From my experiments, it's neither.
The fail in Shafer's design was not the decision of 1UPT itself. It is surprising how he even fails to understand his own fail in the infamous article he wrote when trying to promote his new "self" for his new game. The core problem is that when implementing 1UPT, he failed to complement the decision with its fundamental companion:
Maneuver room!
THAT is what is lacking in Civ5's daring experiment. If you do not believe me, just modify one variable, ONE (MIN_CITY_RANGE), to make more "inter-city" room; increase it to 5 (default is 3), and you will see that even this AI that we so harshly criticize (with justification) starts to shine. Try it.
What 1UPT needs to work (and not only for the AI, even for us humans) is lots of space. The 1UPT decision should have been strictly tied to a drastic change of the old city-centric model of the franchise. 1UPT cannot, and will not, work when the center of everything in the map is the city (and worse, tightly spaced cities). Granted, we cannot part with cities in Civ, they still have a role to play as local centers of power, but production, resource gathering and even population should be dispersed in "province-like" areas, where one city is the "concentrator" and defines control of the area.
To make it graphic, think the strategy map of Napoleon: Total War. I am not talking about real time, detailed battles here: I am talking about the main map. The city was the center of power, and defined control of the local area (the province in that game), but resources, production and even some buildings were scattered around the province. This provided with two huge benefits: 1) Maneuver room, and 2) an important strategic layer of gameplay.
1) is obviously needed both for human enjoyment and for AI's performance. 2) adds a layer of strategy not seen yet in Civ: you can use warfare not necessarily to take control of the entire province, but also to disrupt the enemy's economy, production and even progress (scientific buildings were also outside the city, as academies are in civ5).
I can only hope that FXS learns the lessons from Shafer's biggest mistake, and from other good examples to emulate, and applies them to Civ 6. To be honest, I don't want 1UPT gone, as it can add much more enjoyment if correctly implemented. I want a 1UPT model that is complete, and includes its main, biggest complement:
Maneuver room!