This isn't a replacement for any native Python data structures. It's a collection of three different approximate data structures for dealing with large data sets where you're ok trading off accuracy for reduced memory usage.

I don't understand why it would be "embarrassing?" Python was designed to have certain characteristics and it has pretty obviously succeeded in terms of implementing design features that attracted a large number of users and a broad community. Nothing to be embarrassed about there. If you need faster processing there are numerous extensions and alternatives.