Context - according to Wikipedia [0], just a few days ago someone posted a proof of a bound of 240 [1].

[0] https://en.wikipedia.org/wiki/Twin_prime [1] https://arxiv.org/abs/2608.31126

Related ongoing thread:

GPT-6 Astra - https://news.ycombinator.com/item?id=49554643

(see also https://news.ycombinator.com/item?id=49555621 from there)

What is this AI-generated gobbledygook actually trying to say? That there are infinitely many pairs of primes p,q with q=p+186?

The first equation under "The result" seems to be saying they found an infinite sequence of primes whose density is forever greater than 1/186, which doesn't match my understanding of how prime density works.

blegh, both the code comments and readme reek of verbose claude puke.

> The registered Lean build passed without errors or warnings.

thanks. i like it when my proofs do not error.

Edit: just saw this was published under openai, yikes.

see also https://github.com/openai/LongGapsBetweenPrimes

who knew the primary utility of pure math would be to juice pre-ipo valuations ¯\_(ツ)_/¯

[dead]