[Cryptography] The race to Quantum machines.

[Peter Gutmann]

Tom Mitchell <mitch at niftyegg.com>​ writes:

>IBM believes that commercial quantum machines will be here in about 3-5
>years.   If encrypted data at rest today has value or is a liability in the
>next 3-5 years quantum resistant keys seem important.

Uhh, you need to read the rest of the article, which you've actually quoted:

  "Starting its R&D on quantum computing as early as in 1996, IBM released a
  5-qubit quantum computer in 2016 and unveiled the world's first 20-qubit
  system, dubbed IBM Q System One, at CES 2019, Morimoto said, disclosing that
  the company will soon launch 58-qubit quantum computers."

From that we have at least a few data points, and there's more from non-IBM
sources, so we can extrapolate over time.  Technically we can't actually do
that because from everything I've read it's nonlinear, the first steps are
relatively easy and then it gets harder and harder [0], but let's say it's
linear just for argument's sake.  Anyway, to break 1kbit RSA you need about a
million qubits.  Soon we'll have a computer with 58 qubits.  Graphing things
and drawing a line to where even 1kbit RSA is at risk is left as an exercise
for the reader.

Peter.

[0] Like getting a manned settlement set up on Mars.  So far we can get the
    whole thing a few hundred meters closer to Mars.  After that, it gets a 
    lot harder.