Stuff and nonsense! You clearly haven't read all the information in this thread. There is a clear statement of a requirement for unpredictability in the key (and a key it most certainly is, because it is what an end user uses to identify the row containing it, conforming exactly to Codd's original definition of a key - and it's because it identifies a row that unpredictability is required).

This unpredictability (randomness) requirement is extremely common in applications that use an RDBMS to store personal information (whether they are web applications or not - unpredictable keys are somewhat older than the web). An index like this one is a component of the security wall that maintains privacy. Of course usually isn't the only such component, but for security and privacy each small addition to an attacker's problems is valuable.