Typically, languages such as Python work in tandem with low-level tools such as C/C++. You are given "a nice, fuzzy, easy-to-use Python library" which, as it so happens, invokes lower-level C/C++ code and external libraries to do the heavy lifting. Meanwhile, for everything else that you need to do, Python itself (which is, of course, also a large and sophisticated C/C++ program ...) is fast-enough, efficient, and readily available.
In this scenario, C/C++ becomes a more specialized tool: used to construct higher-level languages such as Python itself; and used to implement specific, targeted functionality (in the form of libraries and subroutines accessible to the higher-level language tools). In this way, you play to each language's respective strengths (and designs), with no significant loss in computer efficiency (but considerable gains in far-more-expensive human efficiency). So, it's simply not a "versus" issue at all.
Q: Scalpel vs. chain-saw?
A: Well, it depends on exactly what it is you are trying to cut ...
Last edited by sundialsvcs; 07-06-2012 at 09:09 AM.