The Abundance of Trust

The Missing Layer

Originally published on Medium, February 14, 2026.

The Pictures in Our Heads

In 1922, Walter Lippmann published Public Opinion and named a problem we still haven’t solved. “We are told about the world before we see it,” he wrote. “We imagine most things before we experience them.”

Lippmann was worried about newspapers. He understood that the average citizen couldn’t directly verify what was happening in foreign capitals or distant markets. We rely on representations. We respond to pictures in our heads, not to reality itself.

And those pictures are manufactured by someone.

But Lippmann was building on something older. Sixty years earlier, Hermann von Helmholtz had proposed that perception itself is a form of unconscious inference. The brain doesn’t passively receive the world. It models the world, predicts the incoming signal, and updates when the prediction fails. What we experience as “seeing” is actually a hypothesis — the brain’s best guess about what’s out there, refined through constant error correction.

This means the brain was never a camera. It was always a prediction engine.

Lippmann saw what happened when you scaled that insight to mass media. If the brain builds models, and media manufactures the raw material for those models, then whoever controls the media shapes the picture. He was describing a 1922 problem: a few newspapers, a handful of wire services, information moving at the speed of print.

Now there are 6 billion people connected to the Internet. The signals are constant, synthetic, adversarial, and faster than human cognition can verify. The prediction engine is still running the same Helmholtz-era inference process.

Our cognition is drowning in inputs it cannot check.

The Prediction Engine

The modern formulation comes from neuroscientists like Karl Friston and philosophers like Andy Clark. They call it predictive processing, and the core claim is elegant: the brain is a hierarchical inference machine that continuously generates predictions about incoming sensory data, compares those predictions to actual input, and propagates the difference (the prediction error) up through layers of processing to update the model.

This isn’t passive reception. It’s active modeling. The brain doesn’t wait for the world to announce itself. It guesses, then corrects. What we call “attention” is really precision-weighting — the brain deciding which prediction errors matter enough to update on. What captures attention shapes calibration. Feed the engine noise, and it calibrates to noise. Feed it threat, and it calibrates to threat.

The system works beautifully at the scale it evolved for.

Robin Dunbar, the British anthropologist, proposed that the human neocortex can maintain stable social relationships with roughly 150 people. The number is contested, but the insight holds: there’s a cognitive limit to how many individuals we can track, model, and trust. Within that 150, the structure is layered. About 5 intimates who receive 40 percent of our social time. 15 close friends. 50 in the broader circle. 150 at the outer edge of genuine relationship.

Beyond 150, we rely on institutions. Reputation systems. Proxies for trust. Rules that let us coordinate with strangers without having to personally verify each one.

The numbers now: 8.2 billion people on Earth. 6 billion connected to the Internet. Projections put us at 10 billion by 2060.

The prediction engine evolved for a village. It’s now processing signals from a planet-scale network, most of which it cannot verify, much of which is designed to exploit the very error-correction mechanism that makes inference possible.

The Minnow

The Rio Grande silvery minnow is three inches long and endemic to New Mexico. For most of human history, it was abundant — a small, unremarkable fish doing what small fish do. Now it’s endangered. As a conservation-reliant species, it is surviving only in a tiny fraction of its original range.

The minnow is an indicator species. When it thrives, the river system is healthy. When it vanishes, the problem isn’t the fish. The problem is the substrate: the flow patterns, sediment loads, temperature regimes, and seasonal rhythms that the minnow’s life cycle requires.

You cannot optimize a minnow into existence. You cannot breed your way to a healthy population while the river runs wrong. You have to restore the conditions. The physics. The actual substrate that makes minnow life possible.

This is not a metaphor. This is how complex systems work.

Trust is the minnow of social systems. When trust is present, coordination becomes possible. People can plan together, build together, solve problems together. When trust is absent, you can mandate all the policy you want — the coordination won’t hold. The collective action will collapse into defection, withdrawal, or tribal sorting.

We’ve spent two decades trying to regulate trust into existence. Content moderation. Fact-checking. Platform governance. Terms of service. Community guidelines. All of it treating trust as a policy problem.

You can’t moderate your way to trust any more than you can legislate a minnow.

The substrate is the problem. The river is running wrong.

The Doctrine

“The very concept of objective truth is fading out of the world.”**— George Orwell, “Looking Back on the Spanish War” (1943)

Here is where it gets uncomfortable.

Some nation-states have recognized that the decisive terrain of the 21st century is not land, sea, air, space, or even cyber infrastructure. The decisive terrain is the inference process itself — the loop that builds the pictures in our heads.

The operations follow a pattern. First, capture the channels through which prediction-relevant information flows. Own the platforms, or own access to the platforms. Second, distort the signal — inject noise, contradiction, and synthetic content that corrupts the update mechanism. Third, overwhelm the system’s capacity to distinguish signal from noise, truth from fabrication, friend from adversary.

“One of the saddest lessons of history is this: If we’ve been bamboozled long enough, we tend to reject any evidence of the bamboozle. We’re no longer interested in finding out the truth. The bamboozle has captured us.”**— Carl Sagan, The Demon-Haunted World (1995)

The goal is not to convince anyone of anything. The goal is to exhaust. To make coherent world-modeling so costly that populations default to tribal priors, learned helplessness, or withdrawal from the information environment entirely. The target is not belief. The target is the capacity to form accurate belief.

This is not speculation. The doctrinal literature is public. Academic journals, military white papers, and strategic communications research have documented these campaigns in granular detail. The substrate being targeted is cognitive. The operations are distributed across every platform, every feed, every notification. The intended outcome is control of decision-making at population scale.

Those tracking the pattern know the sequence. The letters are in the operations.

What does it look like when one trajectory tries to collapse another’s capacity to navigate? Not destruction — disorientation. Not death — capture. The prediction engine keeps running. It just stops producing accurate maps.

The Trust Collapse

The philosopher Annette Baier described trust as a three-place relation: A trusts B with valued thing C.

It’s not abstract confidence. It’s specific vulnerability — you are exposed to another person in a particular domain, and you proceed anyway because you have grounds for expecting they will not exploit that exposure.

Russell Hardin called it “street-level epistemology.” We estimate trustworthiness based on evidence. We update based on experience, reputation, social constraints, and shared stakes. Trust is rational when we have grounds for the estimate and mechanisms for revision.

Both frameworks assume something crucial: that verification is possible. You see the face. You know the history. You have enough context to make the estimate and enough feedback to update it. Trust doesn’t require certainty. It requires a functional loop.

At 6 billion connections going this fast, the loop breaks.

The source is anonymous or synthetic. The signal moves faster than human cognition can evaluate. The feedback is mediated by systems optimized for engagement rather than accuracy — systems that have discovered rage holds attention better than nuance, and novelty better than truth.

The data tells the story.

One in four American men under 35 reports feeling lonely “a lot of the previous day,” according to Gallup. Fifteen percent of men report having no close friends — a fivefold increase since 1990. Forty percent of bars and pubs have closed since COVID. The third places where verification happened, where you saw the same faces and built the update loop through repeated contact — those spaces are disappearing.

The prediction engine adapts to its environment. Feed it signals it cannot verify, and it calibrates to threat. Not because people are paranoid. Because the environment is genuinely adversarial, and the engine is doing exactly what Helmholtz described: inferring the most likely state of the world given available evidence.

Autonomy cannot exceed environment. The engine outputs what the conditions support.

The Missing Layer

Ezra Klein has been making the case for an abundance agenda. America forgot how to build, he argues. Housing is artificially scarce. Energy deployment is strangled by permitting. Infrastructure takes decades instead of years. The problem isn’t technology or resources: it’s vetocracy, NIMBYism, and a political system that makes blocking easier than building.

Get Chris Perkins’s stories in your inbox

He’s right. And his diagnosis is incomplete.

The abundance stack has layers:

• At the top, abundance itself — the material conditions that let people thrive.
• Below that, physical capacity: energy, housing, infrastructure.
• Below that, policy: the rules that permit or obstruct building.
• Below that, coordination: the collective action that makes policy possible.
• And below coordination, the foundation: trust.

You cannot coordinate if you cannot trust.
You cannot build policy if you cannot coordinate.
You cannot expand physical capacity if you cannot build policy.
You cannot have abundance if you cannot expand capacity.

The universe moves forward through computation. Every system (physical, biological, social) processes information and updates state. Choice is direction within the possible. But choice requires orientation. You have to know which way you’re pointed, what the options are, which signals to trust.

Klein is working one layer up. He’s right about the policy problem, the capacity problem, the building problem. But the layer underneath, the trust substrate, is degrading faster than capacity can accumulate.

Destroy orientation, and choice collapses to reflex. Reactivity.

Klein is designing the house. The foundation is washing out beneath it.

Proof of Presence

So what would restore the loop?

The brain’s trust circuitry evolved for physical co-presence. You see the face. You share the space. You know who was in the room. The verification wasn’t abstract — it was embodied, spatial, continuous. You didn’t have to decide whether to trust the signal because the signal was standing in front of you, and your prediction engine had been modeling that person through hundreds of micro-interactions.

This is the primitive that the current environment has severed. Not trust as feeling. Trust as loop — the cycle of exposure, observation, prediction, and update that lets the engine calibrate to another person.

What if you could rebuild that loop at scale?

Not through reputation scores, which can be gamed.
Not through algorithmic verification, which can be captured.
Not through identity documents, which can be forged.

Through presence itself. Proof that you were in the room. Proof that the interaction was embodied. Proof that the signal has a body attached to it.

This is the premise behind Noma — a proximity-based trust network built on Proof of Presence. The mechanism is simple: physical co-location as verification primitive. You don’t trust the platform. You trust the presence, because presence is what your prediction engine evolved to process.

The design principles follow from the physics:

• Exit is always possible. If you can’t leave, it’s not trust — it’s capture.
• Skin in the game is structural, not optional. The verification requires you to show up.
• The network is the trust, not the platform. No central authority decides who’s trustworthy. The loop decides, through repeated presence and accumulated evidence.
• Global trust isn’t a single verification at planetary scale. It’s a chain of local verifications that holds weight — each link tested, each connection embodied.

This is not a solution to the doctrine problem. No app solves population-scale information warfare. But it is a seed — an attempt to restore the substrate that makes trust possible. To rebuild the loop at human scale, so that whatever coordination happens can happen on ground that holds.

Seeds Here

I am not predicting collapse. I am not hoping the systems fail.

I am noticing that the minnow is disappearing and is now, reliant. The solution is not a better minnow. It’s a river that works.

Ten billion people by 2060. Billions more in agents. The question is not whether they’ll be connected — they will be, more densely than we can currently imagine. The question is whether they’ll be able to verify. Whether the prediction engine will have any ground truth to calibrate against. Whether trust will be possible, or whether we’ll have optimized it out of existence in the name of engagement.

The abundance everyone wants (material, social, cognitive) requires coordination. Coordination requires trust. Trust requires a loop: exposure, observation, prediction, update. The loop requires verification. Verification, at the scale the brain evolved for, requires presence.

This is not nostalgia for village life. It’s recognition that the physics hasn’t changed even though the scale has. The brain still infers. It still needs ground truth. It still calibrates to whatever environment it finds itself in.

Build the environment that supports trust, and trust emerges. It’s not magic. It’s not policy. It’s substrate.

The minnow doesn’t need a bigger river. It needs a river that runs. A river that gets lost sometimes.

Verify locally. Coordinate globally. Trust what you can test.

This is a seed. Here.

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 second of four essays. The previous, “The Straightening,” examined the geometry of cognitive fire. The next, “From Building Bunkers to Planting Seeds,” asks what survives it.

Appendix: What the Abundance Agenda Skips

1. Distribution

Abundance is not a production problem. It’s a distribution problem. We already produce enough food to feed 10 billion people. We already have more empty homes than homeless people in the US. Scarcity is structural, not technological.

AI could make production cheaper. It cannot, by itself, solve distribution. That’s political economy, not engineering.

2. Ownership

If AI replaces labor, who owns the AI? If three companies own the models that run everything, that’s not abundance. That’s feudalism with a better User Interface.

The people promising abundance are the same people building the moats.

3. Purchasing power

If AI replaces jobs, where does consumer purchasing power come from? “UBI funded by taxing AI productivity” requires political will to tax the people who own the AI. The same people currently spending billions on regulatory capture and lobbying.

The mechanism for redistribution does not exist. And the people who would need to build it are not building it.

4. Energy

945 TWh for data centers by 2030. A doubling in five years. The “abundance” vision requires energy abundance first. Where’s that coming from? Not fusion (not ready). Not renewables at current deployment rates. Natural gas, mostly. Which has its own constraints.

5. Materials

AI doesn’t mine lithium. It doesn’t smelt copper. It doesn’t build housing. Physical abundance requires physical resources, physical labor, physical logistics. AI can optimize these. It cannot conjure them.