We can use tracking and modern software solutions to plan where to put new satellites.
4500 is far from the maximum possible.
And of course they will do just that.
But satellites carry only relatively little propellant and the bulk of it is used to maintain altitude, especially for LEO/MEO satellites. Serious course changes are going to be pretty rare whereas they are the norm for anything with wings.
On another note, airtravel (not general aviation) tends to happen in 'corridors' with traffic control handing over at the borders of control zones. For satellites the situation is vastly different.
The same technological solutions that we've found for aerospace have almost no application in satellite orbital computations and the management of fleets of satellites.
GGP was trying to make the point that airplanes don't collide and that they travel fast but that's not the whole story, there are a lot of people working hard to keep it that way.
Also, even though NASA is tracking stuff there isn't much they can do about it, it's mostly to help determine new orbits and to make sure that launches do not accidentally intersect with some chunk of space debris.