constant, seconds Keshav Kumar, General Manager - DC Presales at Rahi. “Even today’s advanced infrastructure still struggles to process the data at a necessary pace. While quantum computing can process data at a much faster rate, they are prone to environmental disturbances. A slight change in their environment can affect the qubits.”

“QCaaS enables enterprises to evaluate diverse quantum systems at affordable price points without the overhead of sourcing and procuring a quantum system on premises.”

– Chirag Dekate, VP Analyst, Gartner

Kumar also expands on the noise-hassle and the square-peg-in-a-round-hole problem. “Factors such as temperature and dust can drastically alter quantum computing systems. Due to their fragility, the amount of time the qubit can store data before succumbing to environmental factors (also known as decoherence) is less than a minute. Also, most of the present quantum computing is limited to solving only intractable problems. Not only that, but they fail to save data for more than a year, rendering them useless for many present-day applications.”

Of course, it would be myopic to miss out on the skills and literacy aspects here. L Venkata Subramaniam – Sr. Manager – AI Science and IBM Quantum Ambassador, IBM Research India notes that the greatest challenge for quantum computing to reach enterprise-readiness is education. While quantum computing is still a nascent technology, and there is hardware and software progress to be made before the industry develops applications with a quantum advantage—individuals and entire industries can begin learning about and exploring use cases, today.”

It took 60 years to abstract software in classical computing to the point where users could input a simple line of code into a templated program to build an app or website. Quantum computing is going through a similar process, assures Subramaniam.

And looks like we are getting there.

Breaking the Quantum Dead-end

In an age where our smartphones are as, and if not more, powerful than giant supercomputers sitting in gingerbread facilities; it would be silly to give up on affordability. With time, efforts and lessons learnt from failures – this mountain can also be split into pocket-friendly rocks.

Subramaniam gives a peek into the progress happening in this tunnel. “At IBM, we released a quantum roadmap to signal how we plan to move from today’s small-scale quantum devices to thousand-plus qubit devices that will enable quantum computers to solve hard scientific and business problems. And so far, we have delivered every single goal on time, including the debut of our first 127-qubit processor last fall. And in May of this year, we extended our quantum roadmap even further to clearly lay out how they will blaze a path towards frictionless quantum computing.”

Some quantum computing technology approaches, most notably photonics, can leverage high-volume optics components, and can operate almost entirely at room temperature. These factors reduce capital and operating costs, hopes Khan.

Enterprises can also opt for local simulation as some experts have pointed out. But here they would have to limit the size and compromise with the complexity levels one needs. Renting cloud simulators and hardware is another alternative.

Dekate contends that while quantum computing continues to be nascent, many quantum systems vendors including DWave, Rigetti, IBM, Google, IonQ, Pasqal, Xanadu and several others offer Quantum as a Service capability either directly or through partnership with CSPs including Amazon and Microsoft. “In other words, enterprises seeking quantum exploration, do not have to acquire a quantum system on premises, and can instead use Quantum Computing as a service (QCaaS) approach to build skills and organize around quantum opportunities (a recommended Gartner approach). Co-location providers are also starting to offer quantum computing capabilities as part of their offering suite.”

“With a dearth of skilled quantum computing professionals, organizations should start training their professionals to become quantum ready.”

– Keshav Kumar, General Manager - DC Presales, Rahi

As Khan also reckons, when it comes to accessibility, quantum computers, not just simulators or emulators, are routinely available via in a cloud model.

Although diverse vendors offer QCaaS, pricing models vary, with some charging by the hour, and others charging depending on the nature of quantum algorithms used, weighs in Dekate. “Gartner recommends enterprises to start developing organizational strategies around quantum, by centralizing quantum exploration activities, prioritizing use-case pipeline, and engaging diverse QCaaS providers as a means of developing skillsets and implementing early quantum applications. Doing so enables enterprises to develop actionable plans and create the necessary skills pipeline, and vendor/partner relationships that can help develop sustainable quantum value.” Dekate shares some advice wrapped for enterprises serious about tapping this wave.

“For something as disruptive and novel as quantum computing, it’s imperative to start exploring potential use cases. This involves bringing experts in the most important computational challenges facing the business together with experts on the capabilities of quantum computers to explore together ways that quantum algorithms might provide a better solution.”

– Dr. Michele Mosca, Institute for Quantum Computing, University of Waterloo

In Kumar’s opinion, instead of directly delving into quantum computing, organizations should start to integrate basic usage initially that will not have any drastic impact on their daily operations.

Initiatives like the one from Pawsey Pioneer program, Nvidia’s frameworks and The Quantum Brilliance vision are examples of more work underway here.

Kumar illustrates how back in 2019, IBM unveiled the world’s first commercial quantum computer. “Q System one, while it is still out of the reach of many, the quantum computer entry in the market has shown that commercialization is sooner than once predicted. Industry experts predict that quantum computing can become an integral part of the enterprise ecosystem within 10 years.”

IBM seems to be attacking the noise and awareness parts of the Quantum problem-set. “Our expanded roadmap includes our plans to build a 4,000+ qubit processor by next year, along with significant milestones to build an intelligent quantum software orchestration platform that will abstract away the noise and complexity of quantum machines, and allow large and complicated problems to be easily broken apart and solved across a network of quantum and classical systems,” shares Subramaniam.

This vision, he elaborates, will enable both paradigms of computing—classical and quantum—to work seamlessly together to solve different parts of complex computing problems that are best suited for their respective strengths and capabilities. “Once realized, this era of quantum centric supercomputing will open up new, large, and powerful computational spaces for industries globally.”

With all these alternatives and efforts emerging and accelerating around them, enterprises need to keep discovering new lanes on this road.

Organizations must assess the importance of having a seat at the quantum computing table, suggests Khan. “Do they believe it is an imperative for them to be among the first in their industry to take advantage of quantum computers? Are they able to set a compelling mission? Can they assemble and cultivate a team and the associated deep-science culture? Do they have the means and the resolve to stay in it for a number of years? If so, then they must take steps to participate actively.”

These are crucial questions to ponder. They will take time. But with Schrodinger-like courage, they have to be kept on the table. Affordability, meanwhile, is still a cat in the bag.

By Pratima H

pratimah@cybermedia.co.in