The post didn't ease my concerns either. To further extend your last point, I don't understand why FB would spend so much acquiring whataspp if they didn't plan to mine the data. I told my friends to expect targeted FB ads based on our whatsapp convos in the future.
95% of WhatApp's messages don't event hit a disk and are served from memory. It would take big architectural changes to be able to mine them. (info from the erlang factory slides)
That's an interesting stat. How do you think FB will get an ROI on the 19B ?
EDIT: Also, do you think FB would/could man-in-the-middle the traffic a la NSA? No architectural changes to whatsapp are needed. FB could save and fwd the traffic. Crunch later.