I’m curious if spin-up time per spec conveys an inherent time-to-execution penalty when each spec completes and the next spec file is loaded/executed (as compared to less spec files, more tests per spec)