I've heard a lot of praise for it here and elsewhere, and I looked into it myself before. The praise is actually why I had seriously considered the method, their subreddit is full of success stories and people who absolutely believe in the method. Anyway, I was telling a friend about DS the other day and she was shocked when I said DS recommends at least 600 hours of input before you even start to speak. So it made me think yeah maybe DS is not an efficient method, then I found this thread
Is Dreaming Spanish massively inefficient?
I am convinced that a mixed approach to language learning, one that incorporates Comprehensible Input early on, is probably better than pure Comprehensible Input. Of course DS is a great source for graded listening material, no doubt. I wish something even remotely similar to DS existed for Arabic.
An older discussion from the same forum:
Pablo from Dreaming Spanish: Language can't be explained