Quantum Advantage? I’m Still Asking: Advantage at What Cost?
At a Q&A session for my book last month, someone challenged me on quantum advantage. I have thought about that exchange since, and I still stand by the principle behind my answer, although I would phrase it rather more carefully today.
Demonstrating that a quantum computer can outperform a classical computer on a particular computational task is one thing. Demonstrating commercial quantum advantage is quite another.
For me, the calculation has to include the whole machine: hardware, energy, cooling, control electronics, error management and, importantly, the classical computation surrounding the quantum processor. Researchers have similarly argued that practical quantum advantage should consider factors such as size, weight, power consumption and cost, rather than computational speed alone.
And this has led my team back somewhere rather wonderfully fundamental: logic gates.
We are exploring a deceptively simple question:
How can classical Boolean functions be systematically transformed into efficient reversible quantum circuits, and what determines the quantum resource cost of doing so?
Or, expressed more provocatively (which is what i like to do):
Can better classical logic make quantum algorithms cheaper?
That means examining Boolean logic, reversible computation, quantum gates, circuit depth, ancillary qubits and the cost involved in translating an ordinary computational problem into something a quantum processor can actually execute.
But underneath this lies a misconception about quantum computing that I particularly want to challenge.
Quantum computing is not an all-or-nothing replacement for classical computing.
The interesting future is hybrid.
The problem then becomes almost one of computational triage: given a complex workload, which parts should remain on conventional processors, which genuinely benefit from a quantum processor, and what is the computational cost of passing between the two?
That, for me, is where the really interesting algorithmic work begins.
Perhaps the smartest quantum algorithm of the future will not be the one that does everything quantumly, but the one that knows precisely when not to.
Photo: my team taking what we have built on a test run.
