MIT releases full Multimodal AI course videos with agents, self-evolving AI, touch and smell
JeremyNguyenPhD · x · 2026-09-09
MIT's Paul Liang has publicly released lecture videos and materials for the spring 2026 edition of How to AI (Almost) Anything / Multimodal AI, a follow-up to his widely shared spring 2025 course.
The updated version covers multimodal agents, reasoning, self-evolving AI, and new modalities like touch and smell, plus guest-lecture slides on multimodal AI for health, design, manufacturing, cities, and transportation.
All videos are on a YouTube playlist, with the full course website and materials available online.
Related event: MIT releases full multimodal AI course videos and materials(2 posts)→
More from Companies & People
- Meta exec touts Muse as potentially the best new product of the year — alexandr_wang · 2026-09-09
- ZhenFund portfolio goes all-in on AI: hiring across models, robotics and chips — sujingshen · 2026-09-09
- Terence Tao slams OpenAI for scooping researchers' work, warns of lasting damage to math — Nikvest · 2026-09-09
- New AI podcast interviews attention paper author on its backstory — kchonyc · 2026-09-09
- Lenny shares: the scarce PM skill today is elevating other people's ambition — lennysan · 2026-09-09
- a16z GP: agents can't 'make me a million dollars' — human direction still unlocks the next hill — lennysan · 2026-09-09