Clabby Analytics is now a believer in the commercial viability of quantum computing. IBM’s quantum computing program has progressed a lot further and more rapidly than we expected. Last week IBM updated the research analyst community on its quantum computing strategy. The company described its goal: to reach “Quantum Advantage,” a state where industry problems are solved using quantum circuits that cannot be replicated using traditional computing systems, thus delivering a unique advantage.
IBM further discussed its quantum computing development environment, demonstrating how application programming interfaces and visual management tools have been created, updated, upgraded and integrated with its quantum offerings, including the “IBM Q Experience,” and the “IBM Q Network.” The company’s key point was that it has created a complete environment and software stack for its quantum systems – thus giving developers an integrated toolset for visualizing and building quantum applications.
At the conclusion of this briefing, after considering all the improvements the company has made to its quantum stack, we were left wondering if IBM has reached a new inflection point in quantum computing. What we saw was beta code that provided visual insights into coding for a quantum computer; a highly integrated cloud-based Qiskit development environment; Jupyter Notebook facilities that maintain system configurations and help track progress; and graphical tools that provide deep insights into system behavior and circuit design.
With all of this design and integration work, we were left with the distinct impression that IBM has turned a corner in quantum computing, delivering a fully integrated stack of quantum development and management on real quantum computers (not simulators). These new tools and utilities should make it far easier to improve and extend project development on quantum systems. They also sent a clear message that IBM is “ready-for-business” in the quantum computing world.
IBM’s quantum computing effort to date has been largely experiential and educational. One-hundred-and-twenty thousand people have registered to explore IBM’s Q Experience environment, running their experiments on two 5 Qbit systems and/or a 14 Qbit system. Two-hundred-thousand corporate early adopters, researchers, experimentalists, educators, and students have downloaded the Qiskit development environment that drives quantum computers.
For more serious experimenters, IBM created the IBM Q Network which allows businesses and scientists (who pay a membership fee) to accelerate quantum development on larger systems (20 Qbits.) JP Morgan Chase, Daimler AG, Samsung, ExxonMobil, and JSR Corporation are but a few of the large companies that have signed on to use IBM’s Q Network.
IBM put its first quantum computer on the cloud in 2016, making that computer (and subsequently others) available to the company's clients and the public at that juncture. The company then released open specifications for its quantum cloud, including a slew of application program interfaces (APIs) to simplify development. This effort was followed by the evolving quantum community publishing over 150 third-party papers that shared research results, experiments, and learning methods with IBM Q Experience users. As of today, over 10 million experiments have been run on IBM Q Experience hardware.
First things first: An important insight
Early-on in the analyst briefing, IBM emphasized that “circuits” are the main building block of quantum computing. (A circuit can be created by stringing quantum gates together to perform a computing function. Creating unique problem-solving circuits that can’t be emulated using traditional computers will be the way that nirvana – the Quantum Advantage – will be reached.)
What we think this means from a system/stack/tools perspective is that IBM is planning to strongly focus on building tools that help developers understand circuit behavior. To do this, we expect IBM to build a slew of additional tools and utilities that will visually show circuit representations – and aid in troubleshooting circuits.
In short: when thinking about how quantum computing differs from traditional computing: think “circuits.”
The new news
With IBM Q hardware now readily available, the next order of business for IBM has been to build out the software stack that drives quantum computers. Primary emphasis has been placed on enriching the quantum computing open source Qiskit development environment; introducing APIs; allowing for greater abstraction of the underlying system; graphical management tools to better understand system behavior, and the use of microservices and Docker containers to aid scalability.
The basic software stack on IBM’s Q Experience has been in place for several years. But with the demonstration of new beta software at our analyst briefing, major improvements were evident. IBM introduced an API refactorization that allows optimized services to be created across open quantum systems as well as across the company's premium IBM Q Network systems. New job data management facilities now enable jobs of up to 2 TB of memory to be run. Qiskit extensions enable transpiler circuit remapping, aiding in the deployment of circuits. A new translator has been added to enable Qasm to Qobj translations and includes Qobj validation. And Qiskit also supports three circuit level APIs that can be optimized to a dedicated system.
On another front, microservices have been "Dockerized" for greater scalability. The Jupyter Hub (notebook facility) enables cloud configurations to be saved. And management tools enable realtime connections, notifications, and richer error messages. All of these improvements simplify the configuration, deployment, and management of quantum applications – and this simplification should lead to more aggressive quantum applications development.
The most impressive “new” tool that IBM introduced was its Circuit Composer (this is a re-spin of IBM’s Graphics Composer that was introduced a few years ago.) This new tool has been greatly improved from a graphics perspective, providing a visual representation of how the quantum state of a given quantum computer has changed. It allows developers to create more complex circuits and map new logical circuits to a quantum chip. It aids in troubleshooting. And it has been integrated with Qiskit which assists in this mapping.
In the 1980s, two scientists theorized that if a quantum computer could be designed, it would be able to simulate and model problems that traditional systems could not. There are now great expectations that quantum computing may eventually deliver solutions in medicine and materials (by untangling the complexity of molecular and chemical interactions); in supply chain/logistics (by optimizing fleet operations and logistics); in financial services (by modeling financial data and identifying global risk factors); and in artificial intelligence.
To date, Clabby Analytics has published zero/zed research reports on quantum computing. We’ve been to trade shows and seen the hardware; we’ve listened to numerous marketing pitches; we’ve heard the technical pitches, and we’ve discussed the future of quantum computing with other research analysts – but we’ve been reluctant to publish any research in this field due to the immaturity of the quantum computing ecosystem. We have been skeptics, but we are no longer. With IBM making available a complete and robust quantum systems software/management stack, and especially with maturing graphical development and management tools, we are changing our position on quantum computing. With quantum hardware now supported by IBM in a complete, nicely integrated software stack and development tools, we think quantum computing has reached an inflection point and is now on its way to becoming commercially viable.