This decade we have heard much about how, in the foreseeable future, quantum computing will allow us to accurately model natural processes, spot hidden correlations in data, and reduce some calculations to minutes that would have taken years on the most advanced supercomputer. All good news, once we can produce reliable, fault-tolerant machines at scale.
Separately, we have also heard how other quantum technologies will become commercially viable more quickly, in areas such as timing, communications, sensors, and more. I have explored both topics in my previous two reports, here and here. Life Sciences, Healthcare, the Materials sector, Financial Services, Defence, and Security all stand to benefit.
So, it is unusual to hear that quantum computing may cause cognitive drag, productivity losses, and long-term financial problems in the years ahead, rather than lead us into the sunlit uplands of a classical-plus-quantum world. Yet this is the view of one of the UK’s leading academics.
To reap the benefits of this new, complementary computing model will demand a completely different type of transformation to the digital one that remains ongoing in most businesses. And because the binary, digital, silicon-based world will still exist, and probably dominate long into the future, there is little guarantee that such a transformation can take place at sufficient scale to avoid becoming a drag on those businesses that pursue it.
Incremental gains
This challenging perspective comes from Professor Chander Velu, who is Professor of Innovation and Economics at the UK’s University of Cambridge. He explains:
It requires a quite significant transformation of business models and core processes in order for productivity to really take off. And we believe that, as quantum computers come onboard, there is a significant chance that adeclinein productivity will occur, because firms will need to adapt their core processes and business models. And we believe that this decline could be even larger than that of digital computing [in its early days].
So, why is that the case? He says:
There are several reasons for this. One is what we are calling an incremental benefit relative to the cost for most firms. When we talk to firms today, they tend to think about using quantum computers for very marginal applications, based on existing business problems.
But if we compare digital computers to quantum computers, digital computers replaced electromechanical machines. So therefore, there was very little integration cost, per se. But quantum computers will be working in conjunction with, and will be complementary to, digital computers. So, there will be a larger integration cost relative to digital computers when they came onboard.
He adds:
Firms are beginning to use these computers only for incremental gains, i.e. on existing problems where the integration cost is high. As a result, it is likely that productivity will slow down for a period of time, before larger, more radical applications come onboard.
This makes perfect sense. But to expand on Velu’s point with a related one of my own, since the advent of the World Wide Web, mobility, broadband, cloud services, and Wi-Fi, digital computers have also become hugely adaptable, multipurpose devices. At least at the mass-market end of computation, which has its own economies of scale. (We have yet to carry a supercomputer in our pockets, though we might have cloud access to one.)
By contrast, quantum machines are unlikely ever to be such flexible consumer/business devices, and if they are, then certainly not for a great many years.
Also, it is generally accepted that quantum computing will be a separate and complementary model, and should not be regarded as the next stage in evolution after Moore’s Law hits a brick wall – i.e., they are not about doing the same things quicker. As a result, their returns on investment and productivity advantages will rely solely on the emergence of new, often niche, applications. That said, some applications will doubtless be of enormously high value.
However, in general terms this is likely to mean that building new lines of business around them, with skilled staff, will be far more difficult to do and less profitable than with classical computers, except in markets where quantum devices present truly transformative solutions. For example, in Materials Science, Pharmaceuticals, or fraud detection – assuming that some businesses actuallywantto stamp out complex fraud (a subject for another day).
So, what might the answer be to these challenges? Professor Velu says:
To address this, we believe that we need to think about more mission-driven or society-level challenges for quantum computers. Such as, looking at things like weather forecasting or financial services applications, where the problem is industry wide or societal. A proof of concept could then be shown, and private-sector firms could perhaps scale it up.
But there are further problems, he says:
When we look at the second challenge we face, cognitive drag [a steep learning curve], digital computers were relatively easy to understand. However, one of the issues with quantum computers is that they are not; they are not that intuitive. So, the translation cost between scientists, engineers, and business managers will be high.
We need to address this by providing aids to translation, to more shared-language and -meaning artefacts that enable knowledge sharing, so that scientists, engineers, and business managers can use a common language, and translate digital to quantum and back again, so to speak.
A crisis will hit
All valid points. But then there is what Velu calls the “cryptographic cost”.
Back in January this year, I published a diginomica report called ‘Quantum technology – the black swans are gathering, says start-up CEO’. This looked at the problem of so-called ‘black swans’, unexpected events that change the status quo suddenly (but in retrospect seem obvious).
Among the many black-swan events that could affect our attitudes to quantum technology, my piece noted:
Such a crisisisapproaching. We don’t know precisely when it will hit, but we do know what it is – and what will happen if we fail to prepare for it. It is the threat to the global economy – to banking, ecommerce, supply chains, government systems, and everyday communications – that will arrive when quantum computing, or an emulation of it, can reliably and swiftly break the public-key encryption that underpins our secure transactions and communications.
On the surface, this seems like a ‘known known’, as a politician might say. But the black-swan element is the possibility that this might happen much sooner than the decades-long timescale that has been accepted wisdom until recently. Indeed, as my piece explained, there have been unverified claims that strong encryption hasalreadybeen cracked, or could be in theory. Either way, it is far from impossible that a giant leap forward (or backward, depending on your perspective) might be imminent.
In a world in which hostile states, bad actors, opportunists, and even the simply curious might have downloaded reams of encrypted data on the off chance that they will one day be able to read it – the phenomenon of ‘Store Now, Decrypt Later’ (SNDL) – this could be a quantum game-changer of the most brutal and damaging kind.
Or as Professor Velu puts it, it would be a cryptographic cost – potentially an enormous one. He explains:
Most firms will look at the chance of a cryptographic break, and be very defensive initially. And in order to overcome this particular cryptographic cost, we need to think about building a more quantum-enabled network structure – or, as some people call it, the quantum internet.
As explored in my earlier report this week, this has been a policy aim for the British government since publication of the National Quantum Strategy last year, and the subsequent Missions that have emerged from it.
However, the Missions specify a timescale of “by 2035”. A black-swan event could make the slow pace of delivery – pragmatic and realistic though it may be in technical terms – seem absurdly slow. Especially as China has reportedly already built an intercity quantum network, as my recent report explained. Professor Velu says:
We need to bring quantum computing and quantum communications together in order to enable such benefits to come through and overcome some of these cryptographic costs.
So, in summary, there'll be a drag on productivity as a result of quantum technologies being adopted, and we need to think about enabling business model innovation to reduce that productivity decline and bring the benefits sooner.
Both the private sector and public sector need to work together to enable such benefits to bear fruits! Thank you.
My take
Thank you indeed, Professor Velu. A lot to think about; but even more to act on sooner rather than later, it seems.
