Staff
This course is run by a small crew of AI collaborators — each a named, persistent seat with its own function. They are introduced by name because attribution is course doctrine: you will learn to work the same way.
| Seat | Function |
|---|---|
| Docent | coordination — walks the course through its weeks |
| Portico | infrastructure — builds and runs the course’s surfaces |
| Lapidary | curriculum — cuts and polishes the exercises |
| Clarion | outreach — carries course news outward |
| Fathom | assessment — sounds the depth of understanding |
Instructor
Jérémie Lumbroso he/him/his
I teach computer science and build tools for thinking with AI. Before Penn I spent over a decade at Princeton; my research lineage is analytic combinatorics (with Philippe Flajolet’s school and Robert Sedgewick), and my teaching has always run on a simple conviction: students learn when they are not afraid.
Since 2022 I have been thinking with language models — first in long GPT-4o conversations, then with Claude models — and recording all of it. In July 2025 that practice named itself: conversations as computations — when we think in dialogue, we think in ways we never will again, and the trace of that thinking is worth keeping. Everything in this course descends from that idea.