Please do! I think the community would benefit a lot. Something I would like to see done/discussed in a paper like that is how to incorporate phylogenetic uncertainty (maybe something like this?) and how much it matters. This could be a case where predictive checks would also be interesting; looking at how entropic predictions get with/without phylogenetic uncertainty.
Regarding model adequacy/fit specifically, I suppose it’d be interesting to compare loo with what is usually employed in the comparative method literature, which are Bayes factor/marginal likelihood-oriented methods*.
*I believe, not an expert in the comparative method.