They state replacing backprop is not their goal.

Their goal is to understand how distributed systems which cannot do backprop (the brain) can still do learning.