Read the whole thing tonight. Excellent piece. Without even trying you'll learn how domain driven design works. Damian demonstrates a much more important idea: how get the entire system and how it functions transcribed into definable components and flows. AI can't replace a human's ability to do this, much like how it can't one-shot a $10k MRR SaaS application - the act of gathering the context itself, capturing it, and then translating it into something implementable is incredibly hard to do right. It's the ultimate long-horizon problem in our industry: you never get it "right" the first time. These designs will be reshaped continuously until the software gets retired, if ever. Staying relevant in software in the era of AI is going to be less and less about tools you know versus being able to define a system that correctly models a business problem and communicate its shape to AI and other people who can help you implement and execute it. And knowing how to code / express these models well is still very valuable, i.e. "how can I use the type system to make illegal state unrepresentable inside our system?" - but code monkeys as a profession is over. Deep domain, integration, architecture (not knowing the names of patterns, but understanding where pressure / change sources / etc are likely to occur and designing to accommodate that,) and technical expertise are even more valuable in the era of AI. Adapt and move up the stack.