That was always the main question when interviewing candidates for a job at McKinsey. Not, can they run big divisions in their head, but how do they reason, what logic do they follow, how to decide between multiple options, how/when do they spot errors in their own work (and back up/correct them).
Things are similar with AI now. Eye balling the logs when a model ploughs through a 15 minute job gives (or doesn’t give) you the confidence to give it bigger responsibilities.
SlideMagic: a platform for magical presentations. Free student plan available.