Observing from a respectful distance, it seems to this old cynic that if you have some software/hardware that has twice demonstrated a capability to force a dive into the ground, the solution to the problem lies in removing/changing/improving the software/hardware, not in retraining pilots to deal with its behaviour in a simulator which may well not fully replicate the aberrant behaviour pilots may experience in practice.
Of course, if the need for the software is a design error which causes a pitch up which has to be corrected, the best solution is a redesign. But it would be a pity to scrap all those lovely new airliners, even though that would teach a lesson to manufacturers that speed into the marketplace does not trump getting the design right.
I can think of many situations in global airline operations where a response learned and practiced in a simulator may not solve the problem when it actually strikes. So for my money the industry, ICAO, Boeing, FAA et al, should focus on getting the software/hardware right, regardless of cost, not on training people to deal with it when it goes wrong, as a cheaper option.
And now I'll retreat back to the respectful distance.