A key analogy that went unused in the article is the Red Queen's Race: "it takes all the running you can do to stay in the same place" applies to car traffic precisely because cars are large and powerful and are persistently required to make slow turns and smooth stops relative to their size - think of all the roads where you have to carefully inch out, keep deliberate pace and stopping distance, and so on. Spaces that are "for cars" correlate to spaces that are unsafe "for people". If you add more than a small topping of car traffic to a road, they will impede each other. Therefore, to make
large numbers of cars move quickly you have to add disproportionate amounts of car infrastructure, which creates car-centric land use patterns, which inhibits alternatives and makes cars the default by dint of being the smallest unit that can safely use a road. "If they get bigger, I must get bigger too" is applicable thinking for someone who wants to feel safe on the roads.
If you reduce the scale, the speed, the headways, you have greater ability to adjust. While car pileups are expected, bicycle pileups are a thing usually only seen in races when the limits are being pushed. Ordinary bike commuters can navigate safely without rules when given a mixed bike/pedestrian space of sufficient size. Bikes don't get bigger to feel safer. The resulting infrastructure scales better for the cost, and the Red Queen's Race is averted. When bike lanes reach capacity, the failure mode is crowded-stadium, not stand-still freeway. Traffic slows down gradually instead of creating roadside bloodbaths.
Transit does experience more of a headache from induced demand pressure since it also uses large vehicles, and the entire trip experience is contingent on not holding up that one train or bus - while Tokyo's system works, the crowding is notorious, and the train stations have sizable footprints. But as a collective effort, it can also be much more coordinated and deliver useful results from capacity increases.