Could AI Find Physics We Can’t Understand? @TheoriesofEverything
Could AI Find Physics We Can’t Understand?  @TheoriesofEverything
Uploaded October 2026 | Updated November 2026, 3 days ago
SPONSORS:
- Go to plaud.ai/curt and use the promo code "CURT" to get a Plaud device today
- Our listeners get up to 30% off their entire order at undrdog.com

Every episode days early, ad-free, plus my essays: curtjaimungal.com

This is a podcast about whether AI could discover the laws of physics before we do. Hirosi Ooguri, chair of Caltech’s Division of Physics, Mathematics and Astronomy, joins for his first English podcast to explain why AI has moved faster in mathematics than in physics. He fears it could solve physics in a way we cannot understand. We discuss why quantum field theory lacks a firm mathematical foundation, quantum gravity as the “final boss,” the swampland distance conjecture he proposed with Cumrun Vafa, and his estimate that, left to humans, verifying string theory could take 2,250 years. The conversation also covers the PhD thesis numbers that turned up in a Mathieu group table twenty years later. What an honor.

FOLLOW:
- Spotify: open.spotify.com/show/4gL14b92xAErofYQA7bU4e
- Substack: curtjaimungal.com
- Twitter: https://x.com/TOEwithCurt
- Discord: discord.gg/kBcnfNVwqs
- Crypto: nowpayments.io/donation/TOE
- PayPal: paypal.com/donate?hosted_button_id=XUBHNMFXUX5S4

TIMESTAMPS:
- 00:00 - Measuring Earth in Fifth Grade
- 06:58 - Why Physics, Not Platonic Math?
- 12:34 - Nothing Below the Planck Length?
- 23:15 - Why AI Can't Verify Physics
- 30:47 - Yang-Mills: Hardest Millennium Problem
- 36:31 - Why QFT Resists Rigor
- 46:19 - What If Yang-Mills Fails?
- 52:35 - Are LLMs Just Grad Students?
- 01:01:09 - Swampland and AdS/CFT
- 01:12:57 - Gravity Breaks Wilson's Paradigm
- 01:21:45 - Will AI Surpass Physicists?
- 01:26:58 - 2,250 Years to Verify Strings?
- 01:34:03 - Is String Theory Disappointing?
- 01:40:27 - Distance Conjecture and Inflation
- 01:46:24 - Mathieu Moonshine Discovery
- 01:52:59 - What Would Make Him Quit Strings?
- 02:02:28 - Days of Blank Paper

LINKS:
- Hirosi Ooguri [Caltech PMA]: https://www.pma.caltech.edu/people/hiroshi-hirosi-ooguri-oguri
- On the Geometry of the String Landscape and the Swampland, Ooguri & Vafa [Paper]: arxiv.org/abs/hep-th/0605264
- Non-Supersymmetric AdS and the Swampland, Ooguri & Vafa [Paper]: arxiv.org/abs/1610.01533
- Constraints on Symmetry From Holography, Harlow & Ooguri [Paper]: arxiv.org/abs/1810.05337
- Continuous Family of CFTs and Exactly Marginal Operators, Komatsu et al. [Paper]: arxiv.org/abs/2512.11045
- Universal Bounds on CFT Distance Conjecture, Ooguri & Wang [Paper]: arxiv.org/abs/2405.00674
- Notes on the K3 Surface and the Mathieu Group M24, Eguchi, Ooguri & Tachikawa [Paper]: arxiv.org/abs/1004.0956
- Exploring the Holographic Entropy Cone via Reinforcement Learning, He, Lee & Ooguri [Paper]: arxiv.org/abs/2601.19979
- The Holographic Entropy Cone, Bao et al. [Paper]: arxiv.org/abs/1505.07839
- De Sitter Space and the Swampland, Obied et al. [Paper]: arxiv.org/abs/1806.08362
- Knot Invariants and Topological Strings, Ooguri & Vafa [Paper]: arxiv.org/abs/hep-th/9912123
- Topological Lattice Models in Four Dimensions, Ooguri [Paper]: arxiv.org/abs/hep-th/9205090
- Kodaira-Spencer Theory of Gravity and Exact Results for Quantum String Amplitudes [Paper]: arxiv.org/abs/hep-th/9309140

Full reading list for this episode (all 103 sources): curtjaimungal.com

Guests do not pay to appear. #science
Could AI Find Physics We Can’t Understand?We Solved the Ghost Problem in Three DaysI Quit Physics. Then I Won the Physics Nobel - Geoffrey HintonPain Showed Me How Lightly I Took Others SufferingI Dont Think String Theory Cuts the Mustard - Jim Al-KhaliliI Think Well Soon Have Proofs No One UnderstandsThere Are No Faces on People. You Project Them. - Joscha BachLie 20% of the Time, Become 20% a Liar - Jordan PetersonMy Colleagues Let AI Do the Part I EnjoyString Theorys Predictions Work. In Math. - Richard BorcherdsMy Grant Form Had No Box for Quantum Computing - David DeutschMany Worlds Is Non-Local. Sean Carroll Agrees.
Curt Jaimungal |

Could AI Find Physics We Can’t Understand?

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER