Part of the problem with landfill is that the economics are largely 'profit first, cost later'. We don't really have a good model to ensure that long term maintenance and care of a landfill site post closure is always adequately funded so there's always a risk that an owner who might well have made all his capital investment back, and more, before the landfill closed just runs out of money and walks away no matter how you provide privately. This is a bit like the issue you have with defined benefit pensions and longevity, only you have less certainty on the costs.

Because the potential environmental consequences of just leaving the site unmaintained or poorly looked after are high, in the developed world government almost always ends up providing a backstop to this. This could be described as 'socialised losses and privatised profits' when it gets used, and we would be right to be very wary of this model, even to pay a premium to avoid it.

This one caught my eye also.

I think the software analogy is not especially good, but it feels like there's some poorly-aligned incentives in landfill engineering. The point of landfills is to make trash invisible, so saving money by hiding engineering faults in the landfill itself is easy, they get buried along with the trash, and by the time the leakage becomes noticeable, the builder is long gone (bankrupt, merged/acquired, whatever), and the community is left to spend more on remediation than would have been required for prevention.

The solution to this sort of problem is generally regulation and oversight. I try not to be cynical about the utility of that approach, but it's hard.

The difference from software engineering is that, for landfills, the specifications don't change on the fly, and the engineering best-practices to build them actually exist.

I don’t think that’s analogous. I suspect there are numerous examples of landfills they don’t hurt the environment, and lots of well studied best practices for how to achieve that in others. The equivalent in software is a sort of spherical cow theoretical construct that nobody has seen, let alone knows how to systematically create.

While i 100% agree with your point, the (i'm assuming) highly regulated process of creating and running a landfill site doesnt have the same "fail fast, fail often", "MVP", "rapid development" culture that software does.

I once helped out a project which involved talking to a landfill operator. They mentioned that deer like to drink the water pooling in the landfill and theorized that it tasted "sweet." Of course, hunters then kill and eat these deer.

This also stood out to me, as I've become increasingly aware that most of the landfills around me are not lined.