A set of squares without arithmetic progressions (Q2919672)

In this paper, the authors prove the following result:NEWLINENEWLINE{ Theorem.} For every sufficiently large \(N\), there is a set \(A\subset \{ 1, \dots , N\} \) with NEWLINE\[NEWLINE|A|> cN/\sqrt{\log\log N}NEWLINE\]NEWLINE such that the equation NEWLINE\[NEWLINEx^2+y^2=2z^2,\quad x,y,z\in ANEWLINE\]NEWLINE has only the trivial solution \(x=y=z\), where \(c\) is a positive constant.NEWLINENEWLINEThe authors also pose the following conjecture:NEWLINENEWLINE{ Conjecture.} If the set of positive integers is split into finitely many parts, then the equation \(x^2+y^2=2z^2\) has a nontrivial solution with \(x,y,z\) being in the same part.NEWLINENEWLINEThe following question is still open: Is it true that \(|A|=o(N)\) if \(A\subset \{ 1, \dots , N\} \) and \(x^2+y^2=2z^2\) with \(x,y,z\in A\) has a nontrivial solution ?