← Previous · All Episodes · Next →
Is AI's weakness at "having novel ideas" temporary, or deeper? Episode 13

Is AI's weakness at "having novel ideas" temporary, or deeper?

· 55:54

|

Why "novel ideas" covers at least five different weaknesses moving at different speeds; the 2026 mathematics results and what nine mathematicians said about them; why the tests of AI creativity are weak; the strongest temporary and deeper cases; and what elite human creators actually do, including how fields decide what counted as new.

0:15 The short answer
3:52 1. What exactly is the weakness?
14:45 2. How would we know if it closed?
18:51 3. The strongest case that the weakness is temporary
27:36 4. The strongest case that the weakness goes deeper
34:39 5. What elite humans actually do, and how good they are
42:06 6. Established, disputed, open
45:28 7. What this means for your question
48:28 8. Where to go further
49:34 Appendix A. Models and dates behind the AI studies cited (you can stop here)
54:55 Appendix B. Sources not obtained in full


Subscribe

Listen to Adam's Podcast using one of many popular podcasting apps or directories.

← Previous · All Episodes · Next →