Another reason for the bias toward over-engineering is asymmetric risk. If an over-engineered approach fails, you’ll face less criticism than if a simpler approach fails. As the old saying goes, nobody got fired for choosing IBM. — Scaling up and down