Erdős problem 950
This function was considered by de Bruijn, Erdős, and Turán, who showed that . They gave no proofs, but a proof of the (harder) second claim is given by Gorodetsky here [mathoverflow/508491].
Sources
FormalConjectures/ErdosProblems/
950.lean
Retained formal statement
He notes that if for all for some constant then .
(∃ C > 0, ∃ K > 0, ∀ᶠ (x : ℕ) in Filter.atTop, ∀ (y : ℕ), ↑y < ↑x - Real.log ↑x ^ C → ↑x.primeCounting < ↑y.primeCounting + K * ((↑x - ↑y) / Real.log ↑x)) → Erdos950.f =O[Filter.atTop] fun n => Real.log (Real.log (Real.log ↑n))SolvedStatement only, no proof