Music Since 1900(MUSH 4190)
A 4000-level music-history course normally taught by a musicologist with a PhD — in my own words, “I could not be more meekly qualified to teach this class.” This is where I stopped surviving by improvisation and started building a machine: I authored the course; AI implemented it.
The expertise I brought & the gap
The music is mine; the scholarship was the gap
- Expertise present: a 30-year performance career across all eras — I know the repertoire and the composers in this course cold. The music is mine.
- The gap: how to disseminate it properly, in a scholarly manner, to undergraduate seniors and graduate students. Content mastery present; the pedagogy and scholarly apparatus were the deficit.
How I figured it out — human consultant first
The loop, then the unlock
With my cello colleague I built a rough syllabus and a set-works spreadsheet — candidate works for the course. But it still needed direction and flow. That’s where the origination line did its work:
I created the course architecture and the course design rules. NotebookLM and Gemini implemented them — inside my structure, at speed. I originated; AI executed.
The result was a stable, substantive course ecosystem producing real deliverables: a slide deck, a case study, and a lesson guide. Lacking HTML, I built the course Google Slides via AppsScript — routing around a skill gap with a different tool.
The artifact
The built ecosystem is the proof
This isn’t a story about one good semester — it’s a reusable machine that still runs. The evidence lives as a working repository:
mush-4190 — the course-ecosystem spoke
Where it leads
From private machine to public discipline
The machine I built here was teacher-facing — “for my eyes only.”The next case turns the loop outward: in MUSX 5110 I don’t hand students the machine, I teach them to run the loop themselves.