I took a condensed, single day version of this course internally at Google last year. Afterwards I remember quipping that it was “Too much crash, not enough course.” Generally: the content was good, but I didn’t feel as though it gave me enough time to take each of the lessons on board.
Hopefully this updated and fully self directed version of the course manages to fix some of the issues I had. At the end of the day, there’s no one more qualified to teach you about TensorFlow than the Google Brain team. When I get chance I’m going to look a little closer and do a more complete write up.