I was talking more generally about Web browsers and web APIs, not specifically WebRTC. I still see the web platform as currently the most open, practical solution we have.
However, the reality is that Google IS still a business. And so quite a lot of this key infrastructure, although it is open source, is still being packaged up by a giant private company.
Yes, we can use the web platform including things like WebRTC to do what we want, but I believe the the medium and long-term future will move away from specific implementations by a few companies and towards open machine-processable specifications.
I believe that the way forward will be to move away from textual programming languages and into truly semantic representations of programs, protocols, and implementations. Obviously it will be many years before this is mainstream.
The implementation at webrtc.org gives you cross-platform, cross-device source code in C++. You can build any API on top of that, in any language you want. It's already really easy to use if you an to use the HTML5 WebRTC API built into Chrome (and the same API in Firefox).
It seems to me that you have everything you want, or at least all the pieces of the puzzle. Someone just needs to do the work of assembling them the way you want them, and then of convincing the people with whom you want to talk (your friends, family, or business associates) to use your securely built communication system. Technology-wise it's possible, even fairly easy if you know what you want. But the social aspect (convincing others to use it) is more difficult.
(Full Disclosure: I work at Google, and I have worked both on Hangouts and WebRTC)