How being meta with yourself can help AI be meta with its output for you (what?)
I used Claude and ChatGPT to create the Japanese version of this portfolio (and this English text).
I believe the process of understanding users' mental models (how people naturally think about problems) helps create solutions that feel effortless to users. Not because they're simple, but because they align with how users already navigate the world.
I found something similar when working with AI. The more clearly I could articulate my own thinking (e.g., what matters to me, what "good" means) the closer the outputs got to what I actually wanted. Sharing not just the "what" but the "why" seemed to make the difference. I think of this process as "nurturing your AI" ——building shared understanding through explaining, refining, and iterating together.
This page was also written with AI. But the underlying philosophy and vision the AI models operated under originated from me so I'd like to think this is still primarily my work.