Under the auspices of the Computational Complexity Foundation (CCF)
I give an alternative proof of the xor lemma which may provide a simple explanation of why xor-ing decreases correlation.