I’m curious what the rationale is for not making regions explicit in the source code. It seems like a downside of region inference for a systems language is the unpredictability of the inferred regions.