His observation on code review was that it did help to find low hanging fruit kind of bugs (int overflows and signedness bugs), but that bugs were still easy to find in OpenBSD, even though he felt like the code quality was much better, than in other BSDs.

If you are just reading the slides, they are too out of context.