Nicholas Carr: The Big Switch: Rewiring the World, from Edison to Google.

In his excellent book Holidays in Hell, P. J. O’Rourke visits Future World (an attraction at Disney World) and says that it is "like opening a Chinese fortune cookie to read, ‘Soon you’ll be finished with dinner.’"

I get the same feeling reading Carr’s book (an advance copy) – it is well written, stylish and easily recognizeable like Disney World – and understandable to the masses. The main message of the book is that because of faster networks, computing will be centralized and made accessible like electric power. Carr even draws a line back to the history of electric power provisioning. All very well, we already see this happening with Google applications and Gmail. But I first heard this prediction in 1990, spoken not as a wild speculation of the future but as a likely and not particularly exciting outcome by my thesis advisor, professor Jim McKenney at the Harvard Business School.

The centralized and ubiquitous computing future Carr eloquently predicts is, in principle, a return to shared mainframes accessed over telephone lines, only cheaper and faster by orders of magnitude. The mainframe lost dominance to the PC because people wanted control of their own computing and their own data, so they chose a cheap, weak and unreliable computing platform over one that offered stability, performance (at least in the aggregate) and reliable backups. Otherwise known as a disruptive technology.

Many hard disk crashes and viruses later, a significant portion of the populace have not yet moved their files to Google Docs and are unlikely to do so. For that matter, I would venture that more information and computing is still done on mainframes than on Internet-accessible servers. That is not where the innovation is, true, but new computing platforms come in addition to other platforms, not as replacements.

So we will move into the Cloud, but for social computing, collaboration, and information lookup. People will still want their local storage and (at least perceived) local control. And will end up with a three-tiered personal computer architecture: Traditional centralized computers for transactional systems that demand global recalculation (like airline reservation systems), personal storage and processing for the very personal (where are you going to store those photos, you said?) and cloud-based computing for stuff we want to find and share.

Oh well. This is not news. I know Carr’s book is written for the great unwashed, and I admire his language and clarity of examples, but it is like Tom Friedman’s The World is Flat: If you have been reasonably awake and facing in the right direction the last 10 years or so, you will not find any surprises here.

And that’s a pity, for I read books for ideas, not for summaries. And this one, for all its elegance, had me dozing off more than switching on.