> they are obviously not worse, in some places are much better Not sure why it is obvious, especially after Alex Biryukov et al reverse engineered S-Boxes of S

> they are obviously not worse, in some places are much better Not sure why it is obvious, especially after Alex Biryukov et al reverse engineered S-Boxes of Streebog and Kuznyechik [1]. If you suspect Dual_EC_DRBG kind of weakness, why not use some algorithm without magic constants like Speck [2]? [1] https://eprint.iacr.org/2016/071.pdf https://eprint.iacr.org/2016/071.pdf [2] http://eprint.iacr.org/2013/404.pdf http://eprint.iacr.org/2013/404.pdf