Here is some info I can provide about the GRFP that may or may not help some. I will provide more after the announcements.
1. Very competitive this year:
There were over 16,000 applicants; the most applicants ever. Last year was around 14,000.
2. How well you apply only gets you so far in the process guaranteed.
What do I mean by that? In competitive fellowships, the reviewers are not looking for winners. They are looking for losers (this is sometimes a difficult mindset to keep for people who apply). 16,000 applications is a lot to go through.
Level 1 is seeing who followed instructions. Essays the right length? Number of references in? Margins correct? etc. I will say the other levels some other time, but once you make it to the last level, it is all luck. With so many people left, the reviewers have to be SUPER PICKY. Do you get the reviewer who emphasizes more on letters than personal statement? Or maybe the reviewer who understands your path (coming from a small college and making it through to a better graduate school as you struggled to find research, etc.)
3. Congress did not pass to increase number of winners (increase funding).
NSF GRF wanted to increase the number of winners to ~2,700 instead of ~2,000. It didn't the current or previous application year. The odds this year is far worse than last year.
I tried to list some parts of the NSF GRFP that is something out of anyone's control (except I guess the final reviewers). I hope this helps some realize that whether or not you win doesn't mean anything on you being a good scientist. These reviewers are only seeing a small snap shot of your life that is condensed down to transcripts, two essays, and three letters from people who only see you in a certain light. In the end, what you do during graduate school and your future career matters to you and those who really know you.
Good luck.