Raising the street grade one story has been done before, when flush toilets were introduced in the late 1800's. Notably in Seattle and Chicago but in pockets other low-lying cities as well.
And until 2003, there was even an airstrip only 2mi from downtown.
They're basically the separately-operated basements of stores, hotels, and office buildings networked together. Usually one of them between you and your destination will decide to lock the doors.
I've found myself nearly under my destination, encountering a locked door, and having to backtrack and find my way out through a fitness club.
In NYC at least the lower levels are already present. But the only part where we got the pedestrian level correct so far is the High Line... which has been hugely successful. There's really no reason these types of park-like promenades only have to be built on top of existing disused structures however. More could be built, it's just that people need to understand that there is return on investment for building an elevated pedestrian-only street level.
Not an easy thing to develop certainly, but relative to the cost and level of engineering effort needed to build new subway tunnels, I'm surprised that adding transportation improvements at street-grade and then bumping the street up one level, isn't something that ever gets considered. We could run effective BRT through Manhattan, or surface light rail to solve the crosstown transportation problem.
Imagine how much easier getting from Queens and Brooklyn to the Lincoln and Holland tunnels would be by car or bus if say 39th St, Canal St moved the pedestrians to an elevated walkway. What if Times Square's tourist bottleneck could be eliminated by raising Times Square?
Anyhow, unlikely to happen probably ever, but fun thought excercise
This (sort of) already exists in Minneapolis (and to a much smaller extent, St. Paul and Duluth) courtesy of the skyway system.[1]
It's very well done and has created a completely separate urban ecosystem overlaid on top of the existing, street level, downtown.
That wouldn't remove the need for ventilation and other things you mentioned if that system is underground. But it could be above the ground , on buildings 2nd floors, where the first floor would be retail stores and whatnot
Why not have seats on the pod, that way you don't have to ferry around a 2-tonne car. Indeed if you have solved smart automation, why have rails at all? Presto, you arrive at the idea of self-driving taxis as a public service.
Really, the car would (should!) still be driving itself—it's the thing that gets certified on how to handle safety issues, after all; the car would just be given a dead-reckoning location target by the zone controller, which the car would get to as best it knows how, with the zone controller scanning to "see" when it had successfully boarded, just like with a human-driven car.
If there were few-enough human-driven cars, you could have a team of valet staff accompanying the train, acting as glue for the few human-driven cars in the pack.
If anything, it is much more valuable to segment the city in ways that you don't need subways or roads. You should have districts that are internally navigable strictly by foot, bikes and tram between adjacent districts, and only have rail / road when going beyond the local microcosm of city.
You want greenery anyway, you might as well put it in the bounding lines between regions of your city and have the roads / rail between those to minimize noise pollution outside of the transit hubs.
Unfortunately, it is very expensive to operate that way. The easiest tunnels to build are the ones Paris did for their subway a century ago: dig up the soft limestone under the street, lay tracks, close it up. But you still need to deal with ventilation (especially for non-electric cars), fire suppressant, evacuation, flooding and general maintenance. In the best-case scenario, tunnels are just very expensive.
Then there are places like Atlanta (sitting on granite) or San Francisco (in a fault zone) where things get a lot more complicated.