With the exception of the session that is mentioned in this article, I'm going to have to disagree.
To counter with my own anecdote in here, I work on a few apps (mac and iphone) and every time I went down to a lab to talk to an engineer directly, I was able to get help and have my problem solved. Issues ranged from bugs in Apple's code (with 10.6, and will be fixed by release) to App Portal/App ID issues, to bugs in my own code.
This wouldn't have gotten the love it did if it didn't ring all-too-true to all-too-many people.
I’ve also learned that many of the problems aren’t intentional: many parts of Apple’s internal infrastructure are overloaded or extremely outdated, and they’re scrambling to keep up with their growth, often ineffectively.
One major surprise was when an Apple employee told me that the App Store reviewers are not outsourced: they’re all direct, full-time Apple employees. Many of them work nights and weekends to keep up with the extremely high submission volume, and they’re constantly expanding the staff to keep up with the submission growth.
Source? Why, the very same post that Gruber is quoting in this submission.
Certainly some of the stats I was able to get from testers (thanks PinchMedia) suggest the testers have access to the AT&T network in the US. I could probably find IP addresses from Apple's /8 if people wanted to try to validate further.