Mh. The article is unimpressive. The underlying assumption is that all problems are city-scale problems, which is false. Your 80-20 autonomy above is much better.
There is another piece of vague thinking in here, which is this. When the simple life guys say “city” they seem to think of, I don’t know, Antwerp. Or Parma. Or Dortmund. Or do they think of Beijing, or Cairo, or Mexico City? These are very different objects. Larger cities are (much) more efficient, because their material consumption increases sublinearly, and their knowledge production superlinearly, in city size.
Even your “oh, trade can be less than 5% of GDP, and then we do not need to attract foreign investment anymore” looks completely impossible for Antwerp, and much more realistic for Beijing. But then again, governing Beijing is no longer “local” governance: that city is as large as Belgium, and it has twice the number of inhabitants. So you will still have a more local governance layer, as well as a non-local, metropolis-scale one for metropolis-scale problems (which will be many).
But of course living in Mexico City does not appeal to radical open source digital nomads, who dream of high-tech countryside, so that plan is out the window, too.