Once you've logged into FaceTime you can have a look at all the account settings of the used Apple ID. Username, ID, place and birth date are shown as well as the security question and the answer to it - in plain text, without another password request. To reset the password to an Apple ID, all you need it the exact birth date and the answer to the security question - we tried that out for you, and it worked fine.
The blog post walks through, step by step, how someone without knowing a user's FaceTime password, can change that password. And, according to the report, even logging out of FaceTime doesn't fix the issue as the app keeps the password stored and active, so that it's a snap for anyone to log in to an unattended system.
The upshot in this flaw is that an attacker would need to have physical access to a system pull off these shenanigans. But with more of us computing on mobile devices - notebooks, smart phones, and tablets - flaws that require physical access to a system provide increasingly less risk mitigation.
What troubles me is that Apple doesn't seem to have conducted any kind of threat modeling on this FaceTime beta software. I mean, really, username, Apple ID, and security questions shouldn't be stored in plain text for the world to see. This is especially true when the software doesn't clear the password when one logs out. And it should go without saying that providing the existing password should be a requirement to conducting a password change. With such carelessness, it makes one wonder what other areas where Apple has so blatantly skimped on security. Are they developing so fast now that security is being pushed aside?
For my security and technology observations throughout the day, find me on Twitter.