Crazy or stupid?
Create a system to compete with Microsoft by starting with a stable Linux system, removing the names Linux and GNU from everything, add bug-riddled software and promote the result as a great system.
Crazy or stupid? I know which I vote for.
Shuttleworth chose the wrong path to pursue his dream. The way to go about such an endeavour is to take the time to put together a stable, or relatively stable, system, then promote it. His approach of constantly making changes and introducing them before they are ready is a stupid business model. But the six month release must be met! (How many of Canonical's would be competitors release a new version every half year?) He could even have followed Red Hat's model. A stable retail version and a free testing version. He seems to believe that a pretty interface is all that is needed to attract and keep users. People do not mind bugs and breakage, as long as the screen looks pretty and they get "trophies" for using the system.
Creating their own utilities, such as GUI and an Xorg replacement fit well with the vision of a locked-down OS competing with the big fish, but requires planning and careful implementation. If they succeed in creating this Mir thingie, it will be the same as Unity. Introduced too early in the desire to meet the arbitrary release cycle, make the system even slower and more bug-riddled, and result in even more people abandoning the system. In the end, Shuttleworth's strategy is self-destructive.
Initiatives like Mir typical benefit the larger community, as their innovations propagate into or get forked into other projects.
It will probably benefit the larger community to the same degree as Unity. (Nothing)