Drop a stone into a still pond. The rings spread outward, cross, and fade, each ripple following from the splash in strict obedience to the physics of water. Given the stone, the angle, the depth, it seems the pattern could not have unfolded any other way. Now ask that question about everything. If our universe is a process running on something deeper — a state evolving by fixed rules from a single initial condition — then time itself may be the spreading of the rings: not a stage on which events happen, but the lawful aftermath of one throw. Were the ripples always meant to flow exactly this way? Is every crest and collision — every galaxy, every afternoon, every reader of this sentence — already implicit in the splash? Whether the outcome could be predetermined is a sharper question than it sounds, and modern physics has something precise to say about it.

Notice, though, what the ripple metaphor tells us. Ripples are patterns in something — the water runs deeper than the rings, and its physics, not theirs, does the real work. Carry that image over to the cosmos and you have quietly implied a medium beneath the universe, and perhaps a hand behind the throw. That idea has a name, and a bad reputation.

There is a version of the simulation hypothesis that deserves to be taken seriously, and it is not the one you have heard.

The one you have heard runs on three misunderstandings. First, that empty space seethes with particles flickering in and out of existence, giving reality a provisional, glitchy quality. Second, that quantum mechanics says things only become definite when someone looks, so the universe must render itself on demand, like a video game managing its frame rate. Third, that physicists keep discovering the universe is “pixelated” or “made of information,” so somebody, somewhere, must be running the code.

Each of these is wrong, and wrong in an instructive way. But when you clear them away, what remains is more interesting than the folklore: a set of rigorous results from quantum field theory and quantum gravity showing that particles, space, and even gravity can emerge from something deeper — something that looks a great deal like information processing.

So here is the plan. I am going to build the strongest simulation cosmology that physics permits, using real machinery: lattice gauge theory, quantum cellular automata, holography, quantum error correction. And I am going to mark, precisely, the single sentence where the argument stops being physics and becomes metaphysics. Most writing in this genre hides that seam. This essay is organized around it.

What “simulation” would have to mean

Forget the headset. Forget the render farm. A serious physical definition looks like this:

A universe is simulated if its observable states and laws are implemented by a more fundamental system whose own variables are different in kind from the space, time, fields, and particles experienced inside it.

A hurricane simulated on a computer is not made of air. It is made of numbers and electronic states. Yet within the model, the hurricane genuinely has pressure, temperature, rotation, and turbulence — quantities defined and lawful at the model’s own level. Nothing about the hurricane’s internal meteorology reveals the silicon.

By analogy, our universe might be represented by deeper quantum degrees of freedom that are not themselves electrons, photons, or points in ordinary space. That gives the model two layers:

• The substrate (or host): a fundamental quantum-information system.

• The experienced universe: the spacetime, matter, and forces reconstructed from it.

Note what this definition does not include: a programmer. “Simulation,” in this sense, establishes an implementation layer. Whether anyone built it, and why, are separate hypotheses, and this essay will not touch them.

Two things to unlearn about empty space

The popular picture of the quantum vacuum manages to be wrong in both directions at once. It makes empty space both emptier and busier than it actually is.

First: the vacuum is not nothing. In quantum field theory — the framework underlying all of particle physics — the vacuum is the lowest-energy state of the quantum fields, defined relative to a particular energy accounting and, as we will see, a particular observer. It is what remains after every removable particle has been removed. But the fields themselves remain, and they obey the uncertainty principle. For paired field quantities analogous to position and momentum,

Δq · Δp ≥ ħ/2,

where ħ is Planck’s constant divided by 2π. The vacuum therefore has irreducible fluctuations in its field values, correlations between distant regions, entanglement, and a definite, calculable response to detectors and external disturbances. It is a structured physical state — arguably the most structured object in the theory.

Second: it is not a sea of particles popping in and out of existence. The “virtual particles” of popular accounts are, in general, bookkeeping terms in a calculational expansion — useful fictions of perturbation theory, not short-lived objects flying through space. Producing actual particles from the vacuum requires a physical mechanism — an expanding spacetime, a horizon, an accelerating detector, a strong external field — and the energy always comes from somewhere: the background field or the dynamical geometry. Nothing is created from nothing.

Hold onto both corrections. The simulation cosmology we are building will need the vacuum to be rich, but lawful.

Where particles actually come from

To follow the argument, you need one conceptual upgrade from high-school physics: particles are not tiny balls. They are excitations of fields.

A field is a quantity defined throughout space — the electromagnetic field is the familiar example. Quantum field theory assigns one to every particle type: photons are excitations of the electromagnetic field, electrons of the electron field, quarks of the quark fields, the Higgs boson of the Higgs field. The right analogy is a guitar string. The string is the underlying object; a note is a pattern of vibration in it. The note is perfectly real, but it is not a separate substance added to the string. In the same spirit,

particle = quantized excitation of a field,

and the energy of one quantum of frequency f is E = hf, with h Planck’s constant. A particle is less like a thing and more like a note.

Once you see particles this way, a remarkable family of results becomes intelligible: circumstances under which a state with no particles evolves into a state with particles — lawfully, with the energy bill paid in full.

Expanding universes. Imagine retuning a guitar string while it is still sounding. The old definition of “one clean vibration” and the new one no longer agree. An expanding universe does this to quantum fields: the natural vibration patterns at early times and at late times differ, and there may be no single, God’s-eye definition of “positive frequency” — of what counts as a particle — valid for all epochs. The two definitions are related by a mathematical transformation (named for Bogoliubov) that mixes creation and annihilation:

aₒᵤₜ = α aᵢₙ − β aᵢₙ†.

Here aᵢₙ removes a particle by the early-time definition, aᵢₙ† creates one, and aₒᵤₜ is the late-time notion. When the mixing coefficient β is nonzero, the early-time vacuum contains late-time particles, with expected number

⟨N⟩ ∝ |β|².

The changing gravitational geometry supplies the energy. Leonard Parker derived this mechanism for expanding universes in the 1960s — including the telling precision result that certain highly symmetric (”conformally invariant”) massless fields are not produced by pure conformal expansion. Particle creation is a sharp consequence of specific conditions, not a vague seepage of energy from nowhere. (Parker, Quantized Fields and Particle Creation in Expanding Universes)

Strong electric fields. A sufficiently intense electric field can produce electron–positron pairs, with the pairs’ energy drawn from the field, which weakens accordingly. Julian Schwinger computed the vacuum’s decay probability from the imaginary part of the quantum effective action. (Schwinger, On Gauge Invariance and Vacuum Polarization)

Acceleration. Strangest of all: a uniformly accelerating detector responds to the ordinary inertial vacuum as though immersed in a warm bath of particles at temperature

T = ħa / 2πck,

where a is the acceleration, c the speed of light, and k Boltzmann’s constant — while an inertial observer looking at the very same field state counts zero particles. This is the Unruh effect, and the energy for the detector’s clicks is supplied by whatever does the work of accelerating it. (Unruh, Notes on Black-Hole Evaporation)

Two lessons, and they cut in opposite directions.

The first is radical: “particle” is often an emergent, state- and observer-dependent concept. The field is the invariant reality; the particle count is a derived quantity that can legitimately differ between observers.

The second is deflationary, and it dismantles the video-game intuition: none of this is observer-triggered. The field correlations evolve whether or not anyone measures them. The accelerating detector doesn’t force the universe to render particles; it couples to structure that was in the vacuum state all along. Nothing here supports the claim that reality is computed on demand when looked at.

Five results that make a simulated universe possible

With the vocabulary in place, here are the pillars — five independent lines of research, each establishing that some feature of our world could be implemented by something deeper.

1. A universe on a grid

In 1974, Kenneth Wilson showed that gauge theories — the type of quantum field theory underlying the Standard Model — can be defined on a discrete spacetime lattice while preserving exact gauge invariance. At finite lattice spacing a, continuum momentum is replaced, schematically, by

p → (2/a) · sin(ap/2).

When momenta are small compared to the lattice scale, this reduces to ordinary momentum and you recover ordinary relativistic physics. Near the lattice scale, the relationship between energy and momentum bends, and continuous rotational symmetry degrades to the discrete symmetry of the grid. (Wilson, Confinement of Quarks)

This is normally scaffolding: lattice QFT is a calculational tool, and one takes the continuum limit a → 0 at the end. But Beane, Davoudi, and Savage asked the literal question: suppose the universe is such a lattice — a numerical simulation of the kind our own physicists run, using a specific “early” architecture (unimproved Wilson fermions on a cubic grid). What follows? They derived a lower bound on the inverse lattice spacing of roughly 10¹¹ GeV — about a hundred billion times the proton’s rest energy — from the observed high-energy behavior of cosmic rays, along with candidate signatures: a momentum cutoff in the most energetic particles, and directional effects aligned with the lattice axes. (Beane, Davoudi, and Savage, Constraints on the Universe as a Numerical Simulation)

What this establishes: a lattice simulation could reproduce all of low-energy particle physics and might leave detectable fingerprints in the ultraviolet. What it does not establish: that any lattice is ontologically real — or that a simulator would use a cubic grid at all.

2. Relativity from pure information processing

D’Ariano and Perinotti approached from the opposite end. Start with no spacetime, no fields, no particles — only a network of finite quantum systems updating by local rules: a quantum cellular automaton. Impose principles that read like requirements on a well-formed computation: the update preserves quantum information (unitarity); each cell interacts only with neighbors (locality); the same rule applies everywhere (homogeneity); no direction is fundamentally preferred (isotropy); each cell is as small as possible.

From these assumptions they showed that the automaton’s large-scale, low-energy behavior reproduces the Weyl and Dirac equations — the equations governing electrons and the other spin-½ particles — with related constructions recovering free Maxwell dynamics, the physics of light. Lorentz symmetry, the symmetry of special relativity, emerges at long wavelengths and is slightly deformed near the automaton’s native scale. (D’Ariano and Perinotti, Derivation of the Dirac Equation from Principles of Information Processing; Quantum Cellular Automata and Free Quantum Field Theory)

The chain is constructive:

local quantum update rule → coarse-grained relativistic fields → particles as excitations.

The result is not that our universe has been shown to be an automaton. It is that familiar relativistic physics can arise from quantum-information rules containing neither particles nor continuous space — so the continuum appearance of our world cannot, by itself, rule out a computational substrate. The unpaid bills are real, though: interactions, gravity, and the full Standard Model spectrum must all be recovered without unacceptable artifacts.

3. The world as a hologram

The holographic principle began with a puzzle about black holes: their entropy — their information capacity — scales with the area of the horizon, not the volume inside. That is deeply strange. It suggests that the information content of a region of space is fundamentally two-dimensional, as if the third dimension were bookkeeping. (Susskind, The World as a Hologram)

Juan Maldacena made this precise. In the AdS/CFT correspondence, a gravitational theory in a (d+1)-dimensional “bulk” spacetime is exactly equivalent — dual — to a quantum field theory without gravity living in d dimensions on that spacetime’s boundary. The equivalence is total: every bulk object, black holes included, corresponds to some configuration of the boundary theory. Schematically,

Z(bulk gravity) = Z(boundary QFT),

an identity between the two theories’ complete physical content. (Maldacena, The Large N Limit of Superconformal Field Theories and Supergravity)

For our purposes: a lower-dimensional quantum system can encode, completely, what internal observers experience as a higher-dimensional gravitational universe. An extra dimension of space can be reconstruction rather than raw material.

Two limits, both essential. Duality is not simulation: it is a statement that two descriptions carry the same information, with neither side privileged as “more real.” Simulation is a claim about which description runs on the hardware — and the mathematics is silent on that. And the best-understood case concerns anti-de Sitter space, a spacetime with negative curvature quite unlike our own expanding cosmos.

4. Space, stitched from entanglement

Within holography, something more surprising emerged: the amount of entanglement in the boundary theory is directly geometric. The Ryu–Takayanagi relation states that the entanglement of a boundary region equals the area of a corresponding surface in the emergent bulk, in gravitational units:

S(region) = Area(surface) / 4Għ,

with G Newton’s constant. Entanglement on one side; literal geometric area on the other. (Ryu and Takayanagi, Holographic Derivation of Entanglement Entropy from AdS/CFT)

Push on this and gravity itself falls out. For small disturbances of the vacuum, quantum mechanics supplies an “entanglement first law” relating changes in entanglement to changes in energy, δS = δ⟨H⟩. Faulkner and collaborators showed that this law, combined with the holographic entropy relation, is equivalent — under specified conditions — to the linearized Einstein equations in the bulk: the equations of weak-field gravity, derived from the entanglement structure of a quantum state. (Faulkner et al., Gravitation from Entanglement in Holographic CFTs)

And the encoding has a recognizable engineering signature. Almheiri, Dong, and Harlow showed that the way bulk objects sit inside the boundary theory behaves like a quantum error-correcting code: a single logical object in the bulk has multiple, redundant reconstructions on the boundary, so that damage to part of the encoding leaves the object intact. (Almheiri, Dong, and Harlow, Bulk Locality and Quantum Error Correction in AdS/CFT)

Assembled: space, gravity, locality, and matter can arise as redundantly encoded logical structure within a more fundamental quantum state. But note the limit — “code” names a mathematical structure, not a coder. Nothing in the theorems implies the code was written rather than intrinsic to nature.

5. The universe as its own computer

Finally, the widest lens. Any closed quantum system evolves by the Schrödinger equation,

|Ψ(t)⟩ = e^(−iHt/ħ) |Ψ(0)⟩,

an information-preserving rotation of the system’s state. Quantum computation is also exactly this: unitary evolution. Quantum mechanics sets speed limits on how fast a system can move between distinguishable states, and entropy and gravity bound how much information a region can hold. Seth Lloyd derived these ultimate physical limits on computation and applied the accounting to the observable universe, treating its entire history as a bounded quantum computation with a finite budget of elementary operations and memory. (Lloyd, Ultimate Physical Limits to Computation)

But notice what this most naturally establishes: the universe computes itself. Its lawful evolution is already, formally, information processing. Calling every unitary evolution “a simulation” merely renames physics. An external simulation requires something extra — that the evolution is implemented and interpreted by another system. Physics does not supply that premise.

The ledger

Lay the five pillars side by side and a pattern appears. Each proves a possibility; each requires the same kind of unproven premise to become an actuality.

Framework What follows from physics Extra premise needed Finite lattice model QFT can have detectable cutoff artifacts Our cutoff is a simulator’s grid Quantum cellular automaton Relativistic fields can emerge from local quantum updates Our universe instantiates that automaton AdS/CFT Higher-dimensional gravity can be encoded by lower-dimensional QFT The encoding is an executed simulation Entanglement/QEC spacetime Geometry and matter can be emergent logical variables The code has external hardware Computational universe Physical evolution has information-processing bounds Computation means external simulation

Computability, discreteness, emergence, duality — none of these is simulation. They make a simulated universe physically coherent. They do not produce an external computer. The strongest conclusion the physics alone supports is this:

Quantum field theory and quantum gravity provide several rigorous mechanisms by which particles, locality, and even spacetime can emerge from more fundamental quantum-information structures. They establish that a simulation-like universe is physically coherent and, in architecture-specific cases, potentially testable. They do not establish that our universe is externally simulated.

Why none of this is evidence

It is worth being quantitative about the gap, because this is where most simulation arguments quietly cheat.

Evidence, in the Bayesian sense, is anything more probable under one hypothesis than another. Now take any observation D that is simply a prediction of quantum field theory — vacuum fluctuations, cosmological particle production, the Unruh effect. If both an ordinary universe and a faithfully simulated one obey QFT, then

P(D | simulation) / P(D | not simulation) ≈ 1.

The likelihood ratio is one. The observation cannot move you. Every “quantum weirdness proves we’re simulated” argument founders on this single line: phenomena predicted equally by both hypotheses are evidence for neither. Only simulator-specific deviations — artifacts of a particular architecture — could ever budge the ratio.

One more exclusion, made deliberately. Nick Bostrom’s famous simulation argument is not part of this construction. His is an anthropic, probabilistic argument that runs on assumptions about substrate-independent consciousness, the ambitions of future civilizations, the number of simulations they run, and how to count observers — not on quantum field theory or vacuum physics. (Bostrom, Are You Living in a Computer Simulation?) This essay runs on physics until the precise moment it declares otherwise.

That moment is now.

The axiom, stated out loud

Here is the sentence that does all the work:

Simulation-selection axiom. Whenever observed physics admits an exact realization as quantum-information processing on a more fundamental substrate, treat that realization as ontologically fundamental, rather than regarding the observed laws as brute facts.

Everything before this sentence was physics. Everything after it is physics seen through it.

Be clear about what the axiom is not. It is not Occam’s razor. If anything, ordinary parsimony cuts the other way. Measure hypotheses by description length — K(·) meaning, roughly, the length of the shortest complete specification — and compare:

L_base = K(laws + initial state)
L_sim  = K(host laws + simulator + simulated laws + initial state)

Unless the host and simulator dramatically compress the story — explain our laws more briefly than stating them outright — the simulation hypothesis carries strictly more ontology. The axiom overrides that accounting. It asserts that generative realizability takes precedence over layer count: if the laws can be implemented by something deeper, read them as implemented. That is a coherent stance. It is also a metaphysical selection rule, and it should be named as one.

Before building on it, two corrections to the folk framing.

“Simulated” and “random” are not alternatives. The pop debate — “is the universe simulated, or just random?” — is a category error. Every combination is possible:

Deterministic dynamics Stochastic dynamics Base universe Possible Possible Simulated universe Possible Possible

Unitary quantum evolution is deterministic; measurement outcomes follow probabilistic Born-rule statistics; a simulator could implement genuine host-level randomness or an observationally equivalent pseudorandom process. The live contrast is simulated versus ontologically primitive — implemented by something deeper, versus simply there.

And the logic must be displayed honestly. Let Q be “observed physics is described by QFT and gravity”; C be “those dynamics can emerge from quantum-information processing”; S be “the universe is executed on a deeper substrate.” The five pillars support Q → C. The axiom supplies C → S. Together they yield Q → S. The inference is valid — but its conclusion flows entirely from the second premise. Physics gets you to can. The axiom converts can into is.

With that stipulation made and priced, we are entitled to build. What follows is the cosmology you get if you adopt the axiom — a reading of established physics, not a new theory.

The cosmology

The full hierarchy, from substrate to experience:

fundamental quantum-information substrate
                 │
      local quantum update rules
                 │
entanglement and large-scale field behavior
          ┌──────┴──────┐
   emergent spacetime   quantum fields
     and gravity              │
          │         particles as excitations
          └──────┬──────┘
   atoms, stars, planets, organisms
                 │
   observers and measuring instruments
                 │
classical-looking records and experiences

Each level redescribes the one below, the way temperature redescribes molecular motion. Walk it from the bottom.

The substrate. The most conservative candidate is a local quantum-information system — a quantum cellular automaton, quantum circuit, or tensor network (a structured web of small quantum systems). The D’Ariano–Perinotti results show that unitarity, locality, homogeneity, and isotropy at this level can yield the Dirac and Weyl equations at large scales. Under the axiom, the automaton is fundamental and continuum field theory is its effective, coarse-grained description: substrate update rules → effective fields → particles and classical objects.

The vacuum. The vacuum is the substrate’s reference state — its lowest-energy encoded configuration. Not an absence of computation; not a blank screen awaiting content. It carries the correlations and entanglement we catalogued earlier, and excitations are defined above it: a†|0⟩ = |1⟩, the creation of one quantum. When geometry changes — an expanding universe — the two decompositions of the encoded field disagree about which patterns count as particles, exactly per the Bogoliubov mixing, and particles appear with ⟨N⟩ ∝ |β|². Under the simulation reading, this is not the host allocating particles when observers look. The substrate follows its update rule; the encoded geometry shifts; what counts as an excitation shifts with it. No consciousness required, no rendering on demand.

Space. Under the axiom, the holographic results are promoted from duality to architecture: the lower-dimensional quantum theory is the implementation layer, and the higher-dimensional gravitational world is the environment it encodes. Distance, then, is not fundamental. Two objects are “near” because their underlying quantum information is strongly connected; weakly connected information appears far apart. This gives “simulated space” a precise, non-science-fiction meaning: fundamental quantum state → entanglement relations → effective geometry. And it does not make space an illusion. Temperature is emergent from molecules, and temperature is real. Emergent reality is reality at its own level.

Gravity. General relativity says matter tells spacetime how to curve and curved spacetime tells matter how to move. In this cosmology, that geometry is reconstructed from the substrate’s entanglement — and the Faulkner result shows the reconstruction is not free-form: consistency between entanglement and energy is the linearized Einstein equations. So gravity is not a force separately programmed into the simulation. It is the large-scale consistency condition on how information can be encoded. A planet orbiting its star follows the straightest available path through curved spacetime; at substrate level, that same motion is the lawful evolution of encoded information. The two descriptions relate as water does to its molecules — a flowing liquid at one scale, a swarm of interactions at another, both correct.

Matter. The error-correction structure explains one of the deepest features of experience: stability. In a quantum code, one logical bit is spread redundantly across many physical components, and damage to a few leaves the message intact. If a particle in emergent spacetime is likewise a logical excitation — represented redundantly across many substrate degrees of freedom — then there need be no single substrate cell that “is” the electron. An electron is a protected pattern, not a bead at an address. Hence stable particles despite microscopic churn, robust geometry despite quantum fluctuation, and information reconstructible along multiple routes.

Measurement. Nothing in this picture waits for an observer. An observer is simply another encoded subsystem. A measurement is an ordinary interaction: a system in superposition couples to an apparatus,

|ψ⟩|A₀⟩ → c₁|1⟩|A₁⟩ + c₂|2⟩|A₂⟩,

and then to the air, the stray light, the room — a spreading web of correlation. This is decoherence, and it makes interference between large-scale alternatives astronomically hard to observe, which is why the world looks classical. The host computes nothing “on demand”; the entire coupled state evolves continuously. The defensible analogy is not rendering but decoding: observation extracts and records information already evolving in the system. It does not summon reality into being.

The Big Bang. The Big Bang marks the earliest recoverable state of the emergent spacetime — not necessarily the moment a machine was switched on. A possible history: the substrate begins in a particular quantum state; its internal relations generate a geometry; the geometry evolves rapidly; changing geometry excites the fields (Parker’s mechanism, now read as re-partitioning of encoded excitations); the universe cools; quarks bind into protons; nuclei, atoms, stars, galaxies, and eventually observers condense. Ordinary cosmology survives untouched at the emergent level — the interpretation changes what the fields and spacetime ultimately are, not their equations. One humbling corollary: cosmic time is an emergent variable, so nothing licenses the inference that host time began when ours did. The host might have a different time, or none resembling ours; it might have existed indefinitely; it might hold our entire four-dimensional history as a single mathematical object.

Randomness. Quantum theory predicts probabilities; it generally does not predict individual outcomes. A substrate could implement this several ways: genuine host-level randomness; deterministic hidden dynamics; selection among branches; or a fully branching quantum state with no selection at all. Present experiments do not distinguish these by observing randomness alone. Randomness is therefore evidence neither against simulation nor for it — a symmetry the folk debate persistently misses.

How you could catch it

A simulation hypothesis earns scientific standing exactly when it specifies its architecture — because only architecture leaves fingerprints.

A finite cubic lattice, for instance, predicts a maximum momentum; modified motion at extreme energies; and preferred directions in the sky, aligned with the lattice axes, violating perfect rotational symmetry. Beane, Davoudi, and Savage derived such signatures for their Wilson-lattice model, with ultra-high-energy cosmic rays as the natural detector — the most energetic probes nature provides. (Beane, Davoudi, and Savage)

But mind the asymmetry in what a null result buys. Seeing no lattice artifacts rules out that architecture at that scale — not a finer lattice, a non-cubic network, an automaton preserving the relevant symmetries, a holographic implementation, or an exact continuum simulation. A simulator that reproduces continuum QFT perfectly is empirically indistinguishable from an ontologically primitive universe — which is precisely why, in this cosmology, the final step is carried by the axiom and not by data. The construction is honest about that; it is the genre’s usual bait-and-switch, refused.

What it explains, and what it cannot

As an interpretation, the simulation cosmology unifies a striking list of facts that otherwise sit as separate surprises: that particles are excitations rather than permanent little objects; that empty space has structure; that the definition of a particle can depend on motion and geometry; that spacetime plausibly emerges from entanglement; that gravity follows from consistency conditions on information; that matter is stable in the way redundantly encoded information is stable.

And here is what it does not explain: what the substrate is made of; why it has its particular laws; who, if anyone, initiated it; why this initial state; why conscious experience exists; whether the host is itself simulated. That last question exposes the deepest limit — an explanatory regress:

our universe ← host ← host of the host ← ⋯

At some level, either reality is ontologically primitive after all, or the chain runs forever, or the explanatory structure closes in some way we have not conceived. Simulation relocates the question of ultimate existence. It does not answer it.

Simulated, rather than primitive

Granting the axiom, reality has three levels. At the fundamental level, quantum information evolves by local, lawful rules. At the physical level, entanglement produces effective spacetime, fields, particles, and gravity. At the experienced level, decoherence, chemistry, and biology produce the stable classical world in which observers keep records. The thesis, in full:

The universe we experience is a real emergent world implemented by a deeper quantum-information system. The vacuum is its structured reference state; particles are protected excitations; space is reconstructed from quantum relationships; gravity is the consistency condition on that reconstruction; and observers are internal encoded subsystems — not external spectators, and not triggers.

Every mechanism in that sentence comes from serious physics: lattice gauge theory, quantum automata, holography, entanglement geometry, quantum error correction, cosmological particle production. The one word that does not is implemented — that word is purchased, at full price, by the simulation-selection axiom.

So reject the folk conclusion “simulated rather than random”; the dichotomy was never real. The honest name for the position is simulated rather than ontologically primitive — a universe that is run, rather than merely there. The physics holds both doors open and will not choose for you. The axiom walks through one of them.

At least now you can see the door.

Sources: Parker (1969) · Schwinger (1951) · Unruh (1976) · Wilson (1974) · Beane, Davoudi & Savage (2012) · D’Ariano & Perinotti (2013), (2016) · Susskind (1994) · Maldacena (1997) · Ryu & Takayanagi (2006) · Faulkner et al. (2013) · Almheiri, Dong & Harlow (2014) · Lloyd (2000) · Bostrom (2003)

A case study in what happens when a non-technical operator is given unchecked authority over the e-commerce department of a luxury retailer, and the organization’s immune system does the rest.

The Setup

A premium accessories brand—the kind you see in airport terminals and aspirational gift guides, owned by a global luggage company—runs its DTC e-commerce operation on Salesforce Commerce Cloud. Annual digital revenue is north of $100 million. The platform processes tens of millions of transactions per year across multiple geographies. It is, by any reasonable measure, critical infrastructure for a brand whose retail footprint is increasingly digital.

The department responsible for this platform has no internal engineering. Not “limited” internal engineering. None. Every line of code, every architectural decision, every deployment, every integration is executed by a single systems integrator — one of the large global SIs with a substantial Salesforce Commerce Cloud practice. Five engineers on the account, billing at rates commensurate with a top-tier consultancy. The brand pays for this team as an ongoing engagement, not project-by-project, so the meter is always running.

The SVP of Global Marketing, the P&L owner of the e-commerce department, has organizational oversight of the e-commerce function. The SVP’s background is in brand marketing and retail. She has no technical depth in e-commerce platform engineering and, more critically, has shown no inclination to acquire it or to ensure that someone in the reporting chain possesses it.

She then hires am e-comm director that is best understood as a commercially ambitious operator whose confidence and appetite for authority significantly exceed his demonstrated experience.

He was essentially was promoted from a merchandising manager and business analyst role at a smaller fashion brand. His background is in business operations and platform administration. He can navigate Business Manager—Salesforce Commerce Cloud’s back-office interface—configure promotions, manage content slots, and evaluate whether a feature appears to function when he opens it in a browser. He is, in the most precise and non-pejorative sense of the term, a power user. He is not an engineer. He cannot read code, evaluate architectural decisions, assess vendor deliverable quality, or identify compliance risk in a payment integration.

He presents extremely well. In interviews, executive conversations, and stakeholder interactions he is polished, responsive, energetic, and rarely appears unprepared. He projects certainty, speaks the language of leadership, and instinctively understands how to build relationships upward in the organization. Senior leaders often leave conversations feeling reassured because he has an answer for every question and conveys a strong sense of commitment and availability.

His strengths appear concentrated around coordination, presentation, merchandising-oriented commerce operations, and managing perceptions. He is comfortable discussing e-commerce concepts, platform functionality, customer experience, and business priorities. Having spent substantial time as an end user of Salesforce Commerce Cloud, he understands how commerce teams interact with the platform and can speak fluently about day-to-day digital business operations.

The tension emerges when the discussion moves from commercial ownership into domains requiring deep technical judgment, architecture governance, organizational leadership, or executive-level accountability.

Rather than relying on specialists to extend his capabilities, the newly-hired director centralizes decision-making around himself. He treats expertise less like a strategic partner and more as a service function whose primary responsibility is execution.

In meetings, he tends to occupy the center of the conversation. He prefers being the primary interface to leadership, the SVP, vendors, and key stakeholders. Information flows through him rather than around him. Technical experts are consulted selectively and often after directional decisions have already been formed. As a result, expertise becomes reactive rather than influential.

His management style is highly control-oriented. He seeks visibility into everything, wants involvement in most decisions, and may view independent centers of authority with suspicion. Senior specialists—particularly those whose expertise exceeds his own—can find themselves marginalized, excluded from key discussions, or asked to validate decisions rather than shape them. One-on-one meetings with his direct-reports become status-reporting exercises instead of substantive strategic dialogues.

But he is ambitious, putatively (if not performatively) hardworking, politically aware, and highly motivated to succeed. The challenge is that his model of leadership appears rooted in personal control rather than institutional capability. He gains confidence when decisions depend on him and loses comfort when authority is distributed among experts.

The organization becomes dependent on his ability to personally mediate competing interests rather than creating durable decision structures that function without him.

The narrative arc of such a leader is often predictable. Early performance can appear quite strong because stakeholders experience responsiveness, decisiveness, and high visibility. Features ship, meetings occur, vendors receive direction, and the organization feels active. The longer-term outcome is more bleak: the platform and department gradually accumulates technical debt, vendor dependence, decision bottlenecks, and specialist turnover while maintaining the appearance of strong leadership from the outside.

In short, he resembles a leader who is highly skilled at managing upward, highly motivated to lead, and highly protective of decision authority, but who lacks the executive maturity required to consistently leverage expertise he does not personally possess. The central question is not whether he can make decisions. It is whether he can build an environment in which the best decisions are made even when he is not the most knowledgeable person in the room.

Overseeing this operation, then, is a director

This is the structure. A non-technical director with sole authority over a technical function, reporting to a non-technical SVP who doesn’t monitor the function’s operational health, with a single vendor that owns 100% of the engineering output and has no client-side technical counterparty to evaluate its work.

What follows is a detailed examination of what this structure produces, what it costs, and why the organization’s own immune response ensured that every attempt to fix it was neutralized from the inside.

What Has to Be True for a Technical Hire to Survive

Before examining what happened when this brand attempted to introduce technical oversight, it’s worth establishing the structural preconditions that must be in place for a technical leadership hire to function in an environment like this one. These conditions are not optional enhancements. They are the minimum viable operating environment for the role. Without them, the hire will be neutralized regardless of their competence.

The technical lead must report outside the non-technical director’s chain. If the architect reports to the person whose competency gap the architect exists to cover, the outcome is predetermined. The director controls the architect’s performance reviews, meeting access, project assignments, and ultimately their employment. An architect who surfaces findings that implicitly reveal the director’s limitations — which is the entire point of the role — is structurally dependent on the person most threatened by those findings. The reporting line must go to a CTO, a VP of digital operations, or directly to the executive suite. Someone with no incentive to suppress technical findings and enough organizational authority to back the architect when conflicts arise.

The technical lead must own the release gate. Not advise on it. Not review it after the fact. Own it. Nothing ships to production without his sign-off or the sign-off of someone he’s delegated authority to. The release gate is the single highest-leverage control point in a software delivery organization. It is the mechanism that determines what code reaches customers. In a vendor-dependent model, it is also the primary quality check on vendor output. If the release gate is held by anyone other than the technical lead — a project manager, a department admin, the director himself — then the technical lead’s authority is cosmetic: he can identify problems but cannot prevent them from shipping.

The organizational environment must support the role before the hire is made. A technical lead inserted into a hostile organizational environment — a director who views the role as a threat, a PM who resists changes to the release process, an SVP who doesn’t understand what the role does — will be systematically undermined through a thousand small frictions: exclusion from roadmap discussions, lack of access to vendor contracts, deferral to the PM on process decisions, absence of executive backup when the architect pushes back on a vendor deliverable. Each friction is individually minor and deniable. Collectively, they render the role inert. The leadership and process changes that create a survivable environment for the technical hire must precede the hire itself.

These are not theoretical conditions extracted from a management textbook. They are lessons extracted from what actually happened at this brand when it tried to make this exact hire.

The Immune Response

Under a previous, more technically oriented department head, the brand hired a E-commerce architect. The role was defined correctly: platform architecture oversight, vendor deliverable quality, performance engineering, code review, and technical governance of the SI engagement.

The architect was competent. This is not conjecture — after leaving this brand, he was hired into a commerce engineering role at a well-known consumer technology company, where he now builds commerce platform features. The market had no difficulty recognizing the value that the accessories brand’s director could not.

What happened between the hire and the departure is a textbook case of organizational antibody rejection.

The architect was given no authority over vendor decisions. The SI relationship — the single most important operational relationship in the department — remained under the director’s control. The architect could observe SI output but could not direct it, reject it, or set standards for it.

Release management — the gate that determines what ships to production — was held not by the architect, not by the director, but by a department administrator who had assumed the role through proximity and tenure rather than competency. When the architect attempted to exercise basic quality control over the release process — including something as minor as standardizing JIRA ticket naming conventions — the admin pushed back, and the director sided with the admin. The architect was being overruled on release process by someone whose qualification for the function was being in the room.

The architect was excluded from roadmap discussions. For months, the person hired to provide technical direction for the platform was not included in a single conversation about what the platform should do next. The roadmap was being set by the director, informed by a junior merchandiser-turned-product-manager whose entire technical discovery process consisted of asking the SI vendor’s architect what was possible — meaning the vendor was simultaneously defining the solution space and billing for the implementation.

The previous technical director, who had hired the architect, failed to back him. Whether this was a competency failure (not understanding that the role required delegated authority to function) or a courage failure (understanding but choosing organizational comfort over functional effectiveness), the result was the same. The architect was structurally isolated: no authority, no executive cover, no release gate ownership, no roadmap input.

Then the technical director left for a role at the parent company. The SVP described the departure as the technical director "spreading his wings" — celebrating the exit of the department's only technical leadership as a career growth moment for the person leaving, with no apparent concern for what it meant for the function he was leaving behind. With the technical director gone, the architect lost even the theoretical air cover that the reporting line provided.

The non-technical director, now holding sole authority, moved to terminate the architect. The stated justification included RTO compliance — the architect had been leaving the office an hour or two early on some days, after working nights and weekends from home. This is a director building a paper trail, not managing performance. You don’t build a disciplinary case around office departure times for someone who’s putting in nights and weekends unless you’ve already decided to terminate and need documentation that HR will accept.

The architect was fired. The performance engineering program he was running — Core Web Vitals optimization, JavaScript bundle analysis, critical rendering path improvements, CDN configuration, image optimization pipeline, Lighthouse scoring targets with a regression budget — stopped. Not transitioned. Not reassigned. Stopped.

One month later, the brand posted a job requisition for the same role.

What Walked Out the Door

To understand the cost of the architect’s termination, you have to understand what he was working on and why it mattered.

The performance engineering workstream the architect was running is not exotic or aspirational work. It is the standard, well-understood set of optimizations that every competent e-commerce platform team executes as ongoing operational discipline. The specific initiatives:

Core Web Vitals optimization. Google’s Core Web Vitals — Largest Contentful Paint (LCP), Interaction to Next Paint (INP), and Cumulative Layout Shift (CLS) — are both a search ranking signal and a direct proxy for user experience. Improving these metrics is one of the highest-ROI activities in e-commerce engineering because the benefits compound: better search visibility drives more traffic, faster load times convert that traffic at a higher rate, and the improvements are durable once landed.

JavaScript bundle analysis and optimization. Salesforce Commerce Cloud storefronts, particularly those built on the SFRA (Storefront Reference Architecture) framework, accumulate JavaScript bloat over time as integrations, third-party tags, and custom functionality are layered onto the base. Bundle analysis identifies what’s being loaded, what’s blocking rendering, and what can be deferred, split, or eliminated. On a site processing the traffic volumes this brand handles, trimming even 50KB from the critical JavaScript bundle produces measurable load time improvements across millions of sessions.

Critical rendering path work. The sequence of resources the browser must load before it can display meaningful content to the user. Optimizing this path — inlining critical CSS, deferring non-critical resources, optimizing font loading — directly reduces the time between a customer clicking a link and seeing a usable page. Every 100 milliseconds of improvement in this metric has a documented, published correlation with conversion lift. Google, Deloitte, and Akamai have all published research establishing this relationship across large e-commerce datasets.

Lighthouse scoring targets with a regression budget. Lighthouse is Google’s automated auditing tool for web page quality. Setting scoring targets and establishing a regression budget — a defined threshold below which scores cannot fall without triggering remediation — creates an automated quality gate that prevents performance from degrading as new features are added. Without this, every feature deployment risks introducing performance regressions that erode the gains from previous optimization work. The regression budget is what makes performance engineering sustainable rather than a one-time effort that decays.

CDN tuning and image optimization pipeline. Content delivery network configuration and image serving optimization are infrastructure-level improvements that reduce bandwidth consumption and load times globally, with disproportionate impact on mobile users — who represent the majority of traffic for most consumer brands.

This is the work that was in progress. Not proposed. Not roadmapped. In progress. The architect was actively executing these initiatives, with the technical knowledge to do them properly and the SFCC-specific expertise to navigate the platform’s constraints and extension points.

The director terminated the architect and replaced the entire performance engineering discipline with a Noibu license — a client-side error monitoring tool that detects issues but doesn’t fix them. The performance engineering program was replaced with a dashboard warning light.

The Conversion Gap in Dollars

The brand’s e-commerce conversion rate sits at approximately 1.5–2.5%, against a category benchmark for premium accessories and lifestyle brands of 3–4% on well-optimized sites.

On annual digital revenue of approximately $100 million, the arithmetic is straightforward. Every 50 basis points of conversion improvement on the same traffic volume represents roughly $20–25 million in incremental annual revenue. The full gap to benchmark represents $30–50 million in unrealized annual revenue.

The $30–50 million figure is the theoretical ceiling, and it’s fair to argue that closing the full gap is unrealistic for any single set of initiatives. Fine. Cut it by 90%. A 20–30 basis point improvement — the kind of lift that a sustained performance engineering program demonstrably produces, based on published data from Google, Deloitte, and Akamai spanning thousands of e-commerce sites — represents $2–3 million in incremental annual revenue.

The architect’s fully loaded cost — salary, benefits, overhead — was approximately $200–250K annually. The performance engineering work he was executing is the highest-ROI conversion lever available to an e-commerce platform team: the improvements are measurable, the impact curves are well-documented, and the work compounds over time because faster pages convert better on every visit, not just the first one.

The architect’s annual cost would have been recovered in the first quarter from conversion lift alone, assuming even the most conservative performance-to-conversion impact estimates in the published literature.

Instead, the brand is paying $600K+ per year to an SI vendor that is not doing performance engineering work, is not optimizing conversion, and — as the codebase analysis reveals — is producing output that a competent architect would catch and remediate at the PR review stage.

The performance engineering program wasn’t speculative. It wasn’t theoretical. It was in flight, it was targeting the right metrics, and it was being executed by someone with the platform-specific expertise to deliver. It was terminated by a director who saw Lighthouse scores as “staying in your lane” and replaced it with an error monitoring tool and a product quiz.

A Million Here, a Million There

The conversion opportunity cost is the largest number, but it’s not the only number. The director’s 14 months of sole authority have produced a set of direct costs, wasteful expenditures, and unrealized returns that, individually, each represent a manageable line item. Collectively, they exceed the department’s own operating cost.

Vendor rework waste. A detailed commit log analysis of the platform’s primary release branch over a three-month period reveals systematic rework patterns. A search feature (”Did You Mean” suggestions) required 8+ commits to land, including one with a typo in the feature string itself. A cartridge removal task — fundamentally a deletion exercise — took 10+ commits. An Estimated Delivery Date feature consumed 20 commits across 5 tickets over three months: engineers added a Google Geocoder API integration, debugged it across 4 commits, then removed it entirely; added a server-side cache, then removed it. The surviving output was a browser localStorage cache with no expiration policy and an API retry for ambiguous postal codes. An independent consultant who reviewed the EDD work described it as “churn masquerading as feature work.” Conservatively, the rework waste in the SI billing over a single quarter is $150–200K.

Monitoring tool triage premium. The brand installed Noibu, a client-side error monitoring tool. Noibu generates issue alerts. There is no one on the brand side who triages these alerts — the brand has no internal engineering staff. Every Noibu alert becomes a ticket routed to the SI vendor, who triages it at SI billing rates. The brand is paying premium consultancy rates for tier-1 bug triage generated by a tool it installed. This is a cost generator, not a capability. Annual premium over what an internal resource or a junior offshore triage function would cost: $50–75K.

Payment integration compliance exposure. A CyberSource defect fix that took three months to implement — eliminating the “we were under time pressure” defense — resolved a token-loss bug by persisting a transient payment token at rest on the order object in the platform database. The token, designed by CyberSource to be ephemeral and single-use with a 15-minute validity window, was stored as an unencrypted, indefinitely-retained custom attribute visible and editable in Business Manager. No encryption, no time-to-live, no post-authorization cleanup. An independent consultant characterized it as “a band-aid that adds compliance risk.” Under PCI DSS v4.0 Requirements 3.3 and 3.4, data capable of initiating a transaction must be rendered unreadable at rest or disposed of after authorization. This fix does neither. Additionally, the 3D Secure / Payer Authentication code paths — the exact flows that trigger when a cardholder’s bank requires additional verification — were never updated with the same fix, meaning the original defect still exists for challenged transactions. The remediation cost when this is eventually surfaced — whether by audit or by incident — is $100–200K, scaling significantly if it’s the latter.

Payment cartridge technical debt. The SI vendor touched the CyberSource payment integration multiple times during the director’s tenure, including the defective fix described above. None of these touches modernized the cartridge to CyberSource’s current REST-based integration and Unified Checkout architecture. Each touch added customization on top of an already-outdated foundation, increasing the delta between what the brand has and what the platform vendor supports. The eventual migration cost grows with every quarter of deferred modernization. Current estimated cost: $200–300K and rising.

The product quiz. The department’s flagship feature over 14 months of the director’s sole authority: a product finder quiz. Fifteen-plus commits to land. Proportional allocation of SI billing time: approximately $300K. A product quiz on a luggage site is not a high-leverage conversion play. It is the kind of feature that presents well in a stakeholder meeting and produces negligible lift in an A/B test — assuming anyone is running the A/B test, which, as we’ll see, they are not.

Opportunity cost of 14 months without performance engineering. The architect’s performance program was terminated upon his departure. No replacement program was initiated. No alternative vendor was engaged. The work simply stopped. Fourteen months of foregone performance optimization on a $100M+ digital revenue platform, using even the most conservative published performance-to-conversion impact estimates: $2–3M annually in unrealized revenue.

Total the direct waste and the conservative conversion opportunity cost and you arrive at a figure north of $3 million annually — against a department that costs $800K–$1M per year to run, inclusive of the SI engagement. The department’s fully loaded cost is being exceeded by the cost of its own dysfunction. It is, in a precise financial sense, more expensive to run badly than it would cost to run well.

The Blunder Nobody Taught Him to Avoid

The architect’s termination was a strategic blunder, and the nature of the blunder reveals something about the director that goes beyond the technical competency gap.

A competent political operator in the director’s position — even one who couldn’t read a line of code — would have recognized the architect as the most valuable piece on his organizational board. You don’t have to understand SFCC architecture to understand organizational risk distribution. Any director who has survived a few cycles in corporate life knows the foundational move: keep the subject matter expert close, give them enough room to own the technical narrative, and position yourself as the person who brought in the right talent and empowered them to deliver. When things go well, you take credit for the hire. When things go wrong, you have a technical layer between you and the failure. The architect does the explaining. You do the “I’m holding my team accountable” performance. This is baseline self-interested management — not even good management, just the minimum viable Machiavelli.

The director didn’t do that. He saw the architect as a threat and eliminated him, which means he couldn’t even run the cynical version of the playbook correctly. He couldn’t see one move ahead. A more experienced operator — even a mediocre one, even one acting purely out of self-interest with no concern for the platform — would have kept the architect, constrained his visibility to senior leadership, taken credit for his wins, and used him as a blast shield when things went sideways.

What the director did instead was the move of someone who has never been in a position where accountability finds you. He’s managed in environments where the feedback loop is short and cosmetic — did the promo go live, did the page update, did the ticket close. In that world, the only threat is someone who contradicts you in a meeting. He’s never been through an audit, a breach investigation, a board-level review of engineering delivery quality, or any situation where someone with authority and technical literacy looks under the hood. So he doesn’t model for it. The threat he can see — the architect surfacing his limitations in a standup — is real to him. The threat he can’t see — standing alone in front of findings he can’t explain, with no technical subordinate to absorb accountability — doesn’t exist in his mental model because he’s never experienced it.

The blunder reveals that the director is not just non-technical. He is a poor organizational strategist. He made a decision that was bad for the company and bad for himself, and he didn’t recognize it as either. He traded a visible, manageable internal tension (an architect who might surface his limitations in a meeting) for an invisible, catastrophic external exposure (sole accountability for a platform he can’t technically evaluate, with no one to share the blame). That trade only works in an environment where no one above him ever looks.

The architect, had he still been on staff, would have been the director’s greatest asset when scrutiny arrived. The person who could walk into the room, explain the findings, present a remediation plan, and absorb the technical accountability. The director would have been able to say, “I’m aware of these issues, my architect has been flagging them, here’s our plan.” Instead, when the questions come — and they will come — the director will be standing alone in front of findings he can’t parse, with a vendor who will disclaim responsibility (”we delivered what was scoped and accepted by the client”), and a commit log that speaks for itself.

He optimized for the weekly standup and left himself defenseless against the quarterly review.

And then, one month later, he posted the job requisition for the same role.

That detail alone tells the entire story. If the architect was terminated for performance, you don’t re-open the role in 30 days. If the role wasn’t needed, you definitely don’t re-open it in 30 days. The only scenario where you fire someone and immediately re-post their position is when the termination was personal, not functional. You needed the role. You didn’t want the person. And the reason you didn’t want the person — their expertise, their willingness to push back, their ability to make the director’s limitations visible on a vendor call — was the reason the person was valuable.

The requisition is a confession formatted as a job listing.

The Loudest Voice in the Room

The director is not a passive, delegating manager who trusted the SI vendor and got burned by their quality. That would be a different failure mode — negligent, but at least structurally comprehensible. The director is on every vendor call. He does most of the talking for the brand side. He overrides UX subject matter experts. He overrides data analysts. He interrupts domain specialists to ensure his opinion prevails.

This matters because it eliminates the one defense a non-technical director could plausibly offer: “I relied on the experts and they let me down.” He didn’t rely on the experts. He overruled them. When the UX SME had a better interaction pattern for a checkout flow, the director interrupted and redirected the conversation. When the data analyst flagged a concern about an analytics implementation, the director talked over them. The SI vendor’s engineers, who are billing hourly and have no incentive to fight the client, learned to build whatever the loudest voice in the room described.

The result is a platform that is precisely calibrated to the director’s level of understanding. The product quiz gets prioritized because the director understands product quizzes. The CyberSource cartridge doesn’t get modernized because the director doesn’t understand what that means or why it matters. The EDD feature churns through five tickets across three months because the director can’t evaluate the implementation approach and can’t distinguish between progress and rework. The features that ship are the features the director can conceive of, scoped the way he describes them on calls, built to the standard he’s capable of evaluating — which is: does it look right in a browser?

This also reframes the architect’s termination with additional clarity. The architect wasn’t just a threat to the director’s positional legitimacy on an org chart. He was a threat to the director’s control of the room. A Salesforce Commerce Cloud architect on a vendor call with SI engineers can engage at a peer technical level — challenging implementation approaches, proposing alternatives, evaluating trade-offs in real time. The director cannot do any of that, and he cannot override an architect the way he can override a UX designer or a data analyst, because the architect can push back with technical authority the director can’t counter. The architect’s presence on those calls would have exposed the director’s lack of technical depth in real time, in front of the vendor, on every call.

Removing the architect wasn’t just about the org chart. It was about maintaining dominance in the room where decisions are made.

The behavioral pattern has a name in organizational psychology: the competence-threatening authority response. A manager whose authority exceeds their competence in a domain will systematically suppress, override, or remove the people whose expertise makes the competency gap visible — not necessarily out of malice, but because the gap between their authority and their understanding creates a constant status threat that they resolve by controlling the information environment. Every call where a subject matter expert offers a better answer is a moment where the director’s limitations are implicitly surfaced. Interrupting and overriding isn’t a personality quirk. It’s a governance-level failure mode where the person holding decision rights is actively degrading decision quality to protect their own position in the room.

This is the mechanism by which money turns into waste. The commit log shows the output. The feature inventory shows the poverty of the roadmap. But the vendor call dynamic shows how it happens — how a $600K+ annual SI engagement produces a product quiz and a PCI violation. The answer is that every decision passes through a single point of failure: a non-technical director who insists on being the loudest voice on every call, overrides the domain experts who could improve outcomes, and has removed the one technical peer who could have checked his judgment. The vendor builds what he describes. What he describes is limited by what he understands. What he understands is merchandising configuration, not platform engineering.

The Potemkin Conversion Program

The director, if confronted with the absence of conversion-impacting work over 14 months, would cite three things: Noibu for site performance, a co-branded payment option for payment expansion, and Monetate for A/B testing. Let’s take each seriously and then examine what’s behind the label.

Noibu is not a site performance program. It’s a ticket generator. Noibu is a client-side error monitoring tool. It detects JavaScript errors, session replay anomalies, and revenue-impacting bugs. It is a diagnostic instrument — the equivalent of installing a dashboard warning light in a car. A warning light is not an engine repair program. Noibu generates alerts. Someone has to triage those alerts, determine root cause, prioritize remediation, and fix the underlying issues. There is no one at this brand who does this. The alerts become vendor tickets, triaged at SI billing rates, producing a cost loop rather than a performance improvement cycle. A site performance program is Core Web Vitals optimization, JavaScript bundle analysis, critical rendering path work, CDN tuning, image pipeline optimization, Lighthouse scoring targets with a regression budget. The architect was running exactly that program. The director fired the architect and bought Noibu. He replaced the mechanic with the check engine light and called it an upgrade.

A co-branded financing option is not a payment expansion strategy. It’s a single ticket. Adding one payment method is a line item, not a strategy. A payment method expansion program starts with checkout funnel analysis: where are carts abandoning, what’s the payment method mix by geography and device, what’s the decline rate by issuer, where is 3D Secure friction killing conversion, what’s the adoption curve on digital wallets, are buy-now-pay-later integrations pulling their weight relative to integration cost. A co-branded financing option serves a narrow customer segment. It’s fine as one item in a payments roadmap. Cited as the payment expansion initiative, it reveals that no payments roadmap exists. Meanwhile, the primary payment rail — the CyberSource integration that processes the vast majority of the brand’s transactions — is running on an outdated cartridge with a PCI compliance exposure in production and broken 3D Secure authentication paths. The director added a niche payment option while the core payment infrastructure is compromised. That’s not strategy. That’s feature-list padding.

Monetate is not an A/B testing discipline. It’s a SaaS license. Having Monetate installed is the equivalent of having a gym membership. The question is: what’s the testing velocity? What’s the win rate? What hypotheses are being tested? What’s the cumulative conversion lift from the testing program over the past 12 months? Who is designing the experiments? Who is analyzing the results? Who is maintaining a hypothesis backlog informed by analytics and session data? A mature optimization program runs 2–4 concurrent experiments, maintains statistical rigor standards for declaring winners, and produces a compounding conversion lift over time that is tracked and reported quarterly. If those artifacts don’t exist — the experiment log, the win/loss record, the cumulative measured lift — then Monetate is a line item on an invoice, not a testing program.

This is the director’s pattern in its most revealing form. He acquires tools and integrations that represent capabilities without building the operational muscle that makes them functional. Noibu without a triage function. A financing option without a payments strategy. Monetate without a testing discipline. Each one looks like a line item on a roadmap slide. Each one, under any informed examination, is a label with nothing behind it.

The pattern has a root cause: the director’s mental model of value creation is that of a merchandising operator, not an engineering or product leader. In merchandising, the tool often is the capability. You configure a promotion in Business Manager and it runs. You set up a content slot and it displays. The tool does the thing. The director is applying this mental model to disciplines — performance engineering, payments optimization, experimentation — where the tool is the starting point, not the endpoint. Installation is the beginning of the work, not its completion. The triage function, the analysis discipline, the experimentation rigor, the performance optimization loop — these are the work. The tool is just the instrument. The director doesn’t see the gap because, in every domain he’s worked in before, there was no gap. The tool was the deliverable.

This is why his competitive framing centers on visible features rather than platform capability. His idea of competitive differentiation against rival brands is a product quiz — a visible, demonstrable, screenshot-ready feature that he can present in a meeting and compare against what competitors have on their sites. The actual competitive advantage — a site that loads faster, converts better, and processes payments more reliably than every competitor in the category — is invisible. You can’t demo it. You can’t put it in a deck. But it shows up in the only place that matters: revenue per session. The architect understood this. The director saw a guy fiddling with Lighthouse scores when there was a quiz to build.

Five Days to the Finding

A technical consultant retained by the incoming president to establish a baseline assessment of the e-commerce operation would not need months, or even weeks, to build this case. A competent e-commerce consultant with Salesforce Commerce Cloud experience could surface the complete picture in five business days. Most of that is writing the report, not finding the findings.

Day one is the commit log. Clone the repository, run a commit frequency analysis, categorize tickets by type, identify the contributor list. The single-vendor ownership, the rework patterns, the commit message hygiene issues, and the maintenance-mode velocity profile are all visible from basic git log analysis. A 30-minute review session — literally half an hour with the commit history — is enough to identify the pattern. The consultant will spend a full day being thorough, but the signal is immediate.

Day two is the risk deep-dive. Payments is the obvious first target on any SFCC audit — the consultant goes straight to the CyberSource cartridge, checks the version, examines customizations, traces the token lifecycle. The PCI finding surfaces in hours. The incomplete 3D Secure paths surface in the same session. The cartridge version gap is visible from the manifest. This is a one-day exercise that produces the single most urgent finding in the report.

Day three is the feature-to-cost reconciliation. Pull the JIRA board or project management tool, map tickets to commits, identify what shipped versus what churned. The EDD round-trip, the search feature rework, the product quiz cost profile — all of this falls out of straightforward ticket-to-commit mapping. Cross-reference against the SI vendor’s statement of work or rate card, and you have a cost-per-feature analysis that will make the president reconsider the entire engagement.

Day four is architecture and platform health. SFCC version currency, cartridge dependency map, integration inventory, performance baseline. Core Web Vitals data from Chrome User Experience Report (CrUX) is publicly available — the consultant can pull the brand’s actual field performance data and see whether there’s a measurable degradation that correlates with the architect’s departure. If there is, that’s not an inference. It’s a data point with a timestamp that maps to a personnel decision.

Day five is the report.

The director’s entire body of work over 14 months of sole authority does not survive a week of professional scrutiny. And the director built it that way — he removed the person who could have made the findings more complex or ambiguous, centralized every decision through himself, and left a clean, unobstructed line of accountability from every call, every ticket, every commit back to his desk.

The consultant will also surface the Noibu / Monetate / financing-option pattern — the Potemkin conversion program — simply by asking the second question. The first question — “Do you have site performance monitoring? Do you have A/B testing? Have you expanded payment methods?” — yields yes, yes, yes. Any non-technical reviewer would stop there. The consultant asks the second question: “What’s the Noibu triage SLA? What’s the Monetate experiment velocity and cumulative lift? What’s the financing option’s adoption rate and contribution to overall conversion?” And the answers will either be “I don’t know” or they won’t exist.

The director has been managing to the first question for his entire tenure because that’s the depth of inquiry his previous leadership was capable of. A president with a consultant in the room is going to ask the second question. And the gap between “we have Monetate” and “here’s our testing program’s measured conversion impact over the last four quarters” is the gap that ends his credibility.

The Accountability He Built for Himself

Every organizational decision the director made over 14 months converges on a single structural outcome: he is the sole point of accountability for everything the department has produced, and he has no one to share it with.

He removed the architect who could have absorbed technical accountability. He centralized vendor communication through himself. He overrode the subject matter experts who could have diversified the decision-making. He ensured that the PM and admin who remained were non-technical and deferential to his direction. He made himself the loudest voice on every call, the final decision-maker on every feature, and the sole client-side authority on every vendor deliverable.

He got exactly what he wanted: total control. And total control means total accountability.

When the consultant’s report lands on the president’s desk, the director cannot say “I trusted the vendor and they failed me” — because he was on every call, directing every conversation. He cannot say “my architect should have caught this” — because he fired the architect. He cannot say “the SMEs didn’t raise concerns” — because he overrode them when they did. He cannot say “I was under-resourced” — because five SI engineers is not a skeleton crew, and the problem was direction, not headcount. He cannot say “I’ve only been here 18 months” — because 14 of those months were under his sole authority, which is long enough to own outcomes.

The only honest defense available to the director is: “Nobody told me what good looks like, I was promoted into a role I wasn’t qualified for, and nobody above me caught it either.” Which is true. And which is exactly the case against the SVP who oversaw the department, described the loss of its technical leadership as an opportunity for the director to “spread his wings,” and failed to intervene as the governance gap widened into a compliance exposure, a vendor accountability vacuum, and 14 months of zero conversion-impacting output.

But that defense is an admission, not a rebuttal. It’s the director saying, “I shouldn’t have been in this role” — which is the conclusion the president is going to reach independently once the report is in hand.

The architect is building commerce features at a technology company now. The job requisition is gathering dust. The commit log reads the way it reads. And the director, having ensured that he is the only person in the room who can be held accountable, is about to discover what accountability feels like when someone with authority finally asks the second question.

This is the department-level manifestation of a pattern I’ve written about previously as “Rational Negligence” — the structurally predictable outcome when organizations place non-technical operators in technical leadership positions without governance backstops. The board-level version produces underinvestment in digital infrastructure across entire brand portfolios. The department-level version, as this case illustrates, produces something more intimate and more destructive: an organization that acquires technical capability, rejects it like a foreign body, and then consumes its own budget producing work that wouldn’t survive a week of professional scrutiny. The negligence is rational at every step — each decision the director made was locally coherent from his vantage point — and the accumulated result is a department that costs more to run badly than it would cost to run well.

The Question

On December 9, 2019, two FBI agents walked into a strip-mall computer repair shop in Wilmington, Delaware, presented a federal grand jury subpoena, and left with a water-damaged laptop and an external hard drive. They handed the shop’s owner a property receipt. The subpoena had issued from the United States Attorney’s office for the District of Delaware, where a tax and financial investigation of the machine’s owner had been open since 2018, run by a United States Attorney appointed by the sitting Republican president. The laptop’s owner was Hunter Biden, the son of the man who would, eleven months later, be elected President of the United States.

But is there a case to be made that American journalists abdicated their professional responsibility to interrogate their own premises — to identify their biases, question them, and set them aside in their reporting?

Yes, and the case is serious enough that its most damning versions were written by journalists themselves, in flagship publications, in leaked transcripts, in memoirs of exiled editors. The honest answer also requires discipline, because there are two versions of this indictment in circulation and only one of them survives scrutiny.

I.

The first version is the caricature: that sometime around November 2016, American newsrooms collectively decided they were the resistance, adopted a “we are stopping fascism” mission statement, and converted straight news into propaganda. This version is emotionally satisfying and analytically lazy. It commits the same sin it indicts — fitting every data point to a prior narrative, treating the loudest examples as representative, collapsing the distinction between a columnist’s tweet and a front-page news story. If you believe the caricature, you will be unable to explain why the same press corps savaged the Afghanistan withdrawal, why the Wall Street Journal’s verification work on the laptop’s most explosive claim produced a null result that the Journal published anyway, or why the editorial boards of the major papers were brutal about the Hunter Biden pardon in December 2024. The caricature overpredicts. A model that overpredicts is wrong even when it feels right.

The second version is subtler, better documented, and considerably more disturbing, because it does not require a conspiracy. It goes like this: after 2016, American journalism openly renegotiated its relationship with objectivity; the renegotiation shifted not whether skepticism was applied but to whom; contested empirical questions were prematurely reclassified as settled, and dissent from the premature settlement was pathologized rather than reported; the mechanisms driving all of this were sociological and economic — the demographic homogeneity of newsrooms, the death of local media, and above all the shift from advertising revenue to subscription revenue, which converted readers from an audience to be informed into a constituency to be affirmed; and when the failures became undeniable, the corrections arrived late, quietly, and at a fraction of the prominence of the original errors, which is the specific behavior that converts mistakes into breaches of trust.

That second version is the argument of this essay. It runs from the coverage of the Trump presidency through the pandemic’s origin question, through the summer of 2020, through the domains American media learned to call “the culture war,” and it terminates — deliberately — at the Hunter Biden laptop, because the laptop is the case where every thread of the argument converges and where the documentary record is now, thanks to a federal courtroom, essentially complete. Along the way I am going to state the counter-arguments at full strength, because an indictment that cannot survive its own cross-examination is not worth filing, and because the strongest defense of the press’s conduct in this decade is itself part of the story: the defense explains the mechanism.

One more note on method before we begin. Nearly everything in this essay rests on primary documents and the profession’s own literature: leaked transcripts, sworn testimony, court exhibits, cryptographic verification reports, the Columbia Journalism Review, the Shorenstein Center, peer-reviewed studies of media slant, and the confessional essays of the editors who lived it. Where the record is contested, I will say so. Where a claim depends on an allegation rather than a finding, I will flag the difference. The subject of this essay is a profession that stopped doing those things. The essay should not repeat the offense.

II. The Renegotiation

Begin with what happened in the open, because the most important fact about American journalism’s shift after 2016 is that it was not covert. It was argued, published, institutionalized, and defended — which makes it more interesting ethically, not less. “Abdication” implies a duty quietly dropped. What actually occurred was closer to a public renegotiation of the duty’s terms.

The intellectual groundwork predated Trump. Gaye Tuchman’s 1972 paper in the American Journal of Sociology, “Objectivity as Strategic Ritual,” had argued that newsroom objectivity functioned less as an epistemic method than as a professional defense mechanism — a set of rituals (attribution, the inverted pyramid, the dueling quote) that protected journalists from criticism without guaranteeing truth. Jay Rosen of NYU spent the 2000s and 2010s attacking what he memorably called the “view from nowhere”: the pose of positionless neutrality that, he argued, produced stenography rather than verification. And in 2004, Maxwell and Jules Boykoff published “Balance as Bias” in Global Environmental Change, demonstrating that the press’s mechanical both-sidesing of climate science had systematically misinformed readers by presenting a settled scientific question as a live controversy. The critique was real, scholarly, and in important respects correct: balance is not accuracy, and treating it as accuracy is itself a distortion.

What 2016 did was convert that critique from a scholarly corrective into an operational doctrine, applied to a single subject, under conditions of institutional panic. The clearest single statement of the shift came in June 2020, when Wesley Lowery — a Pulitzer-winning reporter — published an op-ed in the New York Times arguing that newsrooms should abandon the “appearance of objectivity” in favor of “moral clarity.” Lowery’s argument was more careful than its slogan: he was attacking performative neutrality, not verification. But slogans travel and arguments don’t, and “moral clarity” became the banner under which a generation of journalists reconceived their role. Three years later, Leonard Downie Jr. — a former executive editor of the Washington Post, which is to say, a man who had spent his career enforcing the old regime — co-authored a study with Andrew Heyward for Arizona State’s Cronkite School titled Beyond Objectivity. They interviewed more than seventy-five newsroom leaders and reported a finding that would have been unthinkable a decade earlier: many of the people running American journalism no longer regarded objectivity as a governing value at all.

Against this stood a rearguard whose most articulate member was Martin Baron, the editor who ran the Post through the Trump years. Baron’s position, developed at length after his retirement, was that the critics had attacked a straw man: objectivity, properly understood, was never neutrality of soul or false balance between truth and falsehood. It was a method — a discipline of verification applied with equal rigor regardless of whose claims were being tested. This is also the position of Bill Kovach and Tom Rosenstiel’s The Elements of Journalism, the closest thing the profession has to a catechism, which traces the term to Walter Lippmann’s insistence in the 1920s that because journalists cannot be objective, their methods must be. On this account, the choice between objectivity and moral clarity was always false. The method produces the clarity; the clarity without the method produces conviction, which is a different commodity.

The analytical framework that best explains what happened next comes from Daniel Hallin’s 1986 study of Vietnam coverage, The Uncensored War. Hallin observed that journalists implicitly sort all claims into three spheres. In the sphere of consensus (the sky is blue, democracy is good), journalists feel no obligation to present opposing views. In the sphere of legitimate controversy (tax policy, foreign intervention), the balance rituals apply. And in the sphere of deviance (Holocaust denial, flat-earthism), journalists feel not merely permitted but obligated to treat claims as illegitimate — to debunk rather than report. The spheres themselves are invisible to the people using them; sorting a claim into one feels like perceiving reality, not making a judgment.

Here is the crux of everything that follows: the placement of claims into spheres is the single most consequential editorial act in journalism, and it is the one journalists least often recognize themselves performing. Nobody at a news meeting says “let us move the lab-leak hypothesis into the sphere of deviance.” The move happens upstream of consciousness, in the aggregate of a newsroom’s assumptions about what sort of people believe what sort of things. When a newsroom is demographically and politically homogeneous — and we will come to the data on that — the sphere boundaries settle exactly where the room’s shared priors sit, and the room experiences its priors as the shape of reality itself.

The post-2016 renegotiation, then, was not primarily a change in what journalists wrote. It was a change in the sphere map. Claims associated with Trump and his coalition migrated wholesale toward deviance, where the both-sides rituals were suspended and the debunking posture engaged — and, crucially, the migration swept up not only the genuinely false (of which there was an unprecedented supply, as we’ll stipulate shortly) but also the contested, the unresolved, and, in at least one spectacular case, the true. Meanwhile, claims aligned with the newsroom’s own consensus migrated toward the sphere of consensus, where skepticism is not applied because applying it feels like debating whether the sky is blue.

That is the disciplined version of the “abdication” thesis. Not that journalists stopped seeking truth — but that they stopped noticing where they had drawn the boundaries of the thinkable, stopped interrogating the drawing, and reclassified the act of interrogation itself as a form of complicity. The rest of this essay is the evidence.

III. The Asymmetry Engine: Covering Trump, 2016–2020

Any account of press conduct in the Trump era has to begin by stipulating the thing that makes the era genuinely hard to adjudicate: the object of coverage was without precedent. The Washington Post‘s Fact Checker database — whatever one thinks of its occasional pedantry — tallied 30,573 false or misleading claims across four years of the Trump presidency. No prior president generated anything within an order of magnitude of that volume. Sixty-plus courts, including judges the president appointed, rejected the stolen-election claims that culminated in January 6. A press corps that had covered that record with mechanical balance — one voice says the election was stolen, another disagrees; you decide — would have committed exactly the “balance as bias” error the Boykoffs documented in climate coverage. When the New York Times made its deliberate, publicly debated decision in September 2016 to put the word “lie” in a headline about birtherism, that was not activism. It was accuracy, arrived at through exactly the kind of institutional deliberation the method requires. The defenders of the post-2016 posture are right about all of this, and any honest indictment concedes it up front.

The indictment begins where the stipulation ends: with the question of whether the response was calibrated to the object or overshot it, and whether the overshoot was ever measured, acknowledged, or corrected.

Start with the aggregate data, because it exists. Thomas Patterson of Harvard’s Shorenstein Center — a scholar who has studied campaign coverage for four decades and whose central finding, long predating Trump, is that press negativity toward all politicians has been rising since the 1970s — measured the tone of coverage of Trump’s first hundred days in 2017. The finding: roughly 80 percent negative, against 41 percent for Obama’s first hundred days, 57 for George W. Bush’s, and 60 for Clinton’s. On some networks the figure exceeded 90 percent. Patterson himself was careful about interpretation: some of that asymmetry reflects the asymmetry of the underlying conduct — a president under investigation generates investigation stories — and some reflects the secular trend he’d been charting for decades. But the residual is large, it was visible across policy domains where the conduct asymmetry was weaker, and — this is the part that matters for our purposes — the profession showed essentially no curiosity about it. An 80-percent-negative tone distribution is either a faithful measurement of an unprecedented presidency or evidence of a systematic instrument error, and a profession serious about interrogating its premises would have wanted very badly to know which. It did not ask.

Then there is the document that tells you what the instrument was doing from the inside. In August 2019, someone leaked to Slate the full transcript of a New York Times town hall in which executive editor Dean Baquet addressed his newsroom in the aftermath of the Mueller report. The transcript is remarkable for its candor. Baquet told his staff that the paper had “built our newsroom to cover one story” — the Trump-Russia story — and that with that story having failed to deliver its anticipated ending, the paper needed to pivot: the next two years, he indicated, would be organized around race and the president’s character, a framing that took institutional form days later with the launch of the 1619 Project. Read generously, this is an editor doing his job — allocating resources against the biggest stories. Read precisely, it is an editor describing narrative-first planning: the arc is selected, then the coverage is built to it. Every journalist knows the difference between following a story and building a newsroom to cover one. The town hall transcript is the rare artifact in which the second posture is described, by the person doing it, in his own words.

The asymmetry engine had a characteristic failure mode at the micro level too, and it’s the one that most corroded ordinary readers’ trust: the treatment of hyperbole. Trump spoke in a register — salesman’s puffery, comic exaggeration, deliberate provocation — that the fact-checking apparatus was structurally incapable of parsing. The result was a genre of correction that literalized the obviously figurative, “fact-checking” statements no listener had taken as factual claims, which taught audiences that the checkers were scoring points rather than protecting truth. The famous 2016 formulation — that the press took Trump literally but not seriously while his supporters took him seriously but not literally — named the problem without solving it. And in fairness, the problem cut both ways: a president who used “I was joking” as a retroactive escape hatch for statements like his invitation to Russian hackers made literalism a defensible default. But the fact-checkers never developed a doctrine for the distinction, and the absence of doctrine meant the calls defaulted to instinct, and instinct defaulted to the room’s priors. A reader could predict, with depressing reliability, which politicians’ exaggerations would be scored as lies and which would be scored as color.

None of this, so far, is the breach. Aggressive, negative, even instinct-driven coverage of a president is within the historical range; presidents from Adams onward have complained about worse. The breach requires something more specific: the migration of open questions into the sphere of deviance, enforced against dissent, uncorrected when wrong. For that, we need the template.

IV. The Template: Russia

The Trump-Russia story is where the post-2016 press earned both its Pulitzers and its autopsy, and the autopsy was conducted, fittingly, by the profession’s own coroner. In January 2023 the Columbia Journalism Review published Jeff Gerth’s four-part, roughly 24,000-word retrospective on the coverage — the most exhaustive internal examination of a single story the field has produced this century. Gerth, a former Times investigative reporter, is not a Trump partisan, and CJR is not a conservative organ; that is precisely what gives the document its weight.

The findings, compressed: major outlets laundered the Steele dossier — an unverified opposition-research product, paid for by the Clinton campaign and the DNC through cutouts — into their news coverage for years, long after its central claims had failed to corroborate. Stories built on it, or on anonymous characterizations of it, were never retracted. The claim that “seventeen intelligence agencies” had assessed Russian interference — repeated endlessly — was false; the assessment came from three agencies plus the Director of National Intelligence. The Alfa Bank server story, published in the closing days of the 2016 campaign, collapsed entirely; the FBI had dismissed it within weeks, a fact that emerged years later in a criminal trial. The Times‘ 2020 story on Russian bounties for American soldiers in Afghanistan — which dominated a news cycle and a presidential debate — was walked back by the intelligence community itself the following spring, which assessed the claim with only “low to moderate confidence”; the walk-back received a fraction of the original’s play. And Justice Department Inspector General Michael Horowitz’s December 2019 report documented seventeen significant errors and omissions in the FBI’s surveillance applications against a Trump campaign aide — findings that vindicated a set of claims the press had spent two years treating as partisan noise, and that landed, again, with a fraction of the prominence of the coverage they discredited.

Two calibrations, because the template only teaches if it’s rendered accurately. First: Russiagate was not a hoax. The Mueller report documented extensive, real contacts between the Trump campaign and Russian-connected figures, a Russian interference operation the report called sweeping and systematic, and multiple episodes of potential obstruction that Mueller pointedly declined to exonerate. The story deserved aggressive coverage. The failure was not that the press investigated; it was that the press pre-concluded — that a substantial portion of the coverage was built backward from an anticipated ending (collusion proven, presidency ended) rather than forward from verified facts, and that when the ending failed to arrive, the apparatus pivoted rather than audited. That is Baquet’s town hall in a sentence.

Second, and more important for everything that follows: notice what the Russia years built. They built a reflex — a fused association between “damaging material about a candidate, of murky provenance, surfacing near an election” and “hostile foreign information operation.” The reflex had a real referent: 2016’s hack-and-leak was genuine, and the press’s role in amplifying the fruits of it genuinely haunted the institutions that had participated. The lesson the profession drew from 2016 was we were used, and we must never be used again. It is one of history’s neater ironies that this lesson — reasonable on its face, penitential in origin — became the loaded mechanism of the decade’s worst failure. An institution that has decided in advance what the next October surprise will be has already decided how it will cover it. The template was set. It waited four years for a story shaped like its trigger.

The Russia years also established the correction asymmetry that will recur through every case in this essay, so let us name it formally here. Call it the prominence ratio: the ratio between the display given an error and the display given its correction. A front-page story walked back in a paragraph on A17 has a prominence ratio approaching zero. Professional ethics codes — the SPJ’s included — require corrections; none requires prominence parity, and the gap between those two standards is where reputations go to survive. Across the Russia coverage, the ratio was consistently, sometimes vanishingly, low: the dossier arrived with sirens and departed in subordinate clauses. Readers noticed. The profession’s trust numbers — we will come to them — are, among other things, the compounded interest on a decade of near-zero prominence ratios.

V. The Substitution Test, and How to Make It Honest

There is a folk instrument for detecting media bias, and everyone reading this has run it: swap the parties. If a Democrat had done this, would the framing be the same? If the victim were white, would the coverage exist? The instrument has a disreputable air — it lives in comment sections — but its legitimacy is conceded by the profession itself, because journalism’s own reform movements are built on it. When Gwen Ifill gave the phrase “missing white woman syndrome” to the profession in 2004, she was running the substitution test on race and victimhood. When newsrooms flagellated themselves during the Gabby Petito frenzy of 2021 — noting, correctly, that hundreds of Indigenous women had gone missing in the same state without a fraction of the coverage — they were running it again, and publishing the results as self-criticism. Nobody gets to certify the instrument when it runs left and confiscate it when it runs right.

What the profession can say — and what must be said to keep the instrument honest — is that in its folk form the test is methodologically weak. It is an uncontrolled thought experiment: you supply the treatment, you imagine the counterfactual, and you grade your own exam, which means the instrument measures the user’s priors at least as efficiently as it measures the outlet’s bias. The remedy is not to discard the test but to discipline it: replace imagined counterfactuals with matched natural experiments — real pairs of events, similar in kind, differing in the political valence of the principals — and measure the coverage on defined observables.

The record contains such pairs, and they are worth examining closely, because they carry the argument that intuition alone cannot.

The cleanest is Tara Reade against Christine Blasey Ford: the same newspaper, the same category of allegation (a decades-old accusation of sexual assault against a man on the threshold of high office), opposite parties. The New York Times covered Ford’s allegation against Brett Kavanaugh within hours of its surfacing in 2018, as did everyone. In 2020, the paper took nineteen days to cover Reade’s allegation against Joe Biden — and when it did, the story initially contained a sentence noting that the Times had found no pattern of misconduct beyond previously reported unwanted hugs and kisses, a clause the paper deleted after the Biden campaign complained. We know this not through inference but through confession: Dean Baquet described the editing sequence, and the campaign’s role in prompting it, on the record, in an extraordinary interview with his own paper’s media columnist, Ben Smith, in April 2020. Baquet’s explanation — that Kavanaugh was already in a public confirmation process while Biden’s situation differed — is an argument, and readers can weigh it. What cannot be argued away is the pair itself: same paper, same allegation class, nineteen-day differential, and a post-publication edit made at the subject campaign’s request, confirmed by the editor. In any other industry, that last item is called customer influence over the audit.

A second pair, humbler but instructive because it involves no allegation at all, only the framing of frailty. In June 2020, Trump descended a ramp at West Point haltingly and used two hands to drink a glass of water; the Times ran a straight-news piece on the episode and the questions it raised about his health, and cable news chewed it for days. In April 2022, a White House staffer in a full Easter Bunny costume physically steered President Biden away from reporters’ questions on the South Lawn. The episode was covered — the claim that it was ignored is false, and the false version of the claim should be retired, because the true version is more damning — but it was covered as whimsy: a light item, a colorful aide, a chuckle. Same category of event (visible presidential handling/frailty, questions about capacity), same outlets, categorical difference in frame. And the Bunny pair matters more in retrospect than it did at the time, because we now know — from the profession’s own subsequent confession, which we will come to — that the frailty it framed as whimsy was the visible edge of the biggest under-covered story of the decade.

A third pair shows that the instrument is content-neutral, which is what saves it from being a partisan cudgel. In November 2021, a man drove an SUV through a Christmas parade in Waukesha, Wisconsin, killing six people; multiple national outlets rendered the event, in headlines and copy, as a tragedy “caused by an SUV” — agency amputated from the driver. Conservatives caught it instantly, and they were right. The mirror-image construction — “officer-involved shooting,” a phrase engineered to remove the officer from the verb — had been caught by the left for a decade, and they were right too. Each camp has perfect vision for the passive voice when the other side deploys it, and perfect blindness when its own side does. The substitution test, run honestly, indicts everyone’s framing reflexes. That is precisely why it is worth running honestly.

Which brings us to the discipline. If you want the test to survive peer review rather than merely a dinner argument, it needs four features. Use matched pairs, not imagined ones — real events, documented coverage. Measure defined observables: coverage volume and placement; agency language (who is the subject of the verbs); source selection (whose experts frame the event); and correction prominence relative to the original error — the prominence ratio again. Pre-register your predictions: write down what coverage you expect before the outcome is known, because retrodiction is where motivated reasoning lives, on every wing. And accept the literature even when it is inconvenient, because some of it is.

Here is the inconvenient literature, stated at full strength. Gentzkow and Shapiro’s 2010 Econometrica study found that newspaper slant tracks the political composition of the paper’s market — consumer demand — far more than the ideology of its owners; slant, in other words, is substantially a product being sold to an audience that wants it, a finding that will matter enormously when we reach the economics. Budak, Goel, and Rao’s large-scale 2016 study found that on most topics, mainstream outlets’ articles were not strongly ideological in their internal framing — the bias, where it existed, lived predominantly in story selection: which events become stories at all. And Hassell, Holbein, and Miles, in a 2020 Science Advances paper, ran actual correspondence experiments on thousands of journalists and found little evidence of partisan gatekeeping in their willingness to cover hypothetical campaign stories — a result any honest critic of the press has to metabolize rather than ignore, even while noting its scope (local political coverage, hypothetical pitches) leaves the biggest national cases untouched.

Read together, the literature and the matched pairs point at the same refined conclusion, and it is the essay’s hinge: the failure of the post-2016 press was not, in the main, slanted sentences. Reporters mostly wrote careful paragraphs. The failure lived one level up, in the decisions that precede sentences — whether something is a story, when it becomes one, which sphere it is filed in, whose denial ends the inquiry, and how loudly the record is corrected when the filing was wrong. Selection, timing, sphere placement, and the prominence ratio: those are the four dials, they are turned before a single word is written, and they are precisely the dials that a homogeneous newsroom turns in unison without experiencing itself as turning anything at all.

Now watch the four dials operate on the three great stress tests: a virus, a summer, and a laptop.

VI. The Virus: Premature Closure as Policy

The pandemic’s origin question is the strongest single exhibit in the case against the post-2016 press, and it is essential to state the exhibit precisely, because the sloppy version of the complaint (”the media said Covid came from nature and they were wrong”) is both unfair and beside the point. Most virologists genuinely favored a natural, zoonotic origin — a spillover from an animal host, most likely at a wildlife market — and they favored it for defensible scientific reasons. Preferring the zoonotic hypothesis was not the error. Zoonosis remains, as of this writing, a leading hypothesis. The error was procedural, and it maps exactly onto Hallin’s spheres: an open empirical question — whether the virus emerged naturally or escaped from a laboratory studying exactly this family of viruses in exactly the city where the outbreak began — was forcibly relocated from the sphere of legitimate controversy, where it belonged, into the sphere of deviance, where dissent is not weighed but excommunicated. And the relocation was not a passive drift of opinion. It was manufactured, by identifiable people, with identifiable conflicts of interest, and then enforced by platforms and amplified by a press that did not examine the machinery it was amplifying.

The manufacturing has a date and a document. In February 2020 — weeks into the outbreak, when essentially nothing was known — The Lancet published a statement by twenty-seven scientists condemning “conspiracy theories suggesting that COVID-19 does not have a natural origin.” A statement in the world’s most prestigious medical journal, framing lab-origin discussion as conspiracy theorizing, functioned as a starting gun: it told every science journalist on earth which sphere the question belonged in. What the statement did not disclose — what emerged only later, through FOIA’d emails obtained by a nonprofit and by congressional investigators — was that it had been organized largely by Peter Daszak, president of EcoHealth Alliance, an organization that had channeled U.S. grant money to the Wuhan Institute of Virology for coronavirus research. Daszak had a direct, undisclosed professional stake in the conclusion that the lab was irrelevant: if the virus had escaped from research his organization had helped fund, the reputational and legal exposure was enormous. The statement was presented to the public and the press as the disinterested consensus of scientists. It was, in part, the interested product of a man with one of the largest conflicts of interest imaginable. The press printed it as the former for over a year.

The enforcement has dates too. In 2020, when Senator Tom Cotton raised the lab-origin possibility, the Washington Post ran a piece describing the hypothesis with language calling it “debunked” and “fringe” and linking it to conspiracy theories — and then, in 2021, quietly appended a correction softening that framing, an admission that the original “debunked” had itself been unwarranted. PolitiFact rated a lab-origin claim “Pants on Fire” false in September 2020, then retired the rating entirely in 2021, conceding the question was unresolved. Facebook announced in February 2021 that it would remove posts claiming the virus was man-made or lab-made — an actual, enforced ban on discussing a hypothesis — and then reversed the policy in May 2021, within days of President Biden’s order that the intelligence community investigate the lab-leak possibility. The sequence deserves to be stated flatly: a hypothesis was declared false by fact-checkers, banned by platforms, and treated as deviance by the press; then, when the sitting Democratic president ordered a formal intelligence review, the same hypothesis was abruptly readmitted to the sphere of legitimate controversy — the president’s action having functioned as a permission slip. The question had not changed. The evidence had not changed. Only the sphere assignment changed, and it changed on a political cue rather than a scientific one. That is the tell. Sphere assignments driven by evidence move when evidence moves. Sphere assignments driven by tribe move when the tribe’s authorities move.

And the underlying question did not resolve toward the settled answer the enforcers had asserted. By 2023, the U.S. Department of Energy had assessed lab-origin as the most likely explanation, at low confidence; the FBI had assessed it as likely at moderate confidence; other agencies favored natural origin, also at low confidence, and the National Intelligence Council was split. In other words: exactly the condition — genuine, unresolved, reasonable-people-disagree uncertainty — that the sphere of legitimate controversy exists to hold. The press had spent 2020 and much of 2021 insisting that the sphere of legitimate controversy was the sphere of deviance, enforcing the boundary against colleagues and readers, and it turned out the boundary had been drawn in the wrong place by a man with an undisclosed conflict of interest, and printed without the scrutiny that the profession’s entire reason for existing is to supply.

Now the correction behavior — the prominence ratio — because it is here that error becomes something worse. When the lab-leak hypothesis was rehabilitated in 2021, there was no reckoning proportional to the enforcement. Outlets that had spent a year calling the hypothesis a debunked conspiracy theory ran measured pieces noting it now merited investigation, as though the earlier certainty had been someone else’s. The fact-checkers retracted specific ratings but not the framework that produced them. No major newsroom conducted, on itself, the kind of autopsy that CJR conducted on Russiagate. The premature closure was reversed; the practice of premature closure — the reflexive filing of a politically-coded empirical question into deviance on the say-so of interested “consensus” — was not examined, and so it remained available for the next question. Which is exactly what a breach of trust is, as distinct from a mere error: an error is a wrong answer; a breach is a wrong answer plus the refusal to investigate the process that produced it, which guarantees the process will produce more wrong answers and signals to the audience that the institution does not consider itself accountable to them.

There is a defense of the pandemic coverage, and it is not nothing. The information environment was genuinely polluted: some lab-leak proponents were also promoting actual nonsense (bioweapon claims, 5G theories, miracle cures), and distinguishing the serious hypothesis from the surrounding noise in real time, under deadline, was hard. Public-health communication faced a real coordination problem: officials worried, not unreasonably, that entertaining lab-origin would fuel anti-Asian violence or undermine compliance with public-health measures. And some of the “just asking questions” around origins was indeed bad-faith cover for xenophobia. All true. But none of it justifies the specific failure, which was not “the press was uncertain” or “the press was cautious.” The failure was that the press was certain in the wrong direction, enforced that certainty against legitimate scientific dissent, derived the certainty from a source it failed to vet, and declined to audit itself when the certainty collapsed. Caution would have been filing the question in legitimate controversy and covering the debate. What actually happened was the opposite of caution: it was the manufacture of false consensus and its enforcement as deviance. That is not an excess of journalistic scruple. It is its absence.

VII. The Summer: Framing at the Edge of the Fire

The summer of 2020 produces the most-quoted single artifact of the entire media-criticism decade, and it is worth dwelling on why. On the night of August 26, 2020, a CNN correspondent stood in Kenosha, Wisconsin, in front of a building fully engulfed in flames and a landscape of burning vehicles, and the chyron beneath him read: “FIERY BUT MOSTLY PEACEFUL PROTESTS AFTER POLICE SHOOTING.” The image became instantly, permanently iconic — the definitive visual of a press caught, on live television, insisting on a frame its own cameras were contradicting. It is the “the sky is aqua, not blue” of the decade: the moment the gap between the narrative and the visible world became a single shareable screenshot.

And here the honest critic has to do something the sloppy critic won’t, because the truth of this case is genuinely more complicated than the screenshot, and the complications matter in both directions.

Take “mostly peaceful” first, on its own terms. As a statistical claim about the summer’s demonstrations in aggregate, it was true. The Armed Conflict Location & Event Data Project (ACLED), a respected academic conflict-tracking organization, analyzed the more than 10,000 demonstrations associated with the Black Lives Matter movement between late May and late August 2020 and found that over 93 percent involved no serious violence, injury, or property destruction. That is a real finding from a serious source, and any account that pretends the entire summer was a nationwide riot is simply lying about the data. Most of the demonstrations were exactly what “mostly peaceful” said: peaceful. The critics who treat “mostly peaceful” as a self-evidently absurd phrase are, on the aggregate numbers, wrong.

The failure was not the aggregate claim. The failure was, once again, one level up — in framing, in juxtaposition, and in what the aggregate claim was deployed to do. Deploying a true statistic about ten thousand demonstrations as a chyron over a specific building that is specifically on fire is not informing; it is using a true general claim to instruct the viewer to disbelieve a true specific image. It is the passive voice of statistics — a technically accurate frame engineered to direct attention away from the thing in the frame. The 93-percent figure describes the summer; it does not describe Kenosha on August 26, and pretending the general dissolves the particular is its own species of dishonesty, distinct from but symmetrical to the right’s dishonesty of pretending the particular is the general.

And there was a particular, and it was large. The same summer produced what the insurance industry recorded as the costliest civil disorder in United States history: the Property Claim Services division of Verisk estimated insured losses from the unrest at between one and two billion dollars — and insured losses systematically understate total losses, because the uninsured and underinsured, disproportionately the small and minority-owned businesses in the affected corridors, don’t appear in the tally. Multiple deaths were associated with the unrest. Downtown corridors in Minneapolis, Kenosha, Portland, and elsewhere sustained damage that in some cases has still not been repaired. This was real, it was severe, and a press that submerged it beneath “mostly peaceful” was not lying with its statistics but was, through selection and framing and juxtaposition, producing in its audience a systematically false impression of the summer’s texture — an impression that the destruction was marginal noise around a peaceful signal, when for the specific communities in the specific corridors the destruction was the signal.

Then there is the episode that did the most damage to the press’s specifically epistemic authority, because it exposed the sphere machinery operating in real time on a question of physical science. In June 2020, more than 1,200 public-health professionals signed an open letter stating that the anti-racism protests should not be curtailed on epidemiological grounds — that they should be supported as vital to public health — and explicitly cautioning against equating them with the anti-lockdown protests that these same professionals had, weeks earlier, condemned as dangerous public-health threats. Read that sequence slowly, because it is the sphere machinery laid bare: the identical physical act — large numbers of people gathering outdoors during a respiratory pandemic — was assigned to the sphere of deviance (reckless, deadly, condemnable) when performed by one political coalition, and to the sphere of consensus (vital, supportable, necessary) when performed by another, and the reassignment was justified not by any change in the virology of aerosol transmission but by the political valence of the cause. And the press, in substantial part, relayed this as though it were a scientific finding rather than a nakedly political judgment wearing a lab coat. The virus does not read protest signs. A profession devoted to interrogating premises would have interrogated that one loudly. Most of it did not.

The counter-arguments here are real and must be granted. The underlying event — George Floyd’s killing, captured in nine minutes of unbearable video — was genuine, and the grief and rage were genuine, and the vast majority of the people expressing them did so peacefully, and a press that had covered the summer as primarily a riot story would have committed exactly the inverse distortion, submerging a true and enormous signal (a historic civil-rights mobilization) beneath a real but smaller one (the destruction). The framing problem cut both ways, and Fox’s summer — which submerged the 93 percent beneath the fires as reliably as CNN submerged the fires beneath the 93 percent — was the mirror-image malpractice, and no honest account gives one a pass. But “both sides framed selectively” is not exoneration; it is a description of a profession-wide failure to hold two true things at once — historic peaceful mobilization and the costliest civil disorder in the nation’s history — which is precisely the failure a serious press exists to avoid. The both-and was available. The reporting, on both sides, chose the either-or that flattered its audience.

VIII. The Culture War: Deviance Assignment as Default

By the time we reach the domains American media learned to call “the culture war,” the pattern is no longer a set of discrete failures but an operating default, and the mechanism is worth naming: on a defined set of identity-inflected questions, the press ceased treating contested empirical and moral matters as belonging to the sphere of legitimate controversy and began treating one side’s position as the sphere of consensus — which meant the other side’s position, and often the underlying facts themselves, were filed under deviance, reported (when reported at all) in the debunking register, and their proponents characterized rather than quoted. The tell, throughout, is the same as in the virus case: when the sphere assignments move on political cues rather than on evidence, and when the corrections (where reality forces them) arrive quietly and late, you are looking at tribe, not method.

Consider the youth gender-medicine question, because it is the case where the reversal is most fully documented and where an American newspaper’s own reporters revolted against their institution’s coverage. For several years, the framing prevalent in much of the American press treated the affirmative model of pediatric gender medicine — social transition, puberty blockers, cross-sex hormones for minors — as settled best practice, with dissent characterized largely as animus, and the underlying evidence base described as robust. Meanwhile, the health authorities of several European countries with nationalized health systems and no American culture-war stake — Sweden, Finland, the United Kingdom — were independently reviewing that same evidence base and reaching a markedly more cautious conclusion. Finland restricted the affirmative model in 2020; Sweden’s Karolinska and then its national board followed in 2021–2022; and in England, the National Health Service commissioned Dr. Hilary Cass, a senior pediatrician with no dog in the American fight, to conduct a systematic review. The Cass Review, published in April 2024, found the evidence base for pediatric medical transition “remarkably weak,” criticized the quality of the underlying studies, and led the NHS to sharply restrict puberty blockers outside research settings. This was not a fringe conservative document; it was a commissioned review by a mainstream medical body in a country whose politics do not map onto America’s.

The revealing American moment came in February 2023, when the New York Times — which had, to its credit, done exactly the kind of careful, evidence-examining reporting on this subject that the moment demanded, notably Emily Bazelon’s long and scrupulous feature — was met not with praise for that reporting but with an open letter signed by hundreds of contributors and allied organizations attacking the coverage as harmful and demanding it change. The paper’s leadership, again to its credit, held firm and defended the reporting. But the episode is a near-perfect illustration of the pressure that produces sphere-collapse: the demand, from inside the extended institution, was that a genuinely contested empirical question — one that mainstream European medical authorities were actively treating as unresolved — be reported as settled, with the unsettled framing itself denounced as a form of harm. The Times resisted; the fact that resistance required editorial courage, and generated an internal revolt, tells you how strong the default toward sphere-collapse had become.

Or consider the narrower, more concrete matter that stands in here for a whole class of school-policy questions — the design and governance of restrooms and locker rooms in public schools when access is decoupled from biological sex. This is a genuinely hard problem with real and competing interests on every side: the dignity, safety, and inclusion of transgender students, which are serious and legitimate; the privacy and comfort concerns of other students and their families, which are also serious and legitimate; and the practical governance questions that fall on school administrators caught between them. It is, in the truest sense, a matter for the sphere of legitimate controversy — reasonable people, arguing in good faith, weighing incommensurable goods. Yet a substantial part of the coverage filed it under deviance: one set of positions was treated as self-evidently correct and compassionate, the competing set as self-evidently bigoted, the parents raising privacy concerns characterized rather than heard, and the empirical questions (about incidence, about implementation, about the actual experiences of the students on every side) largely unexamined because examining them felt like conceding the deviance framing was wrong. The point here is emphatically not that one side’s answer is correct — the point is that a contested moral and practical question was denied the status of contested, and the denial served the priors of the room doing the reporting.

The unifying diagnosis across the culture-war cases is not that the press adopted progressive conclusions — a press is allowed conclusions on its editorial pages, and individual reporters are allowed private views. The diagnosis is structural: the press repeatedly mistook the boundaries of its own newsroom consensus for the boundaries of legitimate debate, filed the resulting questions accordingly, and then — this is the recurring, load-bearing failure — when reality intruded (when European health authorities reached different conclusions, when the Cass Review landed, when a policy’s real-world tradeoffs surfaced), corrected quietly, incrementally, and without ever auditing the reflex that had produced the misfiling. Selection, sphere, source, and prominence ratio — the same four dials, turned the same way, on question after question, by a room that experienced the turning as simply seeing clearly.

IX. Why It Happened: The Mechanism Is Not a Conspiracy

If the argument so far were merely a catalogue of failures, it would be journalism criticism of the ordinary kind — a bill of particulars, easily dismissed as cherry-picking by anyone inclined to dismiss it. What makes it more than that is the existence of a mechanism: a coherent, boringly sociological and economic explanation for why a profession full of intelligent, well-intentioned people produced this specific pattern of failure without anyone ordering it. The mechanism matters because it is the difference between grievance and diagnosis. Grievance says they are lying to us on purpose. Diagnosis says here are the structural conditions under which sincere people reliably misfile reality, and here is why the misfiling runs in a predictable direction. The second is more useful, more defensible, and — for anyone who cares about the institution’s repair — more actionable. It has three components.

The first is composition. American newsrooms are, and have become progressively more, politically and culturally homogeneous. The recurring surveys of American journalists conducted out of Indiana University over four decades — the closest thing the field has to a longitudinal census — document the trajectory: the share of journalists identifying as Republican, never high, fell into the low single digits by the most recent waves, a figure wildly out of proportion to the roughly even split in the general population. And the homogeneity is not only partisan; it is educational, geographic, and class-inflected, concentrated in a handful of coastal metros, credentialed through a narrow set of institutions, and increasingly detached from the regions and social strata it purports to cover. Now recall Hallin: the sphere boundaries are invisible to the people using them; sorting a claim into consensus, controversy, or deviance feels like perceiving reality. A homogeneous room does not experience its shared priors as priors. It experiences them as the shape of the world. When nearly everyone you work with, socialize with, and respect would find a given proposition obviously true or obviously beyond the pale, the proposition’s sphere assignment is made for you, pre-consciously, by the ambient consensus — and the one corrective that could catch the error, a colleague who holds the other prior and says wait, that’s contested, has been selected out of the building. Composition doesn’t make journalists lie. It removes the friction that would otherwise catch the honest mistake.

The second is the collapse of local news, which removed both a check and a ballast. The extinction of local newspapers over the past two decades — thousands of papers gone, whole counties reduced to “news deserts” — did more than destroy accountability journalism at the county courthouse and the school board. It nationalized the profession’s center of gravity. A reporter who came up covering zoning fights and high-school sports and the local water utility in a mid-sized city was exposed, daily, to the concrete texture of ordinary American life and to sources who did not share her priors. That reporter was ballast. As the on-ramp of local journalism disappeared and the surviving jobs concentrated in a few national newsrooms, the profession lost its regular contact with the country it covers and became a closed conversation among people who increasingly resembled one another, talking, in large part, to one another — on the platform we’ll come to in a moment. The check that local reporting provided against metropolitan groupthink was dismantled not by anyone’s design but by the collapse of the advertising model that had funded it, and its loss removed a second source of friction against the misfiling.

The third component is the deepest, because it changed the profession’s incentives at the root: the shift from advertising to subscription. This is Gentzkow and Shapiro’s finding operationalized by economic catastrophe. For most of the twentieth century, the dominant newspaper business model was advertising, and advertising rewards reach: the paper wants the largest possible audience, which means it has a commercial incentive not to alienate any large segment of potential readers, which means — whatever its editors’ private views — it has a structural reason to be, or at least to appear, broadly nonpartisan. The digital collapse of advertising destroyed that model, and the industry’s survivors replaced it with reader revenue: digital subscriptions, memberships, the direct payment of the audience. And reader revenue rewards something entirely different from reach. It rewards intensity. A subscription business does not need everyone; it needs a devoted core willing to pay, month after month, and the way you cultivate a devoted paying core is by reliably affirming its worldview, confirming its priors, and telling it that its enemies are as bad as it fears. The advertising model punished alienating half your potential audience. The subscription model rewards it, provided you capture and intensify the other half. The incentive did not merely permit audience capture. It selected for it.

This is not speculation; it is the thesis of the single most important internal document the decade produced, and it was written by a man with unmatched standing to write it. James Bennet had been the editorial page editor of the New York Times until June 2020, when he was forced out in the uproar over publishing a Senator Tom Cotton op-ed — an op-ed arguing for the deployment of federal troops to quell the summer’s unrest, whose publication triggered a staff revolt on the grounds that it endangered Black colleagues, and whose editor was gone within days. In December 2023, Bennet published a roughly 17,000-word essay in The Economist that is essential reading for anyone trying to understand what happened to American journalism. His central argument was precisely the audience-capture thesis: that the Times, under the commercial pressure of the subscription transition and the cultural pressure of a young, activist, digitally-native staff, had traded the hard discipline of independence for the easier rewards of affirming a like-minded and paying subscriber base — that it had, in effect, allowed its readers and its youngest employees to become a constituency it served rather than a public it informed. Bennet was not a disgruntled outsider throwing stones. He was the person who had run the institution’s opinion operation, describing, from the inside, the incentive structure that had defenestrated him. The audience-capture explanation is not the critics’ theory imposed on the press. It is the press’s own most senior wounded insider’s theory of himself.

Put the three components together and the mechanism is complete, and it requires no villains. A homogeneous newsroom (composition) that has lost its contact with the country (local collapse) and now depends financially on intensifying the devotion of an ideologically sorted paying audience (subscription economics) will, with total sincerity and no coordination whatsoever, systematically misfile politically-coded questions into the sphere that flatters its audience and its own priors — and will experience the misfiling as simply seeing clearly, because the friction that might have caught it (a dissenting colleague, a source from another world, an advertiser who wanted the other half of the market) has been removed at every level. That is why the failures ran in a consistent direction. Not because anyone decided to fight fascism at the expense of the facts, but because every structural pressure on the modern newsroom pushed the sphere boundaries the same way, and nothing pushed back.

X. Where the Indictment Breaks: The Strongest Case for the Defense

An indictment that only prosecutes is propaganda. The argument of this essay is worth making only because it can survive the strongest counter-case, so here is the strongest counter-case, stated without the hedging that usually accompanies concessions offered in bad faith. If you are inclined to accept everything above too easily, this section is the one to read twice.

First, and most important: the model overpredicts, and the places where it overpredicts are load-bearing. If the post-2016 press were simply an anti-Republican, pro-Democratic affirmation machine, it could not have produced the coverage it in fact produced of the Biden administration’s failures — and that coverage was, at key moments, savage. The chaotic withdrawal from Afghanistan in August 2021 was covered by the mainstream press as the debacle it was, with no evident partisan cushioning; the images of Kabul airport and the deaths at Abbey Gate were not softened. The border and inflation were covered as persistent liabilities for the administration, relentlessly. The Hunter Biden pardon in December 2024 — after the president had repeatedly and explicitly promised he would not issue one — drew editorial condemnation from the major boards as sharp as anything they had aimed at the other party. A pure audience-capture-for-Democrats model cannot explain any of this, and intellectual honesty requires holding it against the thesis rather than explaining it away. The refined thesis survives — the claim was always about asymmetry and sphere placement on politically-coded questions, not about uniform partisanship — but the refinement is a real constraint, and anyone who ignores it has stopped doing analysis and started doing advocacy.

Second, the asymmetry of the object was real and large, and it partially justifies an asymmetry of coverage. This bears repeating because it is the concession critics most often skip: 30,000-plus documented false or misleading presidential claims is not a normal baseline, an attempt to overturn a legitimate election through fake electors and pressure on officials is not a normal event, and January 6 is not a normal Tuesday. A press that had covered these with the same tone it used for ordinary political disagreement would have failed in the opposite direction — it would have normalized the genuinely abnormal, committed the “balance as bias” error at civilizational scale, and been rightly condemned for it. Some meaningful portion of the 80-percent-negative tone distribution is not instrument error; it is faithful measurement of an object that genuinely warranted overwhelmingly negative coverage. The honest critic’s claim is not that the asymmetry was wholly manufactured. It is that the manufactured component was real, was never measured or acknowledged, and — critically — swept up the contested and the true along with the genuinely false, because once a subject is filed under deviance the filing applies to everything associated with it indiscriminately.

Third, mechanical balance genuinely is a form of distortion, and the critics of “both-sidesism” are right about that. The Boykoffs proved it on climate. There are questions — the shape of the earth, the safety of vaccines against the diseases they prevent, whether a particular election was stolen — where “presenting both sides” actively misinforms, because one side is simply factually wrong, and the journalistic obligation is to say so. The renegotiation of objectivity was, in part, a response to a real defect in the old regime: the pretense that truth always lives at the midpoint between two claims is itself false, and sometimes dangerously so. The failure was not the recognition that some questions have a right answer the press should state. The failure was the promiscuous extension of that recognition to questions that did not have a settled right answer — the treatment of the contested (lab origin, youth gender medicine, the tradeoffs of a school policy) as though it were the settled (the shape of the earth). The tool was sound. It was applied to the wrong screws.

Fourth, most reporters, most of the time, wrote careful and accurate paragraphs. The measurement literature supports this: the bias, where it existed, lived predominantly in selection, timing, sphere placement, and correction prominence — the decisions around the sentences — not usually in the sentences themselves. This matters for fairness to individuals: the sweeping claim that “journalists are propagandists” is false and unjust to thousands of people doing careful work under difficult conditions. The structural critique is precisely a structural critique — it indicts a system of incentives and its aggregate output, not the integrity of the median reporter. Anyone who converts “the institution’s incentives reliably distort its aggregate coverage in a predictable direction” into “journalists are liars” has coarsened a true and subtle claim into a false and stupid one.

And fifth — the point that must never be dropped from the mouth of anyone making this argument in good faith, and the reason the substitution test is a mirror before it is a weapon: the exact same analysis, run on the right-leaning media ecosystem, returns a verdict at least as damning. Fox’s coverage of the 2020 election fraud claims was so detached from reality that the network paid $787.5 million to settle Dominion Voting Systems’ defamation suit rather than face a jury, and the discovery in that case revealed hosts privately deriding, in their own text messages, the very claims they were broadcasting to their audience as fact — a form of audience capture so pure it crossed into knowing falsehood, which is worse than anything documented on the other side. The right-leaning ecosystem submerged the 93 percent beneath the fires as reliably as the left-leaning one submerged the fires beneath the 93 percent. If this essay concentrates its fire on the mainstream, legacy, prestige press, it is not because that press is the worst offender in some absolute sense. It is because that press claims the mantle of objective, non-partisan, methodologically rigorous journalism — it asks to be trusted as the neutral arbiter — and the gap between that claim and the conduct is therefore the more consequential breach. Fox never promised to be the view from nowhere. The Times did. A broken promise is a graver thing than a promise never made.

Hold all five of these. They do not dissolve the indictment; the indictment was constructed to survive them, and the surviving core — asymmetric sphere placement on politically-coded questions, driven by a nameable mechanism, uncorrected by a profession that refused to audit itself — stands after every concession is granted. But an indictment that has genuinely metabolized its own strongest counter-case is a different and more dangerous thing than one that hasn’t. It can no longer be dismissed as the other tribe’s grievance, because it has already said, out loud, everything the other tribe would say in rebuttal, and it is still standing.

XI. The Confession: The Age Story

Before the laptop, one more case, because it does something none of the others quite do: it ends in the profession’s own signed confession, and the confession was delivered, without apparent irony, by some of the same people who had enforced the silence it confessed to. The story of President Biden’s cognitive decline, and the press’s coverage of it, is the closest thing this decade offers to a controlled demonstration of the entire thesis — misfiling, enforcement, collapse, and belated confession — with the added feature that the confessors indicted themselves.

For most of Biden’s presidency, questions about his acuity were filed, by much of the mainstream press, under deviance. To raise them was coded as a Fox News talking point, a right-wing smear, “cheap fakes” — a term the White House deployed and much of the press adopted — a genre of bad-faith attack rather than a legitimate line of inquiry into the capacity of the most powerful person on earth. Reporters who pursued the story encountered not only White House stonewalling but peer pressure: to cover Biden’s age seriously was to risk being sorted, by one’s own colleagues, into the wrong tribe. When the Wall Street Journal published a deeply-reported piece in June 2024 documenting, from numerous sources, concerns about Biden’s diminished capacity in private meetings, the story was met not with a collaborative rush to match and extend it — the normal response to a rival’s scoop on a matter of supreme public importance — but with a coordinated dismissal, from the White House and from much of the rest of the press corps, as a partisan hit job built on Republican sources. The sphere machinery again: a legitimate empirical question about presidential capacity, filed under deviance, enforced against the colleagues who tried to report it.

Then came the debate of June 27, 2024, and the sphere collapsed in ninety minutes of live television, in front of tens of millions of people, exactly as the Kenosha frame had collapsed in a single chyron. The reality the frame had been suppressing became undeniable because it was broadcast, unmediated, to the entire electorate at once. And the press pivoted — hard, fast, and without visible reckoning for the years of suppression that had preceded it. Within weeks the same institutions that had dismissed the Journal‘s June reporting were running relentless coverage of Biden’s fitness, coverage that contributed materially to his withdrawal from the race in July. The story that had been deviance in May was wall-to-wall in July, and the evidence had not changed in the interval — only the permission to see it had, granted by an event too public to spin.

And then, in 2025, the confession. Jake Tapper of CNN and Alex Thompson of Axios published Original Sin, a book documenting, in detail, the concealment of Biden’s decline — the managed schedules, the limited exposures, the inner circle’s protective apparatus — and, implicitly and sometimes explicitly, the press’s failure to penetrate it. The book was a bestseller and a sensation. But its deepest significance is in the identity of its authors and the arc of their own conduct. Tapper, accepting an award at the White House Correspondents’ Association dinner years earlier, had rebuked a Trump surrogate on air in 2020 for raising Biden’s cognitive state; he had been, in his own on-air conduct, part of the enforcement. Alex Thompson, accepting a press award in 2025 for his reporting on Biden’s decline, said from the podium, in front of the assembled Washington press corps: “We, myself included, missed a lot of this story, and some people that are still in this room, some people that are still on the payroll, were part of a massive cover-up.” Read that again. The confession was delivered to the profession, by a member of the profession, at the profession’s own annual celebration of itself, and it named the cover-up as ongoing and internal.

Here is why the age story is the decisive bridge to the laptop, and why it belongs immediately before it. It demonstrates, with the profession’s own signature at the bottom, every element of the thesis: a legitimate question misfiled as deviance; the misfiling enforced against colleagues who tried to report accurately; the enforcement sustained by tribal coding and White House pressure rather than by any honest assessment of the evidence; the sudden collapse of the frame when reality became too public to suppress; the pivot executed by the same personnel who had enforced the silence, without proportional reckoning; and finally, uniquely, an explicit confession — years late, delivered by the enforcers themselves, and monetized as a bestseller, which is its own dark comment on the incentives. The age story proves the mechanism is real, because the people inside it eventually said so. And it establishes the essay’s core distinction one final time before the climax: the failure was not the initial misjudgment — everyone misjudges — but the enforcement of the misjudgment against those who saw correctly, and the lateness and self-serving framing of the eventual correction. Lateness is the charge. It is the charge because it is the one the record proves, in the profession’s own hand, at the profession’s own podium.

Now, with every thread drawn — the renegotiation, the asymmetry, the template, the disciplined test, the virus, the summer, the culture war, the mechanism, the counter-case, and the confession — we can finally take up the story where all of them converge, and where the record is not an allegation, or a study, or a memoir, but a federal court exhibit.

XII. The Laptop, Part One: The Timeline the Record Now Fixes

We return to the strip mall in Wilmington, and to the machine the FBI carried out of it under subpoena on December 9, 2019. What follows is the chronology, and it must be laid down with precision, because the singular feature of the Hunter Biden laptop story — the feature that separates it from every other case in this essay and elevates it from error to something without precedent — is that nearly every load-bearing fact in it now rests not on journalism, not on anonymous sources, not on contested studies, but on sworn testimony and physical exhibits entered into evidence in a federal criminal trial. This is the one case where we do not have to argue about what happened, because a United States courtroom established it under the rules of evidence, with the penalties for lying attached.

The provenance chain: In April 2019, a water-damaged MacBook Pro was left at a computer repair shop in Wilmington, Delaware, under a signed work-authorization form containing a standard clause providing that equipment not reclaimed within ninety days would be treated as abandoned. It was never reclaimed. The shop’s owner, in the course of attempting to recover the failing drive’s data for the customer, encountered its contents. In the fall of 2019, alarmed, he brought the material to the attention of the FBI — his father, a retired Air Force colonel, initially making the approach to a field office. On December 9, 2019, FBI agents served a federal grand jury subpoena and took possession of the laptop and an external hard drive, leaving a property receipt. The subpoena issued from the office of the United States Attorney for the District of Delaware, David Weiss — a Trump appointee — whose investigation into the younger Biden’s taxes and finances had been open since 2018.

Every element of that paragraph is now confirmed in the trial record. At Hunter Biden’s federal gun trial in Delaware in June 2024, FBI Special Agent Erika Jensen testified, under oath and subject to cross-examination, that the Bureau had obtained the laptop by grand jury subpoena in December 2019, that its contents were consistent with data from Biden’s own Apple iCloud account, and that there was no evidence the device had been tampered with or its data manipulated. The Justice Department, prosecuting the president’s son, introduced the physical laptop into evidence and relied on data drawn from it to help prove its case. And the defense — Hunter Biden’s own lawyers, who had spent the preceding years insinuating in letters and civil filings that the laptop’s data had been “manipulated” — did not, in the courtroom, where frivolous claims carry sanctions and assertions must be backed by proof, contest the authenticity of the device or its data. When both adversaries in a criminal proceeding treat a piece of evidence as genuine, the authenticity question is not merely answered; it is closed by the most rigorous truth-finding procedure the society possesses.

That closure was not available only in 2024. It is essential to the indictment that the knowability long preceded the courtroom. Consider what was on the public record in October 2020, when the story broke and the suppression began. The subpoena and the property receipt — government documents — were referenced in the New York Post‘s original reporting. The Director of National Intelligence, John Ratcliffe, stated publicly on October 19, 2020, that the laptop was “not part of some Russian disinformation campaign.” The following day, October 20, the FBI wrote to Senator Ron Johnson stating that it had nothing to add to Ratcliffe’s characterization — a conspicuous non-contradiction from the one institution that actually possessed the device and had possessed it for ten months. Tony Bobulinski, a former business associate of Hunter Biden’s, went on the record — on camera, using his real name, describing specific business dealings and the family’s involvement — before the final presidential debate. And the cryptographic means of verifying a large tranche of the emails — the DKIM signatures we will come to — were sitting inside the files the entire time, available to anyone with the will and the technical competence to check them, which every major newsroom possessed.

So the timeline fixes two things at once. It establishes, beyond argument, that the data was authentic and the underlying probe real and predating any political actor. And it establishes that the tools to establish authenticity were available in real time, in October 2020, to any organization that chose to use them. The 2024 trial did not create the truth. It removed the last conceivable excuse for not having found it four years earlier. Which means the failure that occurred in October 2020 was not a failure of available information. It was a failure of will — a choice, made under the influence of the mechanism this essay has spent its length describing, not to look. And that is a different and graver category of failure than any we have examined, because in every prior case the press could at least claim the truth was genuinely uncertain at the time. Here, uniquely, it was checkable, and the checking was declined.

XIII. The Laptop, Part Two: The Machinery of Suppression

What happened when the story broke is best understood as the loaded template from the Russia years firing on a story shaped precisely like its trigger — with the crucial and catastrophic difference that this time the story was true. Recall the reflex the profession had built: damaging material about a candidate, of murky provenance, surfacing near an election, equals hostile foreign information operation, and our sacred duty — learned in the penance of 2016 — is to refuse to be used. Now observe how perfectly the laptop’s surface features fit the trigger, and how that fit pre-coded the entire response before a single fact was examined.

The story arrived through the worst possible courier. It came via Rudy Giuliani, the president’s personal attorney, who had spent the previous year openly seeking damaging material on the Bidens in Ukraine — the very conduct at the center of the first impeachment. Weeks before the laptop surfaced, the Treasury Department had sanctioned a Ukrainian member of parliament, Andrii Derkach, for running an influence operation that laundered anti-Biden material through American intermediaries including Giuliani. The material concerned Burisma and the Bidens — the exact subject of the impeachment that had consumed the prior year. Every surface feature screamed this is the trigger. The courier was a known operative in a foreign influence context; the subject was the impeachment’s subject; the timing was three weeks before an election. If you had designed a true story specifically to be mistaken for a Russian operation, you could not have improved on the laptop’s packaging. And the profession, primed by four years of the Russia template and the penance of 2016, read the packaging and never opened the box.

The suppression then proceeded through an apparatus that had not existed in any prior era — and cataloguing its components is the whole point, because it is the novelty of the apparatus that makes the episode unprecedented. Three layers operated in concert.

The platforms moved first and hardest. Twitter blocked all links to the New York Post‘s story — users could not share it, could not direct-message it; the URL was treated as malware — and then locked the Post‘s own account, the account of the country’s fourth-oldest newspaper, for roughly two weeks, under a “hacked materials” policy. Facebook, acting on generic advance warnings from the FBI about anticipated hack-and-leak operations, throttled the story’s distribution algorithmically while its own third-party review proceeded, suppressing its reach at the decisive moment. Jack Dorsey, Twitter’s CEO, would later call the blocking a mistake; Mark Zuckerberg would later describe, in his own account, the FBI warnings that had prompted Facebook’s throttling. But the mistake was made when it mattered — in the window before an election — and unmade only after, when unmaking it cost nothing.

The intelligence establishment supplied the frame. On October 19, 2020, fifty-one former senior intelligence officials — former CIA directors among them — signed a public letter stating that the laptop story “has all the classic earmarks of a Russian information operation.” The letter’s careful readers noticed what its headline readers did not: it explicitly conceded that the signatories had no evidence of Russian involvement. It was, on its face, a statement of suspicion by people with no access to the actual device, dressed in the authority of their former titles. But the hedge was invisible at the speed of a news cycle, and the press converted “these former officials, lacking evidence, suspect Russian involvement” into “intelligence officials say: Russian disinformation,” a settled descriptor, almost instantly. The letter was, in the most precise sense, laundered authority: it took a political judgment (this hurts our preferred candidate and should be discredited) and passed it through the prestige of the intelligence community to emerge as something that looked like an intelligence finding, which it explicitly was not. We later learned, from Michael Morell’s sworn testimony to the House Judiciary Committee, that the letter’s organization had been prompted by a call from the Biden campaign — from Antony Blinken, then a campaign adviser — and that Morell’s purpose in helping to organize it was, in his own testimony, to aid Biden in the upcoming debate and to give then-candidate Biden a talking point. The frame was not an independent assessment that the press dutifully reported. It was, per the sworn account of one of its own authors, opposition messaging wearing the borrowed uniform of the intelligence services, and the press wore the uniform’s authority without checking who was inside it.

The press itself then completed the circuit. The frame the platforms enforced and the officials supplied was ratified by the coverage. The story was widely characterized across mainstream outlets as “Russian disinformation” — not as an unverified story requiring investigation, but as a known falsehood requiring debunking. The public radio network’s managing editor publicly explained that the outlet would not devote resources to what it characterized as a distraction not worth its listeners’ time. A network anchor, interviewing the sitting president, told him flatly that the laptop material “can’t be verified” — a statement that was true only in the sense that the network had not tried to verify it, and false in the sense that verification was available. The story was filed, wholesale, into the sphere of deviance, and the debunking register engaged, and the reporters who might have opened the box instead spent their energy explaining why the box should not be opened.

And here is the asymmetry that transforms all of this from an ordinary misjudgment into the decade’s signature hypocrisy — the point where the substitution test, run against the profession’s own prior conduct, returns its most damning verdict. The chain-of-custody scruples, the provenance concerns, the sudden fastidiousness about “hacked materials” and unverifiable data that the press and platforms discovered in October 2020 — where had all of that rigor been during the Steele dossier? The dossier was an unverified opposition-research document, paid for by the opposing campaign, of exactly the murky provenance now cited to bury the laptop, and the press had mainlined it into years of coverage without a fraction of this care. The same institutions that laundered an unverified oppo document into the news when it damaged Trump refused to examine an authenticatable laptop when it damaged Biden. That is not a defensible application of consistent standards under difficult conditions. It is the selective deployment of a standard as a weapon — rigor summoned to suppress inconvenient truth and dispensed with to amplify convenient suspicion — and it is precisely the pattern that a decade of near-zero prominence ratios had trained the audience to expect. The laptop did not create the profession’s credibility problem. It proved the audience had been right about it all along.

XIV. The Laptop, Part Three: Authenticity, Forensics, and the Death of the Fabrication Theory

For the indictment to hold at full strength, one objection must be met head-on rather than waved past, because it is the objection the suppression rested on and the one a fair-minded reader will still be harboring: maybe the caution was warranted; maybe the data really was suspect. The forensic record answers this decisively, but answering it requires a distinction that the entire episode blurred and that professional practice keeps rigorously separate — a distinction worth teaching slowly, because it is the crux.

The authenticity of the data — was this really Hunter Biden’s email, his photographs, his text messages? — and the chain-of-custody integrity of a particular physical copy of a hard drive — can we certify that this specific disk image, having passed through these specific hands, was altered by none of them? — are two different properties, established by two different methods, and confusing them is the error at the heart of both the suppression and its later, sloppier defenses. They must be held apart.

Start with the data’s authenticity, because the evidence for it is overwhelming and comes from multiple independent directions. When the Washington Post finally commissioned its own forensic examination — published in March 2022, seventeen months after the story broke — its outside experts, Matt Green of Johns Hopkins and Jake Williams, a former NSA analyst, were able to cryptographically validate nearly 22,000 of the emails on the drive using DKIM signatures. DKIM — DomainKeys Identified Mail — is a system by which a sending mail server attaches a cryptographic signature to each message; the signature can later be checked against the server’s public key to prove the message is authentic and unaltered. Crucially, DKIM signatures survive copying: a validly-signed email remains provably genuine no matter how many times the file containing it is duplicated, because the proof is mathematical and travels with the message. This is why nearly 22,000 emails could be proven authentic even off a drive that had been handled carelessly — and it is why, within days of the original October 2020 story, an independent security researcher had already cryptographically validated the single most important email on the drive, the central Burisma “thank you” message, and published the verification. The tool that settled the question in 2022 existed, and was used by at least one competent outsider, in 2020. The newsrooms that possessed the same capability chose not to deploy it.

The DKIM validation is corroborated from other directions, which is how forensic confidence is properly built — not from one proof but from the convergence of independent ones. Ben Schreckinger, a Politico reporter, independently corroborated key emails through their recipients and, in one case, through a foreign government’s document release, and documented this in a 2021 book. CBS News commissioned its own forensic examination, published in November 2022, using a copy obtained directly from the shop owner — upstream of Giuliani — and its firm found the drive consistent with normal use, with no evidence of tampering. The iCloud match testified to at trial provides yet another independent line: the laptop’s contents aligned with data Apple produced to the government under a separate warrant. And then there is the most quietly conclusive corroboration of all, the one that requires no forensic expertise to grasp: Hunter Biden’s own legal conduct. His lawyers pursued a privacy lawsuit and sent letters to the Justice Department demanding criminal investigation of the individuals who had disseminated his data — a legal posture that is coherent only if the data is his. You cannot sue for the invasion of your privacy through the exposure of data you contend is fabricated. The possessive pronoun in his own lawyers’ filings concedes the entire question of authenticity.

Now the other property — the chain-of-custody integrity of the copies — because this is where honesty requires a genuine caveat, and where the caveat’s precise scope must be understood, because the episode’s tragedy lives exactly here. The Washington Post‘s 2022 examination, having validated the 22,000 emails, also reported that the large majority of the drive’s 217 gigabytes could not be independently verified — not because it showed signs of fabrication (the examiners found no evidence of that anywhere) but because the copies had been handled in ways that destroyed the ability to certify their integrity: the data had been accessed and copied repeatedly without write-blockers (devices that prevent any alteration during forensic copying), metadata was damaged or missing, and some files bore modification dates that postdated Hunter Biden’s possession of the laptop and even the FBI’s seizure of it — artifacts of the careless copying, not evidence of planted content, but fatal to certification nonetheless. The public corpus was, moreover, composite and messy: it descended from a technician’s manual salvage of a failing drive rather than a hashed forensic image; it passed from the shop owner to Giuliani’s associate Robert Costello, then to Giuliani and Steve Bannon, then to the Post, and separately to others — Jack Maxey, the group Marco Polo — who made further copies of copies; it included an encrypted iPhone backup that a third party later cracked with commercial tools; and public discussion intermittently blended in material from an entirely different Biden laptop that the DEA had seized from a doctor’s office in an unrelated matter.

So the honest formulation — the one that neither suppresses nor overclaims, and the one the press owed its readers in 2020 and never delivered — is this: the data was authentic, proven so by cryptography, corroboration, an iCloud match sworn to in federal court, and the subject’s own legal conduct; and separately, the specific nth-generation copies that circulated publicly had been handled so carelessly that no examiner could affirmatively certify every file on them, though no examiner found any file to be fabricated either. “Unverifiable,” properly understood, described the custody of certain copies, not the authenticity of the data — and the collapse of that distinction was the engine of the whole disaster.

Which brings us to the death of the fabrication theory, and to an irony that ought to be dwelt upon, because it inverts the entire moral structure of the episode. Notice whose theory required tampering. It was never the theory of the story’s defenders. The theory that the laptop had been fabricated, doctored, or planted was the implicit theory of the fifty-one officials’ letter (”classic earmarks of a Russian information operation”) and, later, the explicit theory of Hunter Biden’s own lawyers, who used the word “manipulated.” That theory — the suppression’s foundational premise — has now been tested for more than four years, by multiple independent forensic examinations, by adversaries on every side with maximal incentive to prove it, in a criminal proceeding where the defense could have raised it under oath and chose not to. The number of demonstrated forgeries produced by all of that scrutiny stands at zero. Not “few.” Zero. That is about as strong as negative evidence ever gets in the real world: a claim, held by powerful institutions, with every resource and incentive to substantiate it, tested exhaustively for years, yielding not one confirmed instance. The fabrication theory did not merely fail to be proven. It was affirmatively falsified by the totality of the forensic and legal record.

But — and this is the lesson that generalizes beyond the laptop into the whole practice of journalism — the residue of the sloppy custody is permanent, and it functioned as the suppression’s own life-support system. Because a handful of files on the circulated copies can never be affirmatively certified, the phrase “may be manipulated” retained a sliver of technical defensibility, and that sliver was enough to sustain the deviance framing for cycle after cycle. Every unhashed transfer, every copy made at a kitchen table without a write-blocker, every careless handling by Giuliani’s circle bought the dismissers another news cycle of “unverified.” The poor custodianship of the material became, in effect, the story’s own suppressor — the mechanism by which a true story was kept unusable. And that is a media-ethics lesson of the first order, one that indicts the story’s clumsy handlers alongside its motivated suppressors: how you handle true material determines whether the truth is usable. A true story delivered through a contaminated chain of custody, by a compromised courier, into a press primed to disbelieve it, can be true and buried at the same time. The laptop was.

Run the substitution test on the caveat itself, to be sure it is principle and not special pleading — because the test’s final discipline is that it must be applied to one’s own reasoning, not only to one’s opponents’. If a drive belonging to Donald Trump Jr. had spent a year being copied hand-to-hand through a Democratic operative’s circle before publication, every fair analyst would demand exactly this asterisk on the unverified portions — and would be right to. The asterisk is not hedging; it is the consistent application of forensic standards regardless of whose data is at stake. But — and here the stipulation that governs this essay’s conclusion takes hold — the asterisk shrinks to a footnote, and then to nothing of real consequence, the moment an honest observer registers what the courtroom established: that the FBI held the laptop from December 2019, that the Bureau raised no objection to the validity of its data, and that the Justice Department itself put that data before a jury as evidence in a federal criminal trial. Once the government’s own prosecutors are entering the material into evidence against the president’s son, the fastidious concerns about provenance that justified two years of suppression are revealed as what they always were: not a standard, but a pretext.

XV. The Laptop, Part Four: The Third Ring, and the Problem of the State

There is a layer to the laptop story that most media criticism omits, and omitting it flatters the press by making the failure smaller than it was. Without this layer, the laptop is a two-institution problem: the press and the platforms failed. With it, there are three concentric rings of failure, and the innermost ring is the state itself — which changes the episode from a media scandal into something closer to a civic one, and which is necessary to understand before the final judgment, because it both deepens the indictment and complicates the assignment of blame.

For the ten months between the FBI’s seizure of the laptop in December 2019 and the Post‘s publication in October 2020, one institution in America possessed ground truth. The federal government had the physical device. Its own investigators had, according to the later congressional testimony of IRS case agents, treated the data as authentic — matched to Biden’s Apple account — well before the story became public. When the decisive moment arrived — the morning of October 14, 2020, when Twitter’s head of site integrity, Yoel Roth, reached out to his FBI contacts to ask, in effect, is this thing real? — the government that had held the answer for ten months gave, in substance, no comment.

That silence did not stand alone. It coexisted with a season of affirmative shaping by the same government that was declining to confirm. Roth would later state, in a sworn declaration to the Federal Election Commission, that in the run-up to the election the FBI had warned the platforms to expect a hack-and-leak operation and that these warnings had specifically primed him to anticipate a leak involving Hunter Biden. FBI agent Elvis Chan, deposed in the Missouri v. Biden litigation over government-platform coordination, confirmed the cadence of federal warnings to the platforms in the relevant period. In September 2020, a month before the story broke, the Aspen Institute convened a tabletop exercise — a scenario-planning war game — that rehearsed, with platform executives and journalists in the room, the handling of a hypothetical leak of damaging material about “Hunter Biden” involving Burisma. And in August 2020, the FBI had given a defensive briefing to Senators Ron Johnson and Chuck Grassley suggesting that their own Burisma-related inquiry risked being a vehicle for Russian disinformation — a briefing the senators protested in writing as gratuitous and improper, an attempt to chill a legitimate congressional investigation.

Assemble these and a disturbing shape emerges. The one institution with access to the truth was silent about the specific (is this laptop real?) while being voluble about the general (expect a Russian hack-and-leak; here is a war game rehearsing exactly this scenario; beware this line of inquiry). Silence about the specific combined with volubility about the general is not neutrality. It is the manufacture of a frame. By declining to confirm what it knew to be true while broadcasting the expectation that something like it would be false, the government did not merely fail to correct the “Russian disinformation” narrative — it supplied the narrative’s plausibility, and then withheld the one fact that would have collapsed it.

But intellectual honesty requires the strongest version of the state’s defense, and it is genuinely strong — strong enough that the final judgment must be carefully bounded. Justice Department policy prohibits confirming or denying the existence of investigations, for excellent reasons that protect the innocent. And the sixty-day convention — the norm against overt investigative steps that could affect an election in its final stretch — exists precisely to prevent the state from putting its thumb on the electoral scale. The cautionary tale is fresh and points the other way: James Comey’s October 2016 letter reopening the Clinton email inquiry is widely regarded as an improper intervention that may have altered an election, and it was condemned by many of the very people who, four years later, wanted the FBI to speak up about the laptop. So the honest formulation is not that the FBI should have confirmed the investigation three weeks before the election — doing so would have meant the government publicly announcing a criminal probe touching the challenger’s family on the eve of the vote, the exact sin of 2016 run in the opposite partisan direction. The honest formulation is narrower and, I think, unanswerable: an agency that cannot confirm also must not shape. The no-comment was defensible. The priming — the warnings, the war-gaming, the defensive briefing that chilled a Senate inquiry — was not. And the combination of the two was the worst of both worlds: the government secured all the benefits of suppression (the “Russian disinformation” frame took hold) while claiming the protective cover of investigative silence (we can neither confirm nor deny). It got to shape the outcome without owning the intervention.

A necessary caution against overreach, because the difference between analysis and conspiracy theory is precisely the willingness to draw it: none of this establishes a single coordinated plot with a mastermind. The more sober and more likely reading is institutional — norm-following (the genuine no-comment policy) colliding with bureaucratic drift and the same anticipatory-suppression reflex that had captured the press after 2016, producing an emergent outcome that no one fully designed but from which a particular candidate benefited. The IRS whistleblowers’ allegations of a slow-walked investigation are sworn but contested. David Weiss’s later elevation to special counsel sits ambiguously between vindication (an independent prosecutor was empowered) and containment (the elevation came late, after the statute of limitations had lapsed on certain charges). The responsible analyst states the pattern and its most likely institutional explanation without inflating it into a grand design, because inflating it would be to commit, from the other direction, the exact sin this essay indicts: fitting the evidence to the narrative the analyst wishes to prove. The evidence supports institutional failure and improper shaping. It does not, on the public record, support a proven conspiracy, and saying so is the discipline the whole essay has been arguing for.

And now the dramatic irony that can be stated without a shred of overclaim, because every element of it is on the record — the sentence that contains the entire decade in miniature: throughout the first impeachment of Donald Trump — a proceeding whose central charge was that the president had improperly sought a Ukrainian investigation into Burisma and the Bidens — the United States Department of Justice was holding, under grand jury subpoena, in an active criminal investigation, a laptop belonging to Hunter Biden containing Burisma-related material. The two stories were not merely related; they were physically entangled, the arms of the experiment run into each other by history itself, which is why the party-swap counterfactual the substitution test would ordinarily invite is uniquely confounded here — you cannot cleanly reverse the parties in a scenario where the impeachment and the laptop were about the same subject at the same time. Run the substitution test on that entanglement, hold it in the mind alongside the fact that the same Comey-precedent norms a reversed-party critic would invoke are the very norms that make the FBI’s public silence the defensible part of the story, and the episode’s full, vertiginous complexity comes into view. This was not a simple case of bias. It was a case in which bias, institutional norm-following, a compromised courier, careless custody, a novel censorship apparatus, and a genuinely hard set of countervailing principles all combined to bury a true story — and the burial held, through an election, and was corrected only later, quietly, and never with the reckoning it demanded.

XVI. Why the Laptop Is Unprecedented

It is tempting, and common, to file the laptop alongside the other great press failures of the century and call it one more entry in a long ledger. This is a category error, and correcting it is the essay’s central analytic claim about the case. The laptop is not merely a large failure. It is a different kind of failure from almost everything that preceded it, and the difference is the thing that makes it, in a specific and defensible sense, without precedent.

Compare it to the failure that remains, in raw magnitude of consequence, the graver one: the coverage of Iraqi weapons of mass destruction in 2002 and 2003. The WMD failure helped launch a war that killed hundreds of thousands of people; nothing about the laptop approaches that scale of consequence, and any honest ranking by human cost puts Iraq far above it. The Gulf of Tonkin coverage that helped escalate Vietnam belongs in the same graver-by-consequence tier. So the claim is emphatically not that the laptop was the most damaging press failure in American history. It was not.

The claim is about kind, and here the distinction is sharp and, I think, decisive. The WMD failure — and Tonkin before it — was a failure of credulity. The press amplified official falsehood; it was too trusting of power, too willing to relay the government’s claims as fact, insufficiently skeptical of authority. This is the classic form of press failure, the one the profession’s entire apparatus of skepticism is designed to prevent, the one every journalism student is taught to guard against: the danger is that you will believe the powerful when you should have doubted them. Virtually every catalogued failure of American journalism, from the Maine to the Gulf of Tonkin to the WMD, fits this template. The press’s characteristic sin is excessive belief in official sources.

The laptop was the inverse. It was a failure not of credulity but of organized incredulity — a coordinated refusal to believe a true story, an excess of skepticism aimed with precision at an accurate account because that account was politically inconvenient and arrived wearing the costume of the thing the profession had sworn never to be fooled by again. Where WMD was the press believing a falsehood, the laptop was the press disbelieving a truth — and disbelieving it not through passive error but through active enforcement, an apparatus assembled to prevent a true story from reaching the public. That inversion is the first sense in which the episode is unprecedented: it is a failure mode the profession’s defenses were not built to catch, because those defenses were all designed to prevent the opposite failure. Skepticism is the immune system of journalism; the laptop was an autoimmune disorder, the immune system attacking healthy tissue.

The second sense is the apparatus itself, which had no precedent because its components did not previously exist. The WMD failure was executed with the ordinary tools of journalism — credulous sourcing, front-page placement, insufficient challenge to official claims. The laptop suppression required, and deployed, an infrastructure that was historically new: platform throttling, the technical capacity of two or three private companies to make a specific true story physically unshareable across the networks where most Americans now encounter news — a capacity that simply did not exist in 2003; anticipatory government briefings, a channel of coordination between the security state and the platforms, priming the distributors of news to expect and suppress a category of story before it appeared; and a former-officials letter functioning as laundered authority, the conversion of a political judgment into apparent intelligence through the borrowed prestige of the security services. None of these instruments was available to the press or the state during any prior failure. The laptop is the first great press failure of the platform age, and it took a shape that only the platform age made possible: not the amplification of a falsehood through mass media, but the suppression of a truth through the chokepoints of networked distribution, coordinated across private platforms and public agencies, and ratified by a press that supplied the interpretive frame.

And the third sense — the one that matters most for the profession’s capacity to recover, and the one that carries us to the final judgment — concerns the reparability of the two kinds of failure. This is not an abstract point; it is the practical heart of why the laptop is worse than a larger error. When the press fails by amplification — when it relays a falsehood — the failure is, at least in principle, repairable, because there is a false thing to retract. After the WMD coverage collapsed, the institutions could and did perform the repair ritual: the New York Times published its extraordinary editors’ note in May 2004 acknowledging its flawed prewar reporting; the Washington Post ran Howard Kurtz’s long autopsy of its own coverage. There was a specific false story, it could be named, the correction could be attached to it, and the record could be — imperfectly, incompletely, but genuinely — set straight. Amplification failures leave a scar you can point to and treat.

Suppression failures are different, and this is the hinge of the entire argument, so it must be stated with full weight: suppression is the harder failure to repair, because there is nothing to retract — no false story to correct, only an absence. When the failure is that a true story was kept from the public, there is no false article to append a correction to, no fabricated claim to walk back, no discrete error to name and disown. There is only the negative space where the story should have been and wasn’t — the reach that was throttled, the account that was locked, the frame that was relayed, the box that was never opened. You cannot retract an absence. You cannot publish an editors’ note that says “in October 2020, we did not adequately report a true thing, and we algorithmically and editorially prevented others from encountering it.” The machinery of journalistic correction is built entirely around the amplification failure — around the false thing that can be named and retracted — and it has no mechanism, none, for the suppression failure, for the true thing that was buried. Which is precisely why the laptop was never really corrected in the way WMD was corrected. The authentication, when it finally came in 2022, arrived buried in the middle of stories rather than announced; the public radio network never revisited its on-air dismissal; and no major newsroom has ever published, on itself, the kind of institutional editors’ note that the WMD failure eventually produced. Not because the laptop failure was smaller — in its implications for the integrity of an election, it was arguably larger — but because the kind of failure it was does not fit the kind of correction the profession knows how to perform. There was no false story to disown. There was only a true one that had been buried, and you cannot hold a ceremony to un-bury something. The absence just persists, uncorrected, because the profession has no ritual for it.

XVII. The Invoice, and the Partial Repair

Every breach has a price, and the price of a decade of asymmetric sphere placement, premature closure, and uncorrected suppression is recorded with unusual clarity in a single set of numbers, because the profession’s standing with the public is one of the most heavily surveyed quantities in American life. Gallup has asked Americans about their trust in the mass media “to report the news fully, accurately and fairly” for half a century, and the trend line is a straight descent into the basement. Overall trust — those saying they have a “great deal” or “fair amount” of confidence — has fallen to around a third of the public, near the lowest in the history of the measurement. Among self-identified Republicans it has collapsed into the low double digits, and in some recent readings has touched single digits. The partisan gap — the difference between how much Democrats and Republicans trust the press — is the widest ever recorded. And the share of Americans expressing no trust at all — not merely “not very much,” but none — has grown to rival the share expressing a great deal, an astonishing figure for an institution whose entire social function depends on being believed.

These numbers are the invoice. They are what a decade of near-zero prominence ratios costs. They are the compounded interest on every question misfiled into deviance and quietly refiled later without acknowledgment, on every true story buried and never un-buried, on every correction printed in a subordinate clause on an inside page. Correlation is not proof of causation, and the trust decline has multiple contributing causes — a hostile president who spent years attacking the press as “the enemy of the people,” a fragmented media environment that lets citizens curate their own realities, a general collapse of institutional trust across American life that touches churches and universities and Congress as surely as it touches newsrooms. All of that is real and all of it matters. But the trajectory tracks the conduct, the collapse is sharpest in precisely the demographic most alienated by the asymmetric sphere placement, and the profession’s own most senior insiders — Bennet foremost — have identified the audience-capture dynamic as a direct driver. The press did not lose the public’s trust by accident, or solely through its enemies’ attacks. It spent that trust, deliberately, on the affirmation of the audience it chose to keep.

And yet the story does not end in unrelieved indictment, because there is a partial repair underway, and honesty requires reporting it as scrupulously as the failures — indeed, reporting it is itself an application of the essay’s own standard, the standard the essay accuses the profession of abandoning. The signs of correction are real. A.G. Sulzberger, the Times‘ publisher, wrote a long and serious essay in the Columbia Journalism Review in 2023 defending journalistic independence explicitly against the pull of advocacy — a direct, if implicit, acknowledgment that the pull was real and had to be resisted. Joe Kahn, the Times‘ executive editor, has stated publicly and pointedly that the paper must not become an instrument of any political party, even in the service of defeating a candidate its staff and readers may fear — a remarkable thing for an editor to feel he must say, and therefore evidence of what he was pushing back against. The Times held firm against the internal revolt over its gender-medicine coverage. The origins-of-Covid question was reopened and is now covered as the genuinely unresolved matter it always was. And the age-story confession, however late and however monetized, was a confession — the profession did, in the end, say out loud that it had failed, from its own podium, in its own house.

The very existence of the CJR autopsy of Russiagate, of Bennet’s essay, of Beyond Objectivity, of Original Sin — the fact that the most damning documentation of the press’s failures was produced by the press, about the press, in the profession’s own most serious venues — is itself the strongest evidence that the institution retains the capacity for self-correction that this essay accuses it of abandoning. The immune system that attacked healthy tissue in October 2020 has, at least intermittently, begun to recognize the disorder. That capacity is precisely what distinguishes an institution that has failed from an institution that is failing — the first can recover, and the mechanism of recovery is exactly the ruthless self-examination that produced the documents this essay has relied on. Every source in this indictment is, in a sense, a sign of hope: the profession that produced them is not incapable of interrogating its premises. It is capable. It simply, for a decade, on a specific class of politically-coded questions, chose not to — and is now, unevenly and incompletely and under the duress of collapsed trust, beginning to choose otherwise.

XVIII. The Charge, Restated

So return, at the end, to the question with which we began, and state the verdict precisely, because precision was the whole point — precision is the discipline whose absence is the crime, and a verdict rendered carelessly would convict the essay of the very thing it prosecutes.

Have American journalists abdicated their responsibility to interrogate their premises, identify their biases, and set them aside in their reporting? The caricature answer — that the press became a conscious propaganda arm, that newsrooms adopted a “stop the fascists” mission and bent every fact to it — is false, and believing it makes you stupider about the thing you’re trying to understand, because it cannot survive contact with the Afghanistan coverage, the Journal‘s null result on the laptop, the savaging of the Biden pardon, or the profession’s own capacity to produce the documents that damn it. That answer should be retired by anyone serious.

But the disciplined answer is yes, and it survives every cross-examination this essay could mount against it. Not that journalists stopped seeking truth, but that they stopped interrogating where they had drawn the boundaries of the thinkable — stopped noticing that the placement of a claim into consensus, controversy, or deviance is the most consequential editorial act they perform, stopped auditing that placement, and reclassified the act of auditing itself as a kind of complicity with the enemy. The failure was asymmetric by direction, driven by a nameable mechanism — the homogeneity of the newsroom, the collapse of the local ballast, and above all the subscription economics that reward the intensification of a captured audience over the informing of a broad public — and, most damningly, uncorrected: the errors were refiled quietly and late, at a fraction of the prominence of the original certainties, by a profession that for a decade would not perform on itself the ruthless examination it exists to perform on everyone else. Lateness was the charge throughout, because lateness is the charge the record proves, in the profession’s own hand, in CJR and The Economist and Original Sin and from the podium of the Correspondents’ Dinner.

And the laptop is where the charge is proven past any reasonable doubt, because the laptop is the case the profession cannot explain away with the excuse it can offer for all the others. For the virus, for the summer, for the age story, the press can at least plead genuine uncertainty at the time — the truth was contested, the evidence was incomplete, hindsight is unfair. For the laptop, that plea is unavailable, because the truth was checkable in real time and the checking was declined: the subpoena was public, the DNI had spoken, the cryptographic proof sat in the files, the on-the-record witness was on camera, and the one institution that possessed the physical device raised no objection to its data while the press told the country it was Russian disinformation. And then the Justice Department entered that same data into evidence, in a federal courtroom, against the president’s own son, with an FBI agent swearing to its authenticity and the defense declining to contest it — closing, by the society’s most rigorous truth-finding procedure, the question the press had spent two years pretending was open.

The laptop story was genuinely without precedent — organized incredulity toward a true story, suppression rather than amplification, executed through an apparatus that didn’t exist before: platform throttling, anticipatory government briefings, a former-officials letter functioning as laundered authority. Suppression is the harder failure to repair, because there is nothing to retract — no false story to correct, only an absence. And here is the part that vindicates the phrase “breach of trust”: after the amplification failures of the past — the WMD, the credulous relaying of official falsehood — the institutions eventually performed the repair ritual, named the false thing, and disowned it. After the laptop, there was no ritual, because there was no false thing to name — only a true thing that had been buried, and the profession has no ceremony for disinterring truths. The authentication came quietly, mid-story, years late; the dismissals were never revisited; no editors’ note has ever run.

When the media fails to ask a question it committed an error. When the media buries a true story through a novel machinery of suppression, and then declines even the belated, buried, half-hearted correction it managed after its prior failures of credulity then its judgment cannot be trusted—precisely when trust matters most: in the weeks before a national election, on a true story about a candidate, when the whole reason the institution is granted its extraordinary freedoms is to tell the public what it needs to know when it needs to know it.

That is the breach. It is not that they got it wrong. It is that they got it wrong in the one direction their defenses were built to prevent, buried a truth through means that had never before existed, and then would not do for the buried truth even the imperfect penance they had once done for the amplified lie — and the ledger of that refusal is written, in Gallup’s straight descending line, in the trust of a public that concluded, not without evidence, that it had been right to stop believing.

When a request hits an SFCC instance, the platform resolves which file to execute by walking the cartridge path left to right, looking for the first match at the requested file path. This happens at the application server level, per request, at runtime.

Concretely, say the cartridge path is:

app_custom_sharkninja : int_payment_adyen : int_analytics : app_storefront_base

A request to /Cart-Show triggers the runtime to look for controllers/Cart.js in this order:

1. app_custom_sharkninja/cartridge/controllers/Cart.js — found? Use it. Stop.

2. int_payment_adyen/cartridge/controllers/Cart.js — found? Use it. Stop.

3. int_analytics/cartridge/controllers/Cart.js — found? Use it. Stop.

4. app_storefront_base/cartridge/controllers/Cart.js — found? Use it. Stop.

There is no manifest, no import statement, no dependency declaration, and no build step that records which cartridge won. The resolution is the filesystem lookup at request time. The platform doesn’t log which cartridge provided the file unless you enable request-level tracing (which has its own performance cost and isn’t typically on in production).

Why “Implicit” Is the Precise Problem

No static analysis is possible. In a Node.js or Java project, when module A imports module B, that relationship is visible in source code. A tool (compiler, bundler, IDE, linter) can trace the dependency graph, detect circular dependencies, flag missing modules, and tell you exactly which code will execute when a function is called. In the cartridge model, the “import” is the absence of a file in a higher-priority cartridge, causing fallthrough to a lower-priority one. You can’t grep for this. You can’t write a lint rule for it. The dependency is expressed as “this file doesn’t exist here, so the platform will find it somewhere else” — a negative assertion that no static tool can reason about without reconstructing the cartridge path resolution logic externally.

The override is invisible at the call site. When app_storefront_base/controllers/Cart.js calls a helper function via require('*/cartridge/scripts/helpers/cartHelpers'), that * wildcard is SFCC’s module resolution syntax — it means “resolve this through the cartridge path.” The controller author doesn’t know (and can’t know from reading the code) whether cartHelpers will come from app_storefront_base, app_custom_sharkninja, or int_payment_adyen. The behavior depends entirely on which cartridges are in the path and in what order, which is configuration state in Business Manager — not in the codebase. A developer reading the controller in their IDE sees a require statement that resolves to a conceptual path, not an actual file. They must mentally (or manually) walk the cartridge path to know which implementation they’re calling.

Template resolution compounds the problem. ISML templates use <isinclude template="components/header/pageHeader"/> — again, resolved via cartridge path at render time. A controller in one cartridge can render a template that physically lives in a different cartridge, which itself includes a decorator from a third cartridge. The rendering chain crosses cartridge boundaries invisibly. When the page renders incorrectly, you’re reverse-engineering a call chain that was never declared anywhere — it emerged from the cartridge path ordering and the filesystem layout across multiple directories.

The module.superModule pattern makes it worse. SFRA introduced module.superModule as a way to extend a controller rather than fully replacing it. The pattern looks like:

// In app_custom_sharkninja/cartridge/controllers/Cart.js
var base = module.superModule;
server.extend(base);

server.append('Show', function (req, res, next) {
    // Custom logic layered on top of the base Cart-Show
    var viewData = res.getViewData();
    viewData.customAttribute = 'something';
    res.setViewData(viewData);
    next();
});

module.exports = server.exports();

This looks clean in isolation. The problem is that module.superModule resolves to “the next cartridge in the path that has this file.” If int_payment_adyen also has a Cart.js that uses module.superModule, you now have a chain: app_custom_sharkninja → int_payment_adyen → app_storefront_base, where each layer appends or prepends behavior to the route handler. The execution order of the append/prepend hooks depends on the cartridge path order. Reorder the path — which might happen because a different integration requires a specific position — and the execution sequence of these hooks changes silently. There’s no test that catches this without exercising the full request path end-to-end on an instance configured with the exact production cartridge path.

Business Manager configuration as invisible coupling. Beyond code, cartridges depend on Business Manager site preferences, custom object definitions, service configurations, and content slot configurations. These are mutable state in the BM database, not declared in cartridge code. A vendor cartridge’s README might say “create a site preference called AdyenMerchantAccount with value X” — but that preference lives in BM, not in the cartridge. If someone changes it, no version control system records the change. If a different cartridge reads the same preference (or one with a naming collision), there’s no encapsulation. The cartridge model provides no scoping mechanism for configuration — everything is global, everything is mutable, and the only record of who depends on what is documentation that may or may not exist.

What “Compile-Time Visible” Looks Like by Contrast

In a typed service architecture, dependencies are explicit at every level:

// OrderService depends on InventoryService — declared in code, visible to compiler
import { InventoryClient } from '@sharkninja/inventory-client';

export class OrderService {
  constructor(private inventory: InventoryClient) {} // injected, testable, mockable

  async placeOrder(cart: Cart): Promise<Order> {
    const reserved = await this.inventory.reserve(cart.lineItems);
    if (!reserved.success) {
      throw new InsufficientInventoryError(reserved.failures);
    }
    // ...
  }
}

The dependency on InventoryClient is in the source code. The TypeScript compiler verifies the interface contract at build time. If InventoryClient changes its method signature, every consumer fails to compile. The IDE shows you exactly which implementation you’re calling. A dependency graph tool (like madge or nx graph) can map every service-to-service dependency in the system. There is no hidden resolution path, no ordering-dependent behavior, no runtime surprise.

At the API boundary between services, the contract is an OpenAPI spec or a protobuf definition — versioned, diffable, and enforceable via contract testing (Pact, Buf breaking-change detection for proto). If the Inventory Service changes its response schema, the contract test fails in CI before the change reaches production. The equivalent scenario in SFCC — a vendor cartridge changing its helper function signature — is discovered when the storefront throws a runtime error on a customer’s browser.

Why the Service Architecture Is More Maintainable at Year Three

The claim isn’t that microservices are inherently simpler — they’re not. The claim is that the debt profile is different, and the service architecture’s debt is manageable in ways the cartridge model’s debt is not.

Debt Visibility

In a service architecture, when something breaks, the blast radius is traceable. A distributed trace (OpenTelemetry) shows you: request entered the BFF, BFF called Catalog Service (200, 45ms), called Pricing Service (200, 12ms), called Inventory Service (500, timeout). You know exactly which service failed, which API call failed, what the request payload was, and what the error response was. The debugging path is linear: find the failing service, read its logs, identify the root cause.

In the cartridge model at year three, when something breaks: a page renders incorrectly. Which cartridge’s controller handled the request? Check the cartridge path. Which template rendered? Walk the <isinclude> chain across cartridges. Which helper function returned the wrong data? Trace the require('*/...') resolution across the cartridge path. Did a Business Manager configuration change? Check the audit log (if it exists and if the relevant configuration is audited). Did a vendor cartridge update change a shared helper’s behavior? Diff the vendor cartridge against the previous version, if you kept a copy. Each debugging session is an archaeology exercise in a codebase where the execution path was never explicitly declared.

Change Isolation

Service architecture year three scenario: you need to change how promotional pricing is calculated. You modify the Pricing Service. It has its own repository, its own test suite, its own deploy pipeline. You run its unit and integration tests, deploy to staging, run contract tests against its consumers (BFF, Cart Service), validate, and deploy to production. The Catalog Service, Inventory Service, and Order Service are untouched — they don’t know pricing changed because the API contract didn’t change.

Cartridge model year three scenario: you need to change how promotional pricing is calculated. The pricing logic is in app_storefront_base/cartridge/scripts/helpers/pricing.js, but you overrode it in app_custom_sharkninja two years ago. The vendor cartridge for the loyalty program also hooks into pricing via a module.superModule chain. Changing the pricing logic means understanding the three-cartridge chain, testing the interaction between your override and the loyalty cartridge’s extension, and deploying the entire cartridge stack as a unit (because there’s no independent deployment of a single cartridge — you activate a code version that includes all cartridges). The blast radius of a pricing change is the entire storefront.

Team Scaling

Service architecture: at year three, you have 12 engineers. You can assign 2-person teams to own specific services. The Cart team owns cart logic, cart tests, cart deploys. The Search team owns the search index, search relevance tuning, search deploys. Teams work in parallel with minimal coordination because the interface contracts are the coordination mechanism. A team can refactor their service’s internals without coordinating with other teams as long as the API contract holds.

Cartridge model: at year three, you have 12 engineers. They all work in the same cartridge codebase (or a small set of cartridges). Every feature branch potentially touches controllers, templates, and helpers that other branches also touch. Merge conflicts are structural, not incidental — because the override model means customization concentrates in the same set of overridden files. The Cart.js controller that three teams need to modify is a serialization point. You either take turns (slow) or merge frequently (conflict-heavy). There’s no way to decompose the work into independent streams because the cartridge model doesn’t provide real modularity boundaries — it provides a file-override mechanism that collapses into a single execution context at runtime.

Upgrade Path

Service architecture: at year three, you need to upgrade the search engine from Elasticsearch 7 to OpenSearch 2. You modify the Search Service’s infrastructure and its client library. Reindex, validate relevance, deploy. No other service is affected because they consume search results via the Search Service’s API, not via a direct Elasticsearch dependency.

Cartridge model: at year three, Salesforce releases SFRA 7.x with security patches and new features. You’re on SFRA 5.x because the 6.x upgrade was deferred twice — it required re-merging 40+ overridden files, two vendor cartridges were incompatible with the new version, and the QA cycle for a full regression was estimated at three weeks. You’re now two major versions behind the reference architecture, accumulating security exposure and missing platform features. Each deferred upgrade makes the next one harder because the diff between your overrides and the new base grows monotonically. This is the specific debt compounding mechanism that makes the cartridge model unsustainable at scale — the cost of staying current increases over time rather than staying constant.

Failure Mode Comparison

At year three, the service architecture’s failure mode is operational complexity — more services to monitor, more deploys to coordinate for cross-cutting changes, more infrastructure to maintain. These are real costs, but they’re visible, measurable, and addressable with tooling and process (better observability, deploy automation, service mesh).

The cartridge model’s failure mode is invisible coupling and frozen architecture — the team can’t confidently change the system because the interaction effects between cartridges are undocumented and only discoverable at runtime. The response is organizational: change velocity drops, estimates inflate with “impact analysis” buffers, the platform becomes something the team maintains rather than evolves. New features route around the core platform (”let’s build that in a microservice that calls OCAPI”) rather than extending it, creating a shadow architecture that further increases the total system complexity while leaving the cartridge debt untouched.

That shadow architecture pattern — where the team builds new capabilities outside the platform because modifying the platform is too risky — is the clearest signal that the cartridge model’s debt has reached a tipping point. And it’s extremely common in mature SFCC implementations. You end up with the worst of both worlds: a monolithic cartridge stack you’re afraid to touch, plus a constellation of external services that duplicate state and logic because they can’t safely extend the monolith.

The Core Asymmetry

The service architecture’s costs are upfront and linear — you pay to build each service, you pay to operate each service, and those costs are roughly constant per service over time.

The cartridge model’s costs are deferred and exponential — initial development is fast (override a file, add a cartridge, done), but the cost of every subsequent change increases as the override graph grows, vendor cartridges accumulate, and the distance from the base reference architecture widens. By year three, the accumulated cost often exceeds what the service architecture would have cost to build and operate from scratch — and unlike the service architecture, there’s no clean path to reduce it without a replatform.

That’s the structural argument. The cartridge model doesn’t just create debt — it creates debt with a compounding interest rate that accelerates as the system matures.

It’s April 1986, high radiation levels set off alarms at a nuclear power plant in Sweden. The Swedes quickly determine that the radiation most likely originated in Ukraine. After several weeks of denials, the Soviets admit that a large meteor with an unusually large quantity of uranium has crashed near the city of Pripyat in the north of the country.

Coincidentally, this is near the Chernobyl Nuclear Power Plant. The Soviets, of course, deny that there’s been an accident at the plant, but won’t allow outside observers to examine the plant or any evidence from a meteor nor to question or examine any of the victims of the impact.

Western media organizations dutifully report the Soviet version of events, especially after Ronald Reagan, the U.S. President, floats a “nuclear meltdown” theory. Respected nuclear scientists in the west also publish letters of support for the Meteor Theory, advancing the implausibility of a nuclear meltdown scenario at a power plant, and that any release of radiation must have come from a natural source.

Meanwhile, Republicans in the U.S. government advance the possibility that there indeed was a nuclear disaster at Chernobyl. Some of the more conspiratorial ones claim there was a military weapons program at the plant which went awry.

At the same time, Democrats, media companies, and elite institutions refer to such claims of a meltdown at Chernobyl as “racist conspiracy theories”, “anti-science”, and “disinformation”. While TV and Radio broadcasts ban reporting on the Chernobyl Leak Theory.

Thousands of people in Europe subsequently die of radiation poisoning, governments mandate lockdowns to keep people in-doors while medical researchers try to come up with a vaccine for the radiation poisoning. While skeptical protestors are ridiculed as “cherbydiots.”

Of course the events depicted above did not happen, and the Russians copped to an accident at the Chernobyl power plant the same day the Swedes detected the radiation. While an adversarial Western media challenged the Russians on their lack of transparency on the accident.

But just imagine if the West had responded in such a way to claims by the Russians of a “meteor impact.” In other words, imagine that in 1986 Western media would have responded the way they did during the Covid outbreak of 2020: taken the side of a totalitarian government in its PR war to essentially cover-up a man-made disaster, especially after that government banned access to the site of the disaster, any evidence or material from there, or any of its victims.

It’s hard to do so.

                                                                               1.

At first glance, it’s easy to see the immediate appeal of a natural-origin theory for SARS-CoV-2 (Covid): it’s uncomplicated. The natural world is host to a multitude of deadly pathogens always at the ready to cause mayhem for the human race. Previous, though less-deadly pandemics have begun from a natural origin.

In fact, the 2002-2003 SARS epidemic started in China and was genetically traced to a colony of cave-dwelling horseshoe bats in Yunnan province. And similarly to 2020, the Chinese government obfuscated the extent of the outbreak. So the supposition, especially amongst epidemiologists, virologists, and public health officials, that Covid would turn out to be of natural origin is understandable.

What’s not understandable is that given the lack of a smoking gun, the unusual spread of the virus in China and once the circumstantial evidence for a Covid lab-leak started to accumulate, (fairly quickly) after the pandemic started spreading around the world, that the elite media would cling so ferociously to a defense of a natural origin. But more than that, that they would show so much contempt for a lab-leak theory, especially given the likelihood for it being nonzero, is hard to comprehend.

For scientists involved in virology and epidemiology the skepticism is at least understandable. They’re more aware of the possibility of contagion from wild animals, they’re more familiar with lab safety protocols, etc. However, they are not a disinterested party and have a conflict of interest in defending a natural-origin hypothesis. They are also intimately familiar with the advanced genetic manipulation and gain-of-function techniques available to modern virology researchers which could be used to concoct a virus like SARS-Cov-2

So it is ironic that knowing both the advanced techniques available and how a lapse in safety-protocols and pushing the envelope on virology research could easily lead to a lab-leak and cause a global pandemic, that these scientists would be so invested in pushing a natural origin theory. It’s almost as if they were looking for an alibi.

Less understandable is the contempt and skepticism of the elite-media to a lab-leak theory. Journalists are usually trained to be skeptical, especially if a party has a vested interest in pushing a narrative. Obviously scientists, virologists, and researches had a vested interest in promulgating a natural origin theory. If there was even a 1% chance that Covid leaked from a lab, their funding would dry-up, scientists would be blamed for causing one of the deadliest viral outbreaks in history, and their life’s work would be that much harder to undertake and justify.

                                                                       2.

To a skeptical person, the immediate possibility of a lab-leak accident in Wuhan is readily apparent for several reasons:

1) That the first cluster of cases in China, that we know of, centered around the Hunan Seafood Market in densely populated Wuhan, a mere 16 miles north of the Wuhan Institute of Virology—one of the foremost corona virus research centers in the world, is highly-unusual especially for a novel virus. If in fact the virus was the result of animal-based progenitor strains interacting, wouldn’t Covid outbreaks have started near where the animals were caught in the wild, impacting more remote village communities first, rather than appearing “fully-formed” in Wuhan?

2) There is also a question of the virus’ rare structure—the insertion of a furin cleavage site at the S1/S2 boundary in the SARS-CoV-2 spike protein, which allows it to so easily infect humans.

3) Proposals that WIV was working on would have led to a Covid-type virus being created. In 2018, EcoHealth Alliance submitted a grant application titled “Project DEFUSE” to DARPA. It proposed experiments to introduce furin cleavage sites into coronaviruses. Furin cleavage sites are an attribute found in some viruses, including SARS-CoV-2, that can help make these viruses more infectious. But importantly almost never found in naturally-occurring coronovirii.

So then the reasons for elite-media contempt of a lab-leak scenario are essentially puerile, especially for what are supposed to be serious news organizations. The lab-leak hypothesis was, from the very beginning, associated with Donald Trump and his allies. And that was essentially enough.

It’s worth sitting with that for a moment. A hypothesis about the origin of a pandemic that killed millions of people worldwide was dismissed—not on its merits, not after rigorous investigation, not because the evidence pointed decisively elsewhere—but because the wrong people were saying it. The heuristic was breathtakingly simple: if Trump says it, it must be wrong. If Tom Cotton raises it in the Senate, it’s a conspiracy theory. If Steve Bannon amplifies it, it’s disinformation. The proposition wasn’t evaluated; the proposers were.

This is, of course, a textbook genetic fallacy—judging a claim by its source rather than its substance. First-year philosophy students are taught to identify and avoid it. And yet it became the operative editorial framework at The New York Times, The Washington Post, CNN, and virtually every other prestige outlet in the country for the better part of two years.

The irony is almost too neat. These same organizations had spent the Trump years building an institutional identity around being the “adults in the room,” the sober, fact-driven counterweight to populist recklessness. They branded themselves as the arbiters of truth in an era of “post-truth politics.” And when confronted with a genuinely consequential empirical question—one with life-or-death implications for billions—they defaulted to the laziest form of tribal reasoning imaginable: our team says natural origin, their team says lab-leak, therefore natural origin is correct.

There’s a deeper pathology here, too. By 2020, the elite media had developed a reflexive deference to credentialed expertise that was, paradoxically, deeply anti-intellectual. They didn’t engage with the science; they deferred to scientists—specific scientists, with specific institutional affiliations and specific conflicts of interest.

When Peter Daszak organized his Lancet letter declaring that the lab-leak was a “conspiracy theory,” that was treated as dispositive. No mainstream reporter, at least initially, thought to ask why the man whose organization had funneled NIH money to the Wuhan Institute of Virology was being treated as a disinterested authority on whether that same lab might have been the source of the outbreak. The conflict of interest was glaring, and yet it went unremarked upon in newsrooms that supposedly pride themselves on following the money.

This wasn’t deference to science. It was deference to authority—exactly the kind of uncritical, hierarchical thinking these institutions claim to oppose. Real scientific skepticism would have demanded more investigation, not less. It would have treated the absence of an intermediary host as a significant data point. It would have noted the remarkable coincidence of a novel coronavirus emerging miles from one of the world’s foremost coronavirus research laboratories. It would have asked hard questions about the WIV’s database going dark in September 2019, months before the official start of the outbreak.

Instead, what the public got was something closer to a protection racket for the scientific establishment. Reporters who wouldn’t dream of taking a pharmaceutical company’s press release at face value somehow found it perfectly acceptable to take virologists at theirs—virologists who stood to lose everything if the lab-leak turned out to be true: their funding, their reputations, their freedom to conduct the very research that may have caused the catastrophe in the first place.

And the enforcers weren’t subtle. When journalists like Josh Rogin at The Washington Post or science writer Nicholas Wade published careful, evidence-based treatments of the lab-leak hypothesis, they weren’t met with counterarguments. They were met with social opprobrium. The message from the media establishment to its own members was clear: pursue this line of inquiry and you will be placed in the same bucket as QAnon adherents and anti-vaxxers. It was a staggeringly effective form of peer pressure, and it worked precisely because journalists, like most professionals, care enormously about their standing among colleagues.

The result was a kind of informational monoculture at exactly the moment the world needed intellectual diversity most. Millions of people were dying. Governments were making sweeping policy decisions. And the institutions theoretically tasked with holding power accountable—including the power of the scientific establishment—chose instead to function as that establishment’s public relations arm. Not because they’d investigated and found the natural-origin theory convincing on the evidence, but because endorsing it was socially safe, and investigating the alternative was not.

                                                                       3.

In March 2018, EcoHealth Alliance submitted a grant proposal to DARPA under its PREEMPT program (PREventing EMerging Pathogenic Threats). The proposal, titled "Defusing the Threat of Bat-borne Coronaviruses," or Project DEFUSE, laid out a multi-year research plan that, in retrospect, reads less like an abstract academic exercise and more like an inadvertent engineering specification for SARS-CoV-2.

The DEFUSE proposal included several goals: Collect and catalog hundreds of bat coronavirus samples from cave sites across southern China, building a library of novel SARS-related coronaviruses. Construct chimeric viruses—synthetic hybrids assembled from genetic material of multiple different coronaviruses—to test which spike protein configurations were most capable of infecting human cells.

Drafts and notes obtained through FOIA requests revealed that Baric had already generated chimeric spike proteins from bat virus strains that do not appear in any published scientific literature, suggesting preliminary work was underway before the grant was even submitted.

The most consequential of these was to insert furin cleavage sites at the S1/S2 junction of the spike protein of thes virii. A furin cleavage site is a specific molecular feature that allows the spike protein to be cut by furin, an enzyme abundantly present in human cells. This cleavage dramatically enhances the virus’s ability to enter and infect human tissue—it is, in effect, the molecular key that made SARS-CoV-2 so devastatingly transmissible.

The DEFUSE proposal, outlined a process requiring “stitching” together DNA fragments and included plans to introduce “human-specific cleavage sites” (such as the furin cleavage site) into bat viruses using a specific restriction enzyme, BsmBI. BsmBI is part of the “type IIS” restriction enzyme family used in Golden Gate Assembly, which is ideal for such insertions.

Critically, SARS-CoV-2 is the only known virus in its entire genus of SARS-related coronaviruses that possesses this feature. The DEFUSE proposal explicitly described plans to introduce precisely this kind of modification, at precisely this location, into bat coronaviruses.

A 2022 preprint analysis determined that SARS-CoV-2 itself appears to have been assembled in six fragments using this very same enzyme—a correspondence that proponents of natural origin have struggled to explain as coincidence.

Then there is the lack of an Intermediary Host. With SARS-1, researchers identified the intermediary host (civets, then traced to bats) relatively quickly. With COVID, despite years of searching and testing tens of thousands of animals, no intermediary host has been found, and no progenitor virus in nature has been identified. This absence of evidence isn't proof, but it's a conspicuous gap for a virus supposedly originating in nature.

Thousands of demonstrators flooded the streets of Washington, D.C., on Tuesday as lawmakers began certifying the results of the November election. The unrest escalated when some protesters breached the U.S. Capitol and briefly set fire to one of its entrances.

Capitol Police responded with tear gas and gunfire, plunging the capital into chaos. Human rights organizations, including Amnesty International, reported that at least ten protesters were killed and more than 130 injured, although those figures could not be independently verified. Footage released by the American Human Rights Commission appeared to show police officers firing their weapons as demonstrators advanced toward them.

As tear gas drifted through the city, some protesters climbed through Capitol windows after lawmakers voted 195–106 in favor of certifying the election results. Supporters of certification argued that Congress was carrying out its constitutional duty to safeguard democracy.

President Joseph Biden announced the deployment of military forces to assist in restoring order, condemning what he described as “treasonous elements” behind the violence. In a televised address, he said legitimate debate over the election had been “hijacked by dangerous people” whose actions had inflicted a heavy toll on the nation.

“It is not acceptable that criminals posing as peaceful protesters terrorize the public, their elected representatives, and our constitutional institutions and expect no consequences,” Biden said, vowing to hold those responsible accountable.

Defense Secretary Lloyd Austin confirmed that security forces had been deployed to support police responding to what he called a national security emergency. Similar demonstrations were reported in several American cities.

The crisis has shaken the United States, a nation of 334 million people long regarded as a pillar of regional stability. According to Amnesty International, nationwide protests over the previous week left more than 1,000 people dead and 20,000 injured.

The turmoil comes as the United States attempts to project stability abroad. Just days earlier, 400 American police officers had arrived in Haiti as part of an international mission aimed at combating gang violence and restoring order.

The controversy centers on the certification bill introduced by the Biden administration in May. House Majority Leader Nancy Pelosi called the legislation essential to protecting the integrity of elections.

“This decision is not, and never was, about politics,” Pelosi said. “It is about facts, dignity in public service, and honoring those who have fought to protect our democracy.”

Supporters argued that the measure was necessary to ensure the continuity of government. One advocate wrote on social media that lawmakers had already revised the bill in response to criticism, including removing several unpopular taxes.

Opponents, however, accused the government of using the legislation to target Republican voters by classifying certain forms of political support as subversive activity. Critics also pointed to alleged corruption, financial mismanagement, and what they described as the administration’s extravagant spending habits during Biden’s tenure since 2021.

The president now has two weeks to sign the bill into law or return it to Congress with proposed amendments.

Right-wing activist groups that helped mobilize the demonstrations insisted their movement crossed traditional divisions of class, race, and party affiliation. Protesters confronted public officials at community events and religious gatherings, and some carried coffins to the offices of lawmakers who supported the measure.

“The politicians have underestimated our energy and determination,” said Steven Hassel, 26, of San Antonio. “We are demanding answers, and they are shocked that this movement has reached their doorstep.”

Opposition to the bill spread beyond political rallies. Nightclubs across Texas played the national anthem in protest, while church leaders and worshippers criticized the proposed tax increases during Sunday services.

As demonstrations intensified Tuesday, protesters wrapped themselves in American flags, blew whistles and plastic horns, and chanted, “Biden must go.” In Waco, Texas, demonstrators blocked roads with burning tires while shouting “Reject,” the slogan of a popular anti-certification social media campaign.

Government officials accused unnamed foreign actors of fueling the unrest and dismissed many participants as privileged young people detached from the hardships they claimed to oppose.

Protesters rejected that characterization.

“It’s not about how we get here,” said Anita Harris, 19, whose TikTok videos documenting the movement have attracted a large following. “It’s about why we’re in the streets. We see a future slipping away, and we want change.”

Hours before the demonstrations began, several prominent critics of the bill were reportedly abducted, according to the Law Society of Houston. The identities of the abductors remained unknown, though some were suspected of having ties to intelligence agencies. The organization later reported that several detainees had been released.

Other activists said they received threats and intimidating phone calls in the days leading up to the protests and feared for their safety.

Human rights organizations have long accused successive American administrations of abducting and mistreating political opponents. Police declined to comment, but Chief Justice John Roberts condemned the reported kidnappings as “a direct assault on the rule of law.”

International concern grew as violence spread. Ambassadors from thirteen Western nations, including the United Kingdom, said they were shocked by events outside the Capitol and deeply troubled by reports of abductions.

“We condemn the violence reported during protests in Washington and across the United States,” a British diplomatic spokesperson said. “We urge restraint, the restoration of order, and meaningful dialogue.”

Among those caught in the turmoil was Auma Obama, the half-sister of former President Barack Obama. CNN footage showed her struggling through clouds of tear gas while speaking to reporters.

“Young Americans are demonstrating for their rights,” she said before breaking into a cough. “I can’t even see anymore.”

Shortly after 6 p.m., protest organizers urged demonstrators to leave downtown Washington. With public transportation disrupted, many began walking home in groups.

“Go home while it is still safe,” photographer and activist Boniface Mwangi wrote on social media. “The government will try to blame peaceful protesters for any further destruction. They must hear us. Spread the word and travel home together. We will be back.”

Brat Summer

A strange, almost cult-like fervor fell upon progressives on that fateful Sunday in July when Joe Biden, having opted-out of the presidential race, endorsed his vice-president, Kamala Harris, to take his place.

Almost immediately massive donations came pouring-in, to the tune of at least a billion dollars, in the days and weeks following the endorsement1.

And almost immediately the media narrative became that of Harris as the “joy” candidate2; the “coconut tree” candidate; the “Brat Summer” candidate3. In short, the American media was vibin’ with the Kamalanomenon4.

The sudden turn in her popularity is almost astonishing. During her first year as V.P., Harris' public approval ratings sat at 28%, making her one of the least popular vice-presidents in modern history—lower than Dick Cheney, who was reviled by Democrats.

In fact, she gave the “coconut tree” speech itself at a swearing-in ceremony for the commissioners of the Presidential Advisory Commission on Advancing Educational Equity, Excellence, and Economic Opportunity for Hispanics (PACAEEEEOH, for short) in May, 2023—the type of thankless and perfunctory task, that Chief-Executives will use to curry favor with Congressional caucus members, or their constituents. But here performed by the Vice-President, presumably because it was too taxing on Biden.

Whereas in the previous three years, she was a barely-there member of the Biden Administration, suddenly in summer of 2024, she was the happy embodiment of Democrats’ hopes of vanquishing their loathed enemy, Donald Trump5.

2020

The suddenness and ferociousness of Harris’ support was almost like an old Soviet psychological operation or some cultish kirtan, comprised of reality-denial, confirmation bias, historical revisionism, and gas-lighting: that she “was never the Border Czar”6, that she is “among the most qualified” candidates in recent history7, etc.

Millions of progressive supporters, almost overnight, fell in the thrall of the Harris campaign. Whatever her previous faults and deficiencies, they were forgotten in a haze of electoral delirium. Democrats had to believe she was the best person to take on Trump (even if she wasn’t)

Yet Harris was, objectively, amongst the weakest V.P.s turned presidential candidates since Walter Mondale.

Kamala Devi Harris was born 1964 in Oakland, CA to a mother from Madras in India, and a father from Brown's Town in Jamaica.

She graduated from UC Law in San Francisco in 1989, after which she became a Bay-area prosecutor eventually working under Terence Hallinan, the storied San Francisco District Attorney.

Her first run at elected office was to oust Hallinan as the DA in 2003. The election consisted of two rounds. In the first round she received 34% of the vote, while Hallinan received 36%. The third candidate was Bill Fazio, a former prosecutor turned defense attorney, who received 30%. All three were Democrats.

In the second round she beat Hallinan with 56% of the vote.

Seven years later, in 2010, Harris then ran for California Attorney General. That contest was more competitive, barely beating Republican Steve Cooley, with 46% of the vote to Cooley’s 45%. Cooley had the advantage of name recognition as the DA for Los Angeles. But Democrats swept state-wide offices that year combined with support from then-president Obama and she made it over the line.

Harris remained CA Attorney General until 2016, when she ran for U.S. Senate. Her Senate run was fairly mundane. CA has an open primary for that office and she came out ahead of dozens of other candidates, with 40% of the vote. The next closest candidate, Loretta Sanchez, only received 19%.

In the general election she handily defeated Sanchez, a fellow Democrat with over 60% of the vote.

In other words, up until 2019 when she decides to run for president just two years after assuming the Senate office, she exclusively had been a candidate in friendly political territory, with the support of political insiders, including the President himself.

And yet her 2019 candidacy faced challenges from the start. Harris and her closest advisers were indecisive about which states to target, issues to emphasize and which opponents to go after, all the while refusing to make difficult personnel choices and impose order on an unwieldy campaign.

When she entered the race in January of 2019, she wagered that the early primary states of Iowa and New Hampshire would be less significant to her political success than South Carolina, with its predominantly black Democratic electorate. In this view, a strong performance in South Carolina, which votes fourth, would catapult her into Super Tuesday states like California, boosting her momentum.

Harris' campaign message was influenced by extensive polling, which suggested that the word “truth” resonated strongly with people. Consequently, she named her 2019 memoir “The Truths We Hold” and incorporated a similar phrase, “Let’s speak truth,” into her early campaign speeches. However, she eventually abandoned this slogan, believing that voters preferred something more concrete.

Throughout much of the year she concentrated on competing against Biden in South Carolina and beyond. However, her campaign did not foresee that he would maintain strong support among voters.

Then in the first Democratic debate in June, she tried to knee-cap him: “I do not believe you are a racist…but you also worked with them to oppose busing. And, you know, there was a little girl in California who was part of the second class to integrate her public schools, and she was bussed to school every day. And that little girl was me.”

Harris was describing Biden’s remarks in which he recalled his “civil” working relationships with segregationist senators such as James O. Eastland of Mississippi and Herman E. Talmadge of Georgia.

And while she briefly surged to the top of the field, her attack on Biden did little to improve her prospects. In fact it was the high water-mark of her campaign. Five months later, she was out of the race.

Her assumptions about the issues that would inspire Democrats were muddled: she began running on a tax cut aimed at lower- and middle-income voters and then turned to a pay raise for teachers.

However, these proposals failed to captivate voters, particularly those drawn to the policies of Elizabeth Warren and Bernie Sanders. Consequently, Harris soon began to downplay her once-prominent proposals.

At one point, she tried to emphasize a practical agenda, focusing on issues she claimed kept voters awake at 3 a.m. Still, the NY Times referred to her as “an uneven campaigner” who “changes her message and tactics to little effect.”8 Her aides joked about the numerous slogans and catchphrases she used, with one noting how the “speak truth” spring transitioned into the “3 a.m.” summer, leading to Trump-centered “justice” winter.

But somehow, shockingly, Harris’ political future would get a new life as Biden’s choice for V.P.

Biden

When George Floyd died in police custody in the summer of 2020, it became a political liability for a candidate in the Democratic primary to be an elderly white man. Democrats were peak racial revolution then. So the activists within the Democratic party who populate the lower echelons of policy-making and execution9, were skeptical of Biden.

He needed a black woman, if he was to maintain the support of these activists. But ideally someone who was not a purely left-wing ideologue and could demonstrate some more centrist credentials. He got both in Harris.

It’s hard to imagine, though, the Joe Biden of 1988 caving in to pressure from his left wing. That Joe Biden would not have wanted or cared about the support of left-wing activists whose only other choice would have been a distant-second Bernie Sanders. But 2020 Joe Biden was well into his senescence and under the spell of his lefty advisors, so picking the worst candidate in the field, who heavily implied he was a racist to his face, made sense to him somehow.

Yet there was an original sin at the birth of the Biden/Harris administration. On the same afternoon he was sworn in as the nation’s 46^th president, Joe Biden undid several of President Donald Trump’s immigration policies.

The list of orders included renewing DACA, stopping construction on the border wall and putting a moratorium on deportations.

“I’m not making new law. I’m eliminating bad policy,” Biden said while signing the orders.

Subsequently, the Biden/Harris Administration issued a series of immigration enforcement, policy and rule changes that would essentially dismantle immigration enforcement at the southern border in its entirety.

When historians examine the key factors behind Trump’s second victory in 2024, they will likely highlight the Biden administration’s decision to allow over 7.3 million illegal immigrants into the U.S.

These immigration policies negatively impacted Harris’ prospects in two significant ways. Firstly, Biden was elected on the presumption by most voters that he would be a steady hand, wisely guiding the ship of state after the tumult of the Trump years.

By effectively opening the border to anyone who could cross it as one of his initial actions, Biden undermined this perception of competence.

Secondly, the obvious consequences of allowing millions of people from the poorest regions of the world to enter the U.S. without proper documentation, background checks, or means of support, and then transporting them to various towns and cities across the country at taxpayer expense, would make even open border advocates uneasy. This includes increased strain on schools and hospitals, as well as the compounded tragedy when a migrant commits a violent crime, which might not have occurred if Biden had not opened the border in the first place.

Not to mention the safety of our cats and geese!

Combine these along with the chaotic withdrawal from Afghanistan in the summer of 2021, and the grinding effects of inflation on Americans’ pocketbooks and it strains credulity to think that a candidate so closely associated with the Biden Administration would have the slimmest chance of winning without a radical break.

Kamala Harris offered neither.

2024

So why did the Kamalanomenon train go off the rails?

For starters, Harris was a bad candidate. I don’t mean her policies were bad, necessarily. But she clearly had trouble selling them. In interviews, she had a tendency to give word-salad answers while trying to avoid giving concrete ones which might contradict something she said previously10. She famously bombed in an interview with Lester Holt in 2021, in which she struggled to articulate the administration’s strategy for securing the border.

And maybe that works for a lawyer, and certainly many politicians have law degrees. But if you’re even a middling politician, at some level you have to be able to connect to people, especially those whose votes you have to win-over, and who aren’t fanatical ideologues who will enthusiastically vote for you because your name isn’t Trump.

Which leads to the second reason: her dearth of adversarial press appearances, not only in this campaign, but throughout her Vice-Presidency. Almost as soon as she joined the Biden campaign, she was hidden from the press like some secret love-child.

For a candidate who did not go through the traditional primary process, initially shying from the press did her no favors.

Even after her campaign acceded to do more interviews she refused to clarify any serious positions and transitioning to a more open approach with interviews failed predictably.

“Harris was shaky, unreliable, and inexperienced,” said former White House press secretary Ari Fleischer to the Washington Examiner.11 “Walz is a junior varsity player who wasn't ready for the major leagues.”

Harris and Walz have primarily engaged with friendly platforms like The Tonight Show, The View, the Call Your Daddy podcast, and the Howard Stern Show. However, their lackluster performances, even in the face of mostly soft questions, drew criticism and highlighted the reasons behind Harris’ earlier decision to keep out of the spotlight.

Since the summer, when Harris and Walz had their first interview with CNN as the 2024 Democratic nominees, Harris’ media appearances have been carefully managed, while Walz has participated in fewer interviews.

The challenge for Harris and Walz was that, according to Fleischer, “neither is truly prepared to assume the roles of president and vice president.”

Doug Heye, former communications director for the Republican National Committee, believes that Harris’ performance during her media appearances illustrates why she “hasn’t done many interviews.”

“She simply isn’t good at them,” Heye told the Washington Examiner. “In a supportive environment, with a question designed for her to excel, she still fell short.”

Referring to Harris’s appearance on The View, Heye pointed out that she failed to distinguish herself from President Joe Biden, despite polls indicating that voters are seeking change. He argued, “If she can’t provide substantive answers about her plans for the economy or what she would do differently from Biden, then she’s not a strong candidate.”

The third reason: her choice of running-mate. For all the artifice of casting him as “America’s coach,” the truth is that Tim Walz is the liberal governor of a deep-blue state, Minnesota. CNN called him a “progressive champion.”12 And, in fact, it only takes a few minutes of Googling to find that out, it’s not even an open secret. One only has to look at the ideological tilt of the bills he signed as governor, once Democrats had the majority in the state legislature.

Not to mention that he oversaw the destruction of Minneapolis during the summer of 2020. Records and past public statements show that officials in that city described the governor’s initial response as noncommittal.

Records obtained by the NY Times13 show, the mayor considered sending out a news release detailing his request: “I’ve reached out to Governor Walz to request the assistance of the Minnesota National Guard to help de-escalate and prevent any further conflict.”

But Walz did not agree to send the Guard the first night and looting and arson continued.

Ultimately it’s hard to believe that more moderate voters, as one finds in Wisconsin and Michigan, would be overjoyed in Walz’s response to the rioting in 2020. If anything, they would have trepidation with a ticket fronted by a liberal Californian, add a liberal fire-brand running-mate and a vote for Harris-Walz become harder to justify.

Just one of these factors would have imperiled a competant presidentail campaign. Combine them all, and the unfocused and ersatz one run by Harris and you get the pitying results we saw on election night.

1.

The Facts That Aren’t in Dispute

Hunter Biden landed a position on the board of the Ukrainian gas company Burisma in 2014. His annual salary was around $1 million per year, until March 2017, when it was reduced to around $500,000. His father had left the vice presidency in January of that year. The timing may be coincidental. The compensation structure suggests otherwise.

By the time Hunter left Burisma in 2019, his firm had earned roughly $11 million. Burisma was just one of several foreign companies that paid Hunter Biden — an otherwise middling Yale Law graduate with zero experience in the oil and gas industry — millions of dollars in essentially no-show advisory roles during the years his father was Vice President.

The surface-level explanation for why these companies would pay such ludicrous sums for Hunter’s “expertise” is obvious: his immediate connection to the sitting Vice President of the United States.

In Burisma’s case, the calculus was especially transparent. Joe Biden was personally charged with handling anti-corruption efforts in Ukraine on behalf of the Obama administration. Having the Vice President’s son on the payroll wasn’t a staffing decision. It was an insurance policy.

Joe Biden has always maintained that neither he nor Hunter was involved in any wrongdoing — that the sole problem with Hunter’s employment at Burisma was the appearance of a conflict of interest:

“There’s nobody that’s indicated there’s a single solitary thing that [Hunter] did that was inappropriate, wrong . . . or anything other than the appearance. It looked bad that he was there.”

But the record tells a different story. Hunter Biden traded heavily on his surname and used it to extract millions of dollars in advisory and consulting fees from foreign entities with direct interests in U.S. government policy. He was subsequently indicted on multiple felony tax charges related to these financial dealings — charges that included evading taxes on foreign income and funding a lavish personal lifestyle with unreported gains.

The parallel to Paul Manafort is instructive. Manafort — Trump’s former campaign chairman — also ran afoul of the law through his dealings in Ukraine. He pleaded guilty in 2018 to money laundering, tax evasion, and failing to register as a foreign agent. He, too, deceived his tax advisers to avoid paying taxes on foreign income. He, too, spent the proceeds on an extravagant lifestyle. The factual patterns are nearly identical. The political treatment of the two cases could not be more different.

The Legal Standard No One Wants to Apply

Under U.S. federal law, it is illegal to bribe a public official. The statute does not require proof that the official delivered a specific favor in return. In 2008, former Alaska Senator Ted Stevens was convicted for accepting unreported gifts — home renovations paid for by an oil company executive. The government never alleged that Stevens personally returned the favor. The crime was the undisclosed acceptance of value from someone with business before the government.

It strains credulity that if Joe Biden accepted even a dollar of Hunter’s foreign-sourced income, he wouldn’t face exposure under a similar theory. The standard isn’t quid pro quo. The standard is whether a public official received something of value from a party with interests before the government and failed to disclose it.

This is the conversation that Washington has spent years refusing to have honestly. Instead, the defense of the Biden arrangement has cycled through a remarkable sequence of positions, each one abandoned as it became untenable and replaced by the next without any acknowledgment that the previous position was wrong:

Phase One: It’s disinformation. When the New York Post published reporting on Hunter Biden’s laptop in October 2020, the immediate response from Democratic officials, allied media, and former intelligence professionals was to label it a probable Russian disinformation operation. Social media platforms suppressed the story. Major outlets refused to touch it. This position held through the election and for months afterward — until outlet after outlet quietly authenticated the laptop’s contents without ever reckoning with the suppression campaign.

Phase Two: It’s not controversial. Once the laptop was authenticated and the business dealings confirmed, the defense shifted. Hunter Biden was a private citizen. He had every right to pursue business opportunities. The fact that his father happened to be Vice President was irrelevant to his professional qualifications. This position required believing that multiple foreign companies in industries where Hunter had no expertise independently concluded that he was worth seven figures annually for reasons entirely unrelated to his surname.

Phase Three: It’s Hunter’s problem, not Joe’s. When the tax indictments landed and the financial record became undeniable, the perimeter contracted again. Fine, Hunter may have broken the law. But he’s not a public official. His legal problems are his own. Joe Biden had no knowledge of and no involvement in Hunter’s business dealings. This position required ignoring Hunter’s own admissions about putting his father on speakerphone with business associates, the documented meetings between Joe Biden and Hunter’s foreign partners, and the communications on the authenticated laptop referencing “the big guy” and his percentage.

Phase Four: Trump did worse. The final fallback — deployed when all other defenses have been exhausted — is comparative. Even if everything alleged about the Bidens is true, Trump’s conduct was more egregious, more dangerous, more corrupt. This is not a defense. It is an admission dressed up as an argument. “The other side is also corrupt” has never been an exoneration. It is the argument you make when you have no other argument left.

The Question That Remains

The central question was never whether Hunter Biden sold access. By his own admission, he sold at minimum the “illusion of access” to his father. He put Joe on speakerphone to impress clients. He arranged meetings. He traded on proximity to power. None of this is seriously contested by anyone who has reviewed the record.

The real question is where the line falls between the illusion of access and access itself — and whether Joe Biden was a passive bystander to his son’s influence-peddling operation or an active, knowing participant.

The circumstantial evidence is substantial. The salary reductions that tracked Joe’s departure from office. The meetings with Hunter’s business partners that Joe initially denied and later couldn’t explain. The references in Hunter’s own communications to setting aside a percentage for a family member. The exposed financial flows that investigators have traced through a web of Biden family accounts.

None of this constitutes proof beyond a reasonable doubt in a courtroom. But the standard for public accountability is not the standard for criminal conviction. Public officials are expected to avoid even the appearance of conflicts of interest — a standard that Joe Biden himself invoked when dismissing concerns about Burisma.

By his own standard, the Biden arrangement fails. By the standard applied to Ted Stevens, it fails. By the standard applied to Paul Manafort, the underlying financial conduct is nearly indistinguishable. The only reason it has not been treated with equivalent seriousness is that the political coalition with the most to lose from the investigation is also the coalition with the most influence over the institutions responsible for conducting it.

That is not a conspiracy theory. It is an observation about incentives, and it is the kind of observation that corrodes public trust in democratic institutions more effectively than any foreign disinformation campaign ever could.

The question was never really about Hunter. It was about whether the rules apply uniformly or whether they bend based on which team is wearing the jersey. The answer to that question has been obvious for some time. The only remaining question is whether enough people are willing to say it plainly.

There is a political war under way in America, and you don’t need a subscription to a niche newsletter to find the front lines. You need only turn on CNN or NPR at almost any hour of the day. The combatants rarely describe it as a war, because to name it would be to concede that a fight is happening at all. But it is a war nonetheless — a war over political truth itself, and over who gets to decide what counts as true.

The offensive is being waged largely by what an earlier generation of conservatives called the “liberal elite,” and what might now be better labeled the institutional left: the elite newsrooms, the credentialed commentariat, the very-online progressives, and the dense social ecosystem of Media Twitter that sets the day’s acceptable opinions before most Americans have finished their coffee. Their weapons are not tanks or televised debates. They are subtler and, in their way, more powerful: the construction of narratives and counter-narratives, reflexive defensiveness and deflection, and the strategic deployment of “fact-checking” and gaslighting as instruments of control.

The institutional elite — as denoted by Media Twitter, ‘very online’ progressives and the elite media, — have a monopoly on acceptable narratives and “the truth”. Elite media crafts certain narratives favourable to Democrats or parroting left-wing talking points and ideology (“mostly peaceful” protests instead of riots, Hunter Biden’s laptop as Russian “disinformation”, etc) while doing two things at the same time: disingenuously claiming that they’re not crafting left-wing narratives, merely reporting (or fact-checking) “the truth” and secondly deflecting any criticism that they’re just parroting a left-wing narrative by formulating such criticism as a Republican, right-wing, manufactured (as opposed to grass-roots) “assault” on said “truth.”

Take for instance the riots of Summer 2020. There is ample evidence that those riots caused billions of dollars in damage and resulted in dozens of deaths around the country. What’s notable is the response to those riots by progressives, the media, and Democratic politicians.

Firstly the media focused almost exclusively on crafting a narrative of these protests as a broad-based reaction by regular Americans to ongoing racial injustice, when that wasn’t the case at all. Protests were led and populated almost entirely by young white people under 30 who comprised the vast majority of protestors, only made possible because of COVID lockdowns that left them with a lot of time on their hands.

Secondly, the media maintained this narrative even as uneventful protests during the day led to burning, looting and rioting at night. They maintained this narrative even as people were murdered by rioters, businesses looted, and government buildings attacked. Of note, though, is the hypocrisy by the media and Democratic politicians with regards to the protestors: after months of media hyperventilating about people not adhering to social distancing including vile criticism of those protesting lockdowns in front of state legislatures a month prior, those concerns went out the window when literally thousands of people were crowding together yelling BLM slogans in the major cities in Summer of 2020.

Remarkable was any lack of concern about masses of people congregating in such numbers during a raging pandemic, especially after the same politicians and media activists had just spent the last few months prior castigating and chastising anyone congregating in public or flouting public health measures. In fact the media took great glee in labeling right-wing protestors as “idiots” in May 2020 for essentially doing the same thing left-wing protestors were doing in June.

Remarkable too was that any attempt to highlight the disparity between reporting and the ongoing riots was met by deflection and hand-waving by left-wing politicians and the media. They would never concede that rioting was taking place, claiming any criticism was just a “conspiracy theory” by malevolent right-wing reactionary forces who were attempting to slander the “broad-based” protests of “racial reckoning”.

It’s hard to imagine that if similar protests and rioting had taken place in Russia, or China, that coverage of those protests in Western media wouldn’t have included such words as “uprising” or “insurrection”.

Furthermore, these politicians and media-activists claimed, any rioting which was theoretically taking place was entirely justified since rioting was the “language of the oppressed”, and only “property” was damaged, which wasn’t true to begin with: people lost their lives in the riots.

That same deflection and hand-waving by the institutional left can be seen in the debate against Critical Race Theory. The controversy is being presented by the elite media and liberal blogs as a drummed-up culture-war issue by paranoid conservatives who “misunderstand” it, in effect gaslighting them.

During the pandemic, public schools, teachers unions and local schoolboards somehow thought that instead of ensuring children are reading and doing math at grade level, and that they have all the tools to succeed academically while learning from home, they would instead inculcate left-wing racialist and gender dogma in children.

The elite media (NPR, the NY Times, Washington Post, et al.) then try to obfuscate this reality by trying to craft a reality-control narrative, first of all denying that anti-racism is being promulgated at all in public schools: Critical Race Theory, they say, is only taught in certain law-school classes, not in elementary school, furthermore they deflect the indoctrination of children into left-wing political ideology as merely “teaching history”, and thirdly by characterizing grass-roots opposition to this indoctrination as some sort of right-wing astroturf campaign. An episode of NPR’s Fresh Air describes the opposition as being led by “conservative think tanks, media outlets and law firms.” In other words, the elite media is describing this issue as any opposition is fake and artificial, and in any case they merely oppose teaching American history.

The argument by progressives and the media is presented as one where conservatives in state legislatures are imposing bans on the necessary, in their view, teaching of systemic racism, moreover the fact that conservatives are bringing up CRT at all is as a culture-war smokescreen to distract from the usual claims of all the -isms that ail America.

One New Yorker article says that conservatives “have noticed that concepts used in critical race theory — ‘structural racism,’ ‘internalized white supremacy’ — have entered the mainstream, and in some cases have become part of corporate training or public-school education.”

The assumption being that CRT terms, such as “whiteness” are remotely mainstream.

It can never be a reaction to the introduction of “anti-racist” indoctrination by social-justice activists in the various schoolboards and universities. In other words, parents have no legitimate cause to rail against CRT, and any reaction to it is illegitimate if not outright devious.

As proof of this just take the recent letter from the National School Boards Association, where they plead with the Biden admin to have the FBI investigate outraged parents who dare question their school board positions on anti-racism and gender identity.

In it, the NSBA likens parents to “domestic terrorists.”

Now imagine if the same thing had happened under the Trump administration: let’s say the NRA had sent a letter imploring William Barr to pursue activists targeting NRA meetings, and then that the Department of Justice had actually began one? Can you imagine the outrage amongst liberals and in the media!

And yet that’s where we are in the U.S. in 2023: the left claims enemies of the Republic and wrongthingkers are everywhere on the right, whilst simultaneously engaging in the very acts they claim to be against: it’s OK for anarchist rioters try to burn down a courthouse in Portland during the riots of 2020, but when right-wing plebes try to vandalise the U.S. Capitol building, it’s nothing short of an insurrection, according to them.

2. The Double Move: Deny and Deflect

The institutional elite holds something close to a monopoly on acceptable narratives — on “the truth” as officially sanctioned. That monopoly is not maintained by argument, and it does not need to be. It is maintained by controlling the terms on which any argument is permitted to take place: which framings count as neutral and which as “misinformation,” which sources are authoritative and which have been “debunked,” which questions are asked in good faith and which are asked only to sow doubt. The mechanism is worth naming precisely, because it repeats across issue after issue with almost mechanical regularity.

The elite media crafts narratives that flatter Democrats or echo progressive talking points — “mostly peaceful” protests rather than riots, the Hunter Biden laptop story waved away as Russian “disinformation” — while performing two maneuvers at once.

The first is a claim of pure neutrality. The outlet insists it is not crafting a narrative at all; it is merely reporting, or, increasingly, “fact-checking” the plain truth. This is the more sophisticated of the two moves, because it launders editorial judgment through the vocabulary of objectivity. A word choice is presented as a finding. “Mostly peaceful” arrives not as a characterization open to dispute but as an established fact; “disinformation” arrives not as a contested political label but as a verdict, delivered with the full apparatus — the fact-check rating, the credentialed expert, the platform warning label — that lends an opinion the authority of a settled conclusion. The reader is invited to believe not that the outlet has taken a side, but that the other side has taken leave of reality.

The second maneuver activates the instant the first is challenged. Any criticism of the narrative is recast not as legitimate disagreement but as something more sinister: a coordinated, right-wing, manufactured “assault” on the truth itself. The critic is not met on the merits. The critic is diagnosed — as a partisan, a conspiracy theorist, the mouthpiece of some astroturfed campaign traceable to think tanks and dark money rather than to genuine, grassroots concern. This is the elegance of the move: it never has to engage the substance of the objection, because it has redefined the objection as proof of bad faith. To notice the narrative is to reveal yourself as an enemy of the truth.

Taken together, the two moves form a closed loop. Deny that a narrative exists; then, when someone points to it, delegitimize the act of pointing. The first move renders the framing invisible. The second renders any challenge to it disreputable. Between them, they insulate the narrative from the single thing that might dislodge it — scrutiny. Once you see this pattern, you cannot unsee it, and the episodes that follow — the riots of 2020, the fight over Critical Race Theory, the campaign against dissenting parents — are the same machine running on different fuel.

3. Summer 2020: A Narrative That Outlasted the Facts

Consider the riots of the summer of 2020. There is ample evidence that the unrest that year caused billions of dollars in damage and resulted in dozens of deaths across the country. What is most revealing, however, is not the destruction itself but the response to it from progressives, from Democratic politicians, and above all from the media.

From the outset, the coverage was devoted almost entirely to a single storyline: that these were broad-based demonstrations by ordinary Americans reacting to ongoing racial injustice. That framing did not match the reality on the ground. The protests skewed heavily toward the young and the white, with participants under thirty forming a conspicuous share of the crowds — a mobilization made possible in no small part by COVID lockdowns that had left an entire cohort of young people with time on their hands and nowhere to spend it.

More striking still was the media’s refusal to revise the storyline as the facts changed beneath it. The narrative of peaceful protest held firm even as demonstrations that were calm by day gave way to burning, looting, and rioting by night. It held as people were killed, as businesses were destroyed, and as government buildings came under attack. To understand how far the machinery of narrative management would go, it is worth examining three episodes in particular — the sudden disappearance of the pandemic as a concern, the armed occupation of an American city’s downtown, and a single television graphic that captured the whole enterprise in eight words.

The Vanishing Pandemic

For months in the spring of 2020, the public-health establishment and its allies in politics and the press had treated public gatherings as an act of near-homicidal recklessness. When Americans assembled outside state capitols to protest lockdowns and mask mandates, the condemnation was swift and unsparing. Epidemiologists warned that such gatherings would kill. Governors and mayors denounced the protesters. The media reserved a special contempt for them, branding them selfish and stupid — “idiots,” in the framing that circulated widely in May of 2020 — for daring to congregate while a virus raged.

Then the season changed, and so did the science — or so it appeared. When the protests of late spring and summer erupted, drawing enormous crowds into dense proximity in cities across the country, the warnings did not merely soften. They inverted. In June 2020, more than a thousand public-health professionals signed an open letter declaring that the anti-racism protests should not be curtailed on pandemic grounds, and explicitly cautioning against equating them with the earlier anti-lockdown demonstrations, which the letter’s signatories condemned. The same act — a mass of human beings gathered shoulder to shoulder in a plague year — was lethal when performed by the wrong people for the wrong cause, and permissible, even laudable, when performed by the right people for the right one. The virus, apparently, could read a protest sign.

This was not a subtle inconsistency to be teased out by careful analysis. It was a public reversal, executed in the open, by the very authorities whose entire claim to deference rested on the premise that they followed the evidence wherever it led. What the summer of 2020 revealed was that, for a significant portion of the public-health and media establishment, the evidence was downstream of the politics.

A “Summer of Love”: The Occupation of Capitol Hill

Nowhere was the indulgence of the moment more vivid than in Seattle. After police evacuated the department’s East Precinct in June 2020, activists seized several blocks of the city’s Capitol Hill neighborhood and declared them a police-free zone — first christened the Capitol Hill Autonomous Zone (CHAZ), then rebranded the Capitol Hill Organized Protest (CHOP). An American city had, in effect, ceded territory to a self-governing occupation, and the response from its leadership was not alarm but something close to delight.

Seattle’s Democratic mayor, Jenny Durkan, appeared on CNN in June and, pressed by host Chris Cuomo on how long the occupation would last, offered a now-infamous assessment. She described the zone as having a “block party atmosphere” and, asked how long it might persist, replied that the city could see a “summer of love.” She dismissed the notion that this was an armed takeover, likening it instead to the ordinary neighborhood gatherings Seattle held all the time.

The reality diverged sharply from the mayor’s sunny billing. Within roughly ten days of the interview, the zone had been the site of several serious crimes, including an attempted arson, multiple assaults captured on video, and the fatal shooting of a nineteen-year-old. After further bloodshed, Durkan reversed course and ordered the area cleared, later conceding that her “summer of love” remark had been a poor choice of words that she had made “in jest.” The episode is instructive precisely because of its trajectory: an elected official’s first instinct, echoed sympathetically in a national broadcast, was to romanticize the seizure of American streets — and the correction came only after people had died. One is left again to imagine the coverage had a comparable occupation been staged by the right.

“Fiery but Mostly Peaceful”: The Chyron That Said the Quiet Part

If a single image distilled the media’s posture that summer, it aired on CNN in the early hours of August 25, 2020, during the unrest in Kenosha, Wisconsin, that followed the police shooting of Jacob Blake. Correspondent Omar Jimenez reported live from the scene, standing before buildings and vehicles engulfed in flames. Beneath him, the network’s on-screen graphic read: “Fiery but mostly peaceful protests after police shooting.”

The dissonance was total and self-evident. Jimenez himself noted that the fire behind him was one of multiple locations that had been burning across the city over the course of the night. The chyron and the footage told opposite stories in the same frame, and the caption became an instant object of ridicule — a shorthand, ever since, for the media’s willingness to file arson under “mostly peaceful” so long as the fire burned in the service of an approved narrative.

In fairness, the chyron was the work of CNN’s copy desk rather than the correspondent, and Jimenez’s own reporting acknowledged that the daytime demonstrations had given way to violence after dark. But that distinction, while accurate, does not rescue the network; it indicts it. The graphic was not a slip of the tongue from a reporter under pressure. It was a written editorial choice, composed at a desk, by an institution deciding in real time how to characterize what its own cameras were showing the country. The words did not describe the picture. They were meant to instruct the viewer on how to feel about it.

Deflection as Doctrine

Through all of this, any attempt to point out the gap between the reporting and the reality was met not with argument but with the now-familiar deflection. The rioting was denied outright; those who noticed it were accused of peddling a “conspiracy theory” advanced by malevolent reactionaries bent on slandering a “broad-based” movement of “racial reckoning.” It is hard to imagine that comparable scenes in Russia or China would have escaped words like “uprising” or “insurrection” in the Western press. Yet at home, the same events were smoothed into something gentler. When destruction could no longer be denied, it was justified: rioting, we were told, is “the language of the oppressed,” and in any case only “property” had been damaged — a claim false on its own terms, because people had lost their lives.

The pattern, in other words, was seamless. Deny that the violence is happening; when the flames are too obvious to deny, recast them as “mostly peaceful”; when a life is lost, retreat to the language of property and oppression; and at every stage, treat the person pointing at the fire as the real threat to public understanding. This is not journalism failing to describe events. It is a system succeeding at managing them.

4. Critical Race Theory: Denial, Redefinition, and the Astroturf Charge

The same deny-and-deflect pattern structures the fight over Critical Race Theory. The elite media and the liberal commentariat present the controversy as a culture-war panic drummed up by paranoid conservatives who simply “misunderstand” the subject — which is to say, they gaslight the people raising the objection.

The underlying grievance is not mysterious. During the pandemic, at the very moment when public schools, teachers’ unions, and local school boards might have concentrated on getting children reading and doing math at grade level while stranded in remote learning, many chose instead to fold racial and gender ideology into the curriculum. Parents noticed.

The elite institutions — NPR, The New York Times, The Washington Post, and their peers — then set about obscuring that reality through a three-part defense. First, denial: anti-racist ideology, they insisted, is not being taught in public schools at all; Critical Race Theory is a narrow academic framework confined to certain law school seminars, not something a fourth-grader would ever encounter. Second, redefinition: whatever is being taught is not indoctrination but simply “teaching history.” Third, delegitimization: grassroots parental opposition is reframed as an artificial, right-wing astroturf campaign. An episode of NPR’s Fresh Air, for instance, characterized the opposition as the work of “conservative think tanks, media outlets and law firms” — the implication being that the anger is manufactured, and that in any event the objectors merely want to stop the teaching of American history.

From there, the progressive framing writes itself. Conservatives in state legislatures are cast as would-be censors banning the necessary teaching of systemic racism, and the very fact that they raise CRT at all is treated as a smokescreen — a distraction from the litany of -isms said to afflict the country. A New Yorker piece captured the posture neatly, noting that conservatives “have noticed that concepts used in critical race theory — ‘structural racism,’ ‘internalized white supremacy’ — have entered the mainstream, and in some cases have become part of corporate training or public-school education.” The unspoken assumption is that terms like “whiteness” are now simply part of ordinary discourse, and that no reasonable person could object to their arrival in a classroom.

What this framing can never allow is the obvious possibility: that parental anger is a reaction — a response to the introduction of “anti-racist” material by activists inside school boards and universities. Concede that, and the objection becomes legitimate. So the objection is ruled out of bounds in advance. Parents, in this telling, have no valid cause for complaint; any reaction they have is either illegitimate or outright devious.

From “Domestic Terrorists” to the Asymmetry Standard

If proof of that posture were needed, it arrived in the letter from the National School Boards Association urging the Biden administration to bring the FBI to bear on parents who had the temerity to challenge their local boards over anti-racism and gender identity. In that letter, the NSBA reached for the language of “domestic terrorism” to describe the threat such parents supposedly posed.

Now perform the thought experiment. Imagine the same sequence under the Trump administration. Imagine the NRA writing to William Barr to demand action against activists targeting its meetings — and imagine the Department of Justice actually opening such an effort. The outrage from liberals and from the press would have been deafening, and it would not have subsided for weeks.

This is the asymmetry that defines the current moment. The institutional left warns ceaselessly that enemies of the Republic and dangerous “wrongthinkers” lurk everywhere on the right, even as it engages in the very conduct it claims to abhor. When anarchist rioters tried to burn down a federal courthouse in Portland in 2020, it was framed as protest. When a mob of the right stormed the U.S. Capitol, it was, in the same voices, nothing short of an insurrection. The label follows the politics, not the act.

5. The Stakes

None of this is an argument that the right is without sin, or that every conservative complaint is made in good faith. It is an argument about a particular power — the power to decide, before any debate begins, which facts are permitted into the conversation and which are quietly escorted out.

That is the real prize in the war over political truth. Whoever controls the frame controls the range of acceptable conclusions, and controls it invisibly, under the flag of neutral reporting. The elite media’s genius has been to fuse advocacy and adjudication into a single voice — to be both a participant in the argument and the referee who declares it over. A democratic public cannot reason its way to good decisions when one faction owns both the microphone and the fact-check. The first step toward reclaiming that reasoning is simply to name the maneuver for what it is: not the neutral pursuit of truth, but the management of it.