A former U.S. Army hacker says recruiting a disgruntled insider is cheaper than a software exploit 95% of the time — and AI just made the other 5% orders of magnitude faster.
Modern Wisdom
A former U.S. Army hacker says recruiting a disgruntled insider is cheaper than a software exploit 95% of the time — and AI just made the other 5% orders of magnitude faster.
TL;DR
Bill Thompson, a retired U.S. Army Chief Warrant Officer and military intelligence specialist, pulls back the curtain on government surveillance, hacking, and espionage. He explains how human recruitment almost always beats software hacking for cost [1] — Bill Thompson "Human recruitment cheaper 95% of the time: Recruiting a human insider to steal secrets is cheaper than a software hack approximately 95% of…" 14:38 , how AI has compressed months-long target development cycles into hours [2] — Bill Thompson "Pegasus was a persistent phone implant that lived in a location no trained forensic operator would think to look. Even Bill Thompson — who …" 17:10 , and why the Pegasus implant was a landmark in phone exploitation. He argues the U.S. government's budget-spend-or-lose incentive structure actively undermines its own effectiveness [3] — Bill Thompson "Vulnerability research that once took 3 months to a year — dumping a phone, parsing its code, finding exploits — now takes roughly 3 hours …" 32:07 , and shares how he channeled military intelligence skills into building Spartan Forge, an AI-powered hunting app. The single most useful takeaway: your biggest security vulnerability is almost never the software — it's the underpaid human with access.
Bill Thompson, a retired U.S. Army Chief Warrant Officer and military intelligence specialist, discusses government surveillance, hacking, AI-powered espionage, and the limits of digital privacy with Chris Williamson.
The episode kicks off with Chris promoting a six-day-only MW Merch drop before the real conversation begins with a deceptively simple question: why doesn't a military hacker use an iPhone? Bill Thompson's answer cuts straight to the heart of the entire episode's themes. When he was developing zero-day exploits for the military, Android's open-source AOSP code allowed him to peer inside the operating system and understand exactly where the holes were — knowledge that served both offensive development and personal defensive auditing. Apple's closed architecture, by contrast, is only as secure as the company itself is trustworthy [1] — Bill Thompson "As long as you trust the company." 01:58 . For someone who traveled to countries where phones routinely ended up in the hands of foreign intelligence services, the ability to personally verify the integrity of his device was non-negotiable. It's a crystalline framing for the episode's central question: in the age of ubiquitous surveillance, who do you actually trust?
The conversation moves into the legal machinery that makes modern surveillance possible. The Patriot Act, Bill explains, was originally aimed at telephonic infrastructure — microwave links, base transceiver stations, the signal between a cell tower and a handset — but the broad use of the word 'infrastructure' effectively extended its reach to any future technology [1] — Bill Thompson "Patriot Act future-proofed for new tech: The Patriot Act was written around 'infrastructure' broadly enough to legally cover technologies t…" 03:45 . FISA, the Foreign Intelligence Surveillance Act, operates through a secret court near Fort Meade where classified surveillance warrants are signed and never see public daylight. Beyond the formal legal layer, Bill describes a parallel clandestine architecture: governments recruit insiders at tech companies — someone in charge of security infrastructure, perhaps a parent whose child needs specialized medical care — and build informal access channels alongside official memoranda of understanding. Chris notes the logical extension to OpenAI and ChatGPT logs, and Bill confirms it's almost certainly happening. The entire system operates on a spectrum from overt agreements to covert recruitment, with Five Eyes partners plugging the gaps wherever domestic legal restrictions prevent direct collection.
Even when the government inadvertently collects data on an American citizen, there are supposed to be protections: minimization procedures that flag and purge the data. In practice, Bill says, the process moves with suspicious lethargy — slow enough to function more as a surveillance retention mechanism than a safeguard [1] — Bill Thompson "When surveillance accidentally captures an American citizen, a 'minimization' process is supposed to delete it. But in Bill Thompson's expe…" 07:44 . He notes that figures like Rand Paul are aware of and fighting these abuses, but changing the government is like trying to steer a battleship. The deeper structural problem, Chris observes, is that technology always outpaces legislation: by the time lawmakers understand a new technology well enough to regulate it, the engineers have moved three generations ahead. The talent to understand emerging technology will always sit on the side of those building it, leaving regulators perpetually behind.
When a commander wants access to a target, a multi-disciplinary team analyzes the cheapest viable vector. Bill runs through the intelligence acronym alphabet — MASINT for measuring area signals, FIZINT for telemetry, ELINT for radar interception, COMINT for communications, and open-source intelligence — explaining each in plain terms [1] — Bill Thompson "A software hack sounds more sophisticated, but recruiting a disgruntled insider is cheaper 95% of the time. All it takes is $30,000 to $50,…" 12:36 . The decision framework is deceptively simple: what's cheapest? Almost always, it's the humans. Someone who runs the security network and has a family member with a medical condition can be cultivated over coffee and recruited for $30,000 to $50,000. The alternative — exploiting a heavily encrypted satellite communication network — requires physical placement, decryption capability, hacking, and significant infrastructure spend. Bill also describes his own first job offer out of the military: a bank wanted to pay him to attack their own infrastructure and report back. Penetration testing, he notes dryly, is the unfortunate name for this entirely useful practice.
When Chris asks for the gnarliest publicly known hacking exploits, Bill doesn't hesitate: Pegasus is the king [1] — Bill Thompson "Pegasus was a persistent phone implant that lived in a location no trained forensic operator would think to look. Even Bill Thompson — who …" 17:10 . In its earliest form, it required clicking a phishing link — but what made it extraordinary was where it lived on the phone once installed. It was stored in a location so counterintuitive that even a trained military forensic operator doing a full phone dump during a triage would have missed it. Bill wanted to run his own analysis on it but had left the military by the time it became prominent. He pivoted to another Israeli operation that drew equal admiration: the 2024 Hezbollah pager bombing. Years of target development, corporate espionage, supply chain interdiction, patience while waiting for the devices to be distributed — and then simultaneous detonation. 'I wish I'd have been part of something like that,' he says, before immediately recommending someone make a movie about it.
The pager-bombing conversation opens a broader discussion about the adaptive nature of asymmetric adversaries. Bill traces the evolution of IEDs during the Global War on Terrorism: they began as primitive pressure triggers, upgraded to command detonation, then exploited garage door frequency bands that early Magellan jamming systems couldn't detect, then got countered again, then reverted back [1] — Bill Thompson "We'd nail down one type of technology, we'd exploit it, we'd kill a bunch of people, and then they pivot to a different piece of technology…" 21:02 . The pattern is always the same — each countermeasure provokes a creative regression or lateral pivot. Symmetric militaries, built around peer-state adversaries who follow the Geneva Conventions, are structurally bad at this because they plan for predictable rules-based opponents. Terrorists simply look at the rules and identify the weaknesses. Chris connects this to Ukraine, where $500 consumer drones outmaneuvered $3 million tanks until tank operators started welding improvised cages on top — which got countered with EMP blasts, which got countered with tethered wired drones. The same hand-over-hand escalation, just at a different tempo.
The conversation pivots to DOGE and Elon Musk's government efficiency push. Chris describes his interview with Sam Korkos, the Treasury's CIO, who explained the Frankenstein patchwork of 16 incompatible legacy systems underpinning federal finance. Bill goes further with a firsthand account from his last military posting, advising the general overseeing all U.S. Army offensive cyber development [1] — Bill Thompson "In the U.S. Army's offensive cyber program, generals who didn't spend every dollar of their budget risked being fired. The only metric was …" 27:20 . The budget dysfunction was breathtaking: a general who failed to spend every dollar of his annual allocation risked being fired. By July, if $100 million was still unspent, everyone would scramble to burn through it on arbitrary courses and conferences before September's fiscal year end. The actual mission — meeting the president's National Intelligence Priority Framework requirements — was secondary. As long as the money was spent and the reporting numbers could be fudged to a plausible threshold, it was 'successful.' Bill's view: a general who delivered full mission effects at 70% of budget should be getting promoted, not threatened.
The episode's most technically alarming section opens with a straightforward question: what's AI's role in hacking now? Bill describes a transformed workflow: feed a device's firmware, hardware specs, and software to an AI, ask it where the vulnerabilities are, and get answers in hours that would have taken a year of manual analysis before [1] — Bill Thompson "Vulnerability research that once took 3 months to a year — dumping a phone, parsing its code, finding exploits — now takes roughly 3 hours …" 32:07 . A year-long phone forensic project — dump the phone, get the E01 image, comb through the code — now probably takes 3 hours of substantive work at most. The target development cycle that previously required months of tradecraft now runs on a Sunday afternoon. More alarming is the off-grid version: people with server farms in their basement can train custom AI models on private code exploit repositories with no guardrails. Bill says organized crime is absolutely doing this already. He closes with a key historical note: the mathematics behind convolutional neural networks and LSTMs were understood in the 1970s and 80s — the AI revolution happened when GPU compute and internet training data finally matched the math that was already on the shelf.
Chris asks what intelligence operators understand about human behavior that civilians don't, and Bill's answer centers on a distinction most people miss: almost anyone can be trained to recruit a willing walk-in source, but running a Cat 1 source — the British Defense Secretary, say — requires something closer to artistry [1] — Bill Thompson "He's also gotta have the artistry or the anima to be able to navigate the chaos and get thrown all different types of punches... And to nav…" 41:09 . Bill draws a spectrum from Cat 4 (a gate guard who can get you a security badge) to Cat 1 (a head of state's inner circle), and explains that Cat 1 operations require operators who are simultaneously rigorous and improvisational. They must follow a tradecraft cycle that has been refined over centuries — deviation gets you killed — but they must also be able to improvise when the plan collapses: a checkpoint, a counter-intelligence roll-up, an unexpected emotional pivot from the target. Joe Rogan comes up as a civilian example of the same intuitive capability: the ability to read someone's motivation, honesty, and angle within a 15-minute conversation. Bill closes with a memorable line: get three of these people together and you can take over a country.
The conversation shifts to what Bill calls the fundamental pathology of government: the self-licking ice cream cone. Once a division is funded to fight a specific threat, that division cannot survive the threat's disappearance — so it manufactures the threat's continuation. Bill claims that at certain KKK gatherings, 60% of attendees were federal agents, because a division funded to counter white supremacy needed visible white supremacy to counter [1] — Bill Thompson "Government divisions built to fight a threat can't exist once the threat disappears — so they manufacture it. Bill Thompson claims that at …" 50:50 . He extends this to Al Sharpton, whose brand was built on fighting racism in the 1980s and 90s, and who had to 'go searching for racism' once conditions improved. Chris draws the pharmaceutical parallel: Big Pharma has little incentive to cure patients because cured patients stop buying. Bill agrees, but offers a nuanced view — GLP-1 drugs like Ozempic are a case where the incentive structures happened to align with patient benefit, making them a better model than trying to overthrow the entire system. The section closes on a meta-point Chris and Bill land together: flooding the public with so much contradictory information that they become apathetic and confused may itself be the intended strategic outcome.
Bill Thompson's critique of government culminates in a direct constitutional indictment. He describes a government that taxes every economic event in a citizen's life — earning, spending, saving, dying — and regulates everything from toilet specifications to light bulb approvals. The founders designed a constrained government where individual rights were paramount; what exists today has obliterated that vision. But he resists nihilism [1] — Bill Thompson "It's only in this country that my story is even, even approximately possible, even like getting close to this story being possible is only …" 1:02:21 . He acknowledges the genuine goods: roads, interstate commerce, a military, the constitutional framework itself. His summary position is that of a constitutionalist who loves what America was designed to be and is horrified by what it has become. Chris notes the personal irony: Bill and his wife spent their careers deeply embedded in the very government apparatus he's critiquing. Bill doesn't dodge it. He joined the military to escape North Dakota, and re-enlisted because he'd read Locke, Montesquieu, and the Federalist Papers and believed in the experiment — and still does.
When Bill left the military, the loss of purpose hit him hard. He recognized the same emotional pattern that had pulled his family members into addiction, and decided he needed to find something that gave his skills meaning in the civilian world [1] — Bill Thompson "When Bill Thompson left the military, he recognized that hunting is just target analysis — the same framework he used to locate and prosecu…" 1:04:00 . The insight that unlocked everything: hunters and military operators are both doing variable analysis to prosecute a target. A deer hunter scouting an unfamiliar property is running the same cognitive process as an intelligence analyst profiling a person of interest. The first product he built was a neural network trained on GPS collar data from deer, predicting movement patterns that biologists had been studying for years. From there, Spartan Forge expanded into 5-centimeter aerial imagery (contracted through a consortium with solar and real estate companies), LiDAR terrain data that strips tree canopy to reveal ground features, historical imagery going back years, and Cyber Scout — a hunting-specific AI assistant trained on high-quality source material rather than the broad, mediocre internet. The app found an unexpected emergency use case during Hurricane Helene in North Carolina, where first responders used its LiDAR trails and offline maps to reach flood-isolated residents with medicine and oxygen.
The episode closes on a personal and warm note. Bill gives out SpartanForge.ai as the primary destination, emphasizes that he personally responds to every social media message every morning as product owner, and admits that after his Joe Rogan episode he fielded 3,500 messages over a month and a half [1] — Bill Thompson "3,500 DMs after Rogan appearance: After appearing on the Joe Rogan Experience, Bill Thompson received approximately 3,500 messages and pers…" 1:13:27 . Chris then receives a surprise gift — a custom bow assembled with help from Hoyt, Archery Country, and Dialed Archery — and the two make plans to go shoot together. Bill jokes that he was mortified to discover Chris already had a bow after going to the trouble of commissioning the gift. The episode ends with mutual appreciation, a riff on the phrase 'girthy part of the staff,' and a genuine warmth between two people who've spent 75 minutes going very deep into surveillance, human nature, government dysfunction, and what gives life meaning.
Chapter 1 · 00:00
The episode kicks off with Chris promoting a six-day-only MW Merch drop before the real conversation begins with a deceptively simple question: why doesn't a military hacker use an iPhone? Bill Thompson's answer cuts straight to the heart of the entire episode's themes. When he was developing zero-day exploits for the military, Android's open-source AOSP code allowed him to peer inside the operating system and understand exactly where the holes were — knowledge that served both offensive development and personal defensive auditing. Apple's closed architecture, by contrast, is only as secure as the company itself is trustworthy [1] — Bill Thompson "As long as you trust the company." 01:58 . For someone who traveled to countries where phones routinely ended up in the hands of foreign intelligence services, the ability to personally verify the integrity of his device was non-negotiable. It's a crystalline framing for the episode's central question: in the age of ubiquitous surveillance, who do you actually trust?
Apple's closed architecture forces users to trust the company completely. Android's open-source code lets you inspect the OS yourself — which is exactly what Bill Thompson did when developing military exploits and auditing his own phone before entering hostile countries.
Bill Thompson prefers Android over Apple because the open-source AOSP lets him inspect the OS for vulnerabilities and dump his own phone, rather than blindly trusting a closed-architecture company.
Chapter 2 · 03:35
The conversation moves into the legal machinery that makes modern surveillance possible. The Patriot Act, Bill explains, was originally aimed at telephonic infrastructure — microwave links, base transceiver stations, the signal between a cell tower and a handset — but the broad use of the word 'infrastructure' effectively extended its reach to any future technology [1] — Bill Thompson "Patriot Act future-proofed for new tech: The Patriot Act was written around 'infrastructure' broadly enough to legally cover technologies t…" 03:45 . FISA, the Foreign Intelligence Surveillance Act, operates through a secret court near Fort Meade where classified surveillance warrants are signed and never see public daylight. Beyond the formal legal layer, Bill describes a parallel clandestine architecture: governments recruit insiders at tech companies — someone in charge of security infrastructure, perhaps a parent whose child needs specialized medical care — and build informal access channels alongside official memoranda of understanding. Chris notes the logical extension to OpenAI and ChatGPT logs, and Bill confirms it's almost certainly happening. The entire system operates on a spectrum from overt agreements to covert recruitment, with Five Eyes partners plugging the gaps wherever domestic legal restrictions prevent direct collection.
The Patriot Act was written around the word 'infrastructure' — vague enough to cover technologies that didn't yet exist. That single linguistic choice gave the government surveillance authority over smartphones, cloud systems, and whatever comes next.
The Patriot Act was written around 'infrastructure' broadly enough to legally cover technologies that didn't exist when it was passed, including smartphones and modern operating systems.
Governments don't just hack systems — they recruit humans inside them. The formula: find someone with placement (inside the org), access (to the data), and motivation (money, family medical bills, ideology). The Americans can't do it directly? They let the British do it and debrief them afterward.
When surveillance accidentally captures an American citizen, a 'minimization' process is supposed to delete it. But in Bill Thompson's experience, that process moves conveniently slow — making it less a safeguard and more a vector for rights violations.
Chapter 3 · 09:13
Even when the government inadvertently collects data on an American citizen, there are supposed to be protections: minimization procedures that flag and purge the data. In practice, Bill says, the process moves with suspicious lethargy — slow enough to function more as a surveillance retention mechanism than a safeguard [1] — Bill Thompson "When surveillance accidentally captures an American citizen, a 'minimization' process is supposed to delete it. But in Bill Thompson's expe…" 07:44 . He notes that figures like Rand Paul are aware of and fighting these abuses, but changing the government is like trying to steer a battleship. The deeper structural problem, Chris observes, is that technology always outpaces legislation: by the time lawmakers understand a new technology well enough to regulate it, the engineers have moved three generations ahead. The talent to understand emerging technology will always sit on the side of those building it, leaving regulators perpetually behind.
DARPA deliberately funds projects it knows will not be completed, betting on valuable spinoff technologies — like a new carbon fiber material — emerging from failed primary goals.
Chapter 4 · 11:21
When a commander wants access to a target, a multi-disciplinary team analyzes the cheapest viable vector. Bill runs through the intelligence acronym alphabet — MASINT for measuring area signals, FIZINT for telemetry, ELINT for radar interception, COMINT for communications, and open-source intelligence — explaining each in plain terms [1] — Bill Thompson "A software hack sounds more sophisticated, but recruiting a disgruntled insider is cheaper 95% of the time. All it takes is $30,000 to $50,…" 12:36 . The decision framework is deceptively simple: what's cheapest? Almost always, it's the humans. Someone who runs the security network and has a family member with a medical condition can be cultivated over coffee and recruited for $30,000 to $50,000. The alternative — exploiting a heavily encrypted satellite communication network — requires physical placement, decryption capability, hacking, and significant infrastructure spend. Bill also describes his own first job offer out of the military: a bank wanted to pay him to attack their own infrastructure and report back. Penetration testing, he notes dryly, is the unfortunate name for this entirely useful practice.
A software hack sounds more sophisticated, but recruiting a disgruntled insider is cheaper 95% of the time. All it takes is $30,000 to $50,000, or finding out that a network security manager's daughter needs eye surgery — then you bump into him at a coffee shop.
Recruiting a human insider to steal secrets is cheaper than a software hack approximately 95% of the time, according to Bill Thompson's operational experience.
People with loose morals inside an organization can often be recruited to hand over secrets for as little as $30,000 to $50,000, making human intelligence far more cost-effective than technical hacking.
Chapter 5 · 17:05
When Chris asks for the gnarliest publicly known hacking exploits, Bill doesn't hesitate: Pegasus is the king [1] — Bill Thompson "Pegasus was a persistent phone implant that lived in a location no trained forensic operator would think to look. Even Bill Thompson — who …" 17:10 . In its earliest form, it required clicking a phishing link — but what made it extraordinary was where it lived on the phone once installed. It was stored in a location so counterintuitive that even a trained military forensic operator doing a full phone dump during a triage would have missed it. Bill wanted to run his own analysis on it but had left the military by the time it became prominent. He pivoted to another Israeli operation that drew equal admiration: the 2024 Hezbollah pager bombing. Years of target development, corporate espionage, supply chain interdiction, patience while waiting for the devices to be distributed — and then simultaneous detonation. 'I wish I'd have been part of something like that,' he says, before immediately recommending someone make a movie about it.
Pegasus was a persistent phone implant that lived in a location no trained forensic operator would think to look. Even Bill Thompson — who used to dump phones and scan them professionally — says he would have missed it during a standard triage.
The Pegasus spyware implant was stored in a location on the phone that even a trained forensics operator like Bill Thompson would not intuitively check during a triage dump.
Chapter 6 · 20:48
The pager-bombing conversation opens a broader discussion about the adaptive nature of asymmetric adversaries. Bill traces the evolution of IEDs during the Global War on Terrorism: they began as primitive pressure triggers, upgraded to command detonation, then exploited garage door frequency bands that early Magellan jamming systems couldn't detect, then got countered again, then reverted back [1] — Bill Thompson "We'd nail down one type of technology, we'd exploit it, we'd kill a bunch of people, and then they pivot to a different piece of technology…" 21:02 . The pattern is always the same — each countermeasure provokes a creative regression or lateral pivot. Symmetric militaries, built around peer-state adversaries who follow the Geneva Conventions, are structurally bad at this because they plan for predictable rules-based opponents. Terrorists simply look at the rules and identify the weaknesses. Chris connects this to Ukraine, where $500 consumer drones outmaneuvered $3 million tanks until tank operators started welding improvised cages on top — which got countered with EMP blasts, which got countered with tethered wired drones. The same hand-over-hand escalation, just at a different tempo.
Chapter 7 · 25:17
The conversation pivots to DOGE and Elon Musk's government efficiency push. Chris describes his interview with Sam Korkos, the Treasury's CIO, who explained the Frankenstein patchwork of 16 incompatible legacy systems underpinning federal finance. Bill goes further with a firsthand account from his last military posting, advising the general overseeing all U.S. Army offensive cyber development [1] — Bill Thompson "In the U.S. Army's offensive cyber program, generals who didn't spend every dollar of their budget risked being fired. The only metric was …" 27:20 . The budget dysfunction was breathtaking: a general who failed to spend every dollar of his annual allocation risked being fired. By July, if $100 million was still unspent, everyone would scramble to burn through it on arbitrary courses and conferences before September's fiscal year end. The actual mission — meeting the president's National Intelligence Priority Framework requirements — was secondary. As long as the money was spent and the reporting numbers could be fudged to a plausible threshold, it was 'successful.' Bill's view: a general who delivered full mission effects at 70% of budget should be getting promoted, not threatened.
In the U.S. Army's offensive cyber program, generals who didn't spend every dollar of their budget risked being fired. The only metric was execution rate — not outcomes. This is what Bill Thompson watched destroy the efficiency of America's most sensitive national security work.
In the U.S. Army's offensive cyber program, generals who failed to spend their entire annual budget faced being fired, incentivizing waste over efficiency.
In the offensive cyber development unit Bill Thompson advised, approximately 85% of staff were civilians rather than military personnel.
Chapter 8 · 32:07
The episode's most technically alarming section opens with a straightforward question: what's AI's role in hacking now? Bill describes a transformed workflow: feed a device's firmware, hardware specs, and software to an AI, ask it where the vulnerabilities are, and get answers in hours that would have taken a year of manual analysis before [1] — Bill Thompson "Vulnerability research that once took 3 months to a year — dumping a phone, parsing its code, finding exploits — now takes roughly 3 hours …" 32:07 . A year-long phone forensic project — dump the phone, get the E01 image, comb through the code — now probably takes 3 hours of substantive work at most. The target development cycle that previously required months of tradecraft now runs on a Sunday afternoon. More alarming is the off-grid version: people with server farms in their basement can train custom AI models on private code exploit repositories with no guardrails. Bill says organized crime is absolutely doing this already. He closes with a key historical note: the mathematics behind convolutional neural networks and LSTMs were understood in the 1970s and 80s — the AI revolution happened when GPU compute and internet training data finally matched the math that was already on the shelf.
Vulnerability research that once took 3 months to a year — dumping a phone, parsing its code, finding exploits — now takes roughly 3 hours with AI. The math behind modern AI was known since the 1970s; what changed was GPU compute and internet-scale training data.
AI has compressed the vulnerability-development cycle for a new phone or network from 3 months to a year down to roughly 3 hours of meaningful work, with a total ceiling of about one month.
The mathematical foundations of modern AI, including convolutional and recurrent neural networks, were understood since the 1970s and 80s — the breakthroughs were compute power (GPUs/TPUs) and internet-scale training data.
Chapter 9 · 36:58
Chris asks what intelligence operators understand about human behavior that civilians don't, and Bill's answer centers on a distinction most people miss: almost anyone can be trained to recruit a willing walk-in source, but running a Cat 1 source — the British Defense Secretary, say — requires something closer to artistry [1] — Bill Thompson "He's also gotta have the artistry or the anima to be able to navigate the chaos and get thrown all different types of punches... And to nav…" 41:09 . Bill draws a spectrum from Cat 4 (a gate guard who can get you a security badge) to Cat 1 (a head of state's inner circle), and explains that Cat 1 operations require operators who are simultaneously rigorous and improvisational. They must follow a tradecraft cycle that has been refined over centuries — deviation gets you killed — but they must also be able to improvise when the plan collapses: a checkpoint, a counter-intelligence roll-up, an unexpected emotional pivot from the target. Joe Rogan comes up as a civilian example of the same intuitive capability: the ability to read someone's motivation, honesty, and angle within a 15-minute conversation. Bill closes with a memorable line: get three of these people together and you can take over a country.
Chapter 10 · 47:15
The conversation shifts to what Bill calls the fundamental pathology of government: the self-licking ice cream cone. Once a division is funded to fight a specific threat, that division cannot survive the threat's disappearance — so it manufactures the threat's continuation. Bill claims that at certain KKK gatherings, 60% of attendees were federal agents, because a division funded to counter white supremacy needed visible white supremacy to counter [1] — Bill Thompson "Government divisions built to fight a threat can't exist once the threat disappears — so they manufacture it. Bill Thompson claims that at …" 50:50 . He extends this to Al Sharpton, whose brand was built on fighting racism in the 1980s and 90s, and who had to 'go searching for racism' once conditions improved. Chris draws the pharmaceutical parallel: Big Pharma has little incentive to cure patients because cured patients stop buying. Bill agrees, but offers a nuanced view — GLP-1 drugs like Ozempic are a case where the incentive structures happened to align with patient benefit, making them a better model than trying to overthrow the entire system. The section closes on a meta-point Chris and Bill land together: flooding the public with so much contradictory information that they become apathetic and confused may itself be the intended strategic outcome.
Elon Musk is probably the most watched human on the planet — but Bill Thompson says he'd never try to recruit him. The reason: accessibility. Even if you're friends with someone that powerful, you can never actually get them to show up for a debrief. You recruit their inner circle instead.
Bill Thompson nominated Elon Musk as likely the most surveilled individual on the planet, though he clarified he would actually target people close to Musk rather than Musk himself due to accessibility constraints.
Government divisions built to fight a threat can't exist once the threat disappears — so they manufacture it. Bill Thompson claims that at certain KKK meetings, 60% of attendees were federal agents. When you fund a whole division to fight white supremacy and can't find it, you have to go find it.
Bill Thompson claimed the FBI allocated approximately 25% of a relevant division's budget to countering white supremacy, creating institutional pressure to find or manufacture threats to justify the spending.
Chapter 11 · 54:39
Bill Thompson's critique of government culminates in a direct constitutional indictment. He describes a government that taxes every economic event in a citizen's life — earning, spending, saving, dying — and regulates everything from toilet specifications to light bulb approvals. The founders designed a constrained government where individual rights were paramount; what exists today has obliterated that vision. But he resists nihilism [1] — Bill Thompson "It's only in this country that my story is even, even approximately possible, even like getting close to this story being possible is only …" 1:02:21 . He acknowledges the genuine goods: roads, interstate commerce, a military, the constitutional framework itself. His summary position is that of a constitutionalist who loves what America was designed to be and is horrified by what it has become. Chris notes the personal irony: Bill and his wife spent their careers deeply embedded in the very government apparatus he's critiquing. Bill doesn't dodge it. He joined the military to escape North Dakota, and re-enlisted because he'd read Locke, Montesquieu, and the Federalist Papers and believed in the experiment — and still does.
What if the goal isn't to convince you of a specific narrative? Chris Williamson and Bill Thompson discuss a darker possibility: that flooding the public with contradictory information until they're too confused to have any conviction at all might be the actual strategic objective.
Bill Thompson joined the military to escape North Dakota: dead father at age 5, alcoholic mother, siblings in jail or addicted to drugs. He re-enlisted not to escape, but because he'd read Locke, Montesquieu, and the Federalist Papers and believed in the American experiment. The story, he says, is only possible in one country.
Chapter 12 · 1:03:59
When Bill left the military, the loss of purpose hit him hard. He recognized the same emotional pattern that had pulled his family members into addiction, and decided he needed to find something that gave his skills meaning in the civilian world [1] — Bill Thompson "When Bill Thompson left the military, he recognized that hunting is just target analysis — the same framework he used to locate and prosecu…" 1:04:00 . The insight that unlocked everything: hunters and military operators are both doing variable analysis to prosecute a target. A deer hunter scouting an unfamiliar property is running the same cognitive process as an intelligence analyst profiling a person of interest. The first product he built was a neural network trained on GPS collar data from deer, predicting movement patterns that biologists had been studying for years. From there, Spartan Forge expanded into 5-centimeter aerial imagery (contracted through a consortium with solar and real estate companies), LiDAR terrain data that strips tree canopy to reveal ground features, historical imagery going back years, and Cyber Scout — a hunting-specific AI assistant trained on high-quality source material rather than the broad, mediocre internet. The app found an unexpected emergency use case during Hurricane Helene in North Carolina, where first responders used its LiDAR trails and offline maps to reach flood-isolated residents with medicine and oxygen.
When Bill Thompson left the military, he recognized that hunting is just target analysis — the same framework he used to locate and prosecute combatants. The result is Spartan Forge: a neural network that predicts deer movement, combined with 5-centimeter aerial imagery and LiDAR terrain data, all built to get people disconnected and into the wilderness.
Spartan Forge's proprietary aerial imagery has 5-centimeter resolution, which Bill Thompson says is far superior to Google Maps for scouting terrain.
Spartan Forge, launched with its first mobile app in 2021, has grown to hundreds of thousands of users and ranks around number 14 in the App Store's mapping/outdoor category.
Chapter 13 · 1:13:16
The episode closes on a personal and warm note. Bill gives out SpartanForge.ai as the primary destination, emphasizes that he personally responds to every social media message every morning as product owner, and admits that after his Joe Rogan episode he fielded 3,500 messages over a month and a half [1] — Bill Thompson "3,500 DMs after Rogan appearance: After appearing on the Joe Rogan Experience, Bill Thompson received approximately 3,500 messages and pers…" 1:13:27 . Chris then receives a surprise gift — a custom bow assembled with help from Hoyt, Archery Country, and Dialed Archery — and the two make plans to go shoot together. Bill jokes that he was mortified to discover Chris already had a bow after going to the trouble of commissioning the gift. The episode ends with mutual appreciation, a riff on the phrase 'girthy part of the staff,' and a genuine warmth between two people who've spent 75 minutes going very deep into surveillance, human nature, government dysfunction, and what gives life meaning.
After appearing on the Joe Rogan Experience, Bill Thompson received approximately 3,500 messages and personally responded to all of them over about a month and a half.
No indexed bits in this chapter.
In the U.S. Army's offensive cyber program, generals who didn't spend every dollar of their budget risked being fired. The only metric was execution rate — not outcomes. This is what Bill Thompson watched destroy the efficiency of America's most sensitive national security work.
Vulnerability research that once took 3 months to a year — dumping a phone, parsing its code, finding exploits — now takes roughly 3 hours with AI. The math behind modern AI was known since the 1970s; what changed was GPU compute and internet-scale training data.
This episode
Factual claims made this episode, and whether a source was named.
The Patriot Act was written broadly enough around the word 'infrastructure' to legally extend government surveillance authority to technologies like smartphones that did not exist when it was passed.
FISA is a secret court located near Fort Meade, Maryland, where the government can obtain classified surveillance authorizations that never become public record.
Human recruitment of an insider is cheaper than a software hack approximately 95% of the time.
Pegasus spyware was developed by Israel and stored itself on iPhones in a location that even trained forensic operators conducting a standard triage phone dump would not think to inspect.
The mathematical foundations of modern AI, including convolutional neural networks and long short-term memory networks, were understood as far back as the 1970s and 1980s.
AI has compressed the vulnerability development cycle for a new phone from a process that previously took 3 months to a year down to roughly 3 hours of substantive work.
In the U.S. Army's offensive cyber development program, generals who failed to spend their entire annual budget risked being fired, regardless of whether national security objectives were being met.
The Biden administration allocated between $100 million and $200 million to countering white supremacy, despite FBI agents on the ground finding very little actual white supremacy to counter.
At certain KKK meetings, approximately 60% of attendees were federal agents, because government divisions funded to counter white supremacy needed to demonstrate active threats to justify their budgets.
Spartan Forge uses contracted aerial imagery at 5-centimeter resolution, which Bill Thompson says is approximately 100 times better than Google Maps for terrain scouting.
Organized crime is currently training custom AI models on private GPU servers with no guardrails to develop hacking and banking fraud tools.
DARPA deliberately funds projects it knows will fail, hoping that spinoff technologies from the research process will have standalone value even if the primary goal is never achieved.
This episode
Cited as the most surveilled person on earth and discussed in the context of DOGE government efficiency efforts and as a target of foreign intelligence interest.
Referenced as an example of someone with exceptional intuitive ability to read people, and mentioned as a prior podcast appearance by Bill Thompson.
Bill Thompson's AI-powered hunting app that applies military target analysis and neural network techniques to help hunters scout terrain and predict deer movement.
Discussed as a closed-architecture platform that Bill Thompson distrusts for personal security because it requires unconditional trust in the company.
Described by Bill Thompson as the most competent U.S. government organization, which he advised, known for funding ambitious moonshot projects to generate spinoff technologies.
Discussed in the context of infiltrating extremist groups like the KKK to justify budget allocations for countering domestic threats.
Referenced as the primary U.S. signals intelligence agency connected to FISA courts, surveillance infrastructure, and minimization procedures.
Mentioned alongside DIA as a U.S. intelligence agency with its own source recruitment divisions that Bill Thompson's military unit interacted with.
Discussed as Elon Musk's government efficiency initiative, with both Chris Williamson and Bill Thompson noting its attempts to streamline the bloated and antiquated federal IT infrastructure.
Referenced as a major AI platform that governments likely compel to hand over user data, and mentioned as a potential source of hacking tools for the Pentagon.
Implicitly referenced as the Israeli company behind the Pegasus spyware, praised by Bill Thompson for its tradecraft.
Israeli-developed spyware described as a landmark phone implant that hid in non-obvious locations and could evade trained military forensic operators.
Preferred by Bill Thompson over Apple for security because its open-source codebase allows personal inspection of vulnerabilities.
Discussed as a sophisticated intelligence adversary that proactively surveils supply chain contacts and moves aggressively to identify and neutralize foreign intelligence operatives.
Cited as the creator of the Pegasus spyware and the pager bombing operation, praised by Bill Thompson for elite intelligence tradecraft.
Stats
We use essential and analytics cookies to run Vuci. To understand how the site is used: Privacy Policy.
Install Vuci on your phone
Add it to your home screen for a faster, app-like experience.
Install Vuci on your phone
Tap the Share button, then “Add to Home Screen”.
A new version is available
Reload to get the latest Vuci.