Parallelizing `loo`: Status Update 02

I’ve been stuck on a problem I ran when trying to integrate mori with mirai—specifically, if a user has a compute profile with mixed remote and local daemons, what is the cleanest way of mori sharing arguments to just the local daemons and copying data to the remote ones. Beyond researching this, I’ve not made any major changes. I extended the report’s section on mori a little but still need to write a lot more about how mirai and mori work. I also profiled loo::loo() and it seems like the recently released debrief⁠ package could be used to analyze those results further using an LLM.