Six sessions, each a set of decks that can be read in a sitting, and each also offered as one
combined PDF. The syllabus says what every deck covers. Jump to
Session 1 · 2 ·
3 · 4 ·
5 · 6.
Session 1 — Orientation
AI as tool, as object, and as an agent with incentives; the workbench every lab runs on. 5 decks, 140 pages.
All 5 decks in one PDF, with a bookmark per deckDownload PDF
1.1 — AI in Science: What Has Already Happened (21 pp.)Download PDF
1.2 — The Economics of Producing Intelligence (22 pp.)Download PDF
1.3 — What Is AI? Foundations (29 pp.)Download PDF
1.4 — Workhorse Macroeconomic Models (29 pp.)Download PDF
1.5 — The Workbench: Python, VS Code, Git and GitHub (39 pp.)Download PDF
Session 2 — Deep learning
Functions become trainable; equations become losses. 6 decks, 172 pages.
All 6 decks in one PDF, with a bookmark per deckDownload PDF
2.1 — The Learning Problem: a Handwritten Digit, and the Neuron (21 pp.)Download PDF
2.2 — Nonlinearity: the Activation, XOR, and What a Network Can Approximate (28 pp.)Download PDF
2.3 — Forward Propagation, Backpropagation, and Training (31 pp.)Download PDF
2.4 — PyTorch, Curve Fitting, and Yield Forecasts (31 pp.)Download PDF
2.5 — Solving the Growth Model as Supervised Learning on the Euler Equation (29 pp.)Download PDF
2.6 — Architecture as a Constraint on W, and What More Parameters Buy (32 pp.)Download PDF
Session 3 — Reinforcement learning
Sequential decisions are learned from interaction. 6 decks, 162 pages.
All 6 decks in one PDF, with a bookmark per deckDownload PDF
3.1 — The Growth Model, With One Thing Hidden (28 pp.)Download PDF
3.2 — The Simplest Actor–Critic: How It Works, Why It Works, and the Wall It Meets (30 pp.)Download PDF
3.3 — Monte Carlo: Evaluation, Control, and the Data (26 pp.)Download PDF
3.4 — Temporal Difference: TD(0), SARSA, Q-Learning (26 pp.)Download PDF
3.5 — From Tables to Networks: Function Approximation, Policy Gradient, Actor–Critic (26 pp.)Download PDF
3.6 — The Growth Model Across Algorithms (26 pp.)Download PDF
Session 4 — Heterogeneous agents
Distributions and equilibrium become state variables. 4 decks, 168 pages.
Decks 4.1, 4.2 and 4.4 include pages marked Preliminary: they are based on research in progress by Zhigang Feng, Jiequn Han, Thomas J. Sargent and Shenghao Zhu (September 2026). Please read them for this course and do not circulate them further.
All 4 decks in one PDF, with a bookmark per deckDownload PDF
4.1 — Aiyagari, Solved Twice: A Fixed Point, Then a Law of Motion (30 pp.)Download PDF
4.2 — Krusell–Smith, Solved Three Ways (55 pp.)Download PDF
4.3 — Cohorts Under Aggregate Risk (45 pp.)Download PDF
4.4 — Optimal Taxation Without Commitment (38 pp.)Download PDF
Session 5 — Text and LLMs
Language becomes data, measurement, and generation. 6 decks, 208 pages.
All 6 decks in one PDF, with a bookmark per deckDownload PDF
5.1 — Text as Economic Data: From Documents to Counts (38 pp.)Download PDF
5.2 — Meaning in Vectors: Embeddings, Attention, the Transformer (43 pp.)Download PDF
5.3 — Pretrained Language Models: BERT, GPT, Alignment, and Five Limits (37 pp.)Download PDF
5.4 — The LLM as a Measurement Instrument (30 pp.)Download PDF
5.5 — Simulated Agents, Prompts as Design, and Reproducible LLM Research (30 pp.)Download PDF
5.6 — Text Signals and the Bond-Yield Forecast (30 pp.)Download PDF
Session 6 — RAG and agents
Retrieval, tools, traces, and governance in one workflow. 5 decks, 176 pages.
All 5 decks in one PDF, with a bookmark per deckDownload PDF
6.1 — From Chat Window to Agentic AI: What an Agent Is, and Where It Pays (36 pp.)Download PDF
6.2 — How Agentic AI Works: A Loop Built From Scratch, the Harness, the Session, and the Team (35 pp.)Download PDF
6.3 — How to Build With Agentic AI: The System, the Homotopy Method, a Research Case, and the Checks (35 pp.)Download PDF
6.4 — Build Your Own RAG: Give Your Agent a Library (35 pp.)Download PDF
6.5 — GRAM: A Retrieval System Built for One Literature –- and What Agents Do to Research (35 pp.)Download PDF
Earlier version. The ten-lecture version of this course, taught at Zhejiang
University in Summer 2026, is kept with its decks on the
Summer 2026 page.