The argument can be made that you can't test in quality. On the other hand, based on speaking with some Apple developers that I know personally, it would seem that automated testing is not encouraged. Neither unit testing, nor a dedicated team of SDETs for many groups. Contrast that to someplace like Microsoft, where I've worked as an SDET and a skilled, dedicated test team for a product is just the way it is.
Hmm, interesting to see such different approaches.
I just don't see how one can guard against regression bugs without having automated tests, especially when software gets complex.
And these days, all software is complex :)