A different kind of ambition
Google blew it because they only shipped half a product and every developer with any sense spotted that straight away.
The crucial missing bit is a client-server protocol, despite it being obvious that Google has written one for their own client. The impact of that is to crudely, very crudely, push developers towards writing server add-ons (robots etc) because that is the path of least resistance. Anyone wanting to innovate at the client has to write some serious code before they get anywhere near innovation.
To make it even worse, the Google Wave client is closed and designed around a single communications paradigm that has little relevance to most people. The contrast between the open server and the tightly controlled client could not be more noticeable.
If Google really wants Wave to be a success then it would take very little effort, open the client and the c-s protocol and let people develop innovative ways to use the server, but that is clearly not the intent. Google's decision to kill Wave speaks volumes about Google. Not about their willingness to take risks though, but about their unwillingness to lose control even if that means the project dies and their reputation suffers.