That doesn't matter though, as there is a set of people who value their own data and will want to keep control of it - they will build the first usable versions.
I do think that it could take off quickly though if a group of gaming companies got together to have users gaming characters transferable to other platforms via blockchains.
Avid gamers would queue up for it if they could bring their awesome weapon they earn't from 300 hours playing a FPS to whatever new game comes along. could very quickly change
All it takes is one and with the open-source nature and open ended time frame I believe it is inevitable, the question is "when". However I must say that I am more than likely excessively optimistic on that question.
Not only are they usually possible to run offline, if they are not blockchain-based, but they tend to be faster and scale better than traditional web applications.
People usually move through three layers or more of "web" on a daily basis. We have the personal web at home, one at our office and one global one. But moving between these three usually requires specialized applications that are built with roaming in mind, if you want this.
If we instead would build our applications with something like IPFS, we can have applications that transparently move between these domains and works everywhere, without having separate applications or extra development effort.
I think at this point, most technology people have realized the faulty model we built the web on, compared to the internet. Slowly the rest is waking up to the fact that the web could be better in every way, but it'll take time to move everything.