Arguably this is saying that machine learning does not easily and directly yield consciousness as one might expect. If consciousness instead is focused on time

Arguably this is saying that machine learning does not easily and directly yield consciousness as one might expect. If consciousness instead is focused on time and sequences then the machine learning needs to be used more to construct and compare narratives of sequences of observations and actions. That would enable machine learning to derive representations of causality and potentially reason that itself exists. Most machine learning focuses more on raw data and the environment rather than the nature of change over time. Just some thoughts.