Tiny changes to a matrix can flip signs while breaking it into simpler parts, due to floating point quirks—causing big differences in results even with the same random seed. Safer methods exist, but tradeoffs apply. Full deep dive:
https://blog.djnavarro.net/posts/2025-05-18_multivariate-normal-sampling-floating-point/

When good pseudorandom numbers go bad – Notes from a data witch
Multivariate normal sampling can be wildly irreproducible if you’re not careful. Sometimes more than others. There are eldritch horrors, ill-conditioned matrices, and floating point nightmares in here. Teaching sand to do linear algebra was a mistake