A thinking partner, not a shortcut
AI has changed how I build software — not by replacing the craft, but by sharpening it. I treat AI as a genuine collaborator, and I approach that collaboration deliberately, using the four competencies of Anthropic's AI Fluency Framework as my guide: Delegation, Description, Discernment, and Diligence.
Delegation
Deciding what's genuinely worth handing to AI and what belongs with me. The architecture, the trade-offs, and the calls that need real-world judgment stay mine; the repetitive scaffolding and exploration are where AI earns its keep.
Description
Communicating clearly. Good results come from good context — framing the problem, the constraints, and the goal precisely, the same way I'd brief a thoughtful colleague.
Discernment
Evaluating what comes back with a critical eye. Every suggestion gets the same scrutiny I'd give a teammate's pull request: does it fit the system, handle the edges, and hold up in production?
Diligence
Collaborating responsibly — honestly, transparently, and with accountability for what ships. AI can draft a function; owning the outcome is still very much a human job.
Experience makes the collaboration better
The four competencies rest on a simple truth: your own expertise is the foundation everything else builds on. AI amplifies good judgment — it doesn't manufacture it.
Effective & efficient
Knowing the right approach means I ask for the right thing — and recognize a good answer when I see one. Less guessing, more shipping.
A careful eye
Years of debugging tuned me to the subtle, confident mistakes AI can make. I catch them before they become someone's incident at 2 a.m.
Ethical & accountable
I stay responsible for what I deliver — reviewing, testing, and standing behind the work, exactly as I would without an AI in the room.