I guess you’re onto something, that combined with reductive, concave, inference architecture gains and new flux capacitor designs we are on 10x performance maximum optimisation curve.

I can taste that robot barista cappuccino now.