I had a reply from them in October 2013, saying:

I inspected the files we could have returned in early summer, and found that they did not distinguish between high-probability reference genome matches and low-certainty sites. While this is fine for some research purposes, it would have resulted in nearly useless reports from third-party tools. We're currently working on fixing this problem, but we can't yet give a precise ETA (the website's mention of "July" has, sadly, become a placeholder for "I don't know yet").

We will update the website and send you another email when we are ready.

I believe some other people have received the same. When I log in to their website and look at my status page, it says:

We're pleased to report that sequencing has completed. We are now extracting high-probability genotype calls from the raw data, and will send you an email when the process is complete. (This is taking longer than we anticipated; apologies for the delay.)