There is something that I find deeply unsatisfying about this. For me, the first draft is as much about working through the problems space and considering possibilities. If I rely on a chat bot, then I am more apt to become anchored to whatever the chat bot spits back out at me. Even if what it produces is good enough, I do not benefit from the drafting process in the way that I would if I did it myself. Sometimes maybe this is a good enough shortcut but I generally don't believe in shortcuts.
Actually it works not as the first draft, or the final draft - it is the middle draft. In stage one you just drop a bunch of bullet points, ideas, short notes. In stage 2 the model writes your article or paper. In stage 3 you fix it.
Almost 100% next iteration will sport a fact checker, powerful style and format controls, and a much larger context. The development of advanced fact checkers will have a big impact on anything propagated online.
