After quickly skimming it it seems well done. Leans towards mathematical rigor, which may not be ideal for some people. On the other hand, it'd be pretty hard to avoid math: quantum computing is linear algebra incarnate.

The only thing that caught my eye as off was totally minor. They say the many-controlled-Z gate used by Grover's algorithm can be done in O(n^2) constant-sized gates with an argument-by-reference, but with that type of argument you might as well give the tight bound of Θ(n).