"For example, English-speaking WikiTree users know William the Conqueror. But French-speakers know Guillaume le Conquérant. Even if 90% of current WikiTree users speak English, William should be Guillaume in our database because he himself spoke French. We can all share one profile even though we don't share one language."
Modern French speakers know him as that but his name at the time was not spelt in modern French, or for that matter spelt consistently or always in French. This is a clear case of a little learning being a dangerous thing.
It would be more sensible for Wikitree to allow for names to appear in more than one language especially for historical figures. This might involve some additional name categories in the forms.
On the other hand, it should be noted one wouldn't and doesn't find historians writing in English referring to King William I a.k.a. William the Conqueror or William the Bastard or William of Normandy, etc., as Guillaume le Conquérant all the time.
I suppose this is what leads to the nonsensical use of Cyrillic characters to display some of the names of the Duke of Edinburgh's Russian imperial ancestors. Political correctness mixed with a little learning. This is P.C. poseurism and obscurantism when taken to its logical and ludicrous conclusion.
Wikitree naming conventions need to be developed in ways that allow users to be multilingual in a more sensible way than that! After all, we want people to be able to read and understand the biographies entered (I presume).