Thanks for the question! In addition to what Matthias said, the distributions used in the two models and the way hyperparameters are estimated are different, but yes, they’re conceptually similar. I’d be really interested to see people try both approaches on their data and compare how they perform:)