In the late 18th century, the Spinning Jenny allowed one worker to operate 120 spindles instead of one. It shifted value from manual toil to machine oversight.

GenAI is doing something similar.

Scoping and evaluating seem to be the new superpowers.

But both require subject-matter expertise.

You cannot evaluate what you do not understand. You can only know what “good” AI output looks like if you already know the subject deeply.

What do you think are the implications for learning? Is procedural fluency (knowing how to produce the work) a prerequisite for evaluative fluency (knowing how to assess and improve the work), or can we shortcut straight to expert judgment via alternative learning pathways?

Ethan Mollick Daisy Christodoulou The School of Knowledge curious where your thinking is on this. Any recommended reads?