I just finished reading "The Big Switch", Nick Carr's new book. His premise is pretty simple - he posits that computing is moving towards the same centralized, utility approach that power generation did just over a century ago. Given the way things are going - ajaxified web apps, virtualized services like Amazon's EC2 - this is a pretty easy call to make, and he correctly identifies the biggest potential loser in this: Microsoft.

So far as that goes, the argument is fine, and I enjoyed the early alternation between 19th century electrification and the modern build out of data centers. I nodded along as he brought up SalesForce.com as an example of how security and privacy concerns over a cloud based infrastructure are apparently non-worries for most people. Heck, if I hadn't been convinced before, the large scale yawning over Facebook's Beacon would have convinced me.

Where Carr lost me a bit was in the second half of the book, where his more curmudgeonly side came out. I'm completely convinced that there will be side effects we don't like. He brings up the fact that narrow-casting tends to isolate people into like-minded groups, and we are certainly seeing that kind of thing in the political blogosphere. Spend a few minutes on a partisan site, and read the comments (pick a thread, any thread). What you'll notice is that dissent from the common view is rare, and - when it exists - it's shouted down and ridiculed. You can see this in tech blogs as well - just try to get a reasonable conversation going between a Mac fanboy and a PC user, for instance.

Where he lost me was in the last few chapters, where he went down the Ray Kurzeweil singularity street. I'm fairly unconvinced that AI is as close as he thinks it is. Interestingly enough, I think Carr realized that too, because he ended on a much more cautious note. He makes the case that - as with electricity - ubiquitous cloud computing will change society, and after a generation, it will be hard to remember how things used to be. That's a pretty safe bet.

Ultimately, it's an easy read, and Carr draws a decent analogy between 19th century electrification and modern movements to the cloud. One final thought: if you worry about crowdsourcing killing entire fields of endeavor, as Carr does - I'd recommend that you pick up "The Professor and the Madman". Having the public help out with a large task is not new.