I really don't want this to be perceived as a flame ... it's just my two cents:
Arduino is about building robots, weather stations, LED decorations, and ... oh yeah, music stuff. I happen to like that MIDIbox is focused on MIDI and musical devices, in which the timing MIDI i/o or other musical communication is absolutely NEVER compromised, for any reason. I have complete trust in TK and everybody who develops for MIDIbox that this philosophy will always prevail in hardware and software designs, so we have an excellent and future-friendly platform for real-time, embedded musical applications.
I fail to see the value Arduino could bring to MIDIbox, in the context of professional music production. It is always good to stay abreast what's going on in that other realm, but its technology, motivation and culture is precisely that: another realm!
Here's an example: almost everybody in the MIDIbox community is an electronic musician of some sort. Can the same be said for Arduino?
Regarding value in commercial: projects like "Beat707" are proof that cool music apps can be realized on Arduino, but to build it you have to pay somebody for it, and that person had nothing to do with the creation of the Arduino platform! MIDIbox-DMX is a great example of a non-musical MIDIbox project, and to build it all you have to do is download the schematics, documentation and firmware. (Optionally you can buy PCBs at bargain-basement prices.)
It is the good will of great people that has made MIDIbox what it is, not financial gain.
I think the LPC Expresso-based core will pave the way for more shield-like submodules however. The "core" PCB is really a bridge between the LPC and the classic MBHP modules. I envision complete projects as "shields" for the Expresso, and ribbon cable interconnects to other modules becoming less common.