I believe this is already happening with radiology -- robotic vision + edge-detection can identify masses and growths better than human specialists -- medical professionals I trust say "no more radiologists after this generation."

Perhaps the temporary 'trick' is to find sub-disciplines in which "human intuition" still plays a material part -- i.e., not formulaic credit-checks or bid-pricing -- and to acknowledge that we're probably still decades away from true AI singularity, which will be human-disruptive past any predictability.