Last week I told you about MIT’s new report on teaching in the age of AI. This week, it echoed something I’ve been building for over a year.
Justin Reich directs MIT’s Teaching Systems Lab, and I’ve known him since long before “AI in education” was a track at every conference. What struck me most, recording this earlier this year: he arrived, completely independently, at almost the exact same three ideas I built into MasteryFlip.
The moment that’ll stick with you — two computer science professors in Buenos Aires who moved their final exam to a 20-minute oral defense. What happened to their students’ work in the post-ChatGPT era, and why a live, unscripted conversation turned out to be the fix, tells you most of what you need to know about where this is all headed.
This one runs long, about 50 minutes, but it’s one of the richest conversations I’ve had on this show. Grab a coffee.
Key Moments
0:00 – Cold open: Reich on why we should be listening to teachers, not pundits
0:21 – The MIT report, and why it’s a bigger deal than it looks
3:45 – The “Homework Machine” research project — what 120 teachers and students said
6:11 – Jon’s own AI failure story, and the worst reviews of his career
33:25 – The DeepMind story — nobody’s figured out how to train junior engineers with AI either
42:46 – “Get a license” — Reich’s grandfather’s advice
