As an editor, I often gave my writers this admonition: Never promise something without having established at least one feasible means of delivery. A promise should never be something to which one resorts in the absence of reason, otherwise every marriage would be predicated upon perjury. As a writer, I've learned to omit the word "promise" from my commitments, because editors tend to interpret "I promise" more as a statement of desperation than commitment.
It is in that context that we can confidently assert quantum computing has an abundance of promises. If it ends up working, QC will enable the class of large, complex matrix calculations that comprise the bulk of machine learning work, to be executed in moments rather than days, or moments rather than years — or, with enough data in hand, moments rather than centuries. There is a weirdness in physics that we don't really comprehend, but maybe we don't have to comprehend it to acknowledge it. We may very well have to comprehend it anyway, however, in order to make good use of it.
If this is your first ZDNet Status Report, here's how it works: Here, we systematically evaluate the potential for technologies to exhibit genuine influence in their markets, and in the world at large. For each technology, we look at the same ten categories of influence, and rate the progressive and regressive potentials separately on positive-10 and negative-10-point scales, respectively. Then on a 2D Cartesian chart, we give each category its own compass direction, or vector, plot the position of each influence point on that vector, and compute the geometric average location of all points. The distance of that average point from dead center is our final influence score.
We'll begin by summarizing the three main points that our evaluation has made evident:
1) You tend to lose confidence in the goal when someone keeps moving the posts
The quantum computing race, as my ZDNet colleague Daphne Leprince-Ringuet first presented it, has been a global, collaborative, mostly academic effort to build a quantum-mechanical device that not only outperforms classical computers at real-world tasks, but once brought online, is more practical than classical for everyday use with those tasks. The original name for this goal was "quantum supremacy."
Never set a goal for server-based computing and expect Google not to claim it. In October 2019, Google researchers claimed to have reasonably sustained a 53-qubit stack, beating IBM to that number, and planting the first supremacy flag. Other researchers then successfully argued that "reasonably" here is very subjective. Besides, the naysayers added, the problem they solved wasn't a "real problem." In February 2021, Google claimed supremacy again, this time with quantum annealing machine maker D-Link. And again, the cold water taps are turned up to full blast.
One alternative metric first proposed late last year was something called quantum practicality, which aimed to maintain a narrow focus on jobs that humans would actually want done. But then in February, even that goal started to sink backwards. A prominent researcher in the field suggested that, the more one makes an effort to nail down what "practicality" is, the number of coherent qubits needed to meet that objective rises, perhaps exponentially.
Which leads one to rightly ask, why set a business goal for oneself that, by design, cannot be achieved?
2) What Y2K taught us after all
The IT security industry was transformed by an event that actually never took place, and perhaps didn't have to. In 1999, the fear was that any computer system whose interpretation of time was based on a medium-length format with a two-digit year, led people purported to be experts to predict global catastrophes. For example, pondered a Senate subcommittee after hearing so-called expert testimony, date-sensitive microprocessors in life support systems could cause them to shut off during the first second of 2000.
It was the preparation for that non-event that transformed the cybersecurity industry. Today, much of it still relies upon the prediction (some would say the promise) of a dire threat. During last October's Inside Quantum Technology Europe conference, attended mainly by academics and physicists, one astonishing topic to emerge on the docket was, "The Threat of Quantum Systems." Modern-day cryptography would become obsolete like the flip of a switch, it was alleged, the moment the quantum odometers turned over, and QC became reliable enough for everyday use.
This was the very opposite of preaching to the choir — more like preaching abstinence at a brothel. But the presentation's very existence was an indication of some telling facts. One was how a kind of "anti-quantum" subsector of the security industry has already emerged, and that its marketing message was already well-defined. Secondly, though, was that the "pro-quantum" majority found themselves blindsided by the topic being raised in their company — a suggestion that they never took into account the repercussions of their own activities. One wonders whether their digital wristwatches may be set to 1921.
3) Science demands a level of frankness that is unusual for a tech industry
When science gives rise to a new business model, it can only evolve as fast as the pace of scientific discovery. Scientists are, by nature, honest. When the pace of discovery is slow, they say so. They share their own feelings, and validate the feelings their fellow collaborators and even their customers have. Good marketing, by contrast, ensures that the "customer experience" is both positive and predetermined.
One reason for the indeterminate state of consumer and investor confidence in QC is that its leaders share their impressions outright. When they don't know whether they're moving closer to their goals, or instead whether those goals are moving away from them faster, they say so, remarkably candidly. Businesspeople often prefer good news, even when it's fake.
It's very important here, however, to distinguish between the open source business model and the scientific computing model. An open source foundation is quite capable of maintaining the appearance of progress, of innovation, and of goodwill, smearing a thick layer of creamy frosting over the many imperfections and fault lines that emerge whenever multiple groups of human beings collaborate. But a scientific institution loves open debate. And debate, by design, implies the absence of resolution.
Sphere of influence
Now let's look at each category of influence that plays a role in our final score:
There's a fundamental difference between empowerment and enrichment, which this series tries to take into account. You can make a profit from a service or even an idea, but can you use that enrichment to build something new of substance or value? This is one area (among several) where QC presently falls short. Quantum functionality is failing to spawn real-world use cases, just one of which could justify its entire existence. [+3] What's more, in recent months, QC has managed to spawn significant competitive, classical computing efforts in both so-called post-quantum cryptography and in matrix-driven machine learning, all of which are, to some degree, diminishing the urgency of quantum development. [-5, net -2]
Does being involved in the QC industry make any organization more competitive, giving it a method or a technique to which others may not have access? About the only original concepts a QC producer is currently capable of providing, is marketing. [+2] Academia is largely responsible for quantum research, which is nice if you can build a business model purely around the fruits of cooperative labor. That's nice for farming, but not much else. [-7, net -5]
As with any intriguing new idea in technology, quantum computing has triggered the creation of many startup companies. And since World War II, it isn't always the objective of every startup to become "the next IBM," but instead to become part of a company like IBM's or Oracle's or Cisco's broader portfolio. Sustainability is a measure of whether a business built around a technology has a potential for long-term, steady growth, even as a component or an absorbed business unit of some other company.
What's required for sustainability is a cyclical, repeatable formula for business success. As of now, QC does not have such a formula, although many of its ingredients are still simmering. [+3] This does not mean it can't yet be developed. But in the intervening period, the fact that QC's long-term success is dependent upon at least one, perhaps more, world-changing scientific breakthroughs, gives competitive, classical technologies time to build their own, defensive business models. [-4, net -1]
As I've stated before in Status Report, when an organization forms its business model around a product, service, or even an idea it doesn't outright own, all its stakeholders have an obligation to evolve it, or watch it perish. Quantum computing is too lucrative to be owned by any single organization — not in today's business climate. In another era, perhaps IBM or Sperry or GE might have laid claim to its underlying physics. Today, so much depends on the tireless efforts of academics, whose level of success is directly proportional with their degree of collaboration. The incentive to evolve QC is most of its entire story. [+8]
But in the intervening period before QC reaches a state of reasonable practicality, there are active efforts to build high-performance, science-oriented computing services — evolving them to eventually dampen the demand for what quantum services could offer. [-4, net +8]
This is one of the key metrics that determines whether QC's final net influence score is positive or negative. There is a kind of market for quantum services, but just marginally. [+2] Today, its customers appear to be limited to organizations willing to participate in experiments. You know a market is underdeveloped when "standby" markets emerge on the side — in this instance, Noisy Intermediate-Scale Quantum (NISQ) — to entertain future customers until a real-world use case does emerge. Until any single provider breaks the sustainability logjam and sets an example for others, the market for advanced services such as post-quantum cryptography (PQC) is actually advancing more rapidly than the growth of QC itself. [-4, net -2]
For a technology to score highly in this category, it needs to be producing a defined product or service, on a regular cycle, that a clearly defined plurality of customers genuinely need, and hopefully also want. We don't really know, as of this moment in time, what a quantum computing service will do. We have promises of its capacity, and estimates of its capability, but nothing defined that addresses customer needs. [+1] A similar situation befell the early stages of the emerging "serverless computing" industry, where services were defined early on with phrases such as "lambda functions," that may have resonated with a small niche of developers, but not with the people who would actually pay fees to use these functions. Meanwhile, the longer this technology fails to meet its objectives, the more opportunity it has to fall into what I'll call the "5G Trap:" an extended period of under-fulfillment. [-6, net -5]
What passes today for a QC industry does employ academics, engineers, and programmers. They're not yet contributing to gross domestic product, but they are making positive contributions to the global workforce. [+3] The "anti-quantum market," if we can call it that, comprised of NISQ applications and defenses against the "quantum threat," is not only getting its act together faster, but has a reasonable chance of sustaining itself should the QC prospect collapse entirely. [-4, net -1]
Quantum computing has become a topic of significant curiosity, but the momentum may be slowing. The public discourse today is clogged with too many topics that aren't substantive, and facts that are imaginary if not hallucinations, for science to produce something else incapable of holding water. [+2] The public isn't tired of QC yet, but 2021 is still young. [-1, net +1]
What we do learn from QC research, whether or not it becomes an industry, will have a lasting and permanent impact on how we frame the problems we have yet to solve. The algorithm is an art form. It is math in motion. So much of artificial intelligence steers its users away from the algorithms at the heart of its processes, as though it were a radioactive core requiring shielding from living things. Quantum has reinvigorated some old and atrophying muscles in our minds, and any feeling of rebellion the "quantum threat" stirs up, can't quite compete with this. [+7 | -2, net +5]
By definition, an ecosystem produces resources for its own sustenance. The academic community is best, among any group of qualified and experienced people, at assembling intellectual talent. The people needed for a QC ecosystem are all here. [+4] A reason for them to stay there for too long after, say, 2023 — something they could provide to each other to keep their group together and focused — has yet to be determined. And there isn't really anything yet to guarantee that once even the lowest "quantum practicality" bar has been cleared, the stuff of which ecosystems are made will magically manifest itself. [-2, net +2]
Final score: -1.23
What determines whether a final influence score is positive or negative is which half of the graph its average falls on — respectively, the top or bottom. At one level, this might not seem fair. If the net difference in just one of the influence scores tips from negative to positive, the final score itself could flip signs. A strong negative becomes a strong positive with one recalculation. In some cases, a positive drift could send the final score down, not up. A good mathematician might call this system a fallacy on my part.
But a good historian would spot an analogous trend in the everyday world: Our best intentions work against us, when they fail to work concertedly and with a coordinated plan. Among other influence scores, the absolute value of quantum computing's final score is relatively strong. And because its average falls just below the Y-axis line, it's just barely negative. This is indicative of a real-world phenomenon: If the QC industry were to devote less effort to building an academic collaboration team than to coming up with a single, working use case that customers could actually want, and that could justify a QC business model, that sign could indeed flip.
Maybe simpler, less spectacular, use cases aren't as fantastic as self-driving cars. But the last snooze alarm has sounded for quantum computing. It's time for the big girls and boys to hop on out of bed now, out of their jammies and into their britches, get their hands dirty, and build something lasting and permanent. The time for promises expired at "00."