Thinking about how underfunded infrastructure (in the tech sense although probably also other senses) creates a ripple effect that eventually leads to total deadlock. Infrastructure uses infrastructure, so eventually the Goose Plumper team is going to run into a serious bug in the Whatsits they depend on. In an ideal world, the Whatsit team has the spare bandwidth to prioritize a fix for this bug quickly, unblock the Goose Plumpers, and everything moves on relatively smoothly.
But if the Whatsit team is underfunded, they may have to delay fixing this bug or even push back on being able to fix it at all. This means the Goose Plumpers have to find their own workaround, causing delays in their schedule that ripple out to all of their users. This is an issue they had no way of seeing coming, and it's on top of whatever additional delays come from them being underfunded. If there's a dependency cycle—maybe the editor the Whatsit team uses is built on the Goose Plumpers' framework—these delays can echo around indefinitely.
Eventually, even if each infrastructure team is only a little underfunded, these problems build and reinforce each other until no one can reliably get anything done. Work estimates either get longer or longer to account for the increasing number of unknown-unknowns, or they refuse to account for them and get more and more inaccurate. Large infrastructure projects begin to become infeasible, then medium ones that were initially thought to be feasible end up stretching on years longer than anyone expected. Eventually you reach infrastructure deadlock, once all the underfunded bandwidth is taken up desperately trying to hack through the thicket of blockers for these ancient unfinished projects while more blockers blossom all around you.
One Hundred Percent what's going on with the North American freight and rail industry, and also why the SimCity traffic advisor was never lying when he said "You'll regret this!" when you cut funding
