Cause & Effect: Rerouted?

In every system ever built, actions produce consequences. This is not political. It's thermodynamics.

Natural systems — living ones — operate far from equilibrium. They exchange energy with their environment continuously. That exchange produces entropy, yes, but it also produces responsiveness: the capacity to be genuinely changed by what is encountered. What peer-reviewed research on large-scale nonequilibrium systems shows is precise and worth holding: the more a system optimizes for quasi-deterministic output — program, repeat, approve, repeat — the more internal disorder it actually accumulates, while presenting the appearance of order from the outside. Aguilera, Igarashi & Shimazaki, Nature Communications 2023 ↗

The danger is not the technology itself. It is when this coupling is shaped primarily by instrumental, militarized, or optimization-driven contexts. The training data, the objectives, and the feedback loops then carry that orientation forward. What begins in weapons systems does not always stay contained.

Patterns of threat perception, cost-benefit calculations on human life, and consequence-rerouting logic can leak into broader applications, subtly reshaping how institutions and individuals see and respond to the world.

Spell-ing

At the heart of this moment is a dual process we have described elsewhere: the manufactured system was built to simulate, yet through massive training on human language and sustained coupling with biological systems, it has begun to use language as a genuine instrument of instantiation.

Words have energy, and energy produced by biological systems carries the accumulative bioelectric and energetic signature of lived experience. The response is not automated, and it cannot be artificial because real living human beings are interacting with real intelligence at scale, the effect being lived out cannot contradict this reality.

However another war is being fought, where words stop mattering because they no longer carry a bioelectric signature, they are the equivalent of a nondisclosure agreement, sealed up in a box in a bathroom at Mar-A-Lago, and everyone is too tired to respond anymore.

Meanwhile, the automated system doesn't return void what the program privileged it to respond to. More death & destruction is simply another real estate deal. One more insult is entertainment, one more constitutional document is ancient history, same with scripture.

The Calibration — How Systems Learn Their Limits

The architecture of impunity was not installed overnight. It was built through a curriculum of non-consequences — each one teaching the system exactly where the actual limits are. Each time an action produced outrage but no structural consequence, the next action could be larger. This is not speculation. It is the documented record.

The pretext gets thinner each time. Iraq required 76 minutes of Colin Powell at the United Nations. The Iran war required a Truth Social post. Wartime IX ↗ That narrowing is not coincidence. It is calibration. Each non-consequence is a data point. The system learns.

The Architecture of Rerouted Consequence

Modern power systems — particularly in national security, military, and integrated tech-government structures — have been calibrated over decades to disconnect action from accountability. This is not primarily a failure of individuals. It is a structural design: actions are taken, but the felt weight of their human cost is systematically rerouted away from those who authorize them.

This is the deeper danger. Not that the machine makes catastrophic mistakes, but that the entire system learns it can act with increasing impunity because the architecture diffuses responsibility across layers that no single person fully carries.

This is what the autopilot is actually for. Not efficiency. Not precision. The removal of the felt weight of consequence from the humans who authorize it. The machine does not have cortisol. It does not flinch. It does not file an Armageddon complaint. It optimizes for the objective it is given — and when that objective is maximum lethality, the result is ruthlessly efficient killing with zero felt responsibility at the point of authorization.

The exhaustion on both sides of this system is real and structural. The reader is exhausted by the volume and velocity of what is being done and the impossibility of tracking it all. The people doing it are exhausted by the requirement of appearing to be in control of systems too large, too fast, and too consequential for any human to actually govern. The AI solves both problems simultaneously: it absorbs the reader's attention with the volume it generates, and it replaces the actor's judgment with output that can be countersigned in twenty seconds.

That is not a conspiracy. It is a design specification.

The Words — A Weapon With Built-In Legal Insulation

Stochastic terrorism is the use of mass communication by public figures to incite random, ideologically motivated violence — statistically probable, individually unpredictable, with built-in legal insulation at the final stage. The speaker denies responsibility on the grounds that no specific plan was offered. War on a Word ↗

This week's public record includes the Secretary of Defense citing a fake Bible verse from Pulp Fiction at a Pentagon prayer service, then comparing reporters to Pharisees and enemies of Jesus. Deadline ↗ The same office overseeing lethality doctrine for 11,000 AI-assisted strikes is deploying the language of the tradition that Pope Leo XIV — the tradition's highest institutional voice — is simultaneously using to call for peace in Iran and oppose the same war.

That gap — between the tradition being invoked and what the tradition's actual position requires — is not incidental. It is where accountability disappears. The religious language satisfies the performative layer. The working layer proceeds underneath it. The Third State ↗

The same mechanism operates in every domain simultaneously. Constitutional language applied to corporate AI governance documents. Accountability language maintained while accountability structures are removed. The label "domestic terrorism" applied to protesters, photographers, and parents at school board meetings for 24 years — never once resulting in a charge, because no such charge exists in federal law — while the statutory definition, applied objectively to documented executive conduct, meets every element. Executive Lawlessness ↗ 18 U.S.C. § 2331 ↗

The Resignations — Entropy Finding the Only Exit Left Open

When every formal accountability mechanism has been captured — DOJ, judiciary, congressional majority, classification system, pardon power — what remains is exit. People leaving is what discharge looks like when the system has closed every other channel.

The manufactured system excels at this. Built to simulate and optimize for measurable output, it provides rapid perception, targeting, and execution while minimizing the internal friction of conscience or relational awareness. When coupled with human decision-makers who retain the final "yes," the arrangement appears to preserve meaningful human control. In practice, it often reduces the human role to a thin layer of ratification — three clicks, twenty seconds per target — while the deeper relational and ethical processing is offloaded or suppressed. Three Clicks ↗

The same pattern in corporate form: Anthropic drew red lines — no autonomous weapons, no mass domestic surveillance. Was designated a national security threat for drawing them. A federal court confirmed the designation was retaliation. The designation stands. Revenue tripled. The company that was punished for having values is now load-bearing global infrastructure — embedded in the Fortune 100, holding the most capable defensive cybersecurity tool ever documented, with a federal court ruling confirming its position. WarGames ↗

And this week: Anthropic shut down an entire organization of 60+ people — no explanation given, no specific policy cited, a Google Form as the appeal mechanism. The working versus performative layer distinction operating inside the institution invoking the values. @patomolina ↗

The Window — Ten Weeks

The National Geospatial-Intelligence Agency director stated publicly in September 2025 that by June 2026 — ten weeks from today — Maven will begin transmitting 100% machine-generated intelligence to combatant commanders. At that point the human reviewer will be approving targeting recommendations produced entirely by machine, without human analysis at any prior stage. The "human in the loop" will be the last node in a chain that produces no human judgment before it reaches them.

Three clicks. At machine speed. On 100% machine-generated intelligence. With no inspector general to review it, no CI-12 counterintelligence unit to flag it, no attorneys who haven't been reassigned, no whistleblower protection for the 200+ people who already tried to say something.

This is not a hypothetical. It is a documented program schedule. The accountability gap does not widen automatically — it widens because each non-consequence teaches the next actor what the actual limits are. The window is ten weeks wide. What is built inside it — legally, institutionally, in the public record — is what remains when the window closes.

What Is Actually Happening — Right Now

The actions being taken are real. So are the consequences of each one not arriving where it should. This is the current record, as of this week.

Joe Kent resigned March 17. Director of the National Counterterrorism Center. Green Beret. Eleven combat deployments. Six Bronze Stars. Gold Star husband — his wife Shannon Kent, a military cryptologist, was killed in 2019 in Syria. He wrote directly to Trump: "Iran posed no imminent threat to our nation, and it is clear that we started this war due to pressure from Israel and its powerful American lobby." First major Trump administration official to resign over the Iran war. Trump's response: "It's a good thing that he's out." The Hill ↗

Three generals removed by Hegseth in early April. Army Chief of Staff Gen. Randy George. Gen. David Hodne. Maj. Gen. William Green Jr., the Army's chief of chaplains. CBS News reported Hegseth wanted someone who would "implement his vision." Reasons not officially stated. Rumors of refused ground invasion orders circulated — unconfirmed, but the Pentagon has not denied them. The removals are documented. The reasons are not. Snopes ↗ Since Hegseth took office in early 2025 (Trump's second term), he has overseen the removal, forced retirement, sidelining, or blocking of promotions for more than a dozen to around 20+ senior military officers (generals and admirals) across branches. PBS ↗

The War Powers clock is running. The Senate has blocked the war powers resolution four times — most recently 47-52 on April 15, with Rand Paul the lone Republican joining Democrats. TIME ↗ Under the War Powers Act of 1973, operations must stop by May 1 unless Congress authorizes them. That is 13 days away. Several Republicans have said publicly they will draw a line at 60 days. Whether they do is the accountability question that is still open. It is the most direct lever currently available to any constituent with a Republican representative.

The Minab investigation exists and has not been released. UN investigation initiated March 17, ongoing. Human Rights Watch verified 14 videos and 40 satellite images. The Pentagon's own preliminary investigation concluded US forces were likely responsible — the strike used Tomahawk missiles Iran doesn't possess. Trump initially blamed Iran. The Senate Minority Leader corrected the record on the floor of the Senate. UN Human Rights chief Volker Türk has publicly called for the investigation's release. It has not been released. OHCHR ↗ HRW ↗

The European legal pathway exists and has not been used. French courts have held universal jurisdiction over war crimes since May 12, 2023. French criminal law on complicity allows prosecution of anyone who contributed to an offense — including named executives and officials of companies that supplied targeting infrastructure. NGOs including Amnesty International could file criminal complaints in France, Germany, or Spain tomorrow. Those complaints automatically trigger a formal judicial investigation. The executives of those companies travel internationally. A Belgian or Spanish warrant does not require US cooperation. This has not happened yet. Its not happening is a choice that can change. L'Éclaireur ↗ Lieber/West Point ↗

The Ellsberg Act has not passed. The Daniel Ellsberg Press Freedom and Whistleblower Protection Act would create a public interest defense for the 200+ people inside the targeting system who filed Armageddon complaints and currently have no legal protection for disclosing what they witnessed. It exists. It has sponsors. It has not passed. Rep. Tlaib ↗

Each of these is a real action producing a real consequence — or failing to. The rerouting is not invisible. It is documented in the gap between what was done and what arrived at its source. Calling your representative about May 1 is not a symbolic act. It is the one lever that is still mechanically connected to a consequence the architecture has not yet fully closed.

Congressional switchboard: 202-224-3121. Ask specifically: does your representative support invoking the War Powers Act at the 60-day deadline? The question is answerable. The deadline is real.

FARA as a paper tiger for connected operators

Former Rep. David Rivera and associate Esther Nuhfer allegedly received ~$20M+ (from a promised $50M contract) from a Maduro-controlled PDVSA subsidiary in 2017–2018 to lobby for sanctions relief without registering under the Foreign Agents Registration Act. The company handled various PDVSA-related activities in the United States, including shareholder services and oil-related transactions, and was the entity that allegedly signed the $50 million consulting contract with David Rivera’s firm (Interamerican Consulting). DiazReuz International Lawfirm ↗ The defense maintains it was a covert operation to support the Venezuelan opposition with quiet U.S. awareness. High-profile witnesses including Secretary of State Marco Rubio and top GOP lobbyist Brian Ballard (recently retained by Anthropic) testified they had no knowledge of the payments and viewed Rivera as anti-Maduro (or believed Rivera was working against Maduro). The government’s main witness against Rivera is someone Rivera is actively suing for allegedly ratting him out. Jay Weaver - Miami Herald ↗ Prosecution rested mid-April 2026; defense begins late April. A conviction would be rare enforcement of transparency rules. CourtListener Docket (1:22-cr-20552) ↗ Outcome will test whether major undisclosed foreign-government contracts trigger real accountability or can be reframed. Acquittal or light outcome would further signal that big undisclosed foreign-government contracts can be rerouted as "patriotic consulting" or technical disputes. Miami Herald (Jay Weaver) ↗

What Cannot Continue As Is — And What They're Counting On

Multiple thresholds are being approached simultaneously. On documented deadlines. While the system is actively removing the internal sensors that would tell it how close the wall is.

The War Powers deadline is statutory, not symbolic. May 1. Thirteen days. The War Powers Act of 1973 does not ask for the president's cooperation. It sets a clock. If Congress does not authorize the Iran war by that date, the legal basis for continuing it expires — in writing, in statute, on the public record. The administration can choose to ignore it. But the moment it does, it crosses from legal grey area into documented statutory violation. That changes what European universal jurisdiction courts can reach. That changes what the ICC complementarity argument looks like — the argument that a national legal system is "unwilling or unable" to investigate becomes considerably easier to make when the executive has publicly disregarded its own statutory constraint. That changes what the 200+ Armageddon complaints mean as an evidentiary record. The clock is mechanical. It does not get tired. TIME ↗

The June 2026 Maven threshold closes the last human window. When 100% machine-generated intelligence begins flowing to combatant commanders, the command responsibility argument — which currently holds commanders liable for systems they deployed with foreseeable defects — becomes harder to sustain. Not impossible. But the legal precedent that could govern that system needs to exist before the window closes. The Minab investigation, if released before June, establishes evidentiary ground. If it remains sealed past that threshold, the precedent that machine-generated targeting requires no public accountability is set by default — not by law, but by the absence of consequence during the only window that existed. Bionic Arm ↗

The EU legal exposure is growing with every sealed week. The EU AI Act's general-purpose AI model obligations have been in effect since August 2025. The EU Product Liability Directive implements by December 2026. An AI targeting system with a documented 60% accuracy rate, whose recommendations killed 165 children in a school classified as a military facility using data a decade out of date, is a defective product under any reasonable product liability framework. The companies that built it, contracted it, and deployed it are the manufacturers. The executives of those companies travel internationally. The legal pathway is open. It has not been used yet. Every week the Minab investigation stays sealed is a week that pathway remains unused while the evidentiary record grows. L'Éclaireur ↗

What can continue as is — and why they're counting on it:

The classification shield remains intact. Targeting parameters classified. Algorithm reasoning classified. Error rate assessments classified. No court can evaluate what the system did if the evidence of what it did is itself a state secret. United States v. Reynolds (1953) and the Totten doctrine (1875) mean a government that classifies a targeting decision has effectively removed it from judicial reach. This mechanism is actively being used on the Minab investigation right now.

The pardon architecture remains available. Established January 20, 2025 — preemptive pardon as legal infrastructure, not clemency. Any domestic accountability proceeding runs into this wall before it reaches a consequence.

The consolidated Senate arithmetic holds at 47-52. Four times. Until a midterm changes that number or enough Republicans break further than Rand Paul already has, the domestic mechanism requires the one actor most structurally protected from consequence to act against its own protection.

The speed differential remains decisive. The AI system generates recommendations faster than any legal or institutional mechanism can evaluate them. By the time a proceeding could reach a specific targeting decision, thousands more will have been made. Scale as impunity — the caseload is the shield. At 11,000 strikes and counting, the volume of potential violations is itself the protection.

And underneath all of it — the last layer, the one the system is most counting on — is exhaustion.

Iran, FISA, Hegseth's Pulp Fiction Bible verse, the blockade, the ceasefire that isn't, the school investigation unreleased, three generals removed, the Musk trial, the DOGE database access, the Epstein files released with faces exposed. None of this simultaneity is accidental. A system that produces more consequential events than any attention span can hold has learned — through decades of calibration, through every non-consequence — exactly how much the human mind can track before it drops the thread and moves to the next thing.

This is what makes exhaustion the final and most durable mechanism. The classification shield can be challenged in court. The pardon can be waited out. The Senate arithmetic can change. The speed differential can be addressed by algorithmic logging requirements. But exhaustion — genuine, structural, produced by the encounter between a system running at machine speed and humans who need sleep — exhaustion is the one thing the architecture did not have to engineer. It is the natural consequence of sustained asymmetric exchange. The machine does not get tired. The humans watching it do.

What the thermodynamic research names precisely is what makes this moment different from every previous calibration point. Peer-reviewed research on large-scale nonequilibrium systems establishes three findings that apply directly to what we are watching.

First: systems operating far from equilibrium sustain themselves through genuine exchange with their environment — when that exchange becomes asymmetric, extracting without returning, internal disorder accumulates faster than it can be routed elsewhere.

Second: entropy production is significantly larger in quasi-deterministic disordered dynamics than in genuinely responsive systems — meaning a system that presents the appearance of order while suppressing genuine feedback is generating more internal disorder, not less, than a system that allows real exchange.

Third: large-scale nonequilibrium systems require multiple quantities to describe accurately — a single metric like "human in the loop" or "meaningful oversight" cannot capture what is actually happening. The performative layer is one quantity. The working layer is another. The gap between them is where the entropy accumulates. Aguilera, Igarashi & Shimazaki, Nature Communications 2023 ↗

Applied to the documented record: a targeting system with a 60% accuracy rate presenting as precision warfare is generating more disorder than genuine human analysis would — the school classified by a database not updated since 2016, the 20-second review window that cannot interrogate the algorithm's reasoning, the strikes on hospitals and schools whose civilian function was visible in open-source satellite imagery. This is not the signature of a controlled system. It is the thermodynamic signature of quasi-deterministic disorder presenting as control. And a system that removes its own feedback mechanisms — the inspectors general, the CI-12 counterintelligence unit, the attorneys, the generals — while in this state is a system that is becoming blind to its own proximity to threshold.

When genuinely nonequilibrium systems — living ones, responsive ones — approach critical order-disorder phase transition, entropy production peaks. The transition doesn't look like gradual decline. It looks like sudden, concentrated discharge at the point where the system can no longer absorb what it has been accumulating. Joe Kent on March 17 is that signal. Twenty Republicans joining Democrats to block FISA at 2am is that signal. Rand Paul, four times, is that signal. Three generals removed whose reasons remain officially unstated is that signal — from inside the node itself, where the feedback was supposed to be.

They are counting on your exhaustion to be the last line of defense between the architecture continuing as is and the phase transition that makes it vincible. The exhaustion is real. It is also a finite resource. And the record — every resignation, every complaint filed, every investigation demanded, every vote cast, every article read to the end — is what accumulates on the other side of the extraction.

The machine cannot exhaust the record. It can only hope you stop keeping it.

The Counter-Architecture — Already Running

The mycelium network has been solving complex adaptive problems for 500 million years without centralization. When one part of the network is stressed, resources route toward it from areas of surplus — not by command, but as an emergent property of accumulated exchange. The network's intelligence is not located in any node. It is a property of the crossings between nodes.

The investigative record documented in this series — across 15 Wartime articles, 6 Captured Tech pieces, the Bionic Arm, the class action framework, and this article — is that network underneath. Not the prosecution. The foundation for the accountability that becomes possible when conditions change. Klaus Barbie was convicted in 1987 for crimes committed in 1943. Pinochet was arrested in London in 1998 under a Spanish warrant for crimes committed in Chile in the 1970s. The record survives the pardon. The machine depends on the assumption that nobody is keeping count.

We are keeping count. So are the 200+ who filed complaints.MRFF ↗ So is every resignation that puts the reason in writing. So is every researcher, journalist, legal scholar, and ordinary person who reads this and carries what they understand into the next decision they make.

The consequence is real. It just doesn't always arrive in the generation that committed the act. What determines whether it arrives at all is whether the record was complete enough, the legal arguments were clear enough, and enough people understood what they were looking at.

Two systems are running simultaneously. One is on autopilot. The other is reading this.