How can businesses use quantum computing?

While most of IBM’s clients tend to be larger enterprises and universities at the cutting edge, according to distinguished engineer and quantum ambassador Richard Hopkins, those involved in quantum computing need to reach out to companies of all sizes.

With a working quantum computing model chirping in the background of IBM’s London offices, he tells me the reasons why.

The need to engage with SMEs boils down to four things: the relatively low barrier to entry; use case generation; quantum safety and the need to start working on a code of related ethics and regulations.

As we revealed in Part 1 of this report, in 2016, IBM put quantum computing in the cloud. Hopkins argues that the barrier for entry is now even lower than it ever was for AI.

“If you’re an SME and you want to solve a problem using quantum you will frame it as an optimisation problem, an estimation problem or simulation problem. You will fill out an API query using the tools [an open-source Python-based SDK called Qiskit] and then send it off and you’ll get your result back,” he says.

The problem is that not many people know about this yet – that quantum computing is not just a tool reserved for physicists or financial whizz kids – but it can be used in logistics, architecture, construction, industry 4.0 – anything that requires complex decision making or simulations.

“I’m desperate to see more industry bodies engaging in quantum,” says Hopkins, who adds that the industry engagement layer appears to be missing from most of the national and international strategies on quantum that he’s seen so far.

“The British Computer Society is doing some work in this area, but there’s a lot of education that needs to be done to say: ‘This is quantum; This is where you apply it; and these are the sort of problems it will solve – why don’t you go away and think about whether that will be an advantage for your business?’

Hopkins would also like to see sector-specific organisations engaging to work out what the use cases are and what the implications for each sector might be – which will also impact ethics and compliance issues within these sectors.

“It’s got to take a per sector view because the implications for chemistry or material science will be different to the implications for banking or construction. And you can’t rely on IT providers to work out what those impacts are – how would we know?” Hopkins adds.

The engineer agrees that, yes, once we start to learn about more commercial success stories that come with quantum, then talk of regulation and ethics will inevitably follow – as it did with generative AI and Chat GPT. But he also points out that, in AI’s case, it was already too late.

“We should be working on these kinds of things now so that we get ahead of the problem rather than what happened with AI, where frankly, the conversation overtook that debate,” he notes.

Quantum safety

One issue that has piqued the interest of some sectors has been quantum safety.

Quantum computers can break down very large prime numbers very quickly, which is bad news for organisations that need to store data securely (given that most systems are secured and underpinned by cryptography which uses binary code).

It’s widely considered that in ten or 15 years, a fault-tolerant quantum computer capable of running crypto-analytic algorithms, will threaten to break the security of the internet and mobile networks (where the whole edifice relies on public cryptography).

Some security experts have dubbed the day when a robust quantum computer will be capable of cracking the large prime numbers that underlie our public encryption systems as Q-Day.

As well as IBM, firms such Vodafone and HSBC – as well as telco organisations such as the GSMA – are actively working on promoting awareness as well as trialling quantum safe encryption methods.

“If you’ve got information that’s still going to be valuable and sensitive in 15 years’ time, you’d better start thinking about Q safe encryption now,” advises Hopkins.

“Because if I exfiltrate your data now, even though it’s encrypted at some point in 2035 I might be able to decrypt that using quantum computing.”

According to the engineer, IBM has been working with standards body NIST and others on algorithms that they think will be quantum safe, which will be standardised and released next year. “And we’ve been building that into products,” he adds.

I ask Hopkins if we should be worried about synergies between AI and quantum computing, but he explains that they can’t be fused in a simplistic or technical way – although he adds that fusing the results further down the line might be helpful.

“Not so much with GPT – but take machine learning for fraud classifications, for instance. So, if you have a conventional fraud algorithm, and a quantum one, you could combine them and then look at the problem in two different ways, so the  result is better than the one either technology would give alone,” he suggests.

Skills and training

As our conversation comes to an end, I circle back to the issue of users, and specifically, what skills people will be needed to use quantum computers.

To appreciate this, it helps to understand the difference between how a classical computer solves problems versus how a quantum computer approaches it, because it involves a change in mindset.

With a classical computer, you can set up some variables up, assign values to numbers and letters; run some processes over the top and then print out the results at the end. Everything is procedural.

Quantum computers are probabilistic – each time you run a simulation of a problem, you get an answer that may be correct, but that might not be your immediate answer.

‘Programming’ a quantum computer involves setting up the quantum system so that the correct answer is the one most likely to appear.

But since there’s a chance that any run of the system produces an incorrect answer, obtaining confidence in the answer requires running multiple simulations until the pattern of results makes it statistically clear which answer is the correct one.

While this may not sound impressive, even if the simulation needs to be run dozens or hundreds of times, it has the potential to be much faster than a classical computer at checking trillions of trillions of possible answers.

According to Hopkins, with 100 qubit computers, this now takes about a day, “but over time this will come down.”

So where does this fit in terms of users and skills?

Earlier Hopkins explained how an SME could solve a problem using an API, an SDK and quantum on the cloud. This works well for known problems where the algorithm has already been developed.

For this skill level you will need to know some Python programming. “We’re going to need millions of people with these skills, every developer will need to know how to do this,” Hopkins predicts.

The next level up is solving an optimisation, estimation or simulation problem that has not been solved before, that may be solved by tweaking an existing algorithm in terms of how the information comes in and how you model that at the top.

“For this you need to know a bit about how the algorithm works and to know what that means in terms of the answer coming out of the bottom. We’re going to need hundreds of thousands of people with this knowledge level,” says Hopkins.

The highest skill level is reserved for the people who need the algorithm to do something more complex and very specific. This user will need to know the maths required to make it efficient and the code needed to change the shape of the microwave pulses that go into the machine, manipulate the qubits and change the algorithm.

“That person needs to understand not just physics but also how to interact with individual atoms using microwave signals –and that’s a rarified skillset, there’s probably only a couple of hundred of people who can do that,” he says.

The reassuring thing, Hopkins adds, is that young people coming through universities today appear to be embracing quantum.

“People with less old brains that mine that have learned this tech from scratch and think of it as second nature will be the ones who take this tech forward,” he predicts.

“Because half of it is unlearning what you’ve already learned and re-learning something new. Getting younger people to do this is a lot easier than people my age,” he admits.

The IBM engineer and tech ambassador adds that the enthusiasm is palpable in the encounters he’s had at various hackathons and ‘Cyber First’ programme events at universities.

“They adore it. There’s always a queue outside the door of people wanting to know how they can become quantum users.

“It’s absolutely captured their imagination and rightly so – because it will be theirs. It’s a largely unexplored area and they know that there’s gold to be found in them there hills if they can come up with a really good idea or a new way of doing things.”

Missed the first part of this report? Click here