Stanford CS 329Z lecture 6: agent scaffolds as design decisions, taught by DSPy and SWE-agent creators
stanfordnlp · x · 2026-10-14
Stanford's CS 329Z released lecture 6, "Agent Design Patterns and Scaffolds," taught by Diyi Yang plus the creators of DSPy and SWE-agent. Built around decomposition, data, and evaluation, it covers the workflow-vs-agent taxonomy, five composable workflow patterns, and a homework where students build an agent by hand, then rebuild it in DSPy and reflect on what the framework hid. The sharer notes weeks of scheduled agent runs showed failures were scaffold failures, not model failures.
More from coding & agent
- Effect ships zero-dependency native test runner, agent-first — samgoodwin89 · 2026-10-14
- Codebases will look a lot like databases in the AI coding era — hargup13 · 2026-10-14
- Agent Argon reads Spot's SDK docs itself, builds trash-sorting skill on Boston Dynamics robot — chris_j_paxton · 2026-10-14
- Handling live web search in MCP agents without flooding the context window — Few_Training_8950 · 2026-10-14
- Developer runs Claude overnight to auto-localize Super Mario Odyssey on macOS — lxfater · 2026-10-14
- Intent ships PR review with inline diffs, Linear integration, board views and more — LukeW · 2026-10-14