Carnegie Mellon 11-768 (Fall 2026) in one page — all 27 sessions with dates, a direct link to every posted recording and slide deck, the six lectures to watch first, a four-week catch-up plan and the full 2026 reading list

The companion to the reel. Carnegie Mellon is teaching a graduate course on AI agents — 11-768, Fall 2026, taught by Daniel Fried and Graham Neubig (the OpenHands co-creator) — and publishing the lectures, slide decks and the first assignment in the open while the semester runs. PAGE 1 — what is actually public and what is CMU-only, with the three links: the course site, the YouTube playlist and the assignment repo. As of September 16, 2026 that is 4 recordings (4h 36m), 7 slide-deck PDFs and assignment 1 on GitHub, with classes every Tuesday and Thursday through December 3. PAGE 2 — the full schedule, all 27 sessions with date, module, lecturer and milestone, each row linking straight to its deck and recording where one exists. PAGE 3 — the catch-up path: the six lectures in the order that pays first, a four-week plan at an hour a day, and how to do assignment 1 (build a coding-agent harness on an open-source LLM) without a CMU account. PAGE 4 — the 2026 reading list per lecture, from Toolformer and ReAct to Cursor's Plan Mode, OpenHands skills, SkillsBench, ReasoningBank, OSWorld and OpenCUA, plus the course themes and what a student should be able to do by December. PAGE 5 — where to go next once the semester ends. Facts verified September 16, 2026 from the course site, the playlist and the repo. Independent guide from Hyperautomation Labs — not affiliated with Carnegie Mellon University.

By submitting, you'll be subscribed to the Hyperautomation AI Report newsletter — three practical AI emails a week. Unsubscribe anytime.

Free. No spam. Unsubscribe anytime.