A remark on Erdős Problem #462: the Riemann Hypothesis suffices
Year: 2026
Type: preprint
Abstract: Let p(n) be the least prime factor of n. Erdős and Graham asked whether there is a constant C > 0 such that the sum of p(n)/n over the integers n in [x, x + C x1/2(log x)2] is ≫ 1 for all large x (Erdős Problem #462). Zhang (arXiv:2608.24930) studied the sum over composite n, with mean value 4C, and gave an affirmative answer under a hypothesis of Cramér type on primes in intervals of length at least a constant times (log y)2. We show that the Riemann Hypothesis is enough: under RH there are absolute constants C0 and c > 0 such that for every fixed C ≥ C0 and all large x the sum over composite n is at least cC. The proof counts products of two primes of size about x1/2 in the interval and uses the explicit formula in both variables. Notes: a conditional result (assuming the Riemann Hypothesis) obtained by standard methods; nothing is proved unconditionally, for small C, or about the conjectured asymptotic 4C; not formalized; not reviewed by independent human experts; the literature search for earlier statements of this result was not exhaustive. Written with extensive use of generative AI tools under the author's direction. Code. The program used for the numerical illustration and its output are archived separately at doi:10.5281/zenodo.24210939.
Author: Mori, Ryota
Language: English
Subfield: Algebra and Number Theory
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7230136896
