These are cities where it is a common experience to arrive at the airport, take the train into town, and never leave cover for the whole duration of a stay.
It may not all the be same roof, but the permanently covered urban life is a reality today.
Perhaps unintentionally, this leads us towards a different model for making and assessing long-term technology predictions in the first place. Instead of making literal predictions based on form ("a roofed city"), we should make predictions based on function or goal ("effective shielding from inclement weather throughout the day"). Of course, even this framework doesn't protect us from the changing nature of needs and functions over time.
Or the MIT campus with its tunnels (you can walk from building 1 to E19 without going outside, which I will probably do when visiting in January to avoid the cold weather).
Instead we achieved an alternative that does the same thing.
A classic example of how these articles miss what in hindsight is obvious.
Mega projects hard.
Small component built projects easier.
You can drive your car from your garage to underground parking at your enormous Apple spaceship campus where you work, eat, and exercise under one roof. Then you can drive to the mall where you work, shop, and watch movies under one roof. Then you drive home. You're never exposed to the rain.
Obviously this doesn't apply to most people but shopping malls is what got me thinking.