Your last sentence is conjecture, and not in the Register article or in the paper. (FWIW, jonbaer submitted a HN link to the paper last week.) The paper points

Your last sentence is conjecture, and not in the Register article or in the paper. (FWIW, jonbaer submitted a HN link to the paper last week.) The paper points that the reason why it works is because the factors are only off by two bits, eg. 3599 = 0b111011 * 0b111101 It also specifically addresses the RSA claim you made: > However, unless we know in advance that the factors will differ at two bits, this reduction will not allow us to crack big RSA codes.