What do I code when I dont know what would benefit me?

First of all, the fact that you handle some criticism constructively shows a lot of hope.
These days, learning is as productive as never before. You should understand that when you talk to modern-day experts, you often talk to people who successfully learned great deal of technology before MSDN and Wikipedia, even before Internet, in the environment with great deficit of documentation and publications, forget about forums like this one. Not many of such expert can take your words like "I don't know where to start" quite seriously.

"Where to start" is just a common mistake. You don't need to think much about where to start, you can start with any random place. It just doesn't matter, as soon as you persistently and methodically study the subject. Can you imaging the excavation of coal from a mine using a jackhammer? This is how learning looks. At first, you don't know where to start, so you try hammering in few different directions, and finally find some vantage points, and eventually big fragments of the wall just fall down after a minimal impact. With experience, you just reach your goals faster...

Anyway, "capable of learning more" and "don't know where to start" apparently contradict. Capability does not appear by itself, it also grows with experience.

Now, I have no idea how did you draw the conclusion that C# is not the most efficient; this is one of the best languages designed so far, but it certainly have some drawback. However, I am not sure that you can identify those drawbacks. Can you name just a few? This is not a very simple question. Anyway, if you can share some problems you can see in C#, it could really be a topic for discussion.

For Android? Well, I don't know... .NET (CLR, more exactly) is not native to Android. I worked only with .NET and Mono, and it gave me excellent multiplatform solutions. Many things depend on your background and skills. At this time, I do some development using Free Pascal, and this is also one of the best languages. I believe it has stronger cross-platform capabilities than CLR, at least when it comes to Android. I still have to try it out on Android and Linux, but Windows results (in cross-platform type of project) look great.

Also, I find it somewhat disturbing that you tend to get something "popular". Long time ago, I derived a law of nature for myself: collective decisions chose the thing (technology, language, product) which is strictly the worst from the acceptable alternatives. Mind you, it's never the worst, otherwise the decision would not survive at all, it's worst from the acceptable. So, I would not chase for the most popular thing and at least look for some compromise.

One more thing: it's not productive to look for the example of "next facebook". This is already the exhausted field. Everything which already reached success should be forgotten. You need you own idea, which nobody knows. Facebook impressed you? Done. It's already there. You have to think it totally different direction. Isn't that pretty obvious? :-)

—SA

Your comparison of a coal mine with learning is really great. I love it! My virtual 5.

Thank you very much, but, frankly, I picked it up somewhere, and probably this is a widespread comparison. I also remember the speech from a movie, where one character admits his role in science in such words, approximately: "The scientists are classified into those who break trough rock walls, and those who pick up the fallen stones and work with them. And I'm the one from those who pick up."

I think it's important to understand that both types are valuable in culture...

But when it comes to individual learning, each individual should do both, in one's own style and strategy, and nobody, absolutely nobody really helps to get to real results. Lecturers only show the right scope and "set the plank", show expected level of knowledge. And the examinations are more important then lectures...

—SA