In medical research, data privacy is paramount. So it is unlikely that we'll be taking advantage of things like cloud computing anytime soon.
However, even if I could, I doubt that I would in the absence of brutally complete back-ups (which would rather eliminate any space savings). Because the world is broken up into two groups of people: those who have lost data due to crashes and those who are yet to do so. When you are in the former group, it seems the height of insanity not to be as a careful as possible about keeping data safe.
Heck, I dislike how the SAS program editor does not have an autosave feature. I've lost code because I walked away from the computer for a meeting and windows decided that was an ideal time for an auto-update. In that sense, the static environment of using the computer as a tool is at war with the dynamic vision of computer developers.
I wonder if, at some point, we'll branch off into two streams? Computers are cheap enough that you no longer need one platform for everything and I mostly use the internet at work for downloading articles (a function that slowly updated libraries could pretty much handle).
It is an interesting thought.