The article argues that when designing software systems, one should do the simplest thing that could possibly work, rather than trying to create an ideal system. This approach involves understanding the current system deeply and then implementing the simplest solution that meets the requirements. The author believes that this approach can lead to great software design, which looks underwhelming but is actually impressive in its simplicity. The article also discusses potential objections to this approach, such as the risk of creating an inflexible system or not anticipating future requirements.