What separates this from continuous integration. As I understand it the article proposes testing the same build over and over again, well the same commit rebuilt and retested over and over again. Where do you benefit from this?
I could see a situation where tests running against a live system would be worthwhile, this is what I though when I saw the title.