Yes, I should have gone to the Google I/O conference: I would have a sandbox Wave developers account right now. Oh well. It almost gives me a headache thinking about the server resources that will be required to support a world wide deployment and large scale adoption of applications built on Wave, end users of the basic Wave platform, etc. That said, I don't have to worry about how Google implements AppEngine, Wave, etc., and the details of how Amazon implements AWS. That is the point:
Cloud resources place an abstraction barrier between developers and deployment concerns.
As someone who actually enjoys dealing with server deployment issues, this is still a very good thing. Anything that lowers cost and makes development faster is a good thing, even when we have to leave doing some fun work behind us.
I like that there are already two other companies besides Google that are implementing Wave protocols and services (I want access to that Emacs Wave client :-) Open source implementations and multiple service vendors make the platform a safer technology choice. I liked the bit in the demo video on privacy and security (related: see Amazon's white paper on dealing with HIPAA privacy laws in EC2 hosted systems).
I'll obviously wait until I get a Wave developer's account and spend a 100 hours (or so) to kick the tires (some good documentation is already becoming available), but I might decide to write a short book on the Wave platform. I usually just write about things that I already use in my work, but since using resources like Amazon EC2, AppEngine, and Wave will be so important to my work and businesses in the future, time spent on Wave should be a good investment. BTW, I am working on a DevX article right now on the Google Java AppEngine. My APress book Scripting Intelligence: Web 3.0 Information, Gathering and Processing (should be in print in about 4 weeks) has some material on Amazon's EC2, S3, and Elastic MapReduce.