C7 / Writing desk
Working theories, teaching cases, and long-form arguments about what people understand when they make with language models.
Concept / draft / revision
A theory of human–LLM co-generative thinking: process traces, consequential artifacts, and demonstrations of understanding as evidence of learning.
Interdisciplinary research on user persona and story development in generative-AI contexts, with Charlie Potter, Justin Young, and Marielle Leijten.
A creative-coding teaching case asking what becomes visible when a required LLM interview, a personal metaphor, the resulting artifact, and the student’s explanation are read together.
For Digital Ecologies: prompting as an epistemic and material practice, with “prompt space” as where reality is negotiated and “gen time” as the interval between intention and materialization.
A generative human as active co-creator rather than passive recipient, developed through Select → Organize → Integrate.
Questions shared across the desk
Evidence
A polished output can no longer stand alone as evidence of learning. Process-centered assessment reads the route, the choices, and the explanation alongside the thing made.
Uptake
Human–LLM co-generative thinking is not measured by whether a model was used, but by curiosity, discernment, synthesis, application, and reflection.
Translation
“AI proposes, students translate” names the moment when a generated suggestion has to survive material, rhetorical, and technical constraints.
Ecology
Prompting is not merely instruction. It shapes what enters the frame, which relationships become legible, and whose model of reality gets materialized.
Existing abstract / publication unverified
An existing five-page, single-authored SIGGRAPH-track abstract defines structural exposure as a condition where the concept being taught becomes load-bearing in the artifact itself.
Understanding has to show up in the structure, not only in the explanation.
Source / claim / uncertainty