They claim that is part of the reason they review apps, and Marco in his article claims they raise the quality of apps.

Both these claims can still be true. I have had a but get through twice, and eventually it was picked up by the app store reviewers, because it was such a hard to replicate bug.

It is certainly true that some QA is better than no QA. Even professional QA's miss some stuff.

No. They are looking for crashers and obvious glitches (as well as malicious code and using APIs that might change and break the app in the future).

While they do try to ascertain that your app does what you claim it does, they are not there to test your logic.

It is possible that their QA involves checking whether an app uses prohibited API or whether it crashes the system or hogs memory etc. It is obviously a high level QA. Just search and see how many "task killer" apps available on Android device.

Also, I think if this app-review process is not beneficial to users, I dont think they would have started this in the first place. Sure, it is an issue to thousands of developers, but it is beneficial to millions of users.

Business related motives could be a big reason as well. But, who is not?