Glad you find it much simpler – that was exactly the intent! I actually had the idea of simplifying the API for MvNormal when I myself went through Rethinking 1st edition 
I wouldn’t say that Cholesky factors and other algebraic tricks are “low-level details” though – they are actually pretty important to sample efficiently and, in the end, get reliable posterior distributions. I like McElreath quote that “the sampler is part of your model”.
That being said, these tricks are definitely overwhelming when starting, so it’s hard to find the sweet spot between making people aware that they shouldn’t trust the samplers blindly while making the API simple enough to use 
Regarding the 3.9.0 release, the milestone was finished yesterday, so we’re aiming for a release next week 
I’ll post back here once this is done and when chapter 14 of Rethinking 2 is merged in the resources repo 