MODULE 6 · LESSON 2
Free — no login requiredSign in to track progress, save quiz attempts and enrol in the full course.
Sign in to track progress / enrolShould Your Organisation Care Yet?
Beyond cryptography, most organisations should do very little, and being clear about that is more useful than a list of possibilities.
Three tiers
Tier one, act now. Everyone. The cryptography migration from the previous lesson. This applies to every organisation holding data with a meaningful confidentiality lifetime, which is essentially all of them. It is not optional and it is not contingent on anything.
Tier two, watch and build understanding. Some. Organisations whose core business involves problems in Module 3's short list. Concretely: pharmaceutical and biotechnology companies, chemical and materials manufacturers, some energy businesses, and a small number of financial institutions with genuinely quantum shaped problems rather than general optimisation. For these, a watching brief and some internal capability are proportionate.
Tier three, do nothing beyond tier one. Most. If your organisation does not develop molecules or materials, and your computational problems are the ordinary ones of running a business, quantum computing has no near term application for you. Reading occasionally is sufficient.
The uncomfortable but honest point: most organisations are tier three, and vendors have a strong commercial incentive to persuade you otherwise, because the tier three population is where most of the budget lives.
Testing for a real use case
Three questions, in order.
1. Does your problem appear in the short list? Factoring and discrete logarithms, unstructured search where candidates are computed rather than stored, and simulating quantum systems. If it does not, ask the Module 2 question: what would make the wrong answers cancel? A use case with no answer is not a use case.
2. Is your problem quantum mechanical in nature? The strongest signal. If you are modelling molecular behaviour, chemical reactions, or materials at the electronic level, the machine's structure matches your problem's structure. If you are scheduling deliveries, it does not.
3. Would an exponential speedup actually change your business? Sometimes the answer is no even where a speedup exists. If your simulations currently take a day and a day is acceptable, faster simulation is a convenience rather than a transformation. The interesting cases are where a calculation is currently impossible and would become possible.
If all three point the same way, you are tier two. If not, you are tier three, and being tier three is a perfectly good place to be.
A proportionate watching brief
For tier two organisations, this means less than vendors will propose.
Assign one person a few hours a month. Not a team. Someone technical who tracks logical qubit counts, error rates and results in your specific problem domain, and who can distinguish a benchmark from an application.
Follow your own field rather than the field generally. Progress in quantum chemistry matters to a chemicals company; progress in optimisation heuristics does not. Following everything produces noise.
Build a small amount of hands on capability. One or two people able to write a circuit and run it on a simulator. This is a modest investment, and it converts your assessment of vendor claims from trust into judgement.
Revisit annually. The field moves fast in research and slowly in deployment, so quarterly reviews mostly generate activity without information.
When a pilot is justified
A pilot makes sense when you have a specific problem from the short list, an internal expert who can assess the result, a budget you can lose without regret, and a defined question you are trying to answer.
It is theatre when the goal is to be seen doing something, when nobody internally can evaluate what comes back, when success criteria are undefined, or when the classical comparison is not being run.
That last one is the reliable test. A pilot that does not benchmark against a well optimised classical solver cannot tell you anything, because the only interesting question is whether the quantum approach beats what you already have. Insist on it, and treat reluctance as informative.
If a vendor approaches your organisation, a short set of questions establishes quickly whether the conversation is worth continuing. None require technical expertise, and all come from earlier modules.
"Is your hardware gate based, or is it an annealer?" Establishes what you are actually being offered and whether general quantum algorithm claims apply.
"How many logical qubits does your system have?" The word logical is the important part. Watch how it is handled. A straightforward answer such as none yet, we are working with physical qubits, is honest and a good sign. Deflection into physical qubit counts is not.
"For our problem, what makes the wrong answers cancel?" The Module 2 question. A vendor with a real algorithm will explain the mechanism, often enthusiastically. A vendor without one will restate that superposition allows many possibilities to be considered.
"What did a well optimised classical solver achieve on this problem?" The single most informative question. If they have not run the comparison, the result means nothing. If they have and it is close or worse, appreciate the honesty and note the finding.
"Which of your published results were peer reviewed, and has anyone reproduced them?" Distinguishes research from marketing.
"What happens if I do nothing for three years?" For an honest vendor discussing anything other than cryptography, the answer is usually that little changes. Anyone claiming otherwise is selling urgency rather than technology.
A good vendor will answer all of these comfortably, and there are good vendors doing serious work. The questions are not hostile; they simply require specificity, and specificity is what separates the two categories.
A logistics company is offered a quantum pilot to optimise its delivery routing. What is the most appropriate response?
Tier one
Click to flipMigrate cryptography. Applies to every organisation with long lived data and is not contingent on hardware timelines.
Click to flip backThree tiers. Everyone migrates cryptography. Some organisations, chiefly those working with molecules and materials, deserve a watching brief and a little internal capability. Most should do nothing further, and most vendors have a commercial interest in telling them otherwise. Test a use case by asking whether it is in the short list, whether the problem is genuinely quantum mechanical, and whether an exponential speedup would actually change anything. A pilot without a well optimised classical benchmark is theatre, because beating what you already have is the only question worth asking.