This is not the conventional analysis of what is wrong with Face Book that describes its business model and decides if it works or not, it doesn't. It is about how that business model drives a bad exterience for its users and how that provides a different opportunity with a totally different approach both to business and to design which could change that. I have used Face Book for two years, still do, and have discussed some of the ideas I am going to give here with my friends, there.
The thing that is most wrong at Face Book is the user interface and why it doesn't change, or why engineering at Face Book is resisting changes to it. There are several in-browser modifications to the UI as served. They involve javascript and style add-ons to that UI and they suffer the disadvantages of a UI design at Face Book that is driven by factors they can't know or control, in fact, it is the users who are most out of the loop.
Face Book just announced that it serves 1 billion users. So what? No one that I know has anywhere near that number of friends on Face Book. My guess is that most people have as a high estimate 100 friends. So who needs 1 billion users? Face Book needs to make that number available to its business partners, its advertizers. That is supposed to justify the cloud that supports the pages of that many users, and as a cloud, we are spared knowing anythng about where our pages are stored, and also like the cloud, whereever we access Face Book we will see the same posts to our page through the same interface. So far, so good. But the problem this model creates is that the users don't get many choices about the interface design and the add-ons, such as Social Fixer, which is a heroic effort to address the most significant flaws in Face Book's UI, don't have a stable platform as Face Book's engineers keep breaking it for them. This is evidenced by the rapid number of updates needed to add-ons.
So the theory of the cloud is that it is supposed to scale, meaning that in addition to not caring where it actually stores your data or where you get at it from, it doesn't care how much data you store as long as that is below a certain quota. The sizes of Face Book pages is not that large, generally much less than 1 GB with posts, images and all. So, if it is the scale of Face Book's back-end that dictates the inflexibility of it's UI, then get rid of the backend and concentrate on giving the users a much better UI.
The issue of UI design hinges on the way people need to communicate on their systems. Notice I specfically avoided using the idion "Social Media" or "Mobil" or even "Web Page" for there is an even older medium that can inform us about how to communicate with one another effectively that should be considered in the User Interface options. So the opportunity that can be provided to the replacement for Face Book, has two distinct parts, and the key difference between this and Face Book, where Face Book is a failure, is to keep them separate. The first part is to allow the users to choose, which part of the cloud to use, how to host their pages, what web address to tell their friends to go to. The second part is to define styles for what they post to their pages. The function at Face Book, could be made much smaller, especially if their business model fails, which I predict will happen, and they will have to downsize, is to maintain the matrix of people contacts and cloud addresses for their pages. But in this small model, Face Book is just the clearing house for contacts. It doens't own the pages or the UI to them, and it is much cheaper to run.
If you are thinking this is just a throwback to listserv or blogging or even just e-mail, you are partly correct. The USENET, which is still around, and e-mail client tools suggest features that would be nice to have for commuication, such as reply in context, threaded replies, and they are not blogging, and in particular Twitter. This is not to say that those styles of communicatig should not be supported, they should, but that well-worn elements of the older technologies could be re-introduced as options to the UI when people want them. One of the major flaws of Face Book is that it is possible to have public pages that get hundreds and even thousands of replies and there is no way to manage them. This is the general down fall in blogs of even less than 100 replies; most people can't say anything intelligent, or mobil technology doesn't really encourage it, and it isn't worth it to wade through all that for a couple of gems. So blogging is self defeating, both because it takes a big effort to keep it up and the rewards aren't there. The old USENET had a number of features that helps get around the defects of blogs, like skipping replies or subthreads, skipping replies from certain users, that resulted in the ability to pursue long lasting conversations with preservation of context, as also in a good e-mail client, These favor larger displays, number of pixels, that may exclude the older mobil devices, but the newer tablets, even the iPhone, and desktops are able to support the return of these older features.