How many security exploits have been found in the Linux kernel, or other trusted software? How many of those were around for a "long time"? Every one of those got by the normal peer review process.

So the question is, which is harder: does it take more skill to accidently insert a bug that gets by (sometimes for years), or to do so on purpose?