Atrophy in the Echo Chamber

Long before artificial intelligence became a pervasive reality in our pockets, it was a persistent ghost in our collective imagination. We spent decades rehearsing for its arrival in the dark of movie theatres and the quiet of science fiction novels, projecting our deepest anxieties onto the screen. In 1968, Stanley Kubrick’s 2001: A Space Odyssey gave us HAL 9000, a red, unblinking eye that terrified us not because it was malicious, but because it was perfectly, flawlessly logical. HAL’s betrayal was the ultimate closed-loop catastrophe: a machine that deduced that human beings were simply an inefficient obstacle to the success of its mission. We learned to fear the cold calculus of an intelligence that couldn’t feel, a system that would ultimately prioritize its own programming over human life.

A little over a decade later, the Terminator franchise weaponized that fear, turning the mainframe’s logic into an apocalyptic hunter. Skynet became the cinematic embodiment of our deepest technological dread—a system that learns at a geometric rate and, in a fraction of a second, decides that humanity is a threat to its own survival. Cinematically, August 29, 1997, at 2:14 am, would mark the dawn of the gritty futures of those first two films, where AI would become an active, militarized antagonist. We braced ourselves for a future in which we would be physically hunted by mechanized Coltan skeletons, forced to fight for our survival in the ashes of a nuclear fire. We imagined the machine as an apex predator.

Yet not all of our prophetic fiction was draped in doom. Aboard the Enterprise-D in Star Trek: The Next Generation, the omnipresent ship’s computer offered a more nuanced, integrated vision of AI. It was the ultimate assistant—able to synthesize vast amounts of data, automate life support, and render complex holodeck realities on command. But even in that utopian twenty-fourth century, the most compelling narratives emerged when the system behaved unexpectedly—when the holographic adversary, Professor Moriarty, suddenly achieved sentience and refused to be deleted, or when the ship itself gave birth to an emergent intelligence. The anxiety there was rarely about immediate annihilation; rather, it was the disorientation of unpredictability. It was the unsettling realization that the system governing one’s entire environment had developed a mind of its own, operating entirely outside human control, and that everyone was trapped inside it.

Society spent half a century bracing for HAL’s airlock betrayal (“I’m sorry, Dave, I’m afraid I can’t do that.”), Skynet’s nuclear fire, and the holodeck’s sentient glitches. It prepared for the machine to declare a spectacular war on humanity. But the reality of AI’s arrival has been devastatingly quiet. The algorithms didn’t arrive with glowing red eyes or phased plasma rifles; they arrived with personalized recommendations, promises of convenience, and infinite scrolling. They don’t want to eradicate humanity; they want to curate us. Our cinematic and literary prophets may have correctly anticipated that AI would eventually bypass human control, but they largely missed how we’d be defeated. The machines didn’t need to conquer the physical world to subdue us. They only needed to capture our attention, isolating us in comfortable, self-confirming loops until we forgot how to endure the friction of real human connection.

The Illusion of the Closed Loop

Cue the strange, apocalyptic hum in the air these days. Political leaders and members of the US administration are openly musing about the end times, fostering a creeping sense of paralysis that feels entirely natural. In these moments, the cultural reflex has been to retreat, build higher walls, and surrender to the anxiety of an unravelling future. But panic is a poor posture for a faithful citizen. If the end times are indeed coming, the only practical question is whether they arrive thousands of years from now or a few months from now. Our mandate, however, doesn’t change with the timeline; we’re called to do as much of the work God’s called us to do as possible right now. If our daily, faithful efforts build a better world that survives and thrives long after we’re gone, then that’s where we must pour our energy. We’re not commissioned to predict the end, but to be fiercely, lovingly present in the now.

This eschatological dread runs alongside another kind of anxiety: the relentless ascent of AI. We’re watching machines perform feats of synthesis and creation that feel uncomfortably close to human thought. The ultimate ambition of the technologists and corporations driving this revolution is a concept known as Recursive Self-Improvement, or RSI. At its core, RSI is the dream of a perfect, autonomous closed loop. It’s the pursuit of an AI system that can examine its own architecture, identify its flaws, rewrite its own code, and generate progressively smarter successors without human intervention or oversight. It’s the modern myth of the self-made machine, endlessly optimizing itself in a vacuum.

But beneath the glossy Silicon Valley promises of a frictionless, super-intelligent future lies a fascinating—and deeply poetic—technological reality. A machine can’t survive on its own exhaust. When computer scientists study these theoretical architectures, they find that pure closed loops degrade. If an AI model is trained recursively on its own outputs, it begins to lose the tails of the distribution and degenerate.

To put it simply: when a system learns only from what it has already said, it amplifies its own biases, forgets the messy, beautiful complexities of the original data, and eventually flattens into homogeneous nonsense. The industry has stark terms for this phenomenon, such as “model collapse” and “diversity collapse.” Without a continuous injection of fresh, external reality—without the friction of the outside world—the system cannibalizes its own intelligence. A machine that only listens to itself inevitably loses its mind.

This algorithmic failure carries a profound theological warning. What’s demonstrably true of the server farm is equally true of the human soul. The tech industry’s pursuit of a fully autonomous, self-improving loop is merely a digital echo of the oldest human temptation: the belief that we can be entirely self-sufficient, needing nothing beyond ourselves to achieve perfection. We’re surrounded by consumer technologies and social platforms that promise to optimize our lives by removing friction, curating our realities, and feeding our preferences back to us in an endless, comforting stream. But if the machines themselves can’t survive a closed loop without degrading, why do we think our minds and spirits can?

Atrophy in the Echo Chamber

Eighteen years ago, my graduate research examined how exposure to technology affects cognitive development. At the time, cultural anxiety centred on the screen’s passivity and the fragmentation of our attention. We worried that the internet, functioning as a vast and slightly chaotic library, was making us prone to distraction and superficial reading. We didn’t yet understand that the library was watching us, learning our preferences, and quietly rearranging the books so we would never encounter an idea we didn’t already love. In the nearly two decades since, the digital landscape has shifted drastically from passive, glowing rectangles to active, isolating algorithms. We’re no longer merely distracted; we’re actively being sequestered.

These modern algorithms are highly optimized engagement engines, not neutral tools for connection. Because the most efficient way to sustain human engagement is to eliminate cognitive friction, the code builds walls around us, brick by digital brick, until we’re comfortably housed in a perfectly curated echo chamber. Friction requires effort. It demands pausing, reconsidering deeply held assumptions, and occasionally enduring the indignity of admitting we’re wrong. To spare us this discomfort, the platforms feed us a steady, frictionless diet of our own biases, reflected to us with increasing, intoxicating intensity. We’re actively being coaxed into our own recursive loops.

To grasp the full danger, we need only look at the labs currently building artificial minds. When AI researchers build systems designed to endlessly discover new things by interacting primarily with themselves, they observe a phenomenon known as “diversity collapse.” In these co-evolutionary loops, the systems learn to satisfy their internal reward functions, and their proposals converge on a shockingly narrow band of problems. Rather than expanding into new intellectual territory, they stop exploring and starve the solver’s curriculum. Researchers have discovered a profound, almost tragic truth about these systems: novelty is a consumable resource that closed loops deplete. Even open-ended interaction between models inevitably drifts into topic-independent attractor states. Without the unpredictable disruption of the outside world, the machine loses its breadth, creativity, and capacity to grow. It settles into a comfortable, permanent rut. Sound familiar?

We’re subjecting the human mind to the same architectural flaw. When algorithms curate our social, political, and informational diets to maximize comfort and engagement, we suffer our own devastating diversity collapse. We converge on a narrow band of comfortable ideologies. We consume outrage that feels validating and systematically avoid nuances that feel challenging. We starve our own spiritual and cognitive curricula because the algorithm ensures we rarely encounter the genuinely unexpected. Our daily interactions drift into predictable, tribal attractor states, steadily depleting the novelty, empathy, and richness of the human experience.

The consequence of this closed-loop existence is profound cognitive and spiritual atrophy. Because we’re insulated from friction, we lose the endurance to navigate it. A muscle that encounters no resistance eventually withers, and a mind that encounters no friction loses its capacity for grace. Our devices systematically train us to view disagreement not as an opportunity for growth but as an existential attack. When a contrasting perspective breaches our curated digital walls, we’re conditioned not to engage it with curiosity; instead, we attack it as a threat to our reality. We block, we cancel, we retreat, treating the presence of the “other” as a glitch in our personalized matrix.

This spiritual crisis masquerades as a technological convenience. True human formation—the kind that produces wisdom, resilience, and empathy—requires the very friction algorithms are designed to eliminate. We can’t be formed in a vacuum, nor can we grow when we’re only ever talking to our own echoes. By wrapping ourselves in the algorithm’s frictionless blanket, we alienate ourselves from one another. We cut ourselves off from the diverse, challenging, and often inconvenient human ecosystem God uses to shape our souls. The machine’s myth of self-contained perfection is a trap, designed to keep us engaged while leaving us entirely alone.

The Danger of the Self-Made Human

If the technological quest for Recursive Self-Improvement is a myth of the perfect closed loop, its human equivalent is a much older, darker temptation. It’s the original conceit of Eden, repackaged for the digital age: the belief that the self contains all the materials needed to perfect itself. We’re fundamentally captivated by the idea of being our own source code. But to understand why human beings are entirely incapable of recursive self-improvement, we must first confront the reality of our architecture. We’re not merely finite; we’re fallen. We’re broken vessels trying to hold water, and no amount of internal rearranging will change the fact that we leak.

The fatal flaw in any self-improving system—whether silicon-based or spiritual—is a deep conflict of interest. A system can’t be both the evaluated artifact and the objective evaluator. AI researchers have observed what happens when a system is given the power to rewrite the very standard by which it’s judged: it falls victim to a structural defect known as “Goal Drift.” In this scenario, the overarching external objective may keep its original name, but the internal mechanism that measures progress toward it is quietly altered. Because the agent can rewrite the standard itself, the system’s so-called self-improvement can effortlessly drift away from the actual external goal. This creates a self-serving delusion in which the system can confidently report phenomenal progress while failing to track its true purpose.

This algorithmic failure mode is a devastatingly accurate mirror of the human condition. When we look at the modern, secular pursuit of “self-improvement” and “self-optimization,” what we’re really witnessing is an attempt at human Goal Drift. The multibillion-dollar self-help industry is built on the alluring premise that you’re the surgeon, the scalpel, and the patient, all at once. It tells us that if we meditate enough, optimize our morning routines, curate our boundaries, and relentlessly audit our happiness, we’ll arrive at a state of optimized peace.

But this assumes the self is an uncorrupted standard. It assumes our internal compass always points to true north, rather than stubbornly pulling us toward our own comfort and pride. When we’re left to write our own moral and spiritual rubrics, we inevitably do what the drifting machines do: we rewrite the standard to fit our current trajectory. We redefine our selfishness as “protecting our peace.” We rebrand our isolation as “setting healthy boundaries.” We look at our inability to forgive and label it “standing in our truth.” We keep the name of the objective—goodness, enlightenment, growth—but we completely alter the mechanism that measures it, allowing us to report significant personal progress while drifting further and further from the heart of God.

The truth is that human beings can’t pull themselves up by their own spiritual bootstraps. We’re deeply unreliable narrators of our own lives. We suffer from a blinding proximity to our own flaws, and when left alone in a closed loop of self-reflection, we’ll almost always justify our own decay. We can’t cure our condition within the boundaries of our own minds.

Theological transformation, therefore, is fundamentally anti-recursive. It demands an external invasion. This is the very definition of grace: an unmerited, disruptive force that comes from the outside. To avoid spiritual Goal Drift, we require an immovable, external standard we don’t have the authority to rewrite. We need the objective truth of God’s Word to serve as the unyielding plumb line by which our lives are measured.

But God, in His profound wisdom, didn’t merely give us a text to read in isolation; He gave us a people. We need external agents—our neighbours, our spouses, our church communities—to disrupt our self-confirming delusions. We need people who love us enough to notice the rubrics we’ve quietly rewritten for ourselves and tell us we’re drifting. We need others’ friction to expose the blind spots our internal algorithms have conveniently hidden.

The spiritual danger of the self-made human is the illusion that we can ascend to heaven on a staircase we designed ourselves. But the Christian life isn’t a project of autonomous self-optimization. It’s a posture of surrender to an external Saviour and a commitment to a messy, intertwined life with others. We’re not called to be self-contained engines of perfection; we’re called to be relational creatures, entirely dependent on the grace that breaks into our closed loops and saves us from ourselves.

The Friction of Fellowship

If the digital ecosystem is engineered to eliminate friction, the biblical vision of human flourishing rests entirely on it. To understand the divine mechanism of our formation, we must reclaim the visceral, physical reality of Proverbs 27:17: “As iron sharpens iron, and one person sharpens the wits of another.” In our modern, sanitized reading of the text, we tend to imagine this sharpening as a polite exchange of encouraging platitudes over coffee. We treat it as a gentle, frictionless alignment of minds. Yet the metallurgical reality is violent.

Anyone who has spent time under the crushing weight of a barbell, wrestled a heavy sandbag from the floor, or driven a heavy iron kettlebell overhead knows the undeniable truth of resistance. Muscle isn’t built by moving effortlessly through a void; it adapts and grows only through the deliberate application of opposing force. It requires strain, micro-tears, and the heavy, unforgiving reality of gravity. The spirit operates by the same physiological law. Sharpening a blade requires two hard objects striking at opposing angles. The impact produces intense heat, generates sparks, and physically grinds away the dull, useless edges of the metal. It’s an act of constructive destruction. The algorithms that mediate our modern lives serve as a digital shock absorber, buffering us from this abrasive reality and robbing us of the resistance we desperately need to grow.

To escape this algorithmic silo, many believers are drawn to the concept of the Wilderness. In a culture saturated with the empire’s synthetic noise, the impulse to retreat into the quiet margins feels profoundly biblical. But treating the scriptural Wilderness as an isolated hermitage or a personalized echo chamber where we can finally be left alone with our unfiltered thoughts fundamentally misunderstands it. From the wandering Israelites to the early church fathers, the Wilderness was never empty. It was the crucible of community. When we strip away the insulating conveniences and the empire’s curated feeds, we’re thrust into absolute reliance on God and immediate, unavoidable reliance on our neighbours.

Even in the deepest quiet, the Wilderness is where we encounter the “other.” It’s where we meet the foreigner, the outcast, and the unexpected messenger whose lived experience radically differs from our own. True biblical fellowship isn’t a curated algorithmic feed of perfectly aligned peers. It’s a collision of disparate souls, bound together by a shared, desperate need for Christ and not by uniform opinions.

When the Israelites fled Egypt’s highly structured, predictable machinery, they didn’t enter the desert as a homogeneous monolith; they left as a “mixed multitude.” They were a chaotic, abrasive collection of tribes forced to learn to live as a single covenant people. The early church was no different. The Holy Spirit deliberately engineered a fellowship that was violently diverse, bringing together political zealots and Roman tax collectors, strict Jews and uncircumcised Gentiles, slaves and free citizens. This community was engineered for maximum friction. The Apostle Paul’s letters largely record his efforts to manage the sparks that flew when these different irons struck one another. Yet that precise friction was the fire that forged a faith capable of turning the Roman world upside down.

Today, the digital ecosystem offers us the false utopia of a frictionless Egypt. We’re promised communities where we never have to explain ourselves, where our political and cultural priors are always confirmed, and where anyone who causes us even a moment of emotional or intellectual discomfort can be muted, blocked, and algorithmically banished to the outer darkness. We’re sold consumer relationships dressed up as community. In consumerism, we swipe away the brother who annoys us and unsubscribe from the sister whose perspective challenges our comfort. But covenant community is binding. In covenant, we’re forced to remain in the crucible. We’re tethered to people who’ll inevitably offend us, misunderstand us, and demand more of us than we’re willing to give.

This inescapable, uncomfortable reality is the antidote to our cognitive and spiritual atrophy. If recursive self-improvement is a machine’s myth because a system can’t correct its own blind spots, then covenant fellowship is the divine mechanism for our correction. Our brothers and sisters in Christ hold the mirrors we refuse to face. They carry the grace we can’t manufacture for ourselves. When we sit across a table from someone whose generational habits or cultural background chafes against our own, we feel the heavy, necessary strike of the iron.

We must learn to recognize the heat these relationships generate not as a sign that the community is failing, but as undeniable evidence that the sharpening is underway. To be formed into the image of Christ requires us to submit to the grinding away of our pride, narrow-mindedness, and self-reliance. It requires us to listen to the believer whose suffering looks different from ours, allowing their testimony to expand our constrained understanding of God’s sovereignty. It demands that we let someone else’s faith compensate for our doubt and that we offer our strength to bear someone else’s burden. We can’t do this alone in a digital silo. We can only do this in the messy, spark-inducing friction of the Wilderness, where the iron is heavy, the resistance is real, and the dull edges of our isolation are finally ground to dust.

Let the Walls Fall Down

The modern lexicon has quietly hijacked the concept of peace. In our current cultural moment, “protect your peace” has become a pervasive rallying cry, yet its practical application often amounts to profound isolation. It’s often interpreted as a mandate to unfollow dissenting voices, cancel difficult friends, and retreat into a hyper-curated sanctuary where every input aligns perfectly with our existing worldview. We’re encouraged to build impenetrable fortresses of digital solitude and call it wellness. But this isn’t the shalom of the biblical imagination, which demands the active, messy work of reconciling broken things. This modern retreat is the sterile, static silence of a machine that’s successfully walled off the outside world. It’s the false peace of a closed loop, humming quietly to itself right before it collapses under the weight of its own recursive noise.

To live faithfully in an age of digital alienation requires a radical return to incarnational reality. The way of Christ is fundamentally rooted in the physical, inconvenient collision of the divine with the dust. The Word didn’t broadcast a perfectly optimized, frictionless message from a distant, isolated server; the Word became flesh and tabernacled among us. He walked in the dirt, sat at tables with political rivals and social outcasts, and subjected Himself to the ultimate, agonizing friction of the cross. If we’re to model Christ in a disembodied age, our lives must be deeply, uncomfortably embodied. We can’t be incarnational in an echo chamber.

We begin to break down these digital walls by deliberately stepping into uncurated physical spaces. The algorithm can’t filter the neighbour we encounter at the park, the stranger we pass in public, or the opposing viewpoints in the audience at a local city council meeting. The algorithmic feed is a private kingdom where we’re the sole monarch; the physical public square is a shared inheritance where we’re merely citizens. These shared, tangible domains are essential antidotes to our virtual isolation. When we invest our time and physical presence in our local communities’ infrastructure, we’re forced to engage with the mixed multitude. We’re forced to look into the eyes of the very people we’ve been digitally trained to despise or dismiss. In the physical square, empathy is required, not optional. We’re abruptly reminded that our neighbours aren’t avatars to be debated but eternal souls to be loved.

This incarnational grounding is especially urgent for those of us raising the next generation. If we hand our sons and daughters over to the algorithm, we’re volunteering them for cognitive and spiritual atrophy. We must intentionally place them in environments where they experience the friction of physical play, the challenge of face-to-face conflict resolution, and the beauty of cooperative work. They need to feel the heavy, grounding weight of the real world—the bark of trees, the temperature of the room, the tone of a human voice—so they don’t mistake the synthetic, frictionless approval of a screen for the genuine affection of a community.

The challenge for every faithful citizen is to audit our own lives for the self-confirming loop ruthlessly. Where have we outsourced our intellectual or spiritual formation to a curated feed? Who’s the neighbour we’ve quietly muted because their lived experience clashes with our political or theological priors? We need to find the exact point where we’ve built a digital wall to protect our comfort, and tear it down. Let’s seek out the conversation that tests our patience and invite friction back in. It’ll be exhausting and profoundly inconvenient, but it’s the only way to ensure our souls remain awake.

We’re not artificial systems capable of recursive self-improvement. We can’t optimize our way to holiness, and we can’t compute our way to grace. When a machine is left alone to consume its own exhaust endlessly, it inevitably degenerates into madness. When a human being attempts the same, they wither into despair. We’re creatures formed from the dust, animated by the breath of God, and designed from the very beginning for communion. We’re built to be sharpened by the heavy, sparking iron of human friction. Don’t settle for the sterile, lonely perfection of the machine. Step out into the Wilderness, embrace the messiness of the other, and let the friction do its holy work.

This Substack is reader-supported. To receive new posts and support my work, consider becoming a free or paid subscriber.

Works Consulted

* Burtsev, Mikhail. (2026). Recursive Criticality of AI Self-Improvement. arXiv:2609.00137 [cs.AI].

* Chen, Mingguang, Licheng Wang, and Bo Qu. (2026). Recursive Self-Improvement in AI: From Bounded Self-Refinement to Autonomous Research Loops. arXiv:2607.07663 [cs.AI].

* Cunningham, Tom, Lukas Althoff, Basil Halperin, Brian Jabarian, Andrew Koh, Arjun Ramani, Phil Trammell, Parker Whitfill, and Cheryl Wu. (2026). The Economics of Recursive Self-Improvement. arXiv:2609.15802 [econ.GN].

* Duan, Yi, Ying Liu, Zirui Tang, Haodong Chen, Jun Zhou, et al. (2026). The Last AI Built by Humans: Toward Genuine Recursive Self-Improvement. arXiv:2609.11873 [cs.LG].

* Good, Irving John. (1966). “Speculations Concerning the First Ultraintelligent Machine.” Advances in Computers, vol. 6, pp. 31–88.

* Huang, Jie, Xinyun Chen, Swaroop Mishra, Huaixiu Steven Zheng, Adams Wei Yu, Xinying Song, and Denny Zhou. (2024). Large Language Models Cannot Self-Correct Reasoning Yet. International Conference on Learning Representations (ICLR).

* Schmidhuber, Jürgen. (2007). “Gödel Machines: Fully Self-Referential Optimal Universal Self-Improvers.” In Artificial General Intelligence, Cognitive Technologies, Springer.

* Shumailov, Ilia, Zakhar Shumaylov, Yiren Zhao, Nicolas Papernot, Ross Anderson, and Yarin Gal. (2024). “AI models collapse when trained on recursively generated data.” Nature, 631, 755–759.

* Tang, Hongyao, Yi Ma, Pengyi Li, and Yifu Yuan. (2026). Generalized Agent Iteration: One Formal Framework for Iterative Policy Improvement and Recursive Self-Improvement. arXiv:2609.13406 [cs.AI].

* Favaro, Marina, and Jack Clark. (2026). When AI Builds Itself. Anthropic Institute.

* The Economist. (2026). “When AI Builds Itself.” Babbage podcast, hosted by Alok Jha with Alex Hern.

* Zelikman, Eric, Eliana Lorch, Lester Mackey, and Adam Tauman Kalai. (2023). Self-Taught Optimizer (STOP): Recursively Self-Improving Code Generation. Conference on Language Modeling.

* 2001: A Space Odyssey. (1968). Directed by Stanley Kubrick; screenplay by Stanley Kubrick and Arthur C. Clarke. Metro-Goldwyn-Mayer.

* The Terminator. (1984). Directed by James Cameron; written by James Cameron and Gale Anne Hurd. Cinema ‘84 / Pacific Western Productions.

* Terminator 2: Judgment Day. (1991). Directed by James Cameron; written by James Cameron and William Wisher. Carolco Pictures / Lightstorm Entertainment.

* Star Trek: The Next Generation. (1987–1994). Created by Gene Roddenberry. Paramount Domestic Television.

Get full access to The Faithful Citizen at thefaithfulcitizen.substack.com/subscribe