Years ago we tried #predictive #test selection to fix an 8-hour CI loop (run test subsets). It worked, until it exposed intra-test dependencies we never knew existed. A #StarEast talk finally gave that problem a name, and detailed four different approaches pro/con
