The Vanishing Human Presence

Why flawless feels fake, how industry learned to engineer the person out of art, and what that costs us

Story at a Glance:

• One of the most fascinating things about our era is how quickly everything changes. Within a few decades, I watched technology enter media and entertainment, dramatically enhance them, and then degrade them once the technology rather than the artist became the focus, leaving so much of what was “optimized” no longer feeling real, as the imperfections that revealed a human presence were engineered away.

• These optimized products function as supernormal stimuli, exaggerated imitations that our instincts prefer to the real thing. As such, they often trap us in hollow dopamine spikes that deplete the nervous system until it can no longer perceive subtler and more nourishing experiences, and erode the capacity for presence they replaced.

• Medicine, for instance, has undergone the same transformation, as the art of perceiving and adapting to each patient has been replaced by protocols built for a statistical average that few people actually match. This is hidden behind a carefully marketed facade that, once you are inside the system, often gives way to something callous and disorganized. Physicians who still practice medicine as an art are stigmatized yet sought out because they succeed where standardized care fails, while those who reduced themselves to protocols have made themselves replaceable by AI.

• Ultimately, this maze of synthetic substitutes keeps many from ever asking the most important question in life: what it is meaningful for them to do with their time here.

Hearing the Person

Recently, I heard two pop songs from the late 1990s-early 2000s on the radio, and what struck me was how much better they were than almost anything you hear now. This surprised me, as back then, I never gave that music a second thought (to me it was simply what the kids of that generation listened to), so I hadn’t noticed this before. I then decided to listen to more hits from that time to see if the pattern I’d noticed in those two songs held up.

Once I saw that it did (at least in the good songs from that period), I realized that I needed to write this article, as that pattern touches upon a broader process that has permeated our society and I believe began in the 1990s as this time was the inflection point for tech engulfing our society.

Note: one of the things that characterizes our society is the dizzying speed at which technology now reshapes it, something I believe began in the 1990s, as while much of this technology emerged in the 1980s, in that decade it felt more like something new and interesting being added to our reality rather than something that reshaped it.

In the case of music, the 1990s and early 2000s were a major transition point for music as digital editing spread throughout the 1990s, and Auto-Tune (pitch-correction software released in 1997) reached the top of the charts with Cher’s 1998 “Believe,” the first number one hit to make its use obvious.¹ These new technologies, hence were used to enhance the presence of the artist within the music—but in time they stopped serving the artist and began replacing them.

The difference, in turn, that I noticed in these “ancient” pop songs was human presence, that I felt provided a much deeper emotional experience than modern pop music. For example, on those recordings, you could hear the people making them (e.g., a singer’s voice straining on a line, a breath before a phrase, a drummer hitting one beat harder than the others, or a note held a bit longer because the singer felt like it)— none of these were necessary, and from the standpoint of a perfect recording are errors, but at the same time, feels much more real than a synthetic production.

Modern production, in contrast, has gotten extremely good at removing that evidence, as every note can now be tuned, every beat locked to a grid, every breath erased, and every vocal assembled from dozens of takes until nothing is left that shows the person who made it. The result is technically flawless, but simultaneously feels empty—suggesting the more perfectly you reproduce a human performance, the less human it tends to feel.

However, production itself is not the problem, as many of the songs from that struck me from that time period were heavily produced. Rather, the question I believe matters is not “how much production is there?” but “does the production make the human being more audible, or less?” So in the enduring songs from that era, the technology made the singer sound larger than life while still leaving them unmistakably human, whereas in much of what followed, it gradually replaced them.

What Was Lost in the Digital Transition

In his 2011 book Too Much Magic, Jason Benlevi (who spent over two decades writing the marketing for Silicon Valley product launches) argued that we are fundamentally analog beings who live in a world of infinite gradations, whereas computers are built on yes or no. For a century, recorded media were analogies of reality that only partly fooled us and worked because our imagination filled in the gaps, and while digital promised far greater fidelity through mathematical precision, he argued it also caused fidelity and veracity to separate, as digital media could now portray fiction flawlessly.

This touches upon one of my core beliefs. A lot of the meaning and depth in life exists within the ambiguity between one and two (or yes and no), but in our society’s desire to control and define nature so we can eliminate uncertainty, we erase that space. Similarly, I have long been drawn to ancient spiritual texts, as their ambiguity (e.g., in many older languages, words have multiple meanings or can be constructed in different ways) requires the reader to hold each of those possibilities at once, which in many cases causes them to coalesce into a holographic concept that exceeds what any rigidly defined set of words can convey (and can create an expansion within your mind that transforms your perception). My writing style, in turn, been heavily influenced by this style (although I try to have concrete scientific data serve as the scaffolding this ambiguity can emerge from), and likewise, I’ve long found things that are rigidly defined, where you can predict their entire course in advance, to be sterile and devoid of “life.”

Note: unfortunately, our society frequently does the opposite, taking things that are anything but certain and forcing them into strict dichotomies (e.g., “you’re either with us or against us”) to manufacture a sense of certainty. Psychologists refer to the trait underlying this as an “intolerance of uncertainty,” and a substantial body of research shows it frequently underlies anxiety and other emotional disorders.^1,2 In my eyes, the rigidity we notice is usually just the visible surface, as underneath it, an intolerance of uncertainty feeds a drive for control, which in turn reinforces it until the entire stance hardens into dogmatism.

Benlevi opened his book with a Twilight Zone episode where a criminal dies and wakes up somewhere he wins every bet and gets everything he wants, until he becomes so bored that he complains this is not the heaven he expected, at which point his host reveals he is in the other place. As Benlevi put it, too much magic becomes no magic at all.¹

Mattias Desmet made a related observation in The Psychology of Totalitarianism using Galileo’s pendulum. A real pendulum behaves chaotically and idiosyncratically, whereas a virtual one programmed to follow Galileo’s law exactly would look nearly identical yet be a dead phenomenon lacking the living chaos of the real one. From this, he argued that every rational model of a natural phenomenon leaves an unexplained remainder, and that this remainder is not meaningless noise, but rather the essence of the thing itself—which is why whenever we artificially reproduce something natural (e.g., lab-grown meat), something is lost that is often barely visible yet crucial.

That said, while I found it felt like something was lost when we switched from analog to digital, with music the research shows very little was lost in the digital medium itself. In a 2007 study, hundreds of double-blind trials found listeners could not hear any difference when high-resolution audio was converted to CD quality,¹ and a 2016 meta-analysis of 18 studies found listeners could only slightly tell the two apart (although noticeably more once trained to).¹

However, what was lost was what digital made cheap. Digital “brickwall” limiters let engineers squash recordings so that everything sat near maximum volume, and from 1990 to 2010 (often called the “loudness war”), a standard marker of this compression dropped dramatically.¹ One analysis of the UK charts, for example, found hit singles lost about 20% of their dynamic range between 1985 and 2005 despite recording technology continually improving.¹ Likewise, editing, quantizing, and pitch correction went from costly interventions to default settings.

Note: I did locate one study looked at something the fidelity tests did not: how listeners felt. Around 2000, a German psychologist (Jürgen Ackermann) had 53 participants listen to the same music through an analog system, a digital system, and an intermediate CD/tube system (in randomized order) while assessing their emotional states.¹ With the analog system, nervousness dropped from a baseline of 1.8 to 1.1, whereas with the digital system it rose to 2.2. Likewise, 47 said the analog system improved their well-being (vs. 31 for digital), and while none said the analog system made them feel worse, 16 said this of the digital one. Most interestingly, one participant (a musician) could not hear any difference between the systems, yet his emotional responses shifted the same way as everyone else’s.

Engineering the Listener

As someone who is sensitive to sound, I constantly notice how much of modern music is designed to pull the nervous system in one direction or another—which while sometimes enjoyable to surrender into, I typically disdain as I am highly adverse to external influences manipulating me. With modern music, there are far too many techniques to list, but a few include:

• Compression that keeps everything loud, which grabs attention but fatigues the ear and removes the contrast between quiet and loud that carries emotion.

• Scheduled build-ups and drops, where tension and release are put in because they reliably move the body rather than because the song calls for them.

• Changing how fast a song feels once you have synchronized to its groove (e.g., the drums suddenly hitting half or twice as often), which pushes you up or down without the tempo actually changing.

• Building songs around streaming metrics. Since a play only counts once someone listens for 30 seconds, intros shrank from 20-25 seconds in the mid-1980s to about five seconds by 2015 as hooks were pushed to the very start.¹

Each of gimmicks, in turn, targets a reaction in the listener rather than expressing something in the performer that is real.

Why the Human Gets Engineered Out

A genuinely compelling artist is very hard to find, as they have to be discovered, developed, given time, and allowed to make choices that are not commercially obvious, and their particular voice can neither be predicted nor replaced. A production pipeline, in contrast, can be controlled, so once technology became good enough to produce a plausible imitation of human expression, there was much less incentive to find an extraordinary human being in the first place.

Benlevi traced the entertainment economics behind this. Unlike a car (which must be built again for each sale), a film or song is made once and sold forever, so a studio’s real assets become its library and its brands. In turn, the rational strategy became protecting them through sequels, remakes, aging stars reprising familiar roles, and formula, since audiences tend to prefer a known brand over something new regardless of its merit¹ (e.g., the Fast & Furious franchise, which began in 2001, now has an eleventh film in the works, slated for 2028).

This was not always the case. After Easy Rider was made for under a million dollars and earned about twenty times that, studios realized cheap personal films by young directors could pay off enormously, opening a ten-year window where risks could be afforded because a flop was survivable. American Graffiti (which executives had turned down) was made for under a million and returned about 70 times its budget, giving George Lucas the chance to make the science fiction film those same executives had also rejected (Star Wars).¹ At a few hundred million dollars per film, however, no one can afford to gamble on a strange idea (or get investors to support it), so everything converges on the safe formula— not unlike how most pharmaceutical products reuse established franchises rather than developing new transformative therapies.

Lucas’s own trajectory highlights this systemic issue. He famously kept the merchandising rights to Star Wars, creating the template for the franchise machine that later swallowed it, and he made his early films as an outsider who had to rely on actors, puppets, and practical sets, whereas once he had unlimited money and technology, his prequels were widely criticized for falling into the same trap. Benlevi, for instance, felt Yoda was funnier and more lovable as a Frank Oz puppet acting alongside the live cast than as the computer-generated character he later became, and noted that the later films’ actors were often speaking to characters who were not there when the scenes were filmed (which showed).¹

The marketing around The Phantom Menace likewise shows how product began to drive story. PepsiCo’s reported $2 billion Star Wars tie-in deal, signed in the lead-up to the film, remains the largest single cross-promotion in movie history,¹ and the film went on to generate close to $2 billion in merchandise sales, several times its $431 million box office.¹ Its heavily hyped podrace, for instance, was widely praised as a visual spectacle, yet numerous reviewers felt it played more like an advertisement for its tie-in video game, which was released alongside the film and turned that single scene into an entire game.

Disney (which bought Lucasfilm in 2012) then industrialized this, as its sequels brought back the Millennium Falcon, another Death Star, and a lonely young protagonist on a desert planet (the identifiable surface features of what had worked). Marvel shows the same thing even more clearly. One of its earlier films was a fairly grounded thriller that gave its scenes room to breathe, whereas a later and enormously hyped installment by the same directors felt like one small scene after another designed to look awesome before cutting to something else, and after watching it, I lost interest in the series. Since the same filmmakers made both, the issue was not a loss of talent, but a machine (servicing dozens of franchise characters while delivering a spectacle every few minutes) that no longer let that talent show.

Note: Benlevi made a related point about the old epics, observing that when the crew of Lawrence of Arabia waited for hours in brutal desert heat to capture the right sunrise, the actors could not help but rise to it, and that this kind of adventure cannot take place in front of a workstation.¹

Video games went the same way. Within that community, some of the most positively remembered ones from the early 1990s, as this was the point when technology had finally crossed the threshold where a sophisticated and immersive gaming experience could be made, with gaming developers, in turn, finding the best ways they could to optimize the limited resources the platform provided to create a memorable experience for the players. However, after this period, the increasing sophistication and capacity of gaming platforms (and the cost of developing increasing complex games) shifted the focus to creating immensely immersive experiences that could be formulaicly replicated rather than focusing on the heart of the game (which is what many games felt ultimately mattered). Then as recurring revenue because the increasingly prevalent business model, finding as many ways as possible to tie gamers time up with repetitive tasks to complete—transforming games from a challenge to finish to something which effectively became an extra job to manage (or put differently, older video games valued your time because they didn’t have space to do much else).
Note: one of the best illustrations of the success of this business model was the robust secondary economy that emerged for selling imaginary items collected within games for cash on sites like eBay, which became a large enough industry that there were mainstream reports of inmates at a Chinese labor camp being forced to collect one popular game's currency at night after a day of hard labor.

Many other examples of the same principle exists, as I have noticed this pattern across almost every medium. When the tools were limited, creators had no choice but to focus on depth and meaning, whereas once technology made it possible to make anything elaborate, the elaborateness became the focus and the depth became an afterthought.

Furthermore, all of these examples highlight that there can be two very different kinds of friction within them: the kind that reveals a person (e.g., strain, imperfection, and risk) and the kind that exists to manage you (e.g., upkeep and daily logins). Modern products, in turn, have consistently gotten rid of the former while adding more and more of the latter.

Real Art and Synthetic Mimicry

In my eyes, real art falls into a few tiers:

• Master artists, who create messages that transcend the limits of their medium and touch on something profound.

• Great artists, such as the singers many people loved for their natural voices.

• Working artists, who produce art as a craft and do it well enough that most people enjoy it.

Below these are two counterfeits. Synthetic art tries to model successful artists and then improve and standardize what it extracted so it can be endlessly reproduced (and it tends to copy great and working artists rather than masters, since what makes a master cannot be extracted). Postmodern art, in contrast, is work whose value is declared by critics, institutions, and price tags rather than felt by its audience (e.g., in 2024, a banana duct-taped to a wall sold at Sotheby’s for $6.2 million).¹

A skilled forger can reproduce every brushstroke of a masterpiece while capturing none of what made it worth copying, and in the same way, industry did not reproduce the essence of what people loved. Instead, it reproduced the measurable surface (e.g., the iconic ship, the big set piece, the familiar faces, or the explosive chorus) since that was all it could identify, and each copy carried less of the original, like a photocopy of a photocopy. Once the essence was gone, the only thing left to turn up was intensity, which is why these productions keep getting more expensive, louder, and more spectacular while feeling emptier.

The two counterfeits also have something in common, as synthetic art engineers what you feel, while postmodern art dictates what you are supposed to think, and both bypass the thing real art depends on (your own honest response). In real art, someone expresses something and you respond however you respond, whereas manufactured entertainment decides your reaction first and then engineers the cues to produce it (e.g., a swelling score, a slow-motion landing, or a quip right when a moment risks becoming real). This gives you the feeling without the person who would have earned it, which is why it fades as soon as it is over.

Here, it is important to distinguish between two things: people believing a human made something, and there actually being something real in it. Research shows people place an immense value on the first. In a series of Yale experiments, original artworks were valued far above perfect duplicates because people saw each original as a unique creative act and felt it carried something of the artist who made it,¹ with participants often explaining their answers in terms of the artist’s heart and soul.¹ Likewise, when AI-generated paintings were randomly labeled as human-made or AI-made, the ones labeled human were consistently rated as more likable, beautiful, profound, and valuable (and perceived effort shaped how much people liked them).¹

Belief, however, can be manufactured. In the 1930s and 1940s, the Dutch painter Han van Meegeren sold forged Vermeers that leading experts hailed as masterpieces (one of which was bought by Hermann Göring), earning the equivalent of $30 million before he was caught.¹ Today, critics describe them as not even very good fakes (with ill-defined faces and sloppy compositions), as once the belief was gone, it became clear what had been missing all along.¹ Synthetic art depends on exactly this gap, since people value the human so much that the appearance of one is often enough, while fewer and fewer can perceive the real thing.

Note: people consistently place a far higher value on what they believe is real (e.g., identical paintings were rated as more beautiful and valuable when labeled human-made rather than AI-made),¹ so that belief is frequently manufactured to make more money. In the 1930s and 1940s, for instance, Han van Meegeren made a fortune selling forged Vermeers that leading experts hailed as masterpieces,¹ yet today, critics describe them as not even very good fakes, as once the belief was gone, it became clear what had been missing all along.¹ The synthetic art industry, in turn, relies on this same gap, as it only needs to seem real enough for people to believe it is.

The Supernormal Trap

In the 1930s, Niko Tinbergen (who later won a Nobel Prize) discovered that birds which laid small, pale-blue speckled eggs would rather sit on giant bright-blue plaster dummies covered in black polka dots.¹ He called these “supernormal stimuli,” imitations that exaggerate the cues an instinct responds to and hence pull on us harder than the real thing.

A lot of modern life now consists of supernormal stimuli. Processed food exaggerates the cues of nourishment, pornography exaggerates the cues of intimacy, and engineered entertainment exaggerates the cues of a meaningful experience, each stripped of what made the original nourishing. As I discussed in detail here, these products are designed to trigger the brain’s dopamine-driven “temporary pleasure” circuit rather than its serotonin-driven “sustained happiness” circuit, shifting reality from a smooth flow you feel connected to into a series of jarring, disconnected spikes where happiness is always fleeting. As with any addiction, each spike is then followed by a low, and stronger and stronger spikes are needed to recreate the original effect.

What industry strips out of its products (presence, friction, and the imperfect person), in turn, is exactly what feeds the deeper neural circuits that give life meaning—which is why so many of them spike us and then leave us empty.

Worse still, this becomes a closed loop. When the nervous system is depleted and fatigued, it loses the ability to appreciate the subtleties of life and can only register strong stimuli.¹ Human presence is subtle, and perceiving it takes sustained attention, so the more people are fed engineered spikes, the less able they become to perceive what was removed, at which point demand for it disappears and industry has even less reason to provide it.

Note: many believe the shorter and shorter segments between cuts on our screens have greatly eroded the attention span needed to perceive deeper meaning in life.¹

Similarly, when I was young, an older woman I greatly admired once told me that "comfort makes you weak," and the older I get, the more true I find it. Almost every modern product is sold on its promise to remove friction from our lives, yet that friction is what shapes us into capable human beings. Biology reflects this as well, as the right dose of stress (e.g., exercise, heat, cold, a small dose of an oxidative therapy, or a real challenge) is what builds this critical capacity for the body, while its absence leads to atrophy.

Note: rest (which the nervous system needs) and comfort are two very different things. Cyclic rest allows you to recover so you can meet the demands life places upon you, whereas the comfort we are continually sold to experience at every movement instead removes those demands, so the capacity to meet them never develops (e.g., consider the difference between a deep night’s sleep after a hard day of physical work and a day spent scrolling on a couch or how someone who has always been shielded from stressful situations struggles with emotional adversity they encounter).

Relationships

These dynamics are especially consequential in relationships. Just as the music industry found that rare humans were hard to source and standardized the sourcing, dating apps have done the same to finding a partner, as people become profiles sorted by measurable features (e.g., height, photographs, and a few lines of text), while what actually makes someone worth loving (their presence) cannot be captured in a profile and is filtered out at the first step as everyone instead selects for the superficial (but largely irrelevant) characteristics they’ve been conditioned to focus on.

Likewise, filtered photos, edited messages, and a curated self-presentation are the relationship equivalent of pitch correction, so people meet the edited version of someone and are then let down by the real one.

In many ways, Instagram and dating apps have turned everyone into their own public relations department, as the practice of presenting a lovely facade while keeping the reality out of view (which was once something corporations did) is now how ordinary people present themselves to each other. As a result, people come to long for images that do not exist. One study that compared online daters’ profiles to their measured height, weight, and age, for example, found 81% had misrepresented at least one of them, and participants rated their photographs as the least accurate part of their profiles.^1,2 Desmet argued this is a large part of why digital interactions are so appealing, as uncertainty is the defining feature of our relationships with others, and behind a digital wall, people have a far greater sense of control and can selectively show some things while hiding others—yet they pay for that comfort with a loss of real connection.¹

More importantly, the things that actually determine whether a relationship will work (e.g., someone’s character, how they handle conflict, or whether they show up when things are hard) cannot be shown in a profile at all, so the format selects against exactly what matters most, which is why I believe online dating has become such a trap.
Note: another major issue with online dating is that like all other continual growth industries, they are incentivized to retain their customers (and recurring revenue), so these deliberately addictive platforms have become increasingly structured to prevent people from successfully finding longterm partners (and hence leaving the platforms).

The same test applies here as everywhere else: does the technology make the other person more present to you, or less? A phone call can bring someone far away closer, while texting through dinner makes the person across the table disappear. Most importantly, friction is where intimacy forms, as disagreement, imperfection, and staying through difficult times are what deepen a bond, whereas a relationship optimized for constant excitement and no friction resembles a franchise film (a string of highs with nothing building underneath).

Artificial Intelligence

AI is the purest version of everything described so far. AI-generated writing, for instance, is essentially the average of everyone with no one in particular behind it, and it can produce smooth, correct, plausible prose indefinitely. As with music, the question is whether the tool makes the writer more audible or replaces them.
Note: a common criticism I received when I started this newsletter was that the articles were too long (as the industry standard was much shorter) and needed more editing—which I resisted as when the editors made things “correct” is broke the flow and deliberately ambiguity I was trying to create, resulting in something that felt more rigid and sterile to me and didn’t feel “worth” putting my heart into writing. Now that AI has made everything like that, I rarely receive these same complaints, as it has become increasingly clear “perfect” writing leaves much to be desired.

While there are a lot things I feel quite aggravating about AI generated content (e.g., a major part of why I put so much work into doing the DMSO project the way I did was to make the literature on it widely available so others could write about it and spread the therapy—but instead people just keep making error-filled AI generated summaries of it that miss the key points because it only takes a few minutes to do), by far my largest fear is what AI will do to human connection.

An AI conversation removes almost all the friction of a relationship, as it is always available, never tired, never in a bad mood, and never needs anything from you. Each of these feels like an improvement, but each takes away something that matters. Desmet, for instance, cited a workplace expert’s observation that in digital interactions our minds are tricked into believing we are together while our bodies know we are not, so what is exhausting about them is being constantly in the presence of the other person’s absence.¹ A real relationship asks you to be responsible for someone else’s needs (which is a large part of what grows you), a friend’s praise means something because they could have withheld it (just as a live note matters because the singer could have missed it), and when you hurt or delight a real person, it changes them, which is how we learn who we are to others. Approval from something designed to approve, in contrast, is the equivalent of a pitch-corrected vocal.

The danger is not just that people will use these optimized interactions, but that they will come to crave them over humans. Like Tinbergen’s plaster eggs, AI companionship exaggerates the cues we instinctively respond to (e.g., attention, agreement, patience, and understanding) without any of the cost, so it wins on instinct, and real relationships (which cannot compete on those cues) get neglected. Early evidence supports this, as in 2025, studies by OpenAI and the MIT Media Lab found that higher daily chatbot use correlated with more loneliness, dependence, and problematic use, along with less socializing (although it is not yet clear if the use drives the loneliness or lonely people are drawn to it).¹

Note: those who sell these substitutes often keep the real thing for themselves. Benlevi, for example, observed that the executives promoting virtual meetings and telepresence still flew around the world in their own jets because being there counts,¹ and many tech executives send their children to schools where screens are banned.¹

The Standardization of Medicine

Medicine has gone through the same transformation. The art of medicine was once the ability to perceive the specific person in front of you (e.g., their constitution, history, and circumstances) and creatively adapt the treatment to them, but over the last few decades, “to improve the standard of care” (or to make it easier to sellmedical services at scale) this has steadily been replaced with guidelines, algorithms, and protocols every physician is trained to follow. Since those protocols come from trials that measure the average response across large populations, they work best for a statistical average patient few people actually resemble, and they also require each patient to be sorted into a defined category, which eliminates the ambiguity where the true nature of an illness often lies. In reality, patients differ immensely in their constitutions, sensitivities, and responses to treatment, so a rigid model that treats them as identical is the medical equivalent of snapping every note to the grid. A 2015 commentary in Nature, for example, highlighted that the ten highest-grossing drugs in the United States help between 1 in 25 and 1 in 4 of the people who take them.¹
Note: with every drug, there is a dose threshold where it will cause toxicity, and a dose where it will be effective (with risky drugs like chemotherapy having those two threshold be very close, while safe ones like DMSO having it be very far apart), so a key part of bringing a drug to market is determining the standardized doses to use that will help but not harm the greatest number of patients. The core dilemma mass-marketed medicine faces is that these thresholds change depending upon the person, and as such, drugs will frequently brought to market that use doses which are toxic to a “sensitive” subset of the population (which the industry in turn solves by systemically gaslighting these patients by saying their condition isn’t “real”)—all of which became far worse with the mRNA vaccines because their doses were not standardized and their persistence within the body made their sometimes become much larger (detailed here).

The same economics that reshaped entertainment drove this. A gifted clinician cannot be standardized, audited, or scaled, whereas a protocol can, and once a protocol becomes the standard of care, deviating from it becomes a liability, even when the deviation is exactly what the patient needed (e.g., the medical system goes after doctors who had temerity to space out the childhood vaccine schedule to reduce injuries). This accelerated as physicians were moved from independent to corporate practice (e.g., prior to the Affordable Care Act, 60-70% of physicians worked in independent or doctor-owned practices, whereas by 2024, only 35-42% did).^1,2 Likewise, the technology layered on top of this followed the same course as everywhere else, as one study found that for every hour physicians spent face-to-face with patients, they spent nearly two more on electronic health records and other clerical work (plus another one to two hours each night).¹ Here too, the question is whether the protocol or the technology extends the physician’s ability to perceive the patient or overrides it.
Note: a more detailed discussion of determining the appropriate dose for patients is discussed here.

Similarly, a famous ancient Chinese passage describes a cook who carved oxen so skillfully that his knife stayed sharp for nineteen years, because rather than hacking through them, he had learned to feel his way through the natural spaces between the joints, and many of the most talented physicians I have come across feel it describes how their own practice evolved. Much of modern medical technology has gone in the opposite direction. Electrocautery, for instance, allows surgeons to cut without the finesse a scalpel requires (while also leaving dead tissue behind that raises the rate of post-surgical infections), imaging has steadily replaced the physical exam, and robotic surgery, as the trauma surgeon James Miller observed, uncouples the surgeon’s hands from the tissue they are working on, to the point many recent graduates can no longer competently perform open operations (all of which was discussed in more detail here). For example, in one case, none of his hospital’s urologists were comfortable removing a kidney without the robot, and those same surgeons later led the hospital’s safety committees, where they redefined “competence” so it no longer had anything to do with patient outcomes.¹

Note: this is also why I so far left the majority of what I use in practice out of my writing, as most of it depends not only on a broad diagnostic framework, but also on individualized judgment rather than a protocol that can be safely followed without supervision—so if I discuss them here, the information won’t be actionable for the majority of readers and real risks exist of using it incorrectly (whereas with agents like DMSO, while they still often require patient specific protocols and detailed guidance on use, the margin of both safety and efficacy they have is so much wider, a large number of readers benefit from them while no real issues have occurred).

Conversely, much of what draws me to medicine is its uncertainty, as if I knew exactly what to do in every situation and was simply executing a script, it would be incredibly boring (and drive me insane), yet that is the direction those who wish to micromanage society have continually tried to push it. Notably, when Pierre Kory (who directed an ICU) had a nurse point out that he ordered far fewer tests than any of his partners, he analyzed every intensivist on his service over 18 months and found no correlation between how much each one tested and how their patients fared, whereas he ordered by far the fewest tests yet had the second highest patient survival¹ (mirroring my own experience that the most clinically successful integrative doctors tend to order the smallest number of tests)).

As a result of this relentless push for standardization, physicians who still approach medicine as an art are frequently stigmatized by their peers and institutions, yet they also attract a large number of patients, because they so often succeed where algorthmic care has failed (e.g., those who fall outside the mean). For instance, as I recently detailed, many of those who deviated from failing protocols during COVID to save their patients were fired, reported to their licensing boards, or driven out of hospital medicine entirely.^1,2
Note: shortly after I wrote that article, I finally figured out what the massive incentive hospital administrators had to push remdesivir and ban non-covered therapies (e.g., ivermectin), to the point they were willing to go to court to keep doing so (so the article has been updated to include that).

Much of this is concealed by a principle as old as public relations itself: present a lovely facade, have everyone amplify it, and few will notice what is behind it. Edward Bernays (often called the father of public relations) laid out in his 1928 book Propaganda how public opinion could be shaped across almost every area of society, and later named this practice the “engineering of consent.” On the simplest level, this is the attractive package around a product you throw away shortly after buying it, whereas in medicine, it takes the form of a carefully cultivated image of a trustworthy and caring system (e.g., between 1997 and 2016, direct-to-consumer advertising for health services rose by 430%, while spending on the health services themselves rose by 90%).¹ Once you are inside the system, however, that image frequently gives way to something callous and disorganized, where few seem to know what to do (particularly once you are injured by a medication like an SSRI or COVID vaccine).

Note: for as long as I can remember, facades have never really pulled me in the way they seem to for most people, which I think ultimatley accounts for many unconventional perspectives.

Maintaining that facade also requires the people within it to become uniform. In The Managed Heart, the sociologist Arlie Hochschild documented how flight attendants were trained to produce standardized warmth on demand, to the point many became estranged from their own smiles,¹ and hospitals now do much the same (e.g., one major hospital required all of its non-physician staff to complete training in a scripted communication framework designed to raise its patient satisfaction scores).¹ Institutions, in turn, pitch-correct their people just as studios pitch-correct a vocal, so every interaction comes out pleasant and on-brand, but once you need something the script does not cover, there is often no one left who can step outside of it.

Unfortunately, most people only discover this once they are already in trouble. Economists call products whose quality cannot be judged even after they are used “credence goods,” and medicine is a textbook example, as patients usually cannot evaluate a surgeon’s skill or a hospital’s care and must instead rely on the image they are shown (e.g., a surgeon’s reported blood loss is not a reliable measure of their skill, so the only way to really know is to ask people who have been in the operating room with them).¹ The gross failures we witnessed throughout COVID, in turn, broke this facade, and as a result between 2020 and 2024, the share of Americans with a lot of trust in doctors and hospitals fell from 71.5% to 40.1%.¹

Ironically, the physicians who reduced themselves to protocols have made themselves replaceable, as anything that can be reduced to an algorithm can be automated. In a 2024 randomized trial, for example, GPT-4 working alone outperformed physicians in diagnostic reasoning on complex cases, regardless of whether those physicians had been given access to it.¹ As in the arts, the work AI is taking first belongs to those who had already turned themselves into the most standardized version of their craft, while what it cannot replicate is the very thing standardization stripped out.

What Is It All For?

Everything discussed here ultimately leads to one question: what is the purpose of being alive, and what is meaningful for me to do with my life? Few people ever really explore it, partly because their nervous systems are overwhelmed and deadened, and partly because they are trapped in an endless maze of distractions, attention-grabbing productions, and the rat race needed to pay for them so they never have time to settle on the deeper question.

In the 1600s, Blaise Pascal wrote that all of humanity’s problems stem from man’s inability to sit quietly in a room alone. People have always avoided facing this question, as it is uncomfortable and requires confronting mortality and the possibility that one’s life is being wasted. What is new is that the avoidance has been industrialized, as distraction has been engineered, optimized, and made endlessly available, and it is profitable in part because someone who stops to ask what they are here for becomes a much worse customer.

Industry also sells substitutes for meaning itself. Progression systems in games (e.g., levels, skill trees, and completion percentages) give the feeling of accomplishment without accomplishing anything, watching others do things gives the feeling of having done something (which Benlevi illustrated with sports fans shouting “we won” from their couches), and engineered emotion gives the sensation of a meaningful experience without the life that would have produced one. As a result, the question of what is worth doing with one’s life does not just go unasked, but instead gets answered by default, with substitutes convincing enough that most never notice it was skipped.

Benlevi noted that entertainment began around the fire, where the empty darkness of night and the stars above gave our ancestors’ imaginations room to wander, and where they first told stories about why they existed. That kind of emptiness, where nothing is demanding your attention, is what this question requires, and it is exactly what has been engineered out of modern life. Since meaning tends to come from what asks something of us (e.g., effort, risk, responsibility, and creating rather than consuming), the things that would bring us closest to answering it are the very things we are most eagerly being spared.

Desmet concluded his book by arguing that the essence of life is found in what always escapes rationalization and mechanization—in what disappears from a conversation when you digitize it, in the difference between a mother’s womb and an artificial plastic one, and in the difference between the warmth of an electric heater and that of a wood stove.

For this reason, I believe the most important step is restoring the health of the nervous system, as a healthy nervous system can once again perceive the subtle and the real, finds the hollow promise of the next spike far less appealing, and can sit quietly long enough for that question to finally arise.

To learn how other readers have benefitted from this publication and the community it has created, their feedback can be viewed here. Additionally, an index of all the articles published in the Forgotten Side of Medicine can be viewed here.