It's obvious (IMO) that the internet and, in particular, the web is the core of the consumer-(and, probably, business-)oriented technology of the future. One has to find a compromise between being too conservative (which is likely to be synonymous with "too late") and being too daring (which might be synonymous with being "too sloppy"). But it needs to happen.
IMO, it would be very nice to see Linux-based (and FOSS-based) solutions that implement this new paradigm appear well before Microsoft has a chance to do anything about it. (Once the desktop [in its current form] becomes obsolete, my hope is that MS will, as well, become obsolete.)
In 5 to 10 years, people will be using systems that are much more integrated than they are today. Joe will pause his TV program (which he is watching either on his "TV", or his "computer monitor", or his "phone", or ...) to check his email, or do some research on the web (on a topic raised by the TV program he's watching), or send a message (whether it be a tweet, an email, an IM, a something-not-invented yet), or check his phone messages or etc. He'll choose a movie while at his computer, pause it to take a break, and finish watching it on his "TV". We'll have a highly-integrated, internet/web/cloud/...-based system that allows us to do many different things from any location and, hopefully, the monopolists will not have figured out how to leverage this system in order to force everyone to use their technology to access it.