> Jokes aside, I guess population density is just not the main factor in internet. It’s competition, it’s regulation, it’s corruption, and pop density is simply not a deciding factor.

Sort of?

If you're going to provide wired service in rural areas at all, doing it with fast fiber isn't a significantly different maintenance cost than using the old stuff, but it has a high one time cost to transition from copper to fiber. The cost of doing that is more like per-mile than per-customer, which makes the per-customer cost a lot higher where there are fewer customers per mile. There are areas rural enough that nobody would spend the money to run fiber even if there were no regulations at all.

Whereas for NYC it's just unambiguously corruption and regulation destroying competition.

Building in rural areas is hard for physical / engineering reasons. There's more cable to lay, more distance to cover, and fewer people to use that cable and offset the costs.

Building in dense cities is hard because we choose to make it hard. We could (and should) choose differently.

Ironically enough, rural areas now have a ceiling on how bad service can get (because Starlink is a viable alternative). That doesn't work for dense EU/Asian style cities where most people live in 5+ story buildings.

Edge cases are hard is not a bad rule of thumb.

High concentration, and you have saturation issues. Extremely low concentration, and there is not much active elements to leverage on.

Yes, as all rules of thumb, it falls apart in many situations too. But in that case at least the rule of thumb kind of recognize it will poorly scale at full generalization level.

[dead]

[dead]