In celebrating Google’s quantum achievement, we are at least celebrating the spirit of experimentation. We may also be celebrating the demonstration of something that cannot be done in any practical sense by a classic computer architecture. Finally, we may be celebrating everything that was learned in the journey to that achievement, including how to achieve low error rates, by quantum computing standards.
Recap, recently a whitepaper was leaked/appeared on a NASA website before being taken down. The whitepaper describes the completion of a task in 4 minutes, that would take a classical architecture supercomputer 10,000 years. Because the paper characterizes this achievement as “quantum supremacy”, the achievement has received both positive and negative analysis. Critics draw attention to the quantum circuit being custom, and the algorithm being of little real-world value: validating that a random number is a random number.
Many valuable circuits and software modules have been custom; we have only to consider the first integrated circuit, and what has come since then. Many invaluable algorithms and proofs have been steppingstones to much more important achievements. Maybe the title “Quantum supremacy using a programmable superconducting processor” was a stretch if a custom circuit was used, but I’ll leave that to the really smart people to fight over.
Science, innovation, and discovery is a journey. A journey that often takes 10,000 wrong paths before it takes the “right” path. We celebrate the achievements along that journey because we know just how important experimentation is, how little achievements accumulate, and how one day you look back on a larger, not foreseen achievement. We have only to ponder that fact that graphical processing units were not designed for AI, and yet have had such a profound impact, to imagine in awe what surprises the future of quantum computing may hold.
We can get wrapped around the axle on whether a concept only defined in 2012 has been realized, or we can simply say, thank you Google, next! I would propose we do that latter. The important is quantum computers are growing in qubit size and also reducing in error rate. Both are important to the realization of practical, real-world quantum computing that will initially augment classical computing but may in the far-off future supplant it completely. In a world of augmented classical computing, custom quantum circuits may be sufficient to provide enormous value.