I was having similar thoughts and questions. How about a comparison on a data science task? Say a side-by-side comparison processing 20-30 million records through a regression analysis consisting of a continuous outcome and 3-4 continuous indy vars? A collab on that?