For instance, one of the most important things the GNOME team will work on is application development and distribution. The team believes that the GNOME APIs are too much of a shifting target, and that application distribution is difficult and fragmented. On top of that, there's really no way for developers to maintain a brand or earn money.
If you already know where this is going, you're not the only one. First, maintain compatibility for years. A no-brainer. Second, sandboxed applications. Red flags going up for some I'm sure, but GNOME is open source, so I wouldn't worry. Third, easier packaging and, not mentioned but I'm sure included, deployment (I smell an application store). This new application framework will be available outside of GNOME OS for regular distributions as well.
They also want to modernise the development tools and aim to include an application development SDK with GNOME OS. The GNOME human interface guidelines will also be updated, which surprised me because I assumed the radically different GNOME 3 already had its own HIG.
This isn't everything they're working on, but it does offer a glimpse of what GNOME is working towards. They clearly want to become a full-fledged platform, more on par with Android or iOS than GNOME 2 or KDE (KDE might actually be going in a similar direction, I don't know). Overall, I think this is a good idea, because a graphical user interface and its applications - what GNOME essentially is to most users - needs deep system integration to work as good as possible.
What's also clear is that the GNOME guys aren't backing down with this whole GNOME 3 thing. Criticism or no, they're sticking to their guns, and I applaud that. These people have a vision, and while I might not always agree with what they've done to my beloved GNOME 2, they are at least trying to move ahead (like KDE has done and is doing with KDE4).