And hard skills are (relatively) easy.
I read this post in Bioinformatic Zen that discussed the development of "hard skills" (programming and other stuff) and "soft skills" (dealing with failure, self-discipline, etc). It is a very interesting post and I really couldn't agree more with it.
In my experience, the "bio" part is much more difficult that the "informatics" part for us bioinformaticians, even for people who came from a biology background. Things that can be measured and dealt with a logic way are relatively much easier, such as learning a programming language or developing a software package. For things that cannot be dealt with a pure logic fashion, such as what constitutes an interesting and important project or what does one do when a paper is rejected, no one can really teach you and you just have to bang your head against the wall until you somehow magically learn the lesson (if ever).