I have to create lots of short clips to post to various platforms (and they often have limits of ~20MB). In order to get the best possible quality within the size limit I have to do math each time (which is a pain when trying to export multiple clips quickly).
This is difficult to predict accurately, and it will generate too many bug reports when it does not work either by going over the size or not getting close enough (if the heuristic is conservative). In other words, it will take a very large amount of time to develop heuristics that work with all the different codecs, rate control methods, and formats. That is time better spent elsewhere in my opinion.