Katalina Hernandez comments on Interpretability Will Not Reliably Find Deceptive AI