Here is an interesting, recent paper on the evolving role of bioinformaticians
Incidentally, the word BioStar is meant to capture this exact sentiment and has recognized this direction eight years ago ;-)
In the information-driven life sciences the bioinformatician it the true star of the show.
I became interested by necessity, as I expect is the case with many scientists over the last ten years. My lab needed someone to take the reigns from a post-doc that was leaving and since I was a bright-eyed first year graduate student with zero computational background, I was apparently the perfect fit. I blundered through the first six months or so, picked up Python, bash, and a bit of R knowledge along the way and managed to cobble together functional pipelines to handle lots of ChIP-seq, mRNA microarray, and RNA-seq data. Turns out trying to learn bioinformatics with little guidance or formal training is a bit of a pain, who would have thought?
I like to think I've come a long way in the last few years, but I still learn new things daily. Reading my code from a few years ago makes me shudder, and doing analyses now only takes a few days rather than the weeks it would have taken me back then. I think part of what allowed me to get through the really frustrating roadblocks was that I could always do bench work for a few hours or days and get my mind off things. Though that can be a challenge in its own right, as jumping back and forth between the two multiple times a day can be jarring.
Overall, bioinformatics is just a great challenge. It requires analytic skills and critical thinking coupled with biological knowledge and the ability to extrapolate what the results of your analyses might mean in reality. It's fun, and the parts that aren't fun (common data munging, file format conversions, etc) can mostly be automated. Which is kind of fun in itself.
I hope I can find a position that allows me to utilize those skill sets even if bioinformatics isn't my main responsibility.