[Cryptography] Claims of factoring 2048-bit RSA

[Ray Dillinger]

On 11/3/23 09:34, Amarendra Godbole wrote:
> https://www.bankinfosecurity.com/blogs/researcher-claims-to-crack-rsa-2048-quantum-computer-p-3536

This man claims to be achieving quantum speedups in calculations on 
non-quantum hardware.  That claim is intrinsically bullshit. There is no 
reason to read further.

Researchers often use non-quantum hardware to run software simulations 
of quantum calculations.  But those simulations do not actually benefit 
from any quantum speedup; they do painfully slow serial modeling of 
eigenvector calculations that would be single-step hardware ops on 
actual quantum hardware, and then tell you how many operations it would 
have taken if it had been done on actual quantum hardware.  Which, 
because this is only a simulation, is not in fact how it was done.

TLDR, factoring a number without using quantum hardware will go no 
faster (in fact quite a bit slower) than it goes using conventional 
factoring algorithms.  Phones cannot do a 2048-bit factorization before 
the sun explodes, therefore no 'quantum' algorithm modeled on a phone 
can do a 2048-bit factorization before the sun explodes.

If he did not use the obvious-bullshit "quantum" to describe his claim, 
and instead had a new and much better serial factoring algorithm that 
*COULD* crack a 2048-bit factorization on a phone? There are well-known 
published factorization targets he could have broken in order to 
conclusively prove that his claim is not bullshit.  The factors to one 
of those targets would have been in the first ten lines of the abstract 
of any non-bullshit paper about such an algorithm, and are conspicuously 
absent here.

We have no reason at this time to believe that such an algorithm 
exists.  Any paper claiming revolutionary factoring capabilities which 
does not include the factors to one of those published targets will be 
assumed to be bullshit. If that assumption is wrong the researcher can 
easily correct their oversight by breaking one of the published targets 
and adding the evidence to the paper.  If the researcher does ANYTHING 
AT ALL ELSE it's an indication that whatever else they're doing is 
easier for them than breaking one of the published targets - which their 
paper claims would be easy for them to do.  That provides evidence that 
the paper's claim is false.

Bear