The 3b1b Podcast podcast

Jordan Ellenberg on Being Dispensable

0:00
1:08:31
Rewind 15 seconds
Fast Forward 15 seconds

Jordan Ellenberg, a mathematician at the University of Wisconsin–Madison and author of "How Not to Be Wrong" and "Shape," joins the podcast for a conversation about what AI-generated proofs mean for the people who do mathematics. Jordan argues that real problems don’t have beginnings and endings, and that answering one question usually opens a better one.They discuss what the recent Siegel zeros proof reveals about the limits of human intuition, why a 20-billion-token proof that no one can read isn’t really a proof, and how graduate school is training in asking good questions. The conversation also covers fashion in mathematics, how finite group theory faded after its biggest problem was solved, why the field has “rabbis but no popes,” and why Jordan finds it reassuring that if he hadn’t proved his theorems, someone else would have.Join the Patreon at https://www.patreon.com/c/3blue1brown/Discover exciting roles for the mathematically curious at https://www.3blue1brown.com/talent


0:00 Introduction

2:50 The Siegel zeros proof and the limits of intuition

5:20 Learning from human math vs. self-play

8:17 Is an unreadable proof still a proof?

11:40 Will the joy of math survive?

13:30 Using machine learning to find examples

17:27 When machine results lead to insight

20:10 "We're all dispensable"

26:08 Fashion in mathematics

28:10 What happened to finite group theory

32:50 Rabbis, not popes: who decides what matters

37:23 Where do good questions come from?

42:34 A course on making conjectures

47:45 Why math isn't for everyone

49:38 Will AI change who becomes a mathematician?

50:27 Math as a social activity

54:25 Cohomology as a pencil-and-paper machine

56:30 How the future will see us

58:07 The general vs. the particular

1:03:18 Teaching with enthusiasm over clarity

More episodes from "The 3b1b Podcast"