2) Most of the breakthroughs were because of new levels of power or precision. Here we're facing a totally different kind problem, one of complexity. That's something we have no tools to deal with and no idea of even how those tools might look like. It's not like flight, or Internet, or artificial viruses.
3) Did you really just link me to a device from a sci-fi show? :-D The thing is that mass production can use optimizations, shortcuts and logistics that can never be used by custom production. Also, injection molding will always produce smoother shapes than 3D-printing, because the latter is discrete.
Also, ecologically speaking, it's better if people just own a set of mass-produced, high-quality goods than continuously print new stuff on their own. There's too much stuff in the world already.
2) Research has been going on in to make this happen for 40+ years. DARPA has invested a significant amount of funding towards the goal of a BCI. All that's really necessary is the connection, that's the hardest part, just how to talk to the brain and make a safe connection to it. Google is working on it pretty heavily also. It'll be done, just a matter of when. Many science concepts people don't know if they are possible because of some unknown limit, but with BCIs we know it's possible, just a matter of figuring it out, which has turned out to be a difficult but certainly not impossible task.
http://en.wikipedia.org/wiki/Brain%E2%80%93computer_interfac...
3) Maybe in the short-term, but in 40-50 years there is no reason that a "Replicator" can't be created. http://en.wikipedia.org/wiki/Replicator_(Star_Trek)