Nicholas Carr is correct re: world wide computing cloud
I used to think we'd end up with something dynamic and heterogeneous — many companies loosely joined. But we're already seeing a great deal of consolidation by companies like Google and Microsoft. We'll probably see some kind of oligopoly, with standards that allow the movement of data among the utilities similar to the way current moves through the electric grid.
Read the entire interview - Carr is right on it. As far as privacy laws, consumer data protection laws, etc. go: I believe that companies like Amazon, Google, Yahoo, and Microsoft (OK, I am being generous including Microsoft :-) stand a better chance of getting it right if one thing happens: quality of service contracts (I am assuming that cloud data services used by businesses and some individuals are paid for on some sort of subscription basis and not free) include guarantees for data security, including sufficient guarentees to satisfy medical record laws, concerns for company proprietary data, etc.
In the future, after the big vendors get the data security requirements down and guarenteed, it will simply be unreasonable to expect the IT department of a 500 person company to have sufficiently talented lead technical architects, sufficient staff, and sufficient redundant infrastructure to compete with Amazon for supplying secure data services. In addition to the big players, small companies like Dabble DB look very interesting but for smaller service companies it obviously becomes even more important to have a separate backup and alternative recovery system (a good idea even if using Amazon and Google!)
Instead of thinking only in terms of building on infrastructures like J2EE, LAMP, relational databases, etc. running on local servers, now architects need to consider the effective use of commercial secure data services and how these services can be used to decrease the cost and increase the reliability and security of custom business applications.