I think that the main line is this:
Someone is concerned that the current calculation method is not conservative. Someone else came and say it is true and the coeficients are wrong and in fact we should consider other bigger coefficients. This will lead to an overall failure. Luckily, someone propose a solution that involve more resources and shows that in fact these coefficients are smaller and this solve your problem. But you have to employ more sofisticated software. One example is SIF determination and the other is the flange method approach.
In my opinion is good to have some official methods and to stick with them.
And to judge by yourself if something that you design is safe or not (by your self best knowledge as an engineer).
I bet a lot of the systems analysed 10 or more years ago will show failure under the new code.

Regards,