Acer

@AcerFur

Furry pure maths student @Cambridge_Uni. | 🇬🇧🇵🇹 22 He/Him | Interested in all things mathematical reasoning @OpenAI | Opinions are my own.

Bay Area, SF

Joined March 2020

Pinned
Acer
@AcerFur
Aug 1
openai.com
Ten advances in mathematics and theoretical computer science
OpenAI shares new results on long-standing open problems in mathematics and theoretical computer science, including advances in geometry, cryptography, and complexity.
Acer
@AcerFur
15h
I suspect it’s not too hard to find a Hadamard matrix when you have supercomputers worth of compute clusters available with a pretty good coding AI lol Still pretty cool to see all cases <2000 done. General construction algorithm for n=4k when though?
Acer
@AcerFur
Aug 10
It is a wide belief that the simple zero conjecture is harder than RH. It would be funny if these are done together in future.
Acer
@AcerFur
Aug 10
People are calling me double-faced about this take. To be clear, I have the same opinion of GPT’s capabilities currently.
Acer
@AcerFur
Aug 10
Replying to @AcerFur
7/7 The machinery is not novel and the idea is natural but maybe slightly novel and certainly clever. So well done to Anthropic, but make sure we understand that Claude here has still “just” recombined known things in a clever way.
Acer
@AcerFur
Aug 10
1/7 Okay, I’ve spent my lunch digesting the paper more and I think Claude didn’t make much of the contribution here. A lot of the key work was of BGSTB24/25 which I was familiar with when these papers released and were pretty big deals.