1: Output of single cities irrelevant in face of sheer numbers. Doesn't matter if you have one really good city you can more than make up for it with multiple cities at a far more efficient cost. Wonders don't really matter either.
Other than too heavily-pushing 'go tall' a la Civ V, which was unpopular (judging not just from comments here but from the direction chosen with Civ VI - while Civ V may still be the most popular game in the series by its player numbers, Civ VI's design wouldn't have abandoned tall altogether if it was seen as a major selling point), it's hard to avoid this in a Civ game. Civ IV allowed a degree of city specialisation essentially absent in Civ VI, which added some identity to cities, and Wonders were of more importance, but then you just spammed the map with cookie-cutter cities that all fit into one of the three stereotyped specialisations. You almost might as well have had a dropdown menu when you settled asking whether you wanted a Great Person (=food) city, a production city or a commerce city and have the AI governor automate the rest.
2: Can be done if Civ 6 stops rewarding short term activity at the cost of long term investments. Making harvestables scale exponentially with Eras for example, makes harvesting a no-brainer compared with investing in productive potential etc.
As I say, that would require rebuilding the game from the ground up. The very existence of eurekas and inspirations is pure short-term advantage - changing the size of the bonus they provide doesn't change that. Similarly to a large extent with harvesting, and the fact that Civ VI allows you to harvest pretty much any landscape feature other than luxury or strategic resources, so that you needn't find focus on woods but can just settle any site with bonus resources that will always be good tiles.
Much of the game's pacing is predicated on the existence of these systems - without them the tech and civics trees would need reworking entirely, for a start. We have the change-on-a-whim policy system - that could be made a bit more long-term by removing the 'free switch when you get a civic' and add a maintenance cost akin to Civ IV, but it still ultimately rewards short-termism. The victory conditions are mostly designed to be achievable from a very late start if you pivot into them - you can get a cultural victory purely on tourism accrued from the Industrial era onwards, without needing specific (or even any) Wonders as in older games, while the absence of specialised branches in the tech tree naturally pushes you towards everything you need to transition into a science victory whatever your original intent.
3: Not denying possibility, denying competitive power relative to conquest.
There's never yet been a Civ game in which competing peacefully was less optimal than conquering the map. For most of the series' lifetime domination and diplomatic victories were just variants of the same thing. It's pretty hard to see how you even make a Civ game where peaceful play is better than aggression given the basic maths underlying the series: the game rewards having more population and more cities rather than fewer, and conquest is the easiest route to both, while at the same time being the major tool the series has for interacting with and disrupting the AI's strategy.. Civ V's 'favour tall' approach was widely disliked, and even there conquest was rewarded.
Overall, it sounds as though you basically want Civ V with districts - more irreversible decisions with long-term consequences, fewer and more individually distinct (yet specialised) cities, and the closest the series has come to making peaceful victory more attractive than warfare. That's not the direction they've chosen with Civ VI, whose design philosophy appears to be to ape the basic structure of Civ IV but with none of the depth.