I am biased because I quite like living in the suburbs, and get quite suspicious when some manifesto comes up that dictates I should feel bad for it. Why is the solution to live in better designed cities instead of better designed suburbs? Are cities still necessary given the modern conditions?
Honestly my take is that we need Rivieras witch means a bit dense BUT not more than a bit suburbs that are NOT only residential but mixed. As a result we can have "near enough" services/people without failing in the density trap.
The big issue is that human settlements never worked well when centrally planned and when they are not if they work more people want to be there, so the initial low density became right and than too much dense. After that density issue start to bite, people to flee and the cycle restart. No recipe to fix it, it's more a social issue than a mere architectural one.
Some, like McKinsey think we can solve some complexity issue with flying vehicles and I agree, but I do not think we are as near as they state https://www.easa.europa.eu/sites/default/files/dfu/uam-full-... and we still have an enormous isse: making VTOLs/STOLs it's relatively easy (behind the BEV part) but they can just move people and little volume/weigh goods so we can't build a home moving raw materials without roads and if we need them we need to build and maintain them...
Keep following the "in medio stat virtus" principle about the density and diversity described above I think we also need to came back to a moderately distributed economy: some activities need big factories, so far we can't do anything different, BUT only some. Many others can be done in smaller settlements witch for instance that instead of build mega-factories for anything we should try to achieve economy of scale with standards and standard tools in many smaller factories. Easy to say, hard to built but... In the end internet was a free scale network of networks and prove to scale well enough. I think the same principle can be generally valid. The VERY COMPLEX and very expensive game is try it in the real word because social experiments can't be done and destroy as software projects. They demand very long time just to build them. Enormous resources etc and simulations can't really help more than a bit. Since in the past we have build many-centuries longs projects I think we can achieve such result anyway, but not in the modern mind settlement of something quick and done. Convincing people to embrace such revolution and remain engaged for few decades is however... Well... In the past we used religion to push people to such efforts...