As far as I can tell, the "most obvious bug" is not a bug, other than reliance on implementation-defined behavior converting uint64_t -> int64_t (namely, clang/gcc -fwrapv).

Overall, the tone of the article is obtuse and full of needless digs that make me just believe the author has an axe to grind.

I stopped after realizing I didn't care about the ABI of 128 bit ints on Windows.