D-Wave Claims Quantum Supremacy

Quantum computing stock market values are up again, as D-Wave got a paper published in AAAS Science, the top American science journal, claiming quantum supremacy. But SciAm reports:

Loud declarations of various types of quantum advantage aren’t new: Google notably made the first such claim in 2019, and IBM made another in 2023, for example. But these announcements and others were ultimately refuted by outside researchers who used clever classical computing techniques to achieve similar performance. In D-Wave’s case, some of the refutations came even before the Science paper’s publication, as other teams responded to a preliminary report of the work that appeared on the preprint server arXiv.org in March 2024. One preprint study, submitted to arXiv.org on March 7, demonstrated similar calculations using just two hours of processing time on an ordinary laptop. A second preprint study from a different team, submitted on March 11, showed how a calculation that D-Wave’s paper purported would require centuries of supercomputing time could be accomplished in just a few days with far less computational resources.

There is also a lot of skepticism about Microsoft's claim of a topological qubit.

Gil Kalai has not conceded, and has doubled down with his Quantum Computing Skepticism.

Let's review the arguments in favor of quantum supremacy. The most common one is that qubits can be 0 and 1 at the same time, just as Schrodinger's Cat can be alive and dead simultaneously. Operations on qubit are thus able to examine an exponential number of possibilities at the same time, leading to an exponential speedup in computation.

Scott Aaronson says that this is wrong, because it misleadingly predicts an exponential speedup where none is possible. Instead he says the speedup comes from negative probabilities.

The QM probabilities are never negative. That is just his way of making destructive wave interference sound mysterious. When you say a computational speedup comes from wave interference, it is harder to understand.

Aaronson falls back on the argument that it is up to the skeptic to prove that quantum computers are impossible, and that would be very interesting, but no one has done that.

The many-worlds folks say that the speedup comes from computation being done in parallel universes. Most people say that there is no way to observe those parallel universes, but we are supposed to believe that they speed up computations somehow.

Feynman's original argumen was that simulating QM can be exponentially slow, so it can be faster by running a quantum experiment. You can do a chemical reaction faster than you can simulate it from first QM principles. Okay, that is true, but it is a big leap to using QM to factor large integers.

Finally, there is the argument that quantum researchers have made so much progress already. Yes, but maybe it is like slimbing trees to make progress towards going to the Moon. There is progress, but the goals seem as far away as ever. Nobody has a convincing experiment showing that quantum computing is possible.