XelaP comments on Utility Maximization = Description Length Minimization

[XelaP]

I think someone should show what P(X|M_2) looks like after a Jeffrey-Bolker rotation, and see if there’s a nice interpretation of what happens.