[eutectic]

I though finite difference differentiation was notoriously unstable.

[legobmw99]

Perhaps the commenter means something like reverse mode automatic differentiation?

Finite differences does indeed have stability issues, and even if you apply some tricks will only give you about half float precision