I think data portability is going to be another victim of the tragedy of the commons. I'm sure that most people in the area of cloud computing would agree that more portability would be a good thing for the market. But when you get down to it, I don't think the risks of allowing all of your customers to switch away from you at will outweigh the benefits.
And let's face it, avoiding lock in just isn't as high a priority as some others to technology companies. If it were, SQL databases would have been dead a long time ago.
Ultimately this is a market-driven thing; data portability will continue to be an important issue and will increasingly become a measure of openness that users weigh when deciding whether or not to participate in a service.
By "the cloud" he really means "any service not run locally" which equates to "the web". You could argue that not all websites lock your data in, but neither do all cloud services, so I don't see the difference between the two. Funny, that word seems to morph meaning all the time. I think at some point it used to mean a bunch of remote computers-for-hire that one could easily push computation of any size off to, and easily increase or decrease the processing power of. Sort of like abstracting out the hardware problem. I thought that definition was the one that warranted a separate name the most, if any of them did. I could have sworn it meant that at some point ...
This is the case where the first competitor to make a standard, and have more than a couple legitimate vendors following a standard, will dominate the market.
And let's face it, avoiding lock in just isn't as high a priority as some others to technology companies. If it were, SQL databases would have been dead a long time ago.