Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Claims
2014_01_01_konyagin: Konyagin's Theorem 1 bounds the number of k-term representations of 1 by distinct unit fractions below by exp(exp(c k / log k)); accepted on the refereed publication, with a known false identity in its printed proof.
2020_12_10_elsholtz_planitzer: Corollary 3(2) of Elsholtz and Planitzer bounds the number of k-term representations of 1 by a constant to the power (2/5 + epsilon) 2^(k-1); an upper bound of the form c^(2^k), accepted on the refereed publication.
2026_09_25_openai: The OpenAI mathematics release's Corollary 1.2 of 25 September 2026, that ck <= log log F(k) <= Ck for all large k, accepted as a partial answer on the Lean declarations the corpus's verification built and audited.