Hey everyone, welcome back to My Weird Prompts. I am Corn, and I am sitting here in our living room in Jerusalem with my brother, the one and only Herman Poppleberry.
That is me, Herman Poppleberry, present and accounted for. And I have to say, Corn, I am particularly energized today. Our housemate Daniel sent over a prompt that is right up my alley. It is about the intersection of high-tech surveillance and the very old-school art of trickery.
It is a fascinating one. Daniel was looking at recent reports about satellite imagery showing that Iran appears to be burying the entrances to some of its nuclear sites, specifically around the mountains near Natanz and the Isfahan facility. It got him thinking about the whole game of cat and mouse between intelligence agencies and the countries they are watching. If everyone knows the satellites are up there, and everyone knows A-I is looking for anomalies, how do you actually hide anything? Or better yet, how do the analysts know when they are being lied to?
It is the ultimate shell game, but the shells are the size of buildings and the table is the entire planet. Deception in warfare is as old as warfare itself, but the medium has changed. We have gone from the Trojan Horse to inflatable H-I-M-A-R-S launchers and fake nuclear bunkers. What I find so compelling about Daniel's question is this idea of verification. In a world where you can buy high-resolution imagery of almost anywhere on Earth for a few thousand dollars, the secret is not necessarily hiding the object; it is hiding the truth of what that object is.
Right, and that democratization of data is a huge part of this. Daniel mentioned Maxar and the fact that commercial resolution has gotten so good that even hobbyists can do some level of intelligence work. We have seen this with open-source intelligence, or O-S-I-N-T, communities on social media. They are tracking troop movements in real-time. But the pros, the agencies with the big budgets, they are looking for things the rest of us cannot even see. Herman, before we get into the crazy tech like thermal sensing and synthetic aperture radar, let us talk about the history here. This did not start with satellites.
Not at all. If you look back at World War Two, you have the classic example of Operation Fortitude. That was the Allied plan to convince the Germans that the D-Day landings would happen at Pas-de-Calais instead of Normandy. They built an entire ghost army, the First United States Army Group, under General Patton. They had inflatable tanks, wooden airplanes, and they even used speakers to play the sounds of troop movements. From the perspective of a German reconnaissance pilot, it looked completely real. But today, a wooden airplane would not last five seconds against a modern sensor suite.
Because a wooden airplane does not look like a metal airplane to a radar, right?
Exactly. And it does not have an engine that gets hot. It does not have the same electronic signature. This is what we call signature management. Back in the nineteen-forties, you only had to worry about the visual signature. If it looked like a tank from five thousand feet up, it was a tank. Today, you have to manage the visual signature, the infrared signature, the radar cross-section, and even the patterns of life around the object.
Patterns of life. That is a phrase I hear intelligence analysts use a lot. What does that actually mean in the context of a satellite looking at a site in Iran or anywhere else?
It is basically the human footprint. Imagine you are an analyst looking at a suspected underground facility. You see a mountain. Nothing looks out of the ordinary. But then you start looking at the patterns. Are there new roads leading to that mountain? Is there a sudden increase in truck traffic? Where are those trucks going? If you see fifty trucks a day driving into a hole in a hill and coming out empty, you know they are excavating something. They are moving dirt, or what the industry calls spoil. That spoil has to go somewhere. If you find a new pile of dirt three miles away that matches the geology of that mountain, you have just found a secret construction project.
So the deception has to be perfect across every dimension. If Iran is burying an entrance to a nuclear site, like Daniel mentioned, they are not just putting a door underground. They are trying to change the narrative of the landscape. But Daniel asked specifically about how A-I and modern tech like thermal sensing play into this. Let us dig into the thermal side. If I am looking at a satellite image, how am I seeing heat?
It is all about the electromagnetic spectrum, Corn. Most of what we see is visible light, which is a very narrow band. But objects emit heat in the form of infrared radiation. Modern satellites, especially the ones used by organizations like the National Reconnaissance Office or even high-end commercial providers like Maxar with their Legion constellation, carry thermal infrared sensors. These sensors can detect temperature differences as small as a fraction of a degree.
So if there is a buried facility, it might still be warmer than the surrounding rock?
Precisely. Think about a centrifuge hall or a data center. Those machines generate a massive amount of heat. Even if you bury them eighty to one hundred meters deep—which is what we are seeing at the new Natanz site—that heat has to go somewhere. Usually, it is vented out through cooling towers or exhaust shafts. To a thermal satellite, those vents look like bright glowing beacons against the cold ground. Even if the entrance is hidden under a fake parking lot or a pile of sand, the thermal bloom tells the story.
That is fascinating. So the analysts are looking for what should not be there. A warm patch in the middle of a desert where there is no visible machinery. But what about the decoys? Daniel mentioned fake airplanes and airbases. I have seen photos of these inflatable tanks that look incredibly realistic. How does an analyst distinguish an inflatable tank from a real one using a satellite?
This is where it gets really fun and where the A-I comes in. An inflatable tank might look perfect in a standard photograph. It has the right shape, the right colors. But an A-I trained on millions of images of real military hardware is looking for things a human might miss. For one, the shadow. An inflatable tank is often lighter and might sit slightly differently on the ground, casting a shadow that is just a few pixels off. Or look at the tracks. A real sixty-ton tank is going to leave deep ruts in the soil as it moves. An inflatable tank is just sat there. If you see a tank in a field but there are no track marks leading up to it, the A-I flags that as a high-probability decoy.
That makes so much sense. It is the context, not just the object. It is like a digital version of those spot-the-difference puzzles. But I imagine the people making the decoys are getting smarter, too. Are they putting heaters inside the inflatables to fool the thermal sensors?
Oh, absolutely. Modern decoys are sophisticated pieces of engineering. Companies like GaardTech or Russia’s Rusbal make decoys with internal heaters to mimic the thermal signature of an engine. They are coated in metallic paints or have corner reflectors inside to give off the same radar return as a steel hull. Some of them even have small emitters that put out radio frequencies to make it look like there is a crew using a radio inside. It is a full-spectrum deception.
So if the decoy has the right look, the right heat, and the right radar reflection, we are back to square one. How do you break the tie?
You use multi-intelligence fusion. You do not just rely on one satellite. You look at the site over time. This is what we call change detection. If I have a satellite passing over a site every ninety minutes, I can see how things move. A real tank needs maintenance. People walk up to it. It moves to a different position. A decoy often just sits there for days or weeks. If an A-I sees a fleet of twenty tanks that has not moved an inch in three weeks despite a supposed military exercise, it starts to get suspicious.
I love that idea of time being the ultimate truth-teller. You can fake a moment, but it is much harder to fake a month. Now, Daniel specifically brought up the Iranian nuclear sites. There have been reports about tunnels being dug so deep that they are basically immune to conventional bunker busters. When you are looking at something like the Natanz or Isfahan sites, what are the specific clues an analyst is looking for to verify what is happening underground?
For a site like that, you are looking at the external infrastructure. You look at the power lines. A facility that is enriching uranium needs a massive, steady supply of electricity. If you see high-voltage lines running into a mountain, and those lines are drawing a load consistent with thousands of centrifuges, that is a huge indicator. You also look at the water. Centrifuges need cooling. If you see water being pumped from a nearby river into the site, and the return water is five degrees warmer, you know there is industrial activity happening inside that mountain.
And I assume you can see that temperature difference in the water using the same thermal sensors we talked about?
Exactly. Thermal sensing is great for tracking liquid flows. But there is another technology that is even more impressive for seeing through deception, and that is Synthetic Aperture Radar, or S-A-R. Unlike a camera, which needs sunlight, S-A-R sends out its own radar pulses and measures how they bounce back. It can see through clouds, through smoke, and even through some types of camouflage netting.
Wait, it can see through camouflage? How does that work?
It depends on the wavelength of the radar. Some wavelengths can penetrate through thin materials like canvas or plastic netting but will bounce off the solid metal of a vehicle or a missile launcher underneath. So you might have a site that looks like a harmless pile of hay to a regular camera, but to an S-A-R satellite, it looks like a battery of S-three hundred surface-to-air missiles.
That is incredible. So you really cannot hide anymore. But then I think about the Iranian example again. If they are burying the entrances, they are trying to hide the point of access. Is that just to protect it from a physical strike, or are they trying to hide the fact that the entrance even exists?
It is both. In the world of intelligence, we talk about denial and deception. Denial is when you stop the enemy from seeing something. Deception is when you make them see something else. By burying an entrance, you are denying the enemy a precise targeting point for a missile. If you do not know exactly where the door is, it is much harder to knock. But you are also trying to deceive them about the scale of the facility. If the visible footprint is small, you might underestimate how many centrifuges are spinning three hundred feet below the surface.
But as you said, the dirt gives it away. The spoil.
The spoil is the Achilles' heel of any underground project. During the Cold War, the Soviets were masters of this. They would build entire cities underground, but they would hide the dirt by dumping it into existing mines or using it to build fake hills and landscaping. Modern analysts use hyperspectral imaging to combat this. This is a step beyond thermal. It looks at hundreds of different bands of light to determine the chemical composition of the surface.
So they can tell if the dirt on the surface matches the rock from deep underground?
Precisely. If you have a limestone mountain, but the dirt on the surface has the chemical signature of deep-seated granite that has been freshly excavated, the jig is up. The hyperspectral sensor can see that chemical mismatch from space. It is like a forensic lab that covers the whole world.
It feels like the more we talk about this, the more impossible it seems to actually keep a secret. But nations still try. Why? Is it just about making it more expensive for the watcher?
That is a huge part of it. Intelligence is a resource game. If I can force an adversary to spend ten times more money and use ten times more analysts to find my secret, that is a win. Even if they eventually find it, I have diverted their attention and resources. And sometimes, deception is meant to be found.
Wait, what do you mean? Like a double bluff?
Exactly. Think about it. If I want you to think I am building a nuclear site in location A, I might intentionally do a mediocre job of hiding it. I will leave some tracks, maybe a bit of spoil, a few thermal leaks. You find it, you feel clever, and you focus all your satellites and analysts on location A. Meanwhile, the real facility is in location B, where I am being incredibly disciplined and using every trick in the book.
That is the classic shell game. You want the mark to think they have spotted the queen. It reminds me of some of the things we discussed way back in episode two hundred and twelve when we were talking about the psychology of misinformation. Once someone thinks they have discovered a secret, they become very invested in that being the truth. They stop looking elsewhere.
It is confirmation bias on a geopolitical scale. If an intelligence agency has a theory that a certain country is doing X, and they find a site that looks like it is doing X, they are much less likely to keep searching for the site that is actually doing Y. A-I is actually helping with this because it does not have the same biases. An A-I model can be programmed to keep searching the entire country with the same level of intensity, regardless of what it has already found. It does not get tired or bored, and it does not get a promotion for finding a "smoking gun."
That is a great point. The A-I is the ultimate objective observer. But I wonder, can you deceive the A-I itself? We have talked before about adversarial attacks on machine learning. Could you paint a pattern on the roof of a building that makes an A-I think it is a forest, even if a human can clearly see it is a building?
You absolutely can. There is a whole field of study around adversarial camouflage. You can create these "dazzle" patterns that exploit how a neural network perceives edges and textures. To a human, it looks like a weird, psychedelic mural. To an A-I, it looks like a cluster of trees or a pond. There was a famous study where researchers created a patch you could wear on your shirt that would make an object-detection A-I think you were invisible. The same principle applies to buildings and military hardware.
So we are entering an era of "A-I versus A-I" deception. One A-I is trying to spot the anomaly, and another A-I is trying to design the perfect camouflage to fool it. It is an arms race of algorithms.
It is, and it is happening in real-time. But here is the thing, Corn. The humans are still the final arbiters. At the end of the day, an A-I flags an anomaly, but a human analyst at the N-G-A or the Mossad or the C-I-A is the one who looks at it and says, "Wait, that doesn't look right." They bring in the other "INTs"—human intelligence, or H-U-M-I-N-T, signals intelligence, or S-I-G-I-N-T. Maybe they have a spy on the ground who can confirm what is in those trucks. Maybe they are intercepting the radio calls from the construction site. The satellite is just one piece of the puzzle.
It is the "all-source" approach. You never rely on just one thing. I want to go back to Daniel's question about the specific clues. We talked about thermal, we talked about S-A-R, we talked about spoil. What about something as simple as the shadows? You mentioned it briefly with the inflatable tanks, but I have heard that shadows are one of the hardest things to fake in satellite imagery.
They are incredibly difficult because they change every minute of the day. If you are building a fake building or a fake airplane, it has to cast a shadow that matches the position of the sun at that exact moment. If an analyst looks at a series of images taken at different times and notices that the shadow of a building is not moving or is the wrong length for the time of year, they know it is a two-dimensional painting or a low-profile decoy.
A painting? People actually paint fake features on the ground?
Oh, all the time. During the Cold War, the U-S and the Soviets would paint fake shadows of airplanes on their runways to make it look like they had more bombers than they actually did. But again, that only works for a static image. As soon as you have high-revisit satellites like we do now, you can see the shadow failing to move as the sun crosses the sky. It looks ridiculous once you see it in a time-lapse.
It is like that scene in a movie where the character realizes they are looking at a photograph instead of the real world because a clock is not ticking. So, Herman, what is the "state of the art" right now? If a country like Iran or North Korea really wants to hide something from the eyes in the sky in February twenty twenty-six, what is their best bet?
Their best bet is deep, multi-layered concealment combined with rigorous operational security. You bury the facility deep in a mountain that has existing industrial activity. Maybe you put your centrifuge hall under a legitimate cement factory. You use the same power, the same trucks, the same workers. You hide the signal in the noise. This is what we call "hiding in plain sight." If you can make your secret facility indistinguishable from a thousand other legitimate factories, you have won.
The "Purloined Letter" strategy. Put it right on the mantelpiece where everyone can see it, but nobody recognizes what it is. That seems much more effective than trying to build a fake mountain.
It is, but it requires incredible discipline. One mistake, one truck driver talking too much on an unencrypted phone, one thermal leak that is not properly managed, and the whole thing comes crashing down. And with the resolution of commercial satellites now hitting fifteen centimeters through advanced processing, you can see individual people. You can see the types of tools they are carrying. You can see the brand of the air conditioning units on the roof.
Fifteen centimeters. That is what, six inches? You can basically see a dinner plate from space.
Just about. At that resolution, you are not just looking for buildings; you are looking for people and their habits. Do the security guards change shifts at the same time every day? Do they carry the same weapons? Is there a specific car that always parks in the "V-I-P" spot? This is where the A-I really shines. It can track thousands of these tiny details across years of data and find the one thing that changed.
It is the ultimate "big data" problem. I am curious about the ethical side of this, too. We are talking about state actors, but this technology is available to almost anyone now. What does this mean for privacy? If I can buy fifteen-centimeter imagery of my neighbor's backyard, or a company's secret testing grounds, where does it end?
That is the million-dollar question. We are moving into a world of radical transparency. It is getting harder and harder for anyone—a government, a corporation, or an individual—to have a physical secret. There are companies now that sell "economic intelligence" based on satellite data. They count the cars in Walmart parking lots to predict quarterly earnings. They measure the shadows of oil tanks to see how much oil is being stored globally. They look at the light emissions from cities to estimate G-D-P growth in countries that might be faking their numbers.
It is a check on reality. You can lie on your balance sheet, but you cannot lie to the satellite about how many trucks are leaving your warehouse.
Exactly. And that is why Daniel's prompt is so relevant. Deception is the response to that transparency. As the watchers get better, the liars have to get more creative. It is a fundamental part of the human condition, I think. We have always tried to see what is hidden and hide what we do not want seen.
It is the eternal game. I want to circle back to the Iranian nuclear sites one more time, because that is such a high-stakes example. When we hear reports that "all entrances have been buried," like Daniel mentioned, is that a sign of desperation or a sign of confidence?
I would argue it is a sign of adaptation. They know they are being watched every second of every day. They know that Israel and the U-S have incredible intelligence capabilities. Burying the entrances is a way to raise the cost of an attack. It says, "We know you can see us, but we are going to make it as hard as possible for you to do anything about it." It is a defensive form of deception. They are not necessarily trying to convince you the site is not there; they are trying to convince you that attacking it is not worth the effort.
Like a porcupine. It is not hidden, but it is not easy to eat.
That is a perfect analogy. And for the analysts, the challenge is to keep looking past the quills. They are looking for the soft underbelly. Maybe it is a ventilation shaft three miles away. Maybe it is a fiber optic cable being laid in a trench. There is always a clue. The earth never stays perfectly still when humans are working on it.
This has been a fascinating deep dive. I feel like I am going to be looking at Google Earth very differently tonight. Before we wrap up, Herman, do you have any final thoughts on where this is going? What is the next big leap in satellite surveillance?
I think the next leap is persistent, all-weather, real-time surveillance. Right now, we still have gaps. A satellite passes over, then it is gone for an hour. But we are seeing the launch of constellations—hundreds of small satellites like those from SpaceX’s Starshield or BlackSky—that work together to provide a constant video feed of the entire planet. Imagine being able to "rewind" the world. An explosion happens in a city, and you can just scroll back the video to see exactly which car planted the bomb and where that car came from three days ago. That is where we are headed.
Total recall for the entire planet. That is both incredible and terrifying.
It really is. And the decoys of the future will have to be just as dynamic. They will have to move, they will have to "breathe," they will have to live. It is going to be a wild ride.
Well, I think we have given Daniel a lot to think about. From inflatable tanks in World War Two to fifteen-centimeter resolution and A-I-driven change detection in twenty twenty-six, the art of deception is more complex than ever.
It definitely is. And hey, if you are listening and you have a weird prompt of your own—maybe something about technology, history, or just a strange observation about the world—we want to hear it. Daniel is the one who usually finds these gems, but we are always open to more.
Absolutely. You can get in touch with us through the contact form at my-weird-prompts-dot-com. And while you are there, you can check out our full archive of episodes. We have covered everything from the ethics of A-I to the secret history of Jerusalem.
And if you are enjoying the show, we would really appreciate a quick review on your podcast app or on Spotify. It genuinely helps other people find us and keeps the show growing. We have been doing this for over five hundred episodes now, and it is the support from you guys that keeps us going.
Yeah, it really does make a difference. Well, Herman, I think that is a wrap for episode five hundred and fifty-six. This has been My Weird Prompts. I am Corn.
And I am Herman Poppleberry. Thanks for listening, and we will catch you in the next one.
Bye everyone.
Goodbye.
