Well, suggested solution of this problem seems kinda backwards to me. I'd much prefer something along the lines of what Stephen Wolfram suggests - offload computation to computers in the first pre-rigorous stage, which may enable building of the foundation in a more rigorous manner right from the start.
Even at rigorous stage, offloading computation increases dramatically exposure to solutions and builds intuition faster by orders of magnitude.
Even at rigorous stage, offloading computation increases dramatically exposure to solutions and builds intuition faster by orders of magnitude.