The car, the horse, and the quantum computer

It’s tempting to start with “a car, a horse and a quantum computer walk into a bar…” but this post is about a more serious subject.

A couple of weeks ago, we were briefing an industry analyst and they said something to the effect of “all the companies that we see right now are ‘playing’ with quantum computers. They take pre-existing data, try to run it through a quantum algorithm and compare the results to what they got on the classical computer. How long before quantum actually does something useful?”

Which got us thinking about horses and cars. The first cars were inferior to horses. These first cars were slower, had a shorter range, were much less reliable, and made more noise. But several entrepreneurs, scientists, and engineers saw the merit in the “horseless carriage” and continued its development. Today, of course, we may take a horse-carriage ride in Central Park, but the car is a much more dominant form of transportation than a horse.

Quantum computers today are not very useful, mostly because they have a small number of noisy qubits. The vast majority of software written for today’s quantum computers can be simulated on a classical computer, and classical computers are much more available, stable, and convenient to work with.

But when quantum computers reach hundreds or thousands of qubits, they can no longer be simulated by classical machines. This is when quantum computing can start delivering real business value, not just ‘simulated business value’. 

How soon will this happen? If you look at the hardware roadmaps of key vendors like IBM or  Honeywell, the answer is “quite soon”, as soon as 1-2 years.

Of course, these higher-end computers will require software because the current way of developing software for quantum machines does not scale to larger machines. But assuming software such as Classiq’s is available for these machines, the car will start overtaking the horse in the very near future.