Remix.run Logo
▲ kadoban a day ago

My understanding is that there is no central, agreed, library of lean formulations of hypothese+axioms for big open problems. So it's not like, when proving one of these you necessarily just take the lean-encoded problem off the shelf and add the "i proved it" parts.

First step seems to be you encode the problem according to your whims. So there's some checking to be done in terms of "does this actually encode the problem?"

For pro mathematicians I wouldn't expect that that's where they run into trouble in things like this, but still needs a look-over.

▲raincole 11 hours ago | parent | next [-]

Claude's proof just uses Mathlib's formalization of Riemann's Hypothesis. It's the closest thing you can find to an "agreed" one. It's not like that Claude tried to formalize it from human language from scratch.

▲twothreeone 21 hours ago | parent | prev [-]

Often (most?) of the time "solving" a hard problem is partly taking it apart into separate pieces and phrasing those pieces in various novel ways that allow one to tackle them individually and then at the end it all magically falls into place (in this case after 900 pages).