One more thing. If there is no warning then I think by default there should be a warning about style inconsistency which people can turn off if they want, maybe the warning says how to turn it off.
As far as user adoption / feedback, I think that the extraordinary efforts that a small number of people have made should be applauded. Like I mentioned my theory is that the core problem is really about a surface level judgement that people have against things that are unfamiliar. Or maybe style inconsistency is just an easy concrete thing to pick when someone asks about adoption. The real answer may be just that it seems different and requires a lot of effort to learn new things. So if my theory is correct then by removing style inconsistency, that would remove that complaint but not necessarily increase adoption. The feedback would probably just pick on something else that is different as an excuse. You could A/B test by giving different talks and then looking at the feedback, but again that could be misleading because measuring actual adoption takes months if it's even possible.
