4pm and on. Slowdown, stepwise. Discussing performance metrics, collecting the wrong data for the right figures, and generally misleading terminology. What's the point in a proof of concept if even a basic concept is still missing? (And back to the drawing board again. A few hours older.)