I agree with your general sentiment but we're still not there yet.
Unfortunately, Skype hold all your keys and Microsoft changed the architecture to make legal intercept and tapping much simpler. I do not believe it is safe to assume that Skype conversations are private.
Current systems really do rely on users having some understanding of how trust in the application works.
For example, TextSecure really requires you to confirm keys in person (or via QR code etc) if you want to be sure you are not MITM'd. This is not obvious to most users I have spoken to.
Real time communication can use ephemeral keys. A Skype-like communications tool could be open, verifiable, as simple as Skype, and at least as hardened as Skype was before it was castrated.
As for web-of-trust for store and forward communication, social networks are an great way to provide secure key-signing.
Ephemeral keys, sure. Negotiated with... well... who exactly? The person you thought you were talking to, or someone else?
The available solutions are shared secrets with zero knowledge proof (like OTR does), voice verification (like various "secure phones", a web of trust, or CA infrastructure.
Crypto everywhere will improve things immensely, but (repeating myself) ultimately the user needs to understand how they can trust that the other party is who they say they are. So far we do not have a magic (automatic) way to do that for the user.
Unless NSA has a great Max Headroom version of me, I think people will trust that they are talking to, or listening to me. That's why I wrote "realtime communication."
For store and forward you need public key exchange and a mechanism for trusting identity. However, in most use cases where you have a mix of realtime and store and forward communication, you have ample opportunity for key signing where you can trust the identity of the person asking for your signature.
Unfortunately, Skype hold all your keys and Microsoft changed the architecture to make legal intercept and tapping much simpler. I do not believe it is safe to assume that Skype conversations are private.
Current systems really do rely on users having some understanding of how trust in the application works.
For example, TextSecure really requires you to confirm keys in person (or via QR code etc) if you want to be sure you are not MITM'd. This is not obvious to most users I have spoken to.