I don't think the article made that conclusion? I see this in there:

> As seen in the benchmarks, the current Rust standard library unstable sort implementation outperforms the C++ standard library counterparts

And I can't find any generalizations like that from the author, though I may have missed it.