Adoption of AI tools is climbing while trust in them falls — and that gap is a spec, not a contradiction. A look at three postures developers can take toward AI, and a bet that the governed, AI-first middle is where the work is heading.
AI-assisted prototyping doesn't just compress timelines — it changes what kind of thinking is valuable. When a wrong direction costs an afternoon instead of weeks, validation moves out of the spec and into the product, and building itself becomes a form of discovery.
AI development isn't about pressing a button to generate code – it requires mastering new skills like context management, prompt engineering, and collaborative debugging. Part 1 shares practical skills I've learned. Parts 2-3 explore how these changes might transform developer roles and team structures. The key: thoughtfully orchestrating AI while preserving human judgment.