What, exactly, are the flaws you are talking about? I'm interested because email is simple and there may be ways, inventive or not, of fixing it without protocol changes, or at most a layer written above email.
I've been thinking about how to fix email as per YCs list of things they'd like to fund (just to get the juices flowing) and it might be cool to build on top of the email protocol rather than replace it - e.g it's trivial to format, say, todos in an email in a human readable format such that they can still be read and returned checked off by users who don't use this additional layer.
There are two different sets of problems with email.
The first, more concrete, set arises simply because of the age of the protocols. They were developed when the internet was a club of like minded people rather than a global public space. Things like: a lack of rock solid authentication, no guarantees of transit privacy (much less at rest privacy), clunky formatting, multimedia and hypermedia support, arcane attachment problems, and only hacked on support for anything beyond 1:N messaging (e.g. mailing lists).
A lot of these problems have good, solid, proposed solutions but inertia and collective action problems mean that near universal support for protocol additions / changes take years if not decades.
The second, more amorphous, set of problems have to do with how some people use email pervasively in their lives in a fashion that (inevitably) would be more efficiently accomplished with a tool specialized to their needs. So for example, people who spend a fair amount of time in a corporate enviorment with Exchange often bitterly miss the calendar management integration. They say that vanilla email is broken because it has no equivalent capability. Although the critique is legitamite I think the target is off, it's not email that's broken, it's federated calendaring in general. But because email can sort-of, kind-of substitute it gets the blame.
I've been thinking about how to fix email as per YCs list of things they'd like to fund (just to get the juices flowing) and it might be cool to build on top of the email protocol rather than replace it - e.g it's trivial to format, say, todos in an email in a human readable format such that they can still be read and returned checked off by users who don't use this additional layer.