Perplexity AI Perplexity-Level AI Math
18 snips
Jan 15, 2026 Discover how AI is transforming mathematics by not only solving complex problems but inventing new methods, like optimizing matrix multiplication. Learn about groundbreaking advancements where models tackle previously unsolved math challenges. Hear about AI’s role in aiding researchers by streamlining literature reviews and validating arguments. The discussion also highlights the potential impact of enhanced AI reasoning across diverse fields, including engineering and medicine. Tune in for a mind-expanding exploration of math's future!
AI Snips
Chapters
Transcript
Episode notes
AI Rewrote Matrix Multiplication
- Jaeden Schafer recounts Mo Gawdat's keynote where AI redesigned matrix multiplication for a 26% performance boost.
- That change cut hundreds of millions in costs and energy for Google by inventing a new math method.
Models Solving High-Level Math
- Jaeden explains that modern models, like GPT-5 Pro, are solving hard abstract algebra problems quickly.
- These models demonstrate deep mathematical reasoning beyond rote calculation.
ChatGPT Solved An Unsolved Problem
- Jaeden shares Neil Somani's experiment where ChatGPT solved a long unsolved problem in 15 minutes and passed verification.
- The solution checked out using Harmonic's tool and surprised the engineer with its correctness.