I do not see a downside to this new approach. If Microsoft were to split out the patches as Josh suggests, it would make for an untested environment - something I do not want for my production servers.
Microsoft certainly cannot be held responsible for third-party functionality - in my mind they are only obligated to adhere to their own published APIs - or give adequate notice that something needs to change. Overall compatibility is part of the calculus when deciding to implement third party software in the first place. Often it is well worth the trade off to get a better price or better features than is available from Microsoft directly.
Keeping the number of released builds to a finite and manageable level makes complete sense. I would offer a "thank you" to Microsoft for continuing to improve processes - not just software.