A-01Positioning AI is a frame, not a replacement for the personA short note on using AI to make human judgment more structured, testable, and communicable without letting the tool become the identity. ↗
A-02Judgment Human context before automationWhy the situation, the people, and the consequences should be named before choosing a model, tool, or workflow. ↗
A-03Verification Evidence needs a boundaryA practical distinction between what the evidence supports, what remains an assumption, and what should stay private. ↗
B-01AI & research From psychology measurement to AI workflow designHow research habits like operational definitions, scales, data cleaning, and validation map into practical AI automation work. ↗
B-02Products Building systems that can verify themselvesHow checks, visible state, and fallback routes turn a useful experiment into a system that can keep improving. ↗
B-03Workflow Designing useful AI handoffsA handoff is part of the product: the next person should know what happened, what was checked, and what still needs judgment. ↗
C-01Design Why a personal site should feel like a room, not a product pageA note on using an interactive desk and mission map to make the visitor explore Leo's actual judgment rather than scroll through generic sections. ↗
C-02Expression Writing in public without turning everything into contentA quieter way to share an unfinished thought: enough structure to be useful, without pretending every idea is a finished lesson. ↗
C-03Interfaces Interfaces that leave room for curiosityHow small prompts, motion, and hidden depth can make an interface feel alive without making it harder to understand. ↗