One design feature of AP classes is to indeed remove some discretion from teachers. This is a drawback when you have amazing teachers, and a godsend if you have bad or just plain lazy teachers. This is the same pattern as checks-and-balances vs dictators.
Since APs are a control mechanism for teachers then, I would take the article's opinion of APs with a grain of salt. A much more neutral party would be that of a student: a bad test causes pain similar to a bad teacher.
The good news about the student point of view is that most people here can draw from personal experience -- real data.
He cites an AP Calc teacher who was really stellar but got punished by not covering enough material by the time of the AP test.
In your experience, what fraction reflects that AP Calc example above (where discretion is used well at the detriment of the test), and what fraction reflect teachers using discretion badly (and getting caught by the test)? For me the ratio is like 1:3 if not 1:5.
I am open to the idea that there may be good ways to give teachers more discretion in positive ways, but I think it really needs to be acknowledged that APs exist in the first place to solve a problem, a problem that can't just be waved away without convincing data that show an alternative system performs better.