In the US, obtaining right-of-way to build bike lanes in metropolitan areas is infeasible. I've sat in neighborhood meetings with my state's Department of Transportation engineers and planners discussing main artery expansions, and they explained that there isn't sufficient demand for bike paths to include them in any new planning funded by taxpayers; whenever they've tried to include dedicated bike lanes in the past, they were shot down first by politicians looking to curry the fiscal conservative votes, then by special interest groups protecting a car-based culture (housing developers, chambers of commerce,
etc. all see dedicated bike lanes as yielding less traffic to their planned projects). However, obtaining right-of-way to privately build an elevated path, they conceded, while not easy, was more feasible because someone else foots the bill. The problem is no one has figured out a design that can feasibly pay off.
If you can obtain the right-of-way for an elevated path, then what would you design for it that makes it pay off at least a little above break-even?
Generally, I take this as a sign that in the US, unless you deliberately attract a demographic to a specific concentrated area that consciously chooses biking as a major transportation mode, namely, a master-planned community/town built around biking and public transit, cyclists are SOL to comprehensively expand biking for the foreseeable future in most US metro areas outside of "bike ghettos" where there is a strong cycling culture. I'd love to see as much biking for normal daily activities as we saw in mainland China before they opened up to the West, but in the US in many areas it is legitimately risky to bike (despite the actual risk of fatality on a bike gradually dropping over the past three decades).
So a related hypothetical is, if you were in a high-density (say, 30K per square mile scale density) city designed from the beginning for walk/run/bike/public-transit transportation, what kind of bike paths would you want?
The article describes intra- and inter-neighborhood scale at best, but doesn't address efforts to solve US commuter distances, which invariably involve highways. That seems an insoluble challenge at the moment in the US.