In the era of AI and vibe coding it's easy to build new software, start new projects. Generated code has been evolving rapidly along with the quality it carries. Meanwhile we've been busy developing Skills, Guardrails and brand new techniques to try and make AI more predictable, more deterministic. So what's the role of a Software Developer or a Software Architect nowadays? And how can we expect to shape the future?

AI-Agents produce a lot of code very fast... much faster than anyone can review it. So quality becomes not only about code ownership, testing and prodding. It turns into guardrails and rule-setting in order to steer agents in the right direction.

But these are not the goal of this talk. In this talk, we are going to go back to the basics that make for a good Software Developer and Architect in the first place. Understanding the business domain, the boundaries that are meaningful to your use case and how to recognize them, what trade-offs are worth taking, and what judgment calls are yours to make. Why? Because automating code output is a given, but thinking isn't. Maybe anyone can write good code, AI-assisted or not. But the knowledge and the tools to shape the code architecturally have to be taught on to the next generation of Software Architects.

So let us embrace AI-assisted development, but not set for poor architecture, big balls of mud, endless rewrites. Let's have a say in what the loop is, and not just be the human in it. Let's reclaim architecture!