I used Tao's books on Real Analysis during college and my feelings are mixed on his capabilities as an educator. On the one hand his books were really great as he gave a lot of proofs (although most proofs were still left to the reader, but I guess that you can't really learn Analysis without proving Rolle's Theorem or the First Theorem of Calculus on your own) and what I also really liked was that he made no assumptions about your knowledge and started from scratch (first set theory, then properties of natural, whole and rational numbers before constructing the real numbers using Cauchy sequences). On the other hand there were a lot of errors in his books and sometimes he is hard to follow, I guess the books contents are too easy for him and he wrote the book without ever reading back. Even though his books weren't perfect I learned so much about pure mathematics as a non-math major and I don't think I could have learned so much through Rudin's books.