It seems like you're trying to criticize the article, but I think the article is in agreement with you. As far as I understand it, the author does not present a solution to the need to trust the processor and the hardware that moderates memory access.
