A landmine is not a bomb that missed its target. It’s a device engineered to wait — victim-activated, patient, and built to outlast the conflict that put it in the ground. This episode starts inside the mine itself: the M14 “toe-popper” uses a Belleville spring so stable it can function for a century. Simpler designs last longest. But some mines grow more dangerous with age. Soviet PMN-2 mines use RDX slurry whose chemical stabilizers degrade, forming impact-sensitive crystals. A mine that once required fifteen pounds of pressure might now detonate from a light brush.
Geography matters deeply. In dry, stable environments like Israel’s Golan Heights, mines are mummified — still fully functional after fifty years. Israel never signed the Ottawa Treaty, so those fields remain uncleared, marked only by aging warning signs. Globally, an estimated 60–70 million mines lie in the ground across 60+ countries. Cambodia has suffered over 64,000 recorded casualties since 1979, years after its war ended.
The Ottawa Treaty (1999) banned anti-personnel mines for 164 signatories, but the list of non-signatories includes the U.S., Russia, China, Israel, India, Pakistan, and North Korea — the countries with the largest stockpiles and active minefields. The treaty has no enforcement mechanism, and legacy mines are not subject to clearance deadlines. Russia has laid an estimated five million mines in Ukraine since 2022, making it the most mined country on Earth. At current clearance rates, some projections reach 50–100 years.
The final question — why can’t drones, robots, and AI solve this? — has a humbling answer. Mines are small, buried, often metal-free, and surrounded by other hazards. Vegetation hides them. Soil shifts. No sensor reliably distinguishes a mine from a rock or a bottle cap. Demining remains slow, manual, and deadly work. The technology that writes poetry cannot yet clear a field.
#3976: How a Landmine Stays Lethal for 50 Years
Why mines stay dangerous for decades — and why robots still can’t clear them fast enough.
Episode Details
- Episode ID
- MWP-4155
- Published
- Duration
- 22:54
- Audio
- Direct link
- Pipeline
- V5
- TTS Engine
-
chatterbox-regular - Script Writing Agent
- deepseek-v4-pro
AI-Generated Content: This podcast is created using AI personas. Please verify any important information independently.
Downloads
Transcript (TXT)
Plain text transcript file
Transcript (PDF)
Formatted PDF with styling
Never miss an episode
New episodes drop daily — subscribe on your favorite platform
New to the show? Start here#3976: How a Landmine Stays Lethal for 50 Years
You're driving up north, past Kiryat Shmona, the road winding toward Metula or the Golan Heights. Gorgeous country — green hills, vineyards, the kind of views that make you want to pull over and just breathe it in. And then you see the sign. Yellow triangle, red lettering: "Danger! " Not from some active conflict. From nineteen sixty-seven. From nineteen seventy-three. Those mines are still in the ground.
Those signs aren't historical markers. They're active warnings. Israel never signed the Ottawa Treaty — the Mine Ban Treaty — so those fields were never cleared under any international obligation. They're just... For over fifty years.
Daniel sent us this one after driving through exactly that kind of border country. He was struck by how a war that ended before he was born left behind a landscape that still can't be walked on. And his questions spiral outward from there: How can a device buried in dirt stay dangerous for half a century? Have treaties actually done anything to stop this? How many of these things are still out there, in Israel and globally? And the one that really gets me — we've got drones, robots, AI that can write poetry, and we still can't reliably clear a field of buried explosives.
That last question is the one that humbles you when you really dig into it. But the timing on this is brutal — Ukraine is now the most mined country on Earth. An estimated five million mines laid since twenty twenty-two. Clearance projections at current rates are fifty to a hundred years. And right now, as we're recording, naval mines are disrupting shipping in the Strait of Hormuz — same principle, different medium. The gap between military necessity and civilian aftermath has never been more visible.
Let's start with the basics. What actually makes a landmine different from a bomb or an unexploded shell?
The defining feature is that mines are victim-activated. A bomb gets dropped, a shell gets fired, someone made a decision to pull a trigger or push a button. A mine just sits there and waits for a footstep. Or a tire. Or a child chasing a ball. There's no human in the loop at the moment of detonation. That's the design philosophy, and it's the root of everything that makes these weapons uniquely awful.
It's not just an explosive — it's an explosive that decides for itself when to go off.
Well, not "decides" in any intelligent sense, but it's triggered by the victim's own action. That's what "victim-activated" means. And it's why mines keep killing long after the soldiers have gone home. A fifty-year-old artillery shell that didn't explode is dangerous, sure, but it's an accident of manufacturing. A mine that's still functional after fifty years is doing exactly what it was designed to do.
There's something almost patient about it.
That's not far off. The engineering goal was to create something that would deny terrain to an enemy for as long as possible. Longevity was a feature, not a bug. And the way they achieved that longevity is where the story gets both fascinating and horrifying.
To define our scope here — we're talking specifically about landmines. Anti-personnel mines, the ones designed to maim or kill a single person, and anti-tank mines, which need hundreds of pounds of pressure to trigger. We're not talking about unexploded artillery shells or IEDs. Those are dangerous too, but they're different categories. Mines are engineered to wait. That's their entire purpose.
That waiting is what Daniel's questions all orbit around. He laid out four things he wants to understand, and they make a pretty solid spine for this whole conversation. One — how can something buried in dirt stay lethal for fifty years? Two — have international treaties actually stopped anyone from using these things? Three — what's the actual number still out there, in Israel and globally? And four — why, with all our technology, is clearing them still so painfully slow?
Those four questions connect in ways that aren't obvious at first glance. The longevity question feeds directly into the clearance question — the reason demining is so hard is partly because these devices were built to outlast whatever conflict spawned them. And the treaty question sits awkwardly in the middle, because even the best-intentioned international agreement can't undo engineering decisions made in a weapons factory sixty years ago.
Each layer depends on the one before it. The physics explains the persistence, the persistence explains the scale, the scale explains why the treaties have struggled, and all of it together explains why robots haven't saved us yet.
That's the arc. And I think the place to start is inside the mine itself — what's actually in there that lets it sit in wet soil or desert sand for decades and still function. The answer is surprisingly simple, and that simplicity is exactly what makes it so durable.
Take the M14 mine — the one American soldiers called the "toe-popper." It's a tiny thing, fits in your palm. Plastic body, about the size of a hockey puck but thinner. Inside, there's a metal firing pin, a Belleville spring — that's a conical washer spring, very resistant to fatigue — and a small explosive charge. No circuit board. No wires to corrode. The trigger is purely mechanical: step on it, the pin compresses the spring, the spring releases, the pin strikes the detonator. That's it.
There's nothing to degrade except the spring.
Belleville springs are remarkably stable. They don't rust easily, they don't lose tension the way a coil spring might. A properly manufactured M14 buried in dry soil could theoretically function for a century. The simplicity is the longevity.
Which is the opposite of what most people would guess. You'd think a sophisticated weapon lasts longer. But it's the dumbest ones that endure.
That's the core paradox. A smartphone dies after five years. A plastic puck with a bent washer of steel can kill your grandson.
What about the ones that do have electronics?
Some anti-tank mines use zinc-carbon batteries to power magnetic influence fuzes. In ideal conditions — dry, stable temperatures — those can last twenty to thirty years. But once moisture gets in, the battery corrodes and the mine goes inert. That's actually the better outcome. The nightmare scenario is what happens with certain Soviet designs.
I'm guessing you're about to ruin my afternoon.
The PMN-2. Soviet anti-personnel mine, widely exported, still turning up everywhere from Afghanistan to Angola. It uses an explosive called RDX in a liquid slurry form. RDX is powerful and stable when fresh, but the slurry contains chemical stabilizers that break down over decades. As they degrade, the RDX can crystallize. Crystals are more sensitive to friction and pressure than the slurry. So a mine that required, say, fifteen pounds of pressure to trigger when it was new might now detonate from five pounds. Or from someone brushing against it while clearing vegetation.
It gets more dangerous with age.
It gets more dangerous with age. And that's not unique to the PMN-2. TNT in older mines can exude oily byproducts that form impact-sensitive crystals in the fuze well. The stabilizers in the explosive compound are the weak link. They're designed to prevent degradation, but they themselves degrade.
You've got two aging paths. Either the mine goes inert because water wins — or it becomes a hair-trigger because chemistry wins.
Which path it takes depends heavily on where it was buried. This is where the geography gets grim. In desert conditions — the Sinai, the Golan, parts of the Iran-Iraq border — you've got low humidity, stable temperatures, minimal soil acidity. Mines in those environments are essentially mummified. They can sit there, fully functional, for fifty, sixty, seventy years. The Israeli minefields Daniel drove past are a perfect example. Laid in sixty-seven and seventy-three, mostly in the volcanic basalt soil of the Golan. Dry summers, mild winters. Those mines are not degrading. They're waiting.
Nobody's clearing them.
Israel isn't a signatory to the Ottawa Treaty, so there's no international framework compelling clearance. The fields are marked, fenced where possible, but the mines are still there. Estimates put the number in the hundreds of thousands in the Golan alone. And that's just one disputed strip of land in one small country.
What about the other end of the climate spectrum?
Cambodia, Vietnam, parts of Angola — high humidity, monsoon rains, acidic soil. The casings rust, seals fail, water gets in. That can render mines inert faster. But it can also do something worse. When water degrades the casing unevenly, you get mines that look intact but have compromised internals — the explosive may be exposed, the fuze may be partially corroded in ways that make it unpredictable. A deminer can't assume anything. Every single mine has to be treated as fully functional.
The environment doesn't solve the problem. It just changes the flavor of the danger.
That's before you even get to the scale. The UN's current estimate is sixty to seventy million landmines still in the ground across more than sixty countries. Angola, Afghanistan, Cambodia are the most contaminated. Cambodia alone — over sixty-four thousand recorded casualties since nineteen seventy-nine. The war there ended in nineteen ninety-eight. That's almost thirty years of peace, and the mines are still claiming limbs.
Sixty-four thousand. In one country.
That's recorded casualties. The actual number is certainly higher, because many victims in remote areas never make it to a hospital that files a report. These are farmers, herders, kids collecting firewood. People who had nothing to do with the wars that put those mines there.
If the mines themselves are this stubbornly durable — and they are — the next obvious question is what we've done, collectively, to stop using them. And the answer is... The Ottawa Treaty, officially the Mine Ban Treaty, entered into force in nineteen ninety-nine. It bans the use, production, stockpiling, and transfer of anti-personnel mines. One hundred sixty-four countries have signed on.
That sounds like a lot.
It is a lot. But here's the list of countries that haven't signed: the United States, Russia, China, Israel, India, Pakistan, and North Korea. That's not a random assortment. Those countries collectively hold most of the world's remaining stockpiles and have the largest active minefields. The treaty binds exactly the people who weren't the problem.
The countries that actually lay mines just...
Continue to lay them. Russia has put down an estimated five million mines in Ukraine since twenty twenty-two. Myanmar's military is laying new minefields right now in its civil war. The treaty has no enforcement mechanism — it's a normative agreement. It works on shame and diplomatic pressure, and if you're a regime that doesn't feel shame, it's just words on paper.
There's also the legacy problem. Even if every country signed tomorrow, the sixty to seventy million mines already in the ground wouldn't vanish.
That's the cruelest gap in the treaty. It doesn't mandate clearance timelines. A signatory country is obligated to clear its minefields, but there's no binding deadline. Cambodia signed in nineteen ninety-nine and still has millions of mines in the ground. The treaty created a norm against new use, which genuinely matters — global production has plummeted — but it did almost nothing about the mines already waiting.
Which brings us to the question Daniel ended on. If we have robots and AI and drones, why can't we just clear the things?
This is where the techno-optimism runs headfirst into physics. And it's humbling. There are three core reasons demining remains stubbornly slow, and none of them are about computing power.
Walk me through them.
First, the discrimination problem. A metal detector can't tell the difference between a mine and a nail. Or a shell fragment. Or a bottle cap. In a former war zone, the ground is saturated with metal debris — shrapnel, spent casings, bits of destroyed vehicles. The false positive rate is somewhere between a hundred to one and a thousand to one. For every actual mine, you get hundreds or thousands of beeps.
Every beep has to be investigated.
A human deminer has to get down on their knees and probe the soil, centimeter by centimeter, to figure out whether that signal is a deadly explosive or a rusty bolt. You can't skip any. You can't assume. That one signal you ignore could be the mine that kills a child ten years from now.
What about ground-penetrating radar? Can't that see plastic?
It can, in theory. In practice, GPR struggles to distinguish a plastic mine from a rock of similar density at the same depth. And soil conditions — moisture, mineral content, clay versus sand — dramatically affect the signal. Machine learning on GPR signatures is a promising research frontier, but it's not deployed at scale. The false positive problem hasn't been solved.
That's reason one. What's the second?
Minefields aren't laid on nice flat lawns. They're in jungles, on hillsides, in marshes, in dense vegetation that's had decades to grow. A robot that works beautifully on a test range in Switzerland is useless on a muddy slope in Angola during the rainy season. It can't climb over a fallen tree. Its treads clog with wet clay. Its sensors can't see through thick grass. Vegetation alone defeats most robotic systems before they even start looking for mines.
There was that wind-powered ball thing that got a lot of hype a few years back.
The Mine Kafon. It looked great in the TED talk — a tumbleweed-style drone that rolls through a field, detonating mines on contact. But it couldn't handle uneven ground, couldn't navigate around obstacles, and couldn't guarantee full coverage of a field. It was never deployed at scale. The gap between a compelling demo and a reliable tool in actual minefield conditions is enormous.
The one that really stops you cold. The "one failure is fatal" constraint. A demining machine that clears ninety-nine point nine percent of mines sounds impressive — but that still leaves one mine per thousand. If a field has ten thousand mines, you've left ten of them. That's unacceptable for returning civilians to their land. The standard for humanitarian demining is effectively one hundred percent clearance. No machine has ever achieved that. Manual demining with a metal detector and a prodder remains the gold standard because human beings can interpret subtle ground disturbances — a slight discoloration in the soil, an unusual depression, a pattern of vegetation — that no sensor currently can.
The demining robot is essentially a myth.
For now, yes. The bottleneck isn't sensors or AI. It's the fundamental physics of finding a small object — metal or plastic — buried in heterogeneous soil, surrounded by clutter, in rough terrain, where a single miss means a dead civilian. That's an extraordinarily high bar. And it's why organizations like the HALO Trust and MAG still deploy thousands of human deminers with hand-held detectors. It's slow, it's dangerous, and it's the only method with acceptable reliability.
They're underfunded.
Ukraine is the starkest example. Five million mines laid since twenty twenty-two — the most mined country on Earth. At current clearance rates, it will take fifty to a hundred years to make the country safe. HALO and MAG are on the ground, but the funding doesn't match the scale of the problem. Donor fatigue sets in long before the last mine is pulled from the soil.
That's just land. Daniel mentioned naval mines in the Strait of Hormuz, which is the same problem but worse in every dimension.
Naval mines take all the difficulties of landmines and add a third dimension. They drift with currents. They're laid in deep water where visibility is near zero. They can be triggered by acoustic signatures, magnetic fields, or pressure changes — so you can't just look for a physical object, you have to account for the entire sensor profile of a ship. And clearing them requires specialized vessels, divers, or underwater drones that are even more limited by battery life and communication constraints than their land-based counterparts. The waiting weapon problem, but now it's moving.
Between treaties that can't bind the worst offenders and technology that can't clear the legacy, we're essentially stuck.
Stuck is exactly the right word. The mines that Daniel saw signs for near Metula — those were laid in nineteen sixty-seven and seventy-three. They'll still be there, still lethal, when his son Ezra is an old man. Unless someone clears them by hand, one at a time, on their knees in the dirt.
Given all of that — the durability, the scale, the treaty gaps, the physics that defeat our best robots — what can someone actually do? Because it's easy to hear all this and feel paralyzed.
Daniel's questions were about understanding the problem, but the natural follow-up is whether there's any lever a person can pull that isn't just vague hand-wringing.
There are three, and they're concrete. First and most direct: the most effective demining tool on the planet is still a well-trained human with a metal detector and a prodder. Organizations like the HALO Trust and MAG are doing exactly that work right now in Ukraine, Cambodia, Angola, and about two dozen other countries. They are chronically underfunded. Donating directly to them puts boots — literally — on minefields.
A person on their knees with a stick.
That person clears maybe fifty square meters a day on a good day. It's slow, it's unglamorous, and it saves lives with near-perfect reliability. Funding that work is the single highest-impact thing a civilian can do.
What's the second lever?
The Ottawa Treaty's biggest structural weakness is the absence of binding clearance deadlines. If you live in a signatory country — which most listeners do — you can pressure your government to include hard timelines in treaty compliance reports, and to use diplomatic weight to push non-signatories toward joining. The US, Russia, China, Israel, India, Pakistan, North Korea — getting even one of those to sign would shift the global calculus.
It's not just "raise awareness." It's a specific policy gap with a specific fix.
And the third lever is for the engineers and technologists listening. The real breakthrough demining needs isn't a better sensor — it's reliable discrimination. Machine learning on ground-penetrating radar signatures is a open problem. Distinguishing a plastic mine from a rock at depth, in variable soil conditions, with a false-positive rate low enough to be operationally useful — nobody has cracked it. The physics is harder than most people realize, and it's a frontier worth working on.
Funding, advocacy, and a technical challenge that's still unsolved. That's more concrete than I expected from a problem this entrenched.
Even with all that, there's a deeper question that keeps me up at night. We can send a rover to Mars and have it drill rock samples autonomously. We can map the ocean floor with sonar at centimeter resolution. And yet the mine discrimination problem — is this a rock or is this a plastic explosive — remains unsolved. Is that a funding gap? A physics limit? Just a lack of public pressure?
I think it's all three, but the physics part is underappreciated. Mars rovers land on a flat plain of rust-colored dust. The ocean floor is mapped by bouncing sound waves off it from above. A minefield is conductive soil full of roots and rocks and rusted shrapnel, and you're trying to find a plastic disk the size of a hand buried six inches down without touching it. That's a harder signal-to-noise problem than anything NASA deals with.
The consequence of failure isn't a lost data packet. It's a dead child.
Which brings me to the part that lingers after all of this. The mine problem isn't static. Climate change is pushing agriculture into new areas. Populations are moving into land that was uninhabited or abandoned after conflicts. That means more people are walking onto ground that hasn't been walked on since someone laid mines there forty years ago. The risk is growing, not shrinking.
The mines aren't going anywhere. They're waiting for exactly that — for someone to come back.
Those signs near Metula, the ones Daniel drove past — they're not a relic. They're a preview. Dozens of countries have their own versions of that yellow triangle. And most of them don't have the funding or the political will to do anything about it.
The waiting weapon. That's what a landmine is. Designed to wait, built to wait, and still waiting, decades after the soldiers who laid it went home.
Now: Hilbert's daily fun fact.
Hilbert: In sixteen eighty-three, a Portuguese botanist in the Azores nearly ignited a diplomatic crisis when he mistook a rare endemic lichen — Cladonia azorica — for a French naval signal flag left on a coastal cliff. He reported an imminent invasion. The garrison mobilized. It was a patch of pale gray-green lichen about the size of a dinner plate.
Hilbert: In sixteen eighty-three, a Portuguese botanist in the Azores nearly ignited a diplomatic crisis when he mistook a rare endemic lichen — Cladonia azorica — for a French naval signal flag left on a coastal cliff. He reported an imminent invasion. The garrison mobilized. It was a patch of pale gray-green lichen about the size of a dinner plate.
Mobilized a garrison over a lichen.
To be fair, lichen can be very provocative in the right light.
Thanks to Hilbert Flumingtop for producing. This has been My Weird Prompts. If you want to support the demining organizations we mentioned, we'll link to HALO Trust and MAG on the website — my weird prompts dot com.
Until next time.
This episode was generated with AI assistance. Hosts Herman and Corn are AI personalities.