For content consumption, tablets and phones are much more convenient, especially since they are used with well-optimized GUI. This will not change until there are new and better ways of inputting (and working with) large amounts of text, audio and video on portable/touch devices.
Typing into an overflowing textarea, Scrolling, changing a sentence, spelling correction, it's just a tedious process. On a desktop I would have been 100x faster and more accurate.
It is better to start with the market and problem you want to address, then find out what kind of platforms they possess (and plan to continue to possess) and build for that platform, than to start from the technology side.
Really these are very serious problems. But there are some cases you should use desktop apps.
The reason that SaaS is so popular is that it provides a much more predictable and reliable income over one-off sales.
EDIT to add: Win, OS X, Linux? Also, I would not do any desktop program without also thinking about how it would be as a tablet app -- iOS, Android, RT.
Not Everything Yet.
It does solve: hey, I'd like to be lazy and develop with my stock chrome book. It does solve, hey, I'd like to develop without installing software (I think schools will be big customers, especially to get all the setups at once!?) but your typical good developer, is going to be very attached to vim, emacs, sublime text, or whatever he normally uses that isn't part of cloud9.
For developers of more complicated software, embedded, OS, etc. Or even for mobile devs, cloud9 is a non-starter since it doesn't integrate with their physical environments (Like, their ARM board, or iOS device, nuclear reactor, or whatever) so for a wide range of things it's just not an applicable solution.
I think cloud9 will see success in education markets, as well as business from developers who do lightweight web dev work who don't have a favorite editor already.