Earlier quoted context omitted.
What about throwing capital at developing a better way to route traffic than BGP? Or trying to further increase bandwidth and be a faster 'dumb pipe'?
That's a really interesting point. If ISPs reinvested some of their revenue into networking research (or even more applied /R&D work), it would go a long way towards both advancing our field and improving their general reputations. Networking is not my area of expertise, so it's possible that some do this already. I guess I always assumed that ISPs had research departments whose members would be involved in drafting…
The interesting thing is that a huge focus appears to be cost-sensitive. As last mile operators, they have relatively low profit-per-customer (compared to a backhaul operator) with which to build each final physical network link.
Hence crazy stuff like "fiber to the box and short-run copper standard to the house".