Remix.run Logo
▲ shric a day ago

Genuine question: if I make 250k pages of lean that compiles, how do we know what it proved?

If it was as trivial as stating the expression and compiling then there would be easy verification and consensus as soon as it compiles on multiple kernels?

▲thesmtsolver2 20 hours ago | parent | next [-]

The result proved is not 250k pages. The theorem should be much smaller and a formal version of RH.

▲icepush a day ago | parent | prev [-]

Among those who have experience with lean there is consensus that it is correct and the Riemann theorem is proven.

▲roenxi a day ago | parent [-]

The "there is a consensus" standard of proof isn't strong when it comes to maths. People are generally looking for a rigorous argument why something must be true.

▲PierceJoy 21 hours ago | parent [-]

> The "there is a consensus" standard of proof isn't strong when it comes to maths.

Says who? I don't mean that in a rude way, it's a genuine question. Is that the opinion of the community of mathematicians who are capable of understanding these type of proofs, or software engineers who believe nothing can be true unless you can read the code yourself and understand it?