logoalt Hacker News

vintermann • today at 3:29 PM • 6 replies • view on HN

At what point do we start worrying that it's found a bug in Lean instead?


Replies

d-u-d-e • today at 4:05 PM

The Lean kernel should be easy to verify by hand. We can ask the same question with humans though. What if a result that was used to prove some other result turns out to be wrong just because whoever reviewed it neglected a small detail?

➕ show 1 reply
thejokeisonme • today at 4:03 PM

Right now.

cma • today at 4:02 PM

Still possible, but Lean ships with formally verified kernels as of a few weeks ago.

➕ show 1 reply
suddenlybananas • today at 3:49 PM

In principle you can run Lean with different kernels and so there would have to be a bug in all of them for the proof to be faulty I believe.

perching_aix • today at 6:06 PM

Identically in all 4 independent checkers it ships with? Basically never.

dgellow • today at 4:28 PM

That already happened multiple times, it’s a known worry

➕ show 1 reply