This page will discuss Robin Dunn's post at pyxides. Here is Robin's original post.
The essence of the "mission impossible" is this: "Create a code editor that will pry emacs and its 25-year-old nearly dead technology out from under my fingers."
Naturally, I think Leo is pretty close to fulfilling all of Robin's requirements, and a few others as well. Indeed, I think it unlikely that an editor will supplant Emacs merely by doing everything Emacs can do. To supplant Emacs, an editor must offer something fundamentally better. Obviously, I think Leo's outline view of programs and data does indeed offers something fundamentally better than the plain-old-Emacs view that everything is a flat file.
The only major unfinished item on Robin's list is autocompletion. Several other items are unpolished, including history list and other relatively-unused features. I expect that Leo will satisfy all items on Robin's list sometime in 2008.