Are we Quantum yet?

NO. #

There isn’t a quantum computer large enough (in terms of logical quantum bits, or qubits) to do anything useful. To be in the age of quantum computing, someone needs to demonstrate quantum supremacy: a quantum computer running a useful algorithm faster than a traditional computer.

My interest is cryptography, and here quantum computer progress can be measured in its ability to factor semiprimes (numbers with exactly two factors, which are both prime numbers). This is how a quantum computer could use Shor’s algorithm to crack RSA encryption.

Web site certificates typically use 2048-bit RSA keys if they haven’t been upgraded to use elliptic curve cryptography.

The largest number factored by a quantum computer is currently 21 (done in 2012), a 5-bit semiprime (10101 in binary). For a great discussion about the current state of quantum computers and factoring semiprimes, see Replication of Quantum Factorisation Records with an 8-bit Home Computer, an Abacus, and a Dog.

The author replicates some sketchy 2048-bit factoring using a Vic-20 in an entirely reasonable amount of time.

Seriously #

In 1991, RSA Laboratories published semiprimes of various sizes as the RSA Factoring Challenge.

As of now (December 2025), the largest value that has been factored (using traditional computers) is the 829-bit RSA250 number.

The smallest RSA number, the 330-bit RSA100 value, was factored in 1991.

(By the way, there are cash prizes for factoring these numbers! The next prize is $75,000US for factoring RSA896, a 896-bit number.)

Stable Goalpost #

When a quantum computer can factor RSA100 faster than a contemporary computer, then we’ll have started towards the magical Quantum Supremacy.

When a quantum computer can factor an RSA number that can’t be factored by a contemporary computer, we’ll be there. Currently that’s the 862-bit RSA260 number.

For yucks, I used CADO-NFS on my AMD Ryzen 7 7840HS laptop to factor RSA100, and it only took 5m16s (vs. several days when it was originally factored in 1991). A current top-of-the-line non-mobile CPU with more cores would do it in less time.