NYU's post-ChatGPT ML course revamp: teaching students to get smarter with AI, not dumber

gregd_nlp · x · 2026-09-13

David Bau set a 2026 teaching goal: help students master 'the difficult art of using AI agents in ways that make students smarter, not dumber,' and asked peers for lesson ideas. Kyunghyun Cho's NYU blog post offers data: he fully revamped the undergraduate ML fundamentals course after teaching it pre-ChatGPT in 2020. An informal survey showed most graduating CS students had never tried coding assistants. His philosophy: give students a safe environment to make mistakes with AI at school rather than on the job. The course focuses on math/stat fundamentals and the algorithms they realize, without going deep into implementation details.

Related event: Professor Bau: Teach Students to Get Smarter, Not Dumber, With AI Agents(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →