Don't have the manner but the answer is yes. Not 'might be yes', but a definite yes. RSA will be broken before then by new mathematics but thats neither here nor there.

@DavidAttenborough Do you expect a classical computer to be able to factor RSA-2048 then?

@TimothyJohnson5c16 Yes, for a variety of reasons. Theres been algorithmic improvements on fast approximations for NP class problems, and thats all it takes. Along with this, theres been new mathematical tools derived in order to solve those problems. Looking at the trend, and considering a fast approximation of an NP problem suggests there exists a fast approximation for all NP problems, there is therefore every likelihood the same is true for modular exponentiation, or more specifically, RSA. I expect researchers to find one fast approximate algorithm for each problem in NP every 1.5 to 2 years, whether that Co-NP or NP-Complete I don't know, but as RSA relies on the apparent most important of them, I see no reason why that isn't already targeted using the newer methods.

@DavidAttenborough made a market about recently.