[dredmorbius]

A challenge which persists for tens of thousands to millions of years is well beyond merely "political". You're going to need to come up with a different term for that.

Keep in mind that for a long-term nuclear regime, you're also dealing with the extreme likelihood that large amounts of not-yet-vitrified wastes in "temporary" storage may face a sudden change in operational procedure and personnel. Say, during the Widespread Global Disagreement of 2639-2645, in which team Allied and team Axis butt heads once again under stars and bars and revived flags of Indian auspiciousness.

What's your technical solution for that? Or any number of comparable scenarios?

What are the comparable failure modes for widespread renewables infrastructures?

[13thLetter]

> A challenge which persists for tens of thousands to millions of years is well beyond merely "political".

This is absurd, and it's absurd for two reasons.

a) The amount of waste which is still dangerous after that time is meaningless, compared to the amount of other dangerous waste we create (including in the process of constructing batteries and solar panels!) which will still remain dangerous indefinitely, and which is far less well secured and understood.

b) For what other endeavor of man are we required to address its consequences literally millions of years in the future? A renewables infrastructure will consume a large amount of rare earth metals, you know: are you prepared to address the awful consequences of the great neodymium shortage of 11,152 AD? If not, then where do you get off asking for similar future-proofing where nuclear is concerned?

[dredmorbius]

For what other endeavor of man are we required to address its consequences literally millions of years in the future?

Tu quoque fallacy.

Several. Just to be clear, I'm not simply holding nuclear to this standard.

Population, energy systems, resource utilisation, topsoil and water use, environmental contamination.

How long you want to consider "long term" is also an open question, though I'll note:

The modern computer age is roughly 50 years old.

The modern age of mass-industrialisation: electricity, automobiles, mass media, roughly a century old.

The Industrial Age itself, 200 years.

Western Civilisation, about 2,500 years.

Civilisation itself, and history, 6,000 years.

Anatomically modern man, about 200,000 years.

Divergence from common ancestors with chimps, 2 million.

Emergence of mammals, very roughly, 150mya.

Looking forward, there's perhaps 500m to 1 billion years in which life resembling that we know can survive on Earth.

Meantime, on the present "business as usual" track, there are numerous challenges which present on the timescale of years to decades -- shorter if you consider the prospect of nuclear annihilation (minutes to hours), somewhat longer for some more-abundant mineral resources. But numerous challenges seem likely to converge between 2020 and 2100 or so, with the implications of several of those including challenges to running long-lived complex systems with profound implications. Such as creating large quantities of nuclear waste and/or facilities which are not likely to be properly decommissioned and remediated. Hell, there's ample existing problems with this ranging from the former USSR/Russia, US, and elsewhere, with only modest amounts of political and economic disruption.

But I'd suggest that:

1. Avoiding making near-term circumstances more complex than they are (10-200 years or so).

2. Considering just what problems it is that nuclear power does and does not address, directly.

3. A view to a 200 year (industrialisation), 6,000 year (history), to 200k-1m year (evolutionary drift) would likely be somewhat useful to keep in considering pretty much all future paths and decisions.

Along with questions like "why are we here", in a thermodynamic/systems sense, and "what are the implications of this", for both us and the systems with which we interact.

Maximising throughput without limit strikes me as potentially problematic.

Something I'm putting a fair bit of thought into: https://reddit.com/r/dredmorbius/wiki/FAQ

[13thLetter]

"Several. Just to be clear, I'm not simply holding nuclear to this standard.

Population, energy systems, resource utilisation, topsoil and water use, environmental contamination."

Okay, I apologize for being a bit unfair. That being said...

You can't optimize all those things at the same time. If you're concerned about carbon dioxide emissions from power generation leading to the most catastrophic possible greenhouse effect, that's far more damaging than a small amount of nuclear waste buried in a mountain somewhere, and the possibility that some small number of people in a nation or culture that doesn't even exist today may be foolish enough to dig it up a thousand years from now. Such a tradeoff would be well worth it -- even to those folks a thousand years from now, because a modern-day Earth that doesn't have to deal with an environmental catastrophe is going to be wealthier and better-ordered a thousand years from now than one that does.

[dredmorbius]

You can't optimize all those things at the same time.

Also to be clear: I see total throughput, itself a function of population and affluence, as the fundamental challenge. It's not a question of optimisation, but of living within the possibility envelope.

And the harder you push up against that envelope the greater your systemic risk.

There are also nonsystemic risks: asteroid impact, nearby supernova or gamma-ray burst, etc. But as I see it now, the biggest risks humans face are systemic and self-induced.

As my earlier nuclear comments have made clear: I'm not anti nuke, but I see substantial problems, enough to wonder if they're worth the trouble.