You must have above average programming peers. It's my experience that architectural decisions that could cut future work by orders of magnitude are quite commonly missed.