The Straightening

On Rivers, Attention, and the Geometry of Fire

Originally published on Medium, February 12, 2026.

I keep thinking about the silvery minnow.

If you’re not from New Mexico, you’ve probably never heard of this particular minnow. The Rio Grande silvery minnow is an unremarkable fish, maybe three inches long, that once spawned by the millions in the river’s slow-moving eddies and oxbow pools. The river was messy then. It meandered. It flooded. Snow melted and created backwaters where cooler water moved at different speeds, where sediment settled in unpredictable patterns, where ecosystems could develop in the quiet shade between the main channel and the cottonwood bosques along the banks. I remember those days.

Then we straightened the river.

We built levees and carved channels. We replaced the river’s wandering path with efficient geometry. Water that once spread across floodplains now moves fast and direct, from point A to point B. The minnow can no longer sustain wild reproduction. The bosques are dying. And if you ask the Army Corps of Engineers whether the project succeeded, they’d say, yes, the water goes where we want it to go.

I’ve been thinking about that logic as I watch my daughter scroll her phone. Not the content she’s looking at, but the geometry of her attention. The way her eyes track down the feed, rarely pausing, never wandering sideways — slowing down. The algorithm and the device have straightened the channel. Backlit, high-resolution information flows laminar and unbroken, from source to brain to reaction, with nothing to slow it down. No eddies where she might sit with an idea. No oxbows where contradictory thoughts could coexist. No bosques, those quiet spaces at the edges where something slower and stranger might take root. It makes me think about presence.

The content moderation debates often miss this deeper layer. We argue about what should flow through the channel while ignoring what happens when you straighten the channel itself.

Here is what I’ve come to understand, and it unnerves me: the same institutional logic that killed the silvery minnow is now being applied to human cognition. We are optimizing the information environment for throughput, for engagement, for velocity, for efficient delivery of stimuli to response. And in doing so, we are eliminating the cognitive equivalent of floodplains: the slow, inefficient, meandering mental processes where wisdom actually develops.

This isn’t a conspiracy. That’s what makes it worse.

The Fire

A straightened river moves water efficiently. But a straightened river also burns efficiently.

The bosques, those cottonwood forests along the riverbanks, evolved with the floods. When the floods stopped, the underbrush accumulated. When fire came, there was nothing to stop it. The same channels that moved water so efficiently now move fire. The infrastructure meant to serve us becomes the infrastructure that destroys us.

Physical fire requires three elements: oxygen, fuel, and heat. Remove any one and the fire dies.

Cognitive fire has its own triangle. Attention is the oxygen. Content is the fuel. Algorithms are the heat — the amplification and targeting that accelerates ignition. Modern platforms optimize all three simultaneously, and the result is an information environment that burns the way a straightened bosque burns: fast, hot, and with nothing left standing.

Platforms compete for attention. Attention responds to outrage. Outrage requires enemies. Enemies require division. Division generates more outrage. The loop is closed. A hard truth for me has been: nobody designed this. The systems produced it. Game theorists call this dynamic Moloch: competition that forces everyone into worse behavior, even when nobody wants the outcome. Platforms optimize for outrage because engagement pays. Creators race toward extremes because attention rewards it. Politicians amplify fear because fear mobilizes. Media frames everything as conflict because conflict clicks.

No one is steering; we are swimming in the same direction.

The Pattern

This isn’t new. We’ve been here before.

In the 1850s, the telegraph promised to shrink distances and democratize information. By 1866, Western Union controlled over 90% of the market and had formed a monopoly alliance with the Associated Press. The straightening served power. The telegraph enabled coordinated surveillance of Indigenous communities. General Nelson Miles combined heliographs and telegraph with the railroad network of New Mexico and Arizona to rapidly shift forces to reported sightings of Geronimo. The information infrastructure that promised connection delivered control. Standardize the channel, and power moves to whoever controls the switches.

Radio followed the same arc. The early airwaves were chaotic, experimental, local. The Federal Radio Commission called this diversity “propaganda” and standardized the spectrum, favoring corporate-sponsored networks over the multiplicity of diverse non-commercial stations. By the television era, three networks held monopoly audience control. The water flowed straight. Diversity becomes noise, and noise becomes justification for consolidation.

The Internet was supposed to be different. Tim Berners-Lee’s original vision: “No permission is needed from a central authority to post anything on the web, there is no central controlling node, and so no single point of failure.” The U.S. Supreme Court emphasized the potential of the Internet to create a radically more diverse public sphere. What happened instead: by 2020, the House Judiciary Committee found that four corporations, Apple, Amazon, Facebook, Google, increasingly serve as gatekeepers of commerce and communications in the digital age.

And now, AI. One industry analysis estimated that OpenAI and Microsoft together held roughly 69% of the foundation-models and platforms segment — a sign of how quickly the core layer concentrates. The easy data is gone; what remains is gated, copyrighted, or low-signal. By the end of 2023, the Reuters Institute found 79% of top U.S. online news sites were blocking OpenAI’s web crawlers.

The infrastructure that promised democratization delivers concentration. Every time. And with each iteration, the straightening moves one dimension closer to the mind. The telegraph straightened information flow. Social media straightened attention. AI straightens thought itself. When the tool you think with has already decided which questions are efficient, which framings are probable, which reasoning paths offer the least resistance, the straightening has moved inside the act of cognition.

You don’t notice the channel because you’re swimming in it.

The Architecture

There is a name for what happens inside a straightened channel. NATO calls it Cognitive Warfare. China’s PLA calls it Cognitive Domain Operations. The technical literature is extensive. The public discussion is almost nonexistent.

The critical distinction, from NATO General Paolo Ruggiero: “Its goal is not what individuals think but rather the way they think.”

Traditional propaganda targets beliefs. Cognitive warfare targets the prediction engine itself — the infrastructure that allows you to evaluate claims, weigh evidence, distinguish signal from noise. It doesn’t try to convince you. It tries to regulate you — like a thermostat.

One common sequence has five phases, described in doctrine and openly discussed in defense research circles.

Sensing. Before the system tries to change you, it measures you. Every pause, every scroll, every angry comment is data. The system is testing your boundaries in real time: which emotions override your critical thinking, what you need, how much complexity you can handle before you shut down.

Control selection. Once the system knows your vulnerabilities, it selects the right script for your emotional state. Fear gets certainty. Anger gets targets. Loneliness gets belonging. Shame gets redemption. This is why logical rebuttals fail. You’re attacking the content of the message. The message was never about content. It was about triggering a specific emotional state that bypasses your analytical mind. You’re playing chess while the system is playing neuroscience.

The effector. Three movements, each building on the last.

• Destabilization: destroy confidence in your own perception. A person who trusts what they see is hard to control. A person who doubts everything will grab onto anything that promises certainty. Firehose of falsehood, gaslighting, chronic cognitive dissonance with no resolution. You’re not confused by accident. Confusion is the point. A disoriented mind doesn’t seek truth. It seeks relief. And conveniently, the system is right there, ready to offer a simple story that makes the chaos make sense. You weren’t convinced. You were rescued.
• Isolation: control external inputs without obvious censorship. Not a wall. A gradient. Make outside information complicated, hostile, painful, disloyal. You can leave. The activation energy becomes massive.
• Re-encoding: install new reality through identity replacement. “I believe X” becomes “I’m someone who understands X” becomes “people like us know X” becomes “only bad people question X.” At that final stage, the belief has become load-bearing. You can’t remove it without collapsing the identity it supports.

Output. Beliefs fade. Actions lock them in. Once you act publicly, cognitive dissonance spikes, identity crystallizes, social bonds harden. The system doesn’t need to convince you anymore. You’re convincing yourself. The moral inversion completes: “This action was necessary” becomes “they deserved it” becomes “I’m brave for doing it” becomes “this is who we are.”

Feedback. The most elegant part: the system gradually withdraws. Because by now, you’re doing the work yourself. You police your own thoughts. You imagine what the group would think. You preemptively self-correct. You punish your own doubt. The surveillance is inside you now.

Each phase disabled the tool you’d need to resist the next one. This is what straightening does. It removes the eddies, the oxbows, the places where you might pause and reconsider.

By the end, escape doesn’t feel wrong. Escape starts to feel impossible.

The Dynamics

Nobody is coordinating this. That’s the hardest part to accept.

Russian information operations have been described as a firehose of falsehood approach. China runs more sophisticated cognitive domain operations. But the domestic engine is the one that should concern us most, because the domestic actors aren't coordinating. They're competing. And their competition can produce the same effect as coordination — because they're all optimizing for the same engagement dynamics.

Political operators deploy fear narratives because they work. Platforms provide the infrastructure because engagement is profitable. Media amplifies because attention pays the bills. Each actor is individually rational. No one can stop unilaterally. Stop deploying these techniques and your competitors don't. You lose.

This is the prisoner's dilemma at civilizational scale.

Complex systems exhibit consistent behavior immediately before they fail. Researchers who study bank runs, ecosystem collapses, autoimmune cascades, and coordination failures find the same structural properties: invisible accumulation, threshold non-linearity, positive feedback, and irreversibility. A bank's solvency depends on a structure so obvious we forget it's there: the shared belief that the bank is solvent. Confidence erodes invisibly. Then someone makes a visible withdrawal, and each subsequent withdrawal makes the next more rational. The positive feedback loop runs faster than any institutional response.

Trust in institutions has declined slowly over decades. Social media creates conditions where small events produce massive response differentials. Distrust feeds distrust as epistemically isolated communities reinforce their isolation. And we know from other domains that rebuilding trust requires far more than removing the causes of distrust.

This doesn't mean collapse is imminent or inevitable. Systems can operate near thresholds indefinitely. Feedback loops can be interrupted. But it does suggest that proportional thinking is precisely the wrong frame; the intuition that small problems require small solutions, that gradual trends will continue gradually, fails here.

The Window

I'm writing this in early 2026. There's a narrowing window where certain things are still possible.

Not because some arbitrary deadline expires. Because system dynamics compound. Because the physics being written now will constrain what's possible later. Because the straightening continues daily.

I want to be clear about what I'm not saying:

• I'm not saying we're doomed. The isomorphism works both ways. Systems that exhibit pre-collapse dynamics can also exhibit recovery dynamics. Negative feedback loops can be rebuilt. Redundancy can be restored.
• I'm not saying there's a conspiracy. There isn't one. There's something harder to fight — emergent behavior from misaligned incentives.
• I'm not calling fire in a crowded theater. The theater isn't on fire yet. But the underbrush is accumulating, and the channels that would normally distribute a fire's energy have been straightened.

What I am saying: the time to build different infrastructure is before you need it. Not after the fire. Not during the fire. Before.

The Physics

Here is what I've learned from the minnow, from the bosque, from the decades of watching systems.

You cannot fight fire with more fire. You cannot fight content with content. You cannot out logic, logic. You cannot fact-check your way out of a system designed to make fact-checking impossible.

You can only build different environments.

When the Rio Grande was unstraightened, when it still had bosques, oxbows, floodplains, fire couldn't spread the way it does now. The environment itself constrained propagation. Not because someone was watching. Because of physics.

The question for our digital infrastructure is the same question the river already answered: what physics would make cognitive fire harder to spread?

The direction is clear, even if the implementation is incomplete.

1. Identity as trajectory, not credential — because a trajectory can't be stolen, manufactured overnight, or faked by a bot network.
2. Federation over centralization, because when the environment is distributed, it can't be fully captured.
3. Context that travels with data, because information stripped of context is fuel for manipulation.
4. And sovereignty protected by physics rather than policy, because some constraints, sacred boundaries, ethical lines, the things communities hold inviolable, cannot be subject to optimization.

The acequias of New Mexico have operated for four hundred years. Water finding its way through channels dug by hands, governed by communities, maintained by shared labor. The system works because the community maintains it. Because the rules emerged from the people they affect. Because participation is not optional.

This isn't only history. Anyone who has replaced an algorithmic feed with an RSS reader, or switched a ranked timeline back to chronological, has felt the difference; the information still flows, but you chose the channels, and the silence between items is yours. A community that governs its own forum, sets its own norms, and decides what pace of conversation it can absorb is building a bosque. These are small acts of unstraightening. They work because they restore friction — not as inconvenience, but as ecology. Even a share button that asks you to open the link first is a tiny levee break. The eddy is where the minnow lives. Every new bosque starts with someone deciding the straight channel wasn't the only option.

The relationship between entity and environment is the law. The minnow doesn't negotiate with the river.

The Recursion

If you've followed this far, you're watching.

You're watching me describe how the system watches you. You're watching yourself understand the five phases. You're watching for signs that you've already been captured, watching whether your watching is healthy or pathological.

I'm watching too. I'm watching myself describe the architecture. I'm watching whether this description helps or whether it becomes another level of the trap.

Here's the question that won't leave me alone: if I'm watching myself watching, if I'm profiling my own vulnerabilities through this lens, am I defending against the system, or have I already become it?

Is the meta-awareness my immune system or the autoimmune disorder?

The structures are isomorphic.

• Gaslighting: can't trust own perception. Must constantly check perception against external reference. Checking never terminates. Authority rests outside.
• Meta-awareness: can't naively trust own perception. Must constantly check perception against... what? Checking never terminates. Authority rests... where?

The Panopticon works because the watching never stops. You internalize it. You watch yourself being watched. You watch yourself watching yourself being watched. The surveillance doesn't need to continue from outside. It runs forever inside.

The honest answer is: I don't know.

And I can't tell you.

If I say "your watching is healthy, their watching is pathological," I become external authority validating your perception. You become dependent on my validation. That's the structure of gaslighting.

If I say "yes, you're gaslighting yourself, you can't trust your watching," I undermine your perception. That's also the structure of gaslighting.

There is no answer I can give that doesn't participate in the problem.

The minnow doesn't watch itself swimming.

The minnow swims.

The recursion terminates in embodiment. At some point, you stop watching and you act. You swim. You move through the world and the world pushes back. The ground truth is not in the watching. The ground truth is in the contact.

The framework itself could become the cage. If every perception must be validated through the physics. If the cosmology becomes load-bearing identity. If the map becomes more real than the territory.

You don't escape the paradox by solving it intellectually.

You escape by swimming.

I think about my daughter again. The straightened channel of her attention, the fire that runs through it.

She doesn't need to understand the five phases. She doesn't need the framework.

She needs to swim.

And so do I. So do you.

The watching is serving if it leads to swimming. The watching is caging if the preparation never terminates in action.

I've written twenty-six RFCs and over 60 articles published. Built communities. Raised daughters. Driven ten thousand miles thinking. Twice. The watching hasn't stopped me from moving. I try to watch while I move.

The minnow is still swimming. For now.

I don't know what happens next. Nobody does.

But I know this: eventually you have to stop watching and start digging, or swimming, or being present, or volunteering, or walking on grass, or cooking, or something that brings you in contact with the world.

The water will find its way.

Chris Perkins writes from Las Cruces in Doña Ana County, New Mexico, where he lives with his wife Nayomi, their two daughters, and three dogs. He is the founder of NMCITRA (New Mexico Cyber Intelligence & Threat Response Alliance) and the architect of the Kinetic Trust Protocol. His work bridges communities and technology through data literacy and data storytelling, empowering people to see the systems that shape their lives.

This is the first of four essays. The next, "The Abundance of Trust," examines the layer beneath coordination.