For command 11 in Section 0
# Draw 2 independent draws from a standard normal using `x = np.random.randn(2)`, and compute `nu` using `L @ x + mu`
I found that this returns a 2x2 matrix because numpy doesn't massage the 1-D x into a 2-D column column vector. x = np.random.randn(2, 1) resolves this.