Internet computing...
The problem is that network rendering (compositing) is largely IO bound. Once the data has made it to the machine, it doesn't matter how fast the render box is. Rendering over the internet, as a result, is impractical. Even if you had 100 super-fast boxes available on the other side of a $1000/month T1, a single lowly ghz machine on a local (hopefully gigabit) LAN will beat them mercilessly.
Likewise, this is why distributed computing, while popular, has a long ways to go before it can become profitable. It only makes sense if the CPU cycles/bandwidth on both ends are free. (i.e. academic projects)
3D rendering, while CPU intensive, falls to the same problem. This is why you won't be helping render Toy Story 3 any time soon.
Anyways, to return to the topic, NAB is April 7-10. FCP4 will be released. (It should have come out six months ago).
My predictions: Film/HD timecode/editing will now be integrated. Cinema Tools will no longer be a stand-alone project.
Hopefully Apple will fix the subclip bug issue plus the other major bugs.
Background rendering is a possibility I'd like to see.
Moving to the Cocoa toolkit would be nice, but probabally impractical at this stage in development.
Network rendering is an interesting possibility. Coupled with a couple of xServes and a xRaid, Apple could begin attacking the mid-range Avid market. (
http://avid.com/products/ds/)
This would be an interesting response to XpressDV that would be in the best interest of us Joe Users. Avid's gone too long without any real competition.