Based on that thought experiment, my opinion is that there is no way this guy is ever going to get a town of 20000 people in one square mile of Vermont and still sustain a modernized economy.
In my image, I tessellated a square mile into 9600 rectangles, each 88'x33'. 928 of those rectangles became 50'-wide arterial roads flanked by 8'-wide sidewalks. 320 of those rectangles became a 176'-wide expressway corridor with 8 double-crossover overpass interchanges to the arterial roads. (These road areas would also have utility conduits underneath the road surface.) The remaining 8352 lots have an identical access easement for 18'-wide streets, 20' on each side for perpendicular parking, and another 4' on each side for sidewalks.
I assumed 60% of building lots would be residential buildings, with average occupancy of 4 people per residential lot, for a maximum of 20045 residents, with a total of 15776 on-street parking spaces, and 223 of the 640 acres left over for parks, commercial buildings, and government buildings.
The image ignores potential geographical obstructions, so the only way you could get all 29 local access streets, all 8 arterial roads, and the expressway to go through is to build on flat, featureless land, as one might find somewhere between Wichita, Minneapolis, and Indianapolis. In Vermont, there's no way.
And if you are intending to build a city in America without reserving at least 1.8 parking spaces per household, you're going to have problems. Some of the non-residential lots would have to be dedicated parking lots or garages, because the non-residential buildings need more than 1.8 parking spaces per unit cell. About a third of your permanent "residents" are going to be cars, and half of your money-importing commercial visitors will be cars. If you plan your whole city around temples, rather than truck routes, sewers, and optical fiber backbones, you're doing it wrong.