The problem with this approach is that you will know the basics of a "hundred languages", but after a certain point, those basics are the same for all of them. So in order to improve, you should focus on just one language and learn as much as can. Then when you know that one language well, most of it will apply to any other language too. So, don't focus on any "my language is better than yours" or "this is faster than that" issues. Just learn a language well.
As others have mentioned, take on a project. Do you have a collection of CDs or Blu-Rays that you want to list online? Make a PHP website where you can enter the artist or director, and the title of it. Then the PHP could retrieve the information for the item online, and you can verify the results with a single click (or edit). Then upgrade it by making it do that automatically by comparing multiple sources (including CaSe SeNsitiVE titles). Then expand it to retrieve the artwork for it too, from online search engines...
Or how about a Windows GUI program, that has a picture of a drumkit, and when you click on a specific drum (area), a sound related to that drum will play? Then make it do the same with MIDI data automatically. Do the same for a full range of instruments.
Perhaps a program that you use to manage your monthly budget, with a certain savings goal?
Anything that will be helpful to yourself, basically. Make it interesting, or make it solve something automatically that you now have to do manually. Make a program that helps you.
Then make whatever you did, do more! Expand it to any ideas you come up with, even if you don't know how to do it. Then start searching for information on how to do it. Try many times in "playground" projects, until you get it. Then implement it into your real program.
The important thing is to make it first work, and not really follow any rigid rules. You will learn those much easier when you first learn how to get the damn thing to work at all (this is my personal obstacle now).
Good luck :)