IBM made a big leap towards useful quantum computing - how far did they go?

We at Kipu Quantum believe that quantum computers can be useful in the near future. To contribute to this milestone happening as soon as possible, we are developing our application- and hardware-specific algorithms. We believe we have a shot with hundreds to thousands of qubits assuming incremental advances in the hardware. In our view, the first demonstration and consolidation of usefulness will rely on a two-fold attack, by bringing down the circuit depth requirements with compressed algorithms, while also improving the robustness of quantum processors and therefore their ability to allow for longer circuits.

In particular we bet on #neutralatoms, #iontraps and #superconductors. After the two former platforms dominating the headlines in the past six months, news from IBM about a new study published in #Nature broke & grabbed everybody’s attention this week -

What did they achieve? IBM in collaboration with University of California, Berkeley outperformed a classical computer when investigating a spin model. IBM used all 127 qubits of their Eagle processor to model the spin dynamics of a quantum system. The team made use of error mitigation, ie modelling of errors with the aim to account for their effects, and achieved improvement of the quantum processors calibration, which lead to better coherence.The results were compared against conventional (classical) supercomputers, that have been boosted with tensor network techniques, the non-quantum gold standard for such tasks. IBM Eagle beat the next best classical alternative and produced a much better accuracy.

This is remarkable because of several reasons:

• A new record in size: This record was achieved on a much larger quantum system, 127 qubits, compared to previous similar realisations done by others, ie well beyond 100 qubits
• A leap towards utility-scale processors: the system the team used is much closer to being commercial, in the sense that it is accessible via the cloud to a potentially wide audience, and not just a lab-scale experiment people cannot easily interact with
• A more interesting problem: previous demonstrations of quantum computers beating classical computers focused on so-called „quantum supremacy“, where a classical computer is not able to compete at all for an abstract, non-industrial task. In the case of IBM's feat, a classical benchmark, albeit with worse results, was achievable. What is interesting is that the nature of the tested use case is closer to being industrially and commercially relevant than previous instances of supremacy, on its way to useful material design, which I consider a worthwhile tradeoff.

So what’s next? IBMs results are an important step towards useful quantum computing, but more work is needed. The leading quantum hardware companies need to continue the path of improving their processors, and algorithm players such as us need and will advance use cases from obscure to industrially relevant and refine algorithms to extract the full potential of the quantum processors.

Kipu Quantum is proud to contribute in this effort by providing application- and hardware-specific algorithms. Our digital & digital-analog compression technology has been designed to reach useful quantum computing soon. We as a proud member of the IBM Quantum Network wholeheartedly congratulate the team at IBM (and UC Berkeley) for this achievement.

Well done and let’s all keep working!