@chipro The last one might refer to the fact that model accuracy does not necessarily mean application impact since production integration is still needed. The others are indeed a mystery :-)
"'Human-imitative AI' isn't necessarily the right goal. Human brains did not evolve to solve many of the problems that arise in large-scale networked systems and we should solve those latter problems on their own terms." - Michael Jordan. Inspirational keynote. #opml19