The appeal to authority in my original comment is not just deferring to trends or whatever; it's leaning on the deep knowledge (and inside-knowledge!) that a core React team member has of not only React's current implementation, but its underlying intents/philosophy (which informs which paths the team will focus on optimizing, for example), and even its future roadmap. I think it's legitimate to take those things into account, especially in the absence of any significant, concrete evidence showing that the practice in question can cause problems.
