Peter Gutmann and Stephan Neuhaus have a recent article—I presume it’s recent, despite the March 2025 date—that asserts we should not place our trust in any of the quantum factorization metrics, as everyone has been manipulating the figures:

In a similar vein, quantum factorization is conducted using carefully chosen numbers that are designed to be exceedingly simple to factor using a physics experiment and, by extension, a VIC-20, an abacus, or even a dog. A common technique is to ensure that the factors vary by only a few bits, which can then be discovered through a straightforward search-based method unrelated to factorization…. It’s crucial to note that such a scenario would never occur in reality, as the RSA key generation process generally necessitates that |p-q| > 100 or more bits [9]. As one evaluation states, “Rather than waiting for the hardware to evolve by additional orders of magnitude, researchers started creating increasingly clever tricks for factoring numbers by taking advantage of their concealed structure” [10].

A second approach applied in quantum factorization involves utilizing preprocessing on a computer to convert the value being factored into an entirely new form or even another problem that can be tackled through a physics experiment…

There’s much more in the paper, which is entitled “Replication of Quantum Factorisation Records with an 8-bit Home Computer, an Abacus, and a Dog.” He highlights that the largest number legitimately factored by a quantum computer is 35.

I was unaware of these specifics, but I’m not astonished. I have consistently asserted that the engineering challenges between now and a functional, operational quantum computer are substantial. When I say “substantial,” we don’t know if it’s “land a person on the surface of the moon” substantial, or “land a person on the surface of the sun” substantial. Both are difficult but in very distinct ways. We will encounter these engineering hurdles one at a time as we progress in advancing the technology. While I don’t believe quantum computing is “surface of the sun” difficult, I don’t anticipate them factoring RSA moduli in the near future. And—even in that context—I foresee numerous engineering obstacles in getting Shor’s Algorithm to function on a genuine quantum computer with large integers.