The software engineering field is experiencing a new wave driven by advances in machine learning and large language models. This wave is promising to redefine the field, but the conversation around it has become extreme, with some arguing that AI will replace engineers entirely. Others believe that AI is just another tool to be used in the integrated development environment. However, both of these views oversimplify the actual impact of AI on software engineering. Historically, shifts in software engineering have not eliminated complexity, but rather relocated it, from assembly to high-level languages and from monoliths to distributed systems. The term architecture refers to the decision-making process about how a structure holds together, not the actual building process. As AI reduces the effort required to produce code, it amplifies the need for design, trade-offs, and system thinking. The reduced cost of generating code increases the cost of making poor decisions, as mistakes can be propagated at scale. This means that the role of engineers is not replaced by AI, but rather transformed to focus more on high-level decision-making. The need for skilled engineers to make informed design decisions and think critically about complex systems is more important than ever, as AI increases the potential impact of good or bad decisions.