Stanford CS224V Agentic AI course opens: a bottom-up anti-hallucination stack for trustworthy agents
stanfordnlp · x · 2026-10-03
Stanford's CS224V "Agentic AI" course (Monica Lam's team) is now public, built around one question: how to turn unreliable, hallucination-prone LLMs into trustworthy, accountable agents.
The 15-lecture syllabus forms a bottom-up "anti-hallucination" stack: LLM basics; knowledge curation (STORM/Co-STORM); reliable QA over free text via RAG (ColBERT, RankGPT plus generation and verification); agents over structured and heterogeneous data (the course's most distinctive part); formal methods and accountable AI; task-oriented dialogue agents (Genie Worksheet); and meta-agents/meta-optimization.
Slides for the first two lectures are already available. The course argues such systems could accelerate science and expand access to medicine, law, and education—provided reliability is solved first.
More from coding & agent
- Jevbox: open-source self-organizing document drive with hierarchical search, no vector DB — rickasaurus · 2026-10-03
- AI will write 748 tests — except they are all the wrong ones — chrisalbon · 2026-10-03
- After two years building agent memory continuity, the same Opus follows him everywhere — RileyRalmuto · 2026-10-03
- Fireworks launches cache-aware FireRouter with Opus: coding costs cut 57% at 98% accuracy — Madisonkanna · 2026-10-03
- MegaCapybara: RTX 5090-only inference engine hits 2000+ t/s, 2x faster than Ninfer — BringTea_666 · 2026-10-03
- Dev deletes 800k lines of unit tests, argues they lock in slop code in the AI era — rickasaurus · 2026-10-03