The Next 20 Weeks of Systems Engineering

For a long time, programming was limited by our ability to produce code. That constraint is changing quickly. AI systems can now generate large amounts of code with ease, while inspection and validation remain hard.

This shift changes the balance of the field. Code is easier to produce than to reason about, and judgment is in a short supply compared to specialized intelligence. The pressure moves from writing programs to understanding systems, their structure, and their behavior over time.

The talk looks at this transition through a mix of history, language design, and recent developments. It argues that the core challenges ahead are less about generation and more about managing complexity at scale.

AI will change how we build software. As before, the outcome depends on how we adapt.