There's a lot of inertia in the system to design that which gets approved from above, as the article says, rather than to solve local issues locally. You can definitely do it - there are examples across the different states of various design choices - but federal funding may be contingent on following the existing rules.
This relates to a general issue with society(in the US, and to varying extents in other industrialized countries) that has gradually crept up over time: in the post-war period, one-size-fits-all bureaucracy was normalized across many institutions. The rules were necessarily kept simple and strict, using a small amount of data, so that they could be followed accurately. When an exception was needed you talked to an appropriate contact and hoped you were listened to. It was under this system that you got things like the interstates, redlining and urban renewal projects, because there was arbitrary power to decide what was built where and all you had to pass to get a shovel-ready project was a routine checklist.
Then computerization took over and every bureaucracy started adding more complex data models and rulesets. More choices appeared, forms got bigger, more rounds of approvals were added and everyone lost track of who to talk to when things went wrong. You are more likely to be addressed as a categorical minority(e.g. gender or ethnicity) but aren't allowed to be singularly exceptional because there's too much automation in the way.
And when we look at the state of road infrastructure now there's absolutely a case of that phenomenon: The design premise remains locked in "first design for cars at speed" and then other modes are the exceptions that are harder to access: you have to make exceptions to have less parking space, exceptions to add bike paths, exceptions to try a different intersection design and so forth. Major cities are undergoing reform to a lot of these rules, but at varying rates and levels of pessimization. Once you have an established rule, it's not a career-ender to follow it blindly, so you have to take a risk to not follow it.
This relates to a general issue with society(in the US, and to varying extents in other industrialized countries) that has gradually crept up over time: in the post-war period, one-size-fits-all bureaucracy was normalized across many institutions. The rules were necessarily kept simple and strict, using a small amount of data, so that they could be followed accurately. When an exception was needed you talked to an appropriate contact and hoped you were listened to. It was under this system that you got things like the interstates, redlining and urban renewal projects, because there was arbitrary power to decide what was built where and all you had to pass to get a shovel-ready project was a routine checklist.
Then computerization took over and every bureaucracy started adding more complex data models and rulesets. More choices appeared, forms got bigger, more rounds of approvals were added and everyone lost track of who to talk to when things went wrong. You are more likely to be addressed as a categorical minority(e.g. gender or ethnicity) but aren't allowed to be singularly exceptional because there's too much automation in the way.
And when we look at the state of road infrastructure now there's absolutely a case of that phenomenon: The design premise remains locked in "first design for cars at speed" and then other modes are the exceptions that are harder to access: you have to make exceptions to have less parking space, exceptions to add bike paths, exceptions to try a different intersection design and so forth. Major cities are undergoing reform to a lot of these rules, but at varying rates and levels of pessimization. Once you have an established rule, it's not a career-ender to follow it blindly, so you have to take a risk to not follow it.