DiscoverScience on Player FMCan AI help us solve the hardest problems in Mathematics? (ft. Terry Tao) - Into the Impossible With Brian Keating
Can AI help us solve the hardest problems in Mathematics? (ft. Terry Tao) - Into the Impossible With Brian Keating

Can AI help us solve the hardest problems in Mathematics? (ft. Terry Tao) - Into the Impossible With Brian Keating

Update: 2025-12-30
Share

Description

Answer my survey to get a chance to win a $100 Amazon Gift Card!
👉 https://forms.gle/uWgVRCf3BC2xxR2Y7


Please join my mailing list to get FREE notes & resources from this show! Click 👉 http://briankeating.com/yt


Will AI solve future math proofs?


Every time you type a password, buy something online, or send an encrypted message, you’re trusting an assumption about prime numbers: that they don’t hide an exploitable pattern. Modern cryptography depends on primes behaving “randomly enough,” yet many fundamental questions about primes remain unproven.



In this episode, I’m joined by Fields Medalist Terence Tao to explore what mathematicians can prove, what they strongly suspect, and what could change if unexpected structure appeared in the primes. We discuss pseudorandomness and why it matters for encryption, the twin prime conjecture, and how quantum computing reshapes what is feasible in computation and security.



We also get into AI and mathematics: why large language models can sound convincing even when unreliable, how AI can help with idea generation and literature recall, and why verification and proof assistants will matter if AI is to contribute to real mathematical progress. Along the way, Tao explains proof techniques like proof by contradiction, why complex numbers and the square root of minus one are so central, and how high-dimensional geometry breaks low-dimensional intuition.



Finally, Tao shares a real-world example of how math breakthroughs translate into technology: compressed sensing, which has enabled much faster MRI scans by reconstructing images from far less data.



Timestamps:


    00:00 Discrepancy Theory Explained

    09:35 Induction: Science and Mathematics

    14:38 "Proof Concept Through Play"

    18:43 "Complex Numbers and Completeness"

    25:36 Prime Numbers and Cryptography

    29:51 "Computability and Complexity in Mathematics"

    35:04 AI Discovers New Knot Theory Insights

    42:02 "Elegance in Nature's Laws"

    47:12 "AI as Complementary Research Tools"

    51:27 "Humility in Pursuit of Proofs"

    54:46 "Multiple Approaches to Mathematics"


    01:07:14 "Origins of Compressed Sensing"

    01:10:17 "Shannon Bound and Information Limits"


Follow Terry's Blog: https://terrytao.wordpress.com/


Get My NEW Book: Focus Like a Nobel Prize Winner: https://a.co/d/hi50U9U



-


Join this channel to get access to perks like monthly Office Hours:


https://www.youtube.com/channel/UCmXH_moPhfkqCk6S3b9RWuw/join



My tell-all cosmic memoir Losing the Nobel Prize: http://amzn.to/2sa5UpA




Follow me to ask questions of my guests:


    🏄‍♂️ Twitter: https://twitter.com/DrBrianKeating

    🔔 Subscribe https://www.youtube.com/DrBrianKeating?sub_confirmation=1

    📝 Join my mailing list; just click here http://briankeating.com/list

    ✍️ Detailed Blog posts here: https://briankeating.com/blog

    🎙️ Listen on audio-only platforms: https://briankeating.com/podcast


#universe #podcast #briankeating #intotheimpossible #science #astronomy #cosmology #cosmicmicrowavebackground #intotheimpossible #briankeating #terrencetao

Learn more about your ad choices. Visit megaphone.fm/adchoices

Comments 
In Channel
loading
00:00
00:00
x

0.5x

0.8x

1.0x

1.25x

1.5x

2.0x

3.0x

Sleep Timer

Off

End of Episode

5 Minutes

10 Minutes

15 Minutes

30 Minutes

45 Minutes

60 Minutes

120 Minutes

Can AI help us solve the hardest problems in Mathematics? (ft. Terry Tao) - Into the Impossible With Brian Keating

Can AI help us solve the hardest problems in Mathematics? (ft. Terry Tao) - Into the Impossible With Brian Keating