…it’s good that there is some optimism, because we’re seeing encouraging signs – though incremental and certainly not yet exponential – of progress across the industry. And incremental signs of progress are just that: progress. Iterative, incremental progress. It’s not a light switch. Technology development is almost always more analogous to a dimmer than a switch.
Christopher Savoie, CEO & Founder
Hype has proven to be one of the evergreen issues in quantum computing. While many of the opinion pieces that have published over the last few years take a mostly negative view of what they perceive as industry-led hype, Zapata’s position is more nuanced.
In our experience, instances of industry hype are situational. Some claims are clearly attention-seeking and strain credulity, while others are supported by data and scientific rigor. And many others are somewhere in between.
Healthy debate – such as we engage in at Zapata when encountering industry claims – is our go-to response internally. That said, because we are often asked to weigh in on hype broadly and in response to individual claims, we want to communicate our opinion in this blog post by surfacing the most common hype-related questions we are asked.
It seems that there is by default a negative moral judgement attached to the word “hype,” or whatever the statements being made about quantum computing are which are not 100% scientifically precise and err on the side of optimism.
Many of the “anti-hype” statements about quantum computing are strawman arguments that reduce any semblance of optimism about quantum to all-or-nothing viewpoints where either quantum computers can change everything and we can prove it all mathematically, like Shor’s algorithm, or we are totally clueless about quantum but are simply pretending that we know what we are talking about.
This may be the reality of many in the field, but it is not the reality that Zapata is seeing. We believe there is a middle ground where quantum computers can change some things, but not everything — and definitely not all quantum advantages will be rigorously provable like many would hope. Additionally, we are not clueless about where to find insights on why and how quantum computers can outperform classical algorithms.
In our view, it is because of this hype — or optimism, or whatever you want to call it – that more people are drawn to the field. That means more talented people who, in turn, can advance the real science and engineering.
And it’s good that there is some optimism, because we’re seeing encouraging signs – though incremental and certainly not yet exponential — of progress across the industry. And incremental signs of progress are just that: progress. Iterative, incremental progress. It’s not a light switch. Technology development is almost always more analogous to a dimmer than a switch.
Specifically, today we’re seeing that quantum-inspired approaches can, in some cases, achieve a one or two percent improvement in a process or output. And sometimes even more. It is critical to appreciate that in the commercial context, a single digit percentage improvement can be huge. Think of improving the logistical performance of a fleet of vehicles by one or two percent. That would be pretty incredible, right? That could translate into millions or tens of millions of dollars and tons and tons less carbon emitted into our atmosphere. That’s very real and very meaningful progress.
It’s hard to not get excited about the potential of quantum computing. We read regularly about impossible problems solved in minutes by a quantum computer that would take a classical computer hundreds or even thousands of years. But are we actually “there” yet? No. But we do not need for it to be that much better for quantum computing to start to be extremely impactful.
Machine learning is the most likely use case to deliver near-term value for businesses interested in getting quantum-ready. This is because there are areas where classical machine learning struggles – such as generative modeling and with combinatorial optimization tasks – that could be better suited for quantum devices.
We also know that many enterprises are already investing heavily in machine learning and data analytics problems, and bringing in quantum computing to extend that investment makes sense because enterprises already have the ML talent, algorithms and applications in place.
Although quantum advantage, at least in the commercial sense, is not yet reality, we do have a “real” tenable path to where we can show whether or not quantum-inspired models can outperform state-of-the-art machine learning models. In fact, we are seeing empirical evidence of separation between machine learning models using purely classical statistics and those using post-classical statistics. But — and this is important — not yet in enterprise-scale deployments.
Speaking of enterprise deployments…
Even if we have a perfect, fault-tolerant quantum computer tomorrow, we are not going to have a billion-dollar industry overnight. A lot of “peripheral work” on infrastructure, data, analytics, workforce, compliance (among other areas) — in addition to scientific research — needs to be done before we can plug in a quantum computer and start seeing tangible business results. For global Fortune 100 enterprises, this peripheral work spans massive initiatives such as digital transformation, data transformation, AI/ML transformation, etc. These are initiatives that can happen over a timeline of decades.
This timeline is comparable with the currently believed timeline for fault-tolerant quantum computers. So, that is why we are always saying the time to get quantum-ready is not at some vague point in the near-term, it’s right now.
At Zapata, we strive for a balance of optimism and reality across the quantum computing landscape. We would argue that the smart money says that in five to 10 years we should expect to see some form of fully fault-tolerant quantum computation available – opening up the door to new and powerful commercial quantum applications. Much of quantum computing’s future will be heavily – but not totally – dependent on breakthroughs in the hardware’s stability and reliability as we get closer to fault-tolerant devices.
Frankly, it’s amazing how quickly we’re seeing new achievements from the quantum hardware providers in the space – and many of the world’s brightest minds in technology, academia, business and the government are pushing their chips to the center of the quantum computing table.
If I had to think about my wish list for quantum computing achievements in the next ten years, at the top would be quantum computing hardware and software being used in production to solve classically unsolvable real-world problems. To be honest, I think that will happen well before another decade passes, though it won’t apply to every difficult problem.
Some seemingly intractable problems, such as climate change, are so multi-faceted and complex that even quantum computers will not “solve” them in the foreseeable future. But others, such as supply chain optimization and financial portfolio optimization, will see significant progress from where we are today. Zapata is deep into real world work on both of these challenges with customers, so I have confidence saying they will be in production in less than 10 years.
Sharing our understanding of the current state of AI and quantum.