I've talked with the ask patents folks before, and from what i remember, they send email to the examiners. That's it.
They try to occasionally look and see if the stuff they cite was cited by the examiners, but then they assume that if that is the case, that they were the reason the examiner found it. While I understand how hard it is to know why examiners did things in general (You can see exactly what they searched for through PAIR data, but not what results it turned up), this is a very very suspect methodology.
~~
He is puffing his product in this blog post -- nothing more.
And, this particular blog post is Joel up-talking the new "ask patents" exchange and describing why he feels it's needed.... so... it is a puff piece.