I have a lot of skepticism about anything I read from constructionphysics.substack.com. I'm a mechanical engineer, not a civil engineer, so take my opinion with a grain of salt, but it reads like it's written by someone who does not have a solid understanding of engineering fundamentals. For example, they suggested in their "How to design a house to last for 1,000 years" series that you'd want to build such a house out of stainless steel because it "has extremely high corrosion resistance." Even my introductory courses in materials science it is very clear that stainless steel might last for 1,000 years, but it would depend entirely on the environmental conditions and the type of stainless used.

I mean, it would be exposed to the normal conditions any residential construction is exposed to. How often is your house dunked into seawater?

800 years from now something terrible could happen that turns all the rain into sulfuric acid but then all residential buildings everywhere are screwed.

Not a civil engineer, but I think different areas have very different considerations, and I'm not sure there is a 'normal' residential construction. A beach-side house in Florida probably has very different issues from a house built somewhere like Las Vegas. Salt, humidity, wind, earthquakes, soil, and other considerations vary.

> How often is your house dunked into seawater?

For people hit by hurricane Sandy the answer is at least once.

Over that long a time period it could be damaged just by rain - it's hard to say because corrosion tests aren't done with respect to a 1000 year operating life in mild conditions. I'm sure an expert could figure that out, but it would require looking at the corrosion rate of the specific steel, the exposure to any kind of forces that could damage the oxide layer (e.g. scratches), and the thickness of the member.

> How often is your house dunked into seawater?

Hopefully the house ain't in Florida. Or California's Central Valley.

Living near the ocean means that salt air corrosion will become a problem eventually. It eats away car bodies over time, for example.

> because it "has extremely high corrosion resistance."

Isn't this true though? Especially relative to other materials that can corrode?

It seems like you're reacting as if he said it _cant_ corrode.

In general, it's true compared to other metals. Compared to other materials in general (e.g., glass) it's certainly not true. Over the scale of a thousand years, there could be a big difference between something that is corrosion resistant (stainless steel) and something that doesn't corrode at all.

But are you suggesting a non corrosive material would outlast stainless steel? Otherwise I don’t see the point in your distinction. Obviously your example of glass wouldn’t.

I guess I would have thought stone would last longer.

We have a lot of stone or concrete buildings from 2000 years ago still standing, and even more from 1000 years ago, so those would be safe bets. No surprises there. There's a glass window just shy of 1000 years old in a Cathedral in Germany, and other slightly younger ones scattered here and there, so glass definitely can make it. It just isn't very impact resistant, and it's also a fairly new material so there are fewer examples to draw from. Similarly, materials like Bronze would do just fine over 1000 years, and have done so in the past, but they tend to be appropriated for new building projects even if their original building is still standing.

The really impressive ones haven't even had any maintenance done.

(Of course they have all had lots of maintenance done...)

It's an oversimplification to merely use a single property of a material as justification for choosing it when every material choice involves many factors - hardness vs toughness vs strength, weight, cost, repairability, density etc.

I trust any engineer who responds with an “it depends..” over someone who has an absolute answer. Stainless would probably do ok but I’d assume stone or fiberglass/FRP would be superior over time out in the elements. But not the best stuff to build a house with.

Oh what about titanium?

Edit: lol oh well even timber structures Can last a millenium if maintained

Why not simply use gold?

Gold attracts vermin.

How poignant, that the obvious solution is the worst "solution" of all, due to human nature.

Concrete, bricks and granite can also work pretty well... as the Pantheon in Rome proves.

We don't make concrete like the Romans did though.

Still concrete will still last longer than most other materials.

The biggest issue with modern concrete construction is the usage of reenforced concrete using iron or steel rebar which tends to corrode quickly. Reenforced concrete seems to have a half life of ~50 years.

Will be interesting to see how carbon-fiber reinforced concrete[1] fares in comparison.

[1]: https://www.compositesworld.com/news/tprc-research-studies-v...

stainless steel rebar and polymer coated rebar are also options for reinforcing concrete that have a lot of potential, as the issues is moister induced rusting of the rust rebar expanding and cracking the concrete there are multiple solutions that i don't know that we have data on the long term durability of yet. I wonder how one would go about artificially aging something embedded within a block of concrete to test for that.

It appears that coated rebar just offers a channel for the ingress of water, and thus corrosion, if there is a break anywhere in the coating, like a wick in a lamp.

Basalt fiber rebar might be a better way to go, as it won't oxidize. Studies will be needed to find out if it actually is better in the long run.

We do actually, it's just not as useful and used in more demanding applications.

The biggest problem is rebar spalling though, not the concrete.

Correct, because the way the Romans made concrete wouldn't be suitable for 30 and 70-story buildings.

The way the Romans used concrete to form structures that could be plausibly carved out of raw rock. The dome of the pantheon is an impressive feat of engineering, but the general shape bears similarity to rock caverns. Which is great, because concrete is essentially rock.

Meanwhile, the kinds of beams and walls we make out of concrete wouldn't work if carved out of rock, they are only structurally sound because we put steel rebar in the concrete. But rebar expanding due to rust is how concrete usually fails. Plus using rebar prevents you from using salt water in your concrete, which is one of the ways in which Roman concrete might be better than ours.

What exactly is keeping us from making concrete with salt water, for things like sidewalks?

In the northeast US, sidewalks frequently are in horrible condition after only a few years, because of the salt used in the winter. Sidewalks do not use steel rebar; they're just simple poured concrete. So why can't they make them with a better mixture that doesn't corrode so easily? (I'm guessing the answer is: replacing them frequently makes a lot of money for some concrete company that the town mayor is friends with.)

I think the time it takes for the concrete to set is why, as keeping it dry for a long time would be difficult.

Not to mention the wooden beams, under the portico, which are also original. They of course, are out of the rain; so are a solution to a lesser problem, where aging is concerned.

Stone and adobe, in the right conditions.

https://en.m.wikipedia.org/wiki/Ancestral_Puebloan_dwellings

In those conditions even plain, untreated wood lasts a long time. Many of those structures still have wood beams in pretty good condition.

ie, the durability of those structures is more due to where they are, than what they are made of.

Although (at least the right) stone is almost certainly a pretty good answer in general. There are plenty of historical examples and I'm skeptical that anything else exposed to the elements would do better. And if not exposed to the elements that feels like cheating.

Fair, but the sandstone and clay in that area is generally pretty awful. TBH most of the ancestral Puebloan structure are crumbling apart despite being protected by overhanging rock "roofs".

You're right about the pueblos specifically. They're still extant mostly because the climate is so dry there.

Something like granite would be a much more long-lived structural material in wetter climates. Indeed that's the primary stone used in, say, Machu Picchu.

It depends on whether they are occupied or not. I worked on an upstate New York farm with a barn that was post and beam, with a stone foundation. It was ~350 years old but had a good roof for that entire time.

Fiberglass/FRP is an interesting one. I believe the limiting factor there would be that the polymers used would slowly degrade/weaken over time due to chemical aging. Fiberglass boats last up to 50 years but that is in a harsher environment - not sure how long you could go on land. Perhaps some polymers could be formulated to last that long, I'm not sure.

They can last a lot longer than that but that’s about the duration that frp has existed for boats, probably early ‘60s is when production boats started to use frp hulls.

Frp is great for fuel tanks and exhaust systems - it’s superior to stainless for these applications (corrosion resistance being key factor). But yes as mentioned by sibling UV will definitely degrade

UV from sunlight degrades fiberglass/FRP over time. I believe that boats are painted to protect them from this, and need to be recoated periodically because the uv sunlight will degrade the paint eventually, also.

Seems unlikely that you’d have a reliable source of paint for 1000 years, which might be the limiting factor.

But there’s always some rate of chain scission, even in the absence of UV/water/whatever. I just have no idea whether that has a meaningful impact on material properties over 1000 years.

I feel like if you're really hard pressed for a UV blocker you could just use mud.

Yeah I've seen old boats turn to dust. UV utterly destroys fiberglass given a little time.

>I trust any engineer who responds with an “it depends..” over someone who has an absolute answer.

That heuristic matches on the hordes of liars who try and sound smart while hedging their claims with a liberal application of weasel words to avoid being provably wrong. And those people outnumber the "good engineers" probably by at least an order of magnitude...

"When a measure becomes a target, it ceases to be a good measure."

It is, however, far more interesting to me when people are able to split the hypothesis space well. For instance, take the question "Is drinking 1500 ml of water daily good for you?". Well, that depends on what other fluids you consume, whether or not you are on an IV as well, and whether or not you are drinking it by accidental consumption via waterboarding. It even depends on whether or not Baldur the Interstellar Water Hunter exists and will kill you for every drop over 1499 ml that you consume.

It's true, it does depend. But machines can exhaustively describe the hypothesis space now. Comprehensiveness is not the only virtue of a human.

Not to mention, we know how to build structures that can last 1,000 years because it has already been done before! Using a dubious material in place of a known good one is not the best engineering approach.

Who's "we"? The people who built structures 1000 years ago are all dead (I assume).

Just because some humans somewhere figured out something in the past doesn't mean that humans somewhere else today know how to do it. Lots of knowledge has been lost to time.

We know ways that worked a few times and failed many thousands of times more often.

How many buildings of the same type didn't survive?

If your goal is to use a method that has _any_ chance of success, however small, then the old ways are probably a good bet. If you want the highest chance of success you should probably at least evaluate new methods.

Your best chance would probably be to go primitive - many different pyramids of various designs have stood for over a thousand years.

For each one that is still standing, how many are dust?

Evidence would seem to point toward not many. Even the Bent Pyramid, which may have come before Egyptians perfected the math behind building pyramids, still stands. Throw in the numerous ones found in various jungles by LIDAR and they seem to hold up quite well.

Straying off topic, are there any discussions going the opposite way? Is it possible to build a structure that lasts maybe 10 or 15 years, then either disassembles or decomposes without too much damage to the surrounding area? Something actually livable, and nicer than a tent.

Mud bricks, wattle and daub, thatch, and the like? Then build out the interior with rough sawn planks.

What's old is new again!

I reached a point where I treat anything from substack the same way I treat I links from medium; assume the content is crap and / or useless and move on. Which is a pity, I know, but I simply don't have time to find the rare gems in all the useless "content".

I read it as meaning at least 1,000 years. Modern humans started living in caves in Europe about 100,000 years ago, and their cave-houses still exist. So, if your house is also a cave you should be able to get at least that.

If they want the house to last for exactly 1,000 years then I'm not sure what to tell you. Maybe something based on the milleneum clock? Or explosives and a really long fuse, Wile E Coyote style?

I read it the same way?

It’s funny because many buildings exist for 500-3000 years in various forms, and they are all some variant of brick, compressed earth or stone.

Many, many similar buildings did not last though.

It's quite possible that if you want a newly built structure to last that long with the highest probability, you should pick new materials.

Or, if you need 1 intact building in 1000 years, build many many of them now and hope you get lucky with at least one of them.

Yeah not to mention the construction of the plant itself! I wouldn’t want to be operating the drill that punctures the magma chambers!

I’m optimistic. If we spent less time engineering society and more time engineering the earth, I think we’d all be a lot happier.

* * *

> I wouldn’t want to be operating the drill that punctures the magma chambers!

Years of Dwarf Fortress have prepared me for this moment.

The JPL plan doesn't actually puncture the magma chambers.

I felt the ignorance in my fingers even as I typed it. That's awesome!

Magma drill bits and shank bores need to be made of unobtainium to resist the heat and pressure conditions of the deep earth environment. Edward Brazzelton has developed a new way to synthetize and shape the material. He's also developed a laser-based high-speed drilling array, though he would need billions of dollars to speed up the research and development. I heard he takes checks and credit cards though.

>I wouldn’t want to be operating the drill

This is where AI applications could actually be of benefit. Also, instead of sending Harry Stamper and a bunch of misfit roughnecks to some asteroid of doom, we could just train AI on Harry's knowledge of deep well drilling. It'll be fine.

Why do we need AI for this when remote controlled operations would work just fine, especially terrestrially?

Because you can't get VC funding for your new venture if you don't have AI some where in your slide deck

It doesn't matter whether a human or a robot drills into Yellowstone, the resulting eruption will kill off all humans on earth anyway.

which is exactly what the AI is attempting to achieve

The article mentions one proposal that specifically says they would drill holes around the magma chamber as drilling into the chamber itself could possibly cause an eruption.

Or exactly the other way around.

Without "engineering society" we'd still live in tribes killing each other at first sight. We have reduced that substantially. Some societies are a bit behind (like allowing lethal military-grade weapons at home), some are further advanced. Overall, it seems a good idea to use our intellect to advance societies and we've come a long way from the dark ages.

Our earth engineering has come so far that with the exception of a few national parks and reserves, we've used every little corner to cut off and kill everything that existed on it and turned it into less-and-less usable farmland. A few areas are cities or golf courses or transportation highways, the rest are terrible monocultures or their next stage: deserts. We've extinguished more species than we know and of those that we have not, we have brought lots to close to it. 95% of all fish are dead. Sea levels are rising rapidly. Large areas have water shortages. We need less "earth engineering", not more of it.

All of the problems you listed sound to me like earth-engineering problems, except for the "problem" of private property. I strongly disagree with the premise that social engineering raised us from tribal groups. It was agriculture and domestication and fire. The data suggests global farmland is reducing [1] and becoming more and more productive [2].

It is funny that we see things exactly opposite!

1. https://www.dtnpf.com/agriculture/web/ag/news/article/2021/0....

2. https://ourworldindata.org/crop-yields

> I can imagine that the consequence of this might actually just be strengthening the "barrier" the magma has to erupt through, and when it finally builds up enough energy to do that, it'll be stronger in proportion to how strong the newly cooled down barrier had become?

Certainly the movie version has the one rogue scientist working late into the night redoing the calculations and realizing that as soon as the plant goes live it will trigger the volcano.

In real life, I have no idea how this kind of thing gets analyzed, and what kind of uncertainty is involved. It feels a bit like those "put a bunch of reflectors in space to stop climate change" ideas. On the other hand, removing energy from a system seems like a good way to make it safer...

> Certainly the movie version has the one rogue scientist working late into the night redoing the calculations and realizing that as soon as the plant goes live it will trigger the volcano.

Jor El in any number of them.

> In real life, I have no idea how this kind of thing gets analyzed, and what kind of uncertainty is involved.

This is similar to what I was trying to get across -- how do we know something this complicated will have only the intended effect that the author is describing?

The comparison to reflectors in space to address climate concerns is a great one - something where a first pass/ back of the envelope makes it seem interesting, but any amount of detail added reveals that we'd have no clue what would happen

> On the other hand, removing energy from a system seems like a good way to make it safer...

Nope, that one rule of thumb can't help you. There is way too much energy for that to make any difference.

Just notice how the timescale of an active disaster is completely different from the ones estimated for success.

Betteridge's law of headlines aside. I feel like the scale of the topic is almost completely human inconceivable when trying to tie it to a realistic engineering task. It seems like a fun what-if, but when you do that you are using making giant hand waving excuses, and not anything that looks like a factual claim.

Agreed. Geothermal energy can be cheap and awesome, but there are reasons why it's only practical in certain parts of the world.

The article states that the JPL plan only considers the energy generation to be a useful byproduct.

If the risk were truly imminent and quantifiable, I assume it would be pursued however absurdly exotic and fanciful it seemed on its face.

On the other hand, defusing a volcano on a timeline of a thousand years or longer would be politically challenging. That might see the dissolution and construction of multiple states (here, electric power generation may help assure survival). Add in that civilization can't seem to get its head on straight about comparatively faster-acting climate change, for example, and can barely prepare for an earthquake.

At least imagining an exploding apocalyptic supervolcano may elicit more visceral emotions.

Still, it doesn't appear there's much cause for worry so it's only a brainstorm.

We should be able to use the latest horizontal drilling tech from fracking to tap Yellowstone from miles away

> Betteridge's law of headlines aside.

Godwin-Bettridge-Graham law: Any headline that ends in a question mark which is posted to HN will have Bettridge's law of headlines mentioned at least once in its first dozen comments.

What's the law for when someone has to point this out too, I wonder?

The “Why?” Combinator

The linked JPL proposal is exploratory and, while it has a more serious engineering perspective, it does not have "confidence". https://scienceandtechnology.jpl.nasa.gov/sites/default/file...

The USGS is very unconvinced. https://www.usgs.gov/faqs/can-we-drill-yellowstone-stop-it-e...

From the USGS article:

> Scientific research has proven again and again that depressurization is one of the factors that drives magma toward the surface to erupt.

The first time I heard of this idea my immediate thought was, "so you want to shove a needle into a massive magma balloon." I know I wouldn't.

>Any allowed geothermal extraction would lower the pressure on the existing geysers and hot springs, altering their behavior and, in many cases, causing them to disappear.

Oh no. Saving the lives of ten million people might make a geyser stop working. How awful.

Have you read the whole thing? It says it is not at all clear that this will save lives. It also could cause an eruption.

This is HN. No-one reads the whole thing. Some don't even read the beginning!

Well, consider the fact that the vast majority of the earth surface, in fact, almost the entire surface of the earth is entirely volcano-less. Look under your feet -- no volcanoes! Move over a couple of steps and look again -- no volcanoes.

So it is safe to say that a certain thickness of rock would serve enough to prevent volcanoes. You just have to provide enough cooling to ensure a sufficiently thick layer of magma solidifies into solid rock.

Yeah, but how much? The crust is far deeper than humans are capable of penetrating. It's probably a fair concern that it's also far deeper than humans are capable of cooling.

Well, lets ponder this a little further. Currently there are no volcanoes in Yellowstone. So we know that the rock layer is sufficient. We also know that the rock keeps losing heat with time because it has higher than average temperature. But we also know that the rock keeps gaining heat from magma pushing up from some deep underground crack.

What we have to do is measure the net heat gain and calculate how much additional heat a geothermal plant has to extract from the rocks to turn it into net heat loss. Of course the plant will have to extract it from the hotter spots in order to prevent any local maximums from causing volcanoes. But that is the most efficient option for a geothermal plant anyways.

As far as measuring the net heat gain -- it should not be that hard. All you need is to sprinkle a couple of hundred sensors in the rock, monitor them for a while to be able to remove surface effects (like the sun and ambient air temperature) and you should be able to get a decent estimate. In fact I would not be surprised if seismologist already have a pretty good estimate of the heat gain.

Then, when you build the plant, you can be certain that as long as heat is being drained from the rock, magma will keep solidifying and the rock layer will keep getting thicker, etc.

Of course a good earthquake can screw up all of these plans, but an earthquake can destroy any power plant. But then again Yellowstone is one of the most earthquake prone areas of the world, so that perhaps is the biggest danger of this plan.

TFA gives this number as ~1.5GW per year -- I believe calculated by the JPL proposal.

> Look under your feet -- no volcanoes!

What if whoever built this McDonald's/Chevron amalgamation did so atop a volcano? How would I know?

That's not how this works. Plate tectonics won't stop if you cool down things a little bit. The amount of cooling you'd need to do amount to have that stuff stop is comparable to creating forests and oceans on mars. Not gonna happen, we'll kill off humankind before that.

> Where does the confidence to make this claim come from?

I have no idea, but I can give you a data point from my personal experience: about ten years ago we visited Rotorua in New Zealand, which promotes its geothermal activity as a tourist attraction [1]. We were told by the locals when we were there that the geothermal activity had reduced noticeably since a geothermal power plant was installed some years prior, resulting in controversy over the wisdom of having built it.

Rotorua's geothermal activity area is a lot smaller than Yellowstone, but as an anecdote it indicates that reducing the risk of eruption is not entirely implausible.

---

[1] https://www.waiotapu.co.nz/

Good comment, but had to automatically downvote because of the use of “lol”.

Maybe the earth just burps occasionally and it's not really a gradual buildup of pressure and this would do nothing.