Pip is not fine. Pip doesn't resolve dependencies [0]. It also has no way to sensibly manage the 'semver during development, frozen versions in production' needs of modern development, that both Ruby's Bundler and Node's npm take care of. Those are, in my opinion the 2 biggest requirements of a package manager for non-trivial projects, and it totally fails to handle them.
Here's a post on the latter issue by the developer of pip-tools which aims to solve that problem (I've not tried it yet, but my team is considering it): [1]
It's all good. So in my experience, I had issues getting my development environment up to speed at work. We had dependency files, but weren't well maintained. I made Combo so that a Dev-team could easily share a file with other workers and bring their engineers' dev environments up to speed in a few minutes. Furthermore, open source projects could distribute their "Combofiles" with ease, grow their contributor base, and focus more on development than menial issues.
Besides, the main goal of Combo is to eventually join "pip" or take on all the capabilities of pip.
Here's a post on the latter issue by the developer of pip-tools which aims to solve that problem (I've not tried it yet, but my team is considering it): [1]
[0] https://github.com/pypa/pip/issues/988
[1] http://nvie.com/posts/better-package-management/