I think it's worth having this discussion if we can separate the value (which is arguable) from the process, which is in fact pretty ambitious.
From a technical point of view, I think it is clear that MQA is "not just a codec" by any sense of the meaning. The full scope of it's ambition, to correct all timing errors in the A/D, D/A chain and to do so using equipment specific data is monumental.
Benchmark has one of the best explanations, as well as technical criticisms I've ever read about MQA here:
https://benchmarkmedia.com/blogs/application_notes/163302855-is-mqa-doaThe blog puts to rest any attempt to relegate MQA to the world of "mere codecs."
Is it worth while? On my MQA capable Brooklyn I leave it off. You should do as suits you best.
Best,
E