OpenAI faster Fourier transform questioned over cost and assumptions
What happened
On October 13, an article on the Hacker News front page analyzed the faster Fourier transform claimed in OpenAI's math repository. It argues the asymptotic advantage is limited by huge hidden constants and arithmetic assumptions that don't match real computers, estimating the constant C at 10^100 when N=10^12. The algorithm also assumes arbitrary-precision complex multiplication takes one operation. The analysis does not confirm the theorem's correctness; it questions whether the theoretical result can speed up practical FFTs. No verification of real-world speedup is given.
Written by AI from the coverage · updated 1 hour ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pageA look at OpenAI’s math that’s claimed computes faster Fourier transforms
OpenAI 数学仓库声称给出更快的傅里叶变换，文章分析认为其渐近复杂度优势受巨大隐藏常数和脱离实际计算机的运算假设限制。文中估计，当 N=10^12 时常数 C 为 10^100，并指出算法假设任意精度复数乘法只需一次操作。文章未确认定理的正确性，主要质疑这种理论结果能否解决实际 FFT 加速问题。
Heat over time
Not enough continuous observations to draw a trend yet.