Chapter 43: Enhancement Outside the Clinic — Work, Sport, Military, and the Coercion Question

"A man who has a why to live can bear almost any how." — Friedrich Nietzsche, Twilight of the Idols (1889), on the durability of reasons — the question this chapter asks is what happens when the why belongs to somebody else.

Overview

Every previous chapter in this book had a patient somewhere in it.

Sometimes the patient was explicit — a person with type 2 diabetes, a person with growth hormone deficiency, a person recovering from surgery. Sometimes the patient was implicit, an assumed reader wondering whether a compound might help with something they actually had. But the structure was always the same, and it is the structure medicine spent two centuries building: somebody has a condition, a clinician who owes that person a duty judges whether an intervention is likely to help more than harm, the person consents or declines, and the outcome is somebody's responsibility.

This chapter is about what happens when you remove all of that and keep the molecule.

That is what enhancement is. Not a different drug — frequently the same drug — used by a person who is not sick, chosen without a diagnosis, evaluated without a clinician, and justified by a benefit that is often relative rather than absolute. The compound does not know the difference. Every other part of the situation does.

There is a version of this chapter that would be easy to write and useless to read: the one that concludes enhancement is wrong, or the one that concludes the objections are squeamishness. This book takes neither position. Whether people may pharmacologically improve capacities they were not born with is a values question, and pharmacology does not settle values questions.

What this book does insist on is narrower and more useful. Every position in this debate contains empirical claims, and those claims can be evaluated. "Sanctioned enhancement would be safer than the black market" is an empirical claim. "These compounds meaningfully improve performance in healthy trained adults" is an empirical claim with an answer, and the answer is not the one most people assume. "Enhancement is inevitable" is a prediction about human behavior dressed as an observation about human nature. Sorting the values from the facts is most of the work, and almost nobody does it.

In this chapter, you will learn to:

• Name the four things clinical medicine supplies that enhancement removes, and explain why their absence changes the proposition even when the molecule is identical
• Describe what is documented and what is merely assumed about enhancement in competitive workplaces
• Explain what sport's enforced answer has cost and what it has achieved
• State why "voluntary" means something different inside a hierarchy
• Give the coercion argument formally, and say where it bites hardest and where it barely bites
• Present the level-playing-field argument at full strength before criticizing it
• Explain why enhancement combines maximum uncertainty with minimum oversight, by construction
• Rate the empirical claims embedded in ethical positions, and notice where the rating system strains
• Build the list of conditions that would make enhancement defensible — and count how many are met

Learning Paths

All five paths should read §43.1, §43.5, and §43.7. They are the structural spine, the core argument, and the safety consequence, and nothing else in the chapter makes sense without them.

💊 GLP-1 — §43.2 and §43.5 apply more directly to metabolic drugs than most readers expect. Appearance and body composition are workplace-relevant in some occupations, and the coercion structure does not care whether the goal is a heavier deadlift or a smaller waist. Chapter 44 carries this to population scale. 🏋️ Performance — this is your chapter, and §43.3, §43.7, and §43.8 are the sections to sit with. The rating in §43.3 is the one most likely to be unwelcome; read the reason line twice. 🔬 Science — §43.5 and §43.6 are where the reasoning is densest. The coercion structure is a formal argument with a shape you will recognize from game theory, and §43.6 is an exercise in presenting an opposing position at full strength. 💄 Cosmetic — §43.2 and §43.5. Appearance-based enhancement in customer-facing and appearance-adjacent work is the least-discussed application of the coercion argument and possibly the most widespread. 🏥 Clinical — §43.1, §43.7, and §43.9. You will meet these situations, usually in the form of a patient asking about something you did not prescribe and do not have data on. Chapter 39 is the conversational companion to this one.

43.1 Enhancement when there is no patient

Start with a description rather than an argument, because the argument is downstream of the description and most disputes about enhancement are conducted without it.

Clinical medicine supplies four things. They are so ordinary that they become invisible, which is exactly why removing them goes unnoticed.

A diagnosis. Somebody has determined that a specific thing is wrong. This is not a formality. A diagnosis is what converts a symptom into a target, and it is what makes it possible to say whether a treatment worked. Without a diagnosis there is no defined endpoint, and without an endpoint there is no way to distinguish improvement from the ordinary variation of being a person.

An indication. A regulator, having reviewed trial data, has stated that this drug is appropriate for this condition in this population. An indication is a compressed summary of an enormous amount of evidence, and it carries a specific meaning: in people like the ones studied, at the exposures studied, the benefit exceeded the harm. Note the qualifiers. They are load-bearing. Every one of them fails to transfer when the population changes.

A clinician with a duty. Somebody with training is making a judgment for this specific person, and is professionally and legally accountable for it. This is the piece that is hardest to replace and easiest to forget. The value is not merely expertise — expertise can be found in a textbook. The value is that a fiduciary relationship exists: someone whose obligation runs to the patient rather than to a sale, a sponsor, an employer, or a view count.

A risk-benefit judgment made under that duty. Not a generic judgment about a molecule, but a particular one that incorporates this person's other conditions, other medications, family history, values, and tolerance for uncertainty.

Enhancement removes all four. That is the entire structural point of this chapter, and everything that follows is a consequence of it.

THE SAME MOLECULE, TWO SITUATIONS

  CLINICAL USE                          ENHANCEMENT USE
  ───────────────────────────           ───────────────────────────
  Diagnosis          ✓ present          Diagnosis          ✗ none
  Indication         ✓ approved         Indication         ✗ none, or off-label
  Clinician with     ✓ duty runs to     Clinician          ✗ absent, or duty runs
    a duty             the patient                             elsewhere
  Risk-benefit       ✓ individualized   Risk-benefit       ✗ self-made, from the
    judgment           by someone          judgment             Chapter 6 pipeline
                       accountable

  THE MOLECULE                          THE MOLECULE
  ───────────────────────────           ───────────────────────────
  Identical                             Identical

  Everything that differs is structural. Nothing that differs is chemical.

Read that diagram carefully, because it is easy to misread as a moral scorecard. It is not one. It is an inventory. The right-hand column is a list of absences, and absences are not automatically wrongs — plenty of things people do without medical supervision are fine. The point is that when people compare enhancement to clinical use, they usually compare only the bottom half of the diagram — the molecule — and conclude the situations are similar because the molecule is the same.

They are not similar. The molecule is the least variable element in the entire situation.

This also explains something that otherwise looks like inconsistency in this book. Growth hormone appears in Chapter 14 as a treatment with real, approved, sometimes life-changing uses. It appears here as the archetypal enhancement compound with weak support. Both are correct, and they are not in tension, because a rating attaches to a claim with a population and an endpoint, never to a molecule (Chapter 5, rule 1). "Growth hormone for children with growth hormone deficiency" and "growth hormone for a healthy adult who wants to be stronger" are different claims about different populations with different endpoints and different evidence bases. Compressing them into a single verdict about a molecule destroys precisely the information that would let you decide.

One more consequence, and it is the one people find least intuitive. Because the enhancement user makes their own risk-benefit judgment, the quality of that judgment depends entirely on the quality of their information. Chapter 6 traced how peptide information degrades as it travels from a paper to a press release to a video to a comment, each step dropping a qualifier, a population, or a route of administration; Chapter 42 traced who gets paid at each step. The enhancement decision sits at the end of that pipeline — made with the most degraded version of the evidence, by the person with the most at stake, with no one in the room whose job is to say that study was in rats.

🧬 The Molecule — why these particular compounds, and not others

There is a pattern in which peptides show up in enhancement contexts, and it is not driven by efficacy. It is driven by chemistry and regulatory history — both of which you can read off the compound without knowing anything about whether it works.

They are synthesizable. Chapter 1 made the point that peptides of moderate length can be built step-by-step on a machine. That same accessibility means a compound with no human safety data can be manufactured cheaply and at scale. Low barriers cut both ways.

They are injectable, and injection is a route people already accept. A population that has normalized subcutaneous injection has removed the main practical friction.

They mostly have laboratory codes rather than generic names. Chapter 1 flagged why that matters: naming authorities assign a stem name when a compound enters serious clinical development. A molecule known only by a code, decades after its discovery, is telling you that the development which would have generated human safety data did not happen, or did not finish.

Put those three together and you get the defining feature of this chapter's subject matter: the compounds most available for enhancement are, systematically, the compounds with the least human data. That is not a coincidence and it is not bad luck. It is what the selection process selects for. Availability and evidence are, in this market, close to inversely related.

🔍 Check Your Understanding

1. Name the four things clinical medicine supplies, and say which one you think is hardest to replace outside the clinic. Defend your choice.
2. Why does this book give growth hormone one rating in Chapter 14 and a different one here without contradicting itself?
3. A friend says: "It's the same molecule the doctor prescribes, so it's the same thing." Using the diagram in this section, give the shortest correct reply you can.

43.2 The workplace

Now to the domain that affects the most people and is discussed the least.

Competitive employment produces enhancement pressure for a simple reason: in a great deal of professional life, advancement is relative rather than absolute. There is one promotion. There are a fixed number of partner slots, residency positions, tenure lines, roster spots, contracts. Performing well is not the criterion; performing better than the other candidates is.

Economists call something with this property a positional good — a good whose value depends on rank rather than absolute level. In a positional contest, any improvement adopted by some participants changes what counts as adequate for everyone. Hold that thought; §43.5 is built entirely from it.

First, an honesty obligation. The well-documented version of workplace pharmacological enhancement concerns stimulants, not peptides. Prescription stimulants used by people without a diagnosis have been studied in academic and professional settings for two decades, and there is a real literature — though even there, prevalence estimates vary widely by population, method, and definition, and quoting a single number would misrepresent that spread. This book will not quote one.

Peptide use in workplace contexts is not well characterized. I want to say that as plainly as possible, because the temptation to fill the gap is enormous and everyone gives in to it. There is no good prevalence estimate. There is no representative survey. There are clinics advertising "optimization" to professionals, and there are advertisements addressed to executives, and there are people who will tell you it is everywhere. None of that is measurement. A confident number here would be fabricated, and I would rather leave you with a hole than fill it with something invented.

What can be said with confidence is structural, and structure is often more useful than prevalence anyway.

WHY A PRIVATE CHOICE BECOMES A PUBLIC COST

  ABSOLUTE CONTEST                      POSITIONAL CONTEST
  (your result stands alone)            (your result is compared)

  You improve → you benefit.            You improve → you benefit AND
  Nobody else's situation changes.        everyone else's standing falls.

  Example shape:                        Example shape:
    learning a language                   competing for one promotion
    recovering from injury                a fixed number of roster spots
    lowering your own blood pressure      an admissions cycle

  ENHANCEMENT HERE IS                   ENHANCEMENT HERE IS
  a private matter.                     a cost imposed on non-adopters,
                                        whether or not anyone intended it.

The right-hand column is the whole workplace problem in miniature. Where advancement is relative, an enhancement adopted by some becomes a cost imposed on all — not because anyone acted maliciously, but because in a ranked contest, other people improving is functionally identical to you declining. No one has to coerce anyone. The arithmetic does it.

Three features make some workplaces more susceptible than others, and it is worth being able to spot them:

Observable, comparable output. Where performance is measured on a common scale — billable hours, sales figures, cases closed, publications, response times — comparison is cheap and continuous. Where performance is diffuse and judged qualitatively, the pressure is weaker.

High stakes concentrated in short windows. Examination periods, audit seasons, trial dates, funding cycles, and product launches concentrate the temptation, because a temporary advantage during a decisive window can have permanent consequences.

Weak or absent testing, combined with strong norms of silence. Sport, as §43.3 discusses, resolved this with an explicit rule and an enforcement apparatus. Almost no civilian workplace has anything comparable, and most have strong reasons not to want one.

That last one produces a specific and underrated harm: misperception. When people cannot see what others are actually doing, they estimate from rumor — and rumor systematically overestimates. If you believe your competitors are all enhanced, you experience pressure whether or not the belief is true. A compound with no real performance effect can still generate real competitive pressure, real spending, and real risk exposure, purely through what people believe about what other people are taking.

⚠️ Hype Check — "everyone at the top is on something"

The claim, in its usual form:

"You don't get to that level naturally. Everyone at the top is on something — they just don't talk about it. If you're not, you're competing with one hand tied behind your back."

What's true in it. In some well-documented competitive domains — chiefly elite sport, where testing exists and produces published sanction records — undisclosed pharmacological use is demonstrably not rare. And the silence is real: there are strong incentives not to disclose, which means the visible rate is a floor, never a ceiling.

Where it fails. In four places, and they compound.

First, it is unfalsifiable as stated. "Everyone is on something, they just don't talk about it" cannot be checked, and a claim that cannot be checked is not evidence — it is a mood. Chapter 5's fifth rule requires that a rating be falsifiable; apply the same standard to assertions you encounter, not only to the ones this book makes.

Second, it generalizes from the most-tested, most-scrutinized environment on earth to environments with none of those features.

Third — the one people find genuinely surprising — it assumes the enhancement works. §43.3 rates that claim, and for the growth-hormone and muscle-peptide compounds the rating is ❌. "Everyone at the top is on something" is frequently deployed to explain a performance gap that the compounds in question have not been shown to produce.

Fourth, it functions as a recruitment argument regardless of its truth. Notice who benefits from your believing it. Chapter 42 followed that money; it does not stop at the edge of this chapter.

Verdict: an unfalsifiable claim, generalized from an unrepresentative domain, that presupposes an efficacy finding which has not been demonstrated — and which, believed, produces the pressure it describes.

43.3 Sport, briefly

Chapter 38 owns anti-doping in full: the World Anti-Doping Agency, the Prohibited List and the criteria by which substances are added to it, therapeutic use exemptions, the results-management process, and the specific position of peptide hormones and secretagogues within all of it. Go there for the substance. This section owns exactly one point about sport, and then it gets out of the way.

Sport is the only domain in this chapter that has built an explicit, enforced, adjudicated answer to the enhancement question.

That is genuinely unusual, and it makes sport enormously informative as a natural experiment — not because it settles anything, but because it is the only place where we can observe what an enforced answer actually costs and actually achieves. Everywhere else, the question is open and the consequences of closing it are hypothetical. In sport, the question was closed decades ago, and the consequences are visible.

What the enforced answer achieves. A published, appealable standard that competitors know in advance. A stated principle for why something is prohibited rather than a bare list. A mechanism — the therapeutic use exemption — through which a genuinely sick athlete can receive needed treatment without being treated as a cheat. Detection capability sufficient that violation carries real risk, which changes behavior even when it does not eliminate it. And a body of adjudicated decisions giving the rule meaning in specific cases, which is more than any workplace policy on this subject has.

What the enforced answer costs. A large, permanent, expensive apparatus. Intrusive requirements on athletes' private lives that no other profession accepts. A category of unintentional violations arising from contaminated supplements — which Chapter 34's material on the unregulated supply chain makes entirely predictable, and which lands on athletes who did nothing anyone would recognize as cheating. A perpetual lag between what can be taken and what can be detected. Adjudication that is slow and occasionally ruins a career on the basis of a finding later revised. And an unfairness in the exemption mechanism itself: permitting a therapy only with documentation advantages athletes with better medical access and administrative support, who are not the same population as the athletes with the greatest medical need.

That last cost is the one worth carrying into the rest of the chapter. Even the domain that answered the question, and answered it with real institutional seriousness, produced a system whose fairness properties are contested. Anyone who proposes an enforced answer for a non-sport domain should be asked, first, whether their proposal has thought as hard as anti-doping has — and second, whether they expect a better outcome than anti-doping achieved, and why.

I am not going to say what the balance of that ledger is. Reasonable people who know the system well disagree about it. What I will say is that it is the only ledger we have, and arguments about enhancement that ignore it are arguing from imagination.

💊 In the Clinic — the therapeutic use exemption is where the two worlds touch

The therapeutic use exemption is worth understanding structurally even if you will never encounter one, because it is the seam where clinical medicine and enhancement rules meet, and seams are where you can see how a structure is built.

An athlete with a genuine medical condition may need a treatment that appears on the Prohibited List. The exemption process exists so that treating the illness does not require ending the career. Chapter 38 covers the mechanics; what matters here is what the mechanism concedes.

It concedes that the same molecule can be therapy or enhancement depending entirely on the situation of the person receiving it — §43.1's point, formalized by an institution that had to make it adjudicable rather than merely philosophical. The rule is not "this substance is bad." It is "this substance is prohibited unless there is a patient," and the exemption is the procedure for establishing that there is one.

It also concedes something less comfortable: establishing the presence of a patient requires documentation, and documentation is unequally available. Two athletes with the identical condition and unequal access to specialist care do not have equal access to the exemption. That is a real equity problem inside a system explicitly built for fairness, and it recurs — in a much less examined form — everywhere else in this chapter.

As always: decisions about any actual treatment belong with a clinician who knows the person's history, and here, with someone who also knows the applicable rules.

📊 Evidence Rating

Claim: Peptide enhancement — specifically the growth-hormone-axis and muscle-directed compounds of Part III — produces meaningful performance benefit in healthy trained adults. Rating: ❌ Hype outpaces evidence (as of 2026) Reason: For these compounds there is essentially no outcome evidence in healthy trained adults, and where trials were conducted in disease populations they showed gains in lean mass without corresponding gains in strength or function (Chapter 16) — a dissociation that is precisely the opposite of what the performance claim requires. What would change it: adequately powered, randomized, controlled trials in healthy trained adults, with functional endpoints — strength, power, sport-specific performance, time to exhaustion — rather than body composition, showing benefit that survives blinding.

That rating deserves a sentence of framing, because it is the one readers most often want to argue with, and the arguments are usually about the wrong thing.

The rating describes the evidence, not the molecule (rule 2). It does not say these compounds do nothing. It says the confident version of the performance claim is not supported by human outcome data. And note what it does not rest on: it does not rest on disapproval of enhancement (rule 4 forbids downgrading on distaste), and it would not be upgraded by a persuasive mechanism, however elegant (rule 3). If someone explains to you exactly how a compound ought to increase muscle protein synthesis, they have given you a reason to run the trial, not a result from one.

43.4 Military and occupational contexts

Some institutions have authority over their members. Militaries, obviously — but also emergency services, certain transport and aviation roles, offshore and remote industrial work, and any setting where an organization can lawfully direct what a person does with their body during working hours and sometimes outside them.

In these contexts, the consent question changes character, and the change is worth stating carefully because it is easy to state luridly and hard to state accurately.

I am going to handle this abstractly, and I want to explain why rather than let it look like evasion. Specific claims about what particular armed forces or agencies have administered to their members are either classified, contested, historical, or invented — frequently more than one at once. I cannot source contemporary specifics to the standard this book applies to everything else. So I will describe the structure and assert nothing about current practice. If that is unsatisfying, the alternative was to make something up, and Chapter 5's rules apply to me too.

The structure is this. An institution that can order can also pressure, and the two are not cleanly separable from the inside.

Consider what "voluntary" means in a hierarchy. In ordinary life, a voluntary choice is one where declining carries no penalty imposed by the person doing the asking. Inside an institution with authority over assignments, promotion, evaluation, deployment, and social standing, that condition is extremely difficult to satisfy, and — this is the important part — it is difficult to satisfy even when everyone involved is acting in good faith.

Suppose a program is offered, genuinely optionally, with a genuine right of refusal and genuine assurance that refusal will not be recorded. The person deciding still knows:

• that the offer came from someone with authority over their future
• that their peers can see who participated
• that "declined" and "was not selected" can become indistinguishable in a personnel record over time
• that being seen as unwilling to do what others did is a social fact regardless of whether it is an administrative one

None of that requires anyone to threaten anything. It follows from the shape of the institution. This is why occupational health ethics treats employer-offered medical interventions as a distinct category requiring separate safeguards — the same intervention offered by an employer and by an unaffiliated physician is not the same offer, because the consequences of declining differ.

Two further features are specific to these settings and worth naming.

The institutional interest is real and sometimes admirable. An organization responsible for people's survival under extreme conditions has a legitimate interest in their alertness, endurance, and recovery. The coercion problem does not require bad motives, which is exactly what makes it hard to solve by improving motives.

The risk-bearing is asymmetric and long-tailed. Occupational exposures in these settings can have consequences that appear decades later, after the person has left, at which point the institution that captured the benefit is no longer the party managing the harm. §43.8 is about this asymmetry generally; the history of occupational exposure litigation in other domains — dusts, solvents, radiation — is a preview of how long recognition can take.

One practical addition. Whatever is true about institutional programs, there is a separate and better-attested phenomenon in physically demanding occupations: individual, unsupervised use driven by occupational performance demands and often by injury. That is not an institutional program. It is the workplace pressure of §43.2 in a population with high physical standards, high injury rates, and frequently poor continuity of care — and it belongs with §43.2 and §43.7 rather than here.

🔍 Check Your Understanding

1. Explain, without using the word "order," why an offer made inside a hierarchy is a different offer from the same words spoken by an unaffiliated physician.
2. This section refuses to assert specifics about current military practice. Give the reason, and then say what kind of source would be required for such an assertion to meet this book's standard.
3. Why does the fact that an institution's motives may be genuinely good make the coercion problem harder rather than easier?

43.5 The coercion question

This is the chapter's core, and it is the argument I would most like you to be able to reconstruct from memory a year from now, because it recurs in debates about technologies that have nothing to do with peptides.

An enhancement adopted by enough people stops being an advantage and becomes a baseline.

That sentence contains the whole thing. Everything else is unpacking.

Here is the formal version, in five steps.

THE COERCION STRUCTURE — how an option becomes an expectation

  STEP 1   A contest is positional. Rank matters, not absolute level.
           (One promotion. One roster spot. A fixed number of slots.)

  STEP 2   An enhancement appears that shifts rank. A few people adopt it.
           At this stage it is a genuine ADVANTAGE, and the adopters
           are choosing to gain something.

  STEP 3   Because the contest is positional, non-adopters' standing FALLS
           without any change in their behavior. They have done nothing
           differently and are now doing worse.

  STEP 4   Some non-adopters adopt, to restore their previous position.
           They are not seeking an advantage. They are avoiding a loss.
           ── THIS IS THE HINGE OF THE ARGUMENT ──

  STEP 5   At sufficient adoption, the enhanced state becomes the norm.
           The advantage has been competed away — everyone is enhanced,
           relative rank is roughly what it was — but everyone now bears
           the cost and the risk.
           The remaining choice is not "gain an edge?"
           It is "accept a disadvantage?"

  NOTE WHAT IS ABSENT: nobody compelled anybody. No employer required it,
  no institution ordered it, no rule changed. The structure emerged from
  ADOPTION ALONE.

Step 4 is the hinge, and it is where the moral character of the decision changes without the decision itself looking any different from the outside. The person adopting at step 4 is doing exactly what the person at step 2 did — same compound, same route, same risk — but their reason is different in a way that matters. Step 2 is seeking a gain. Step 4 is avoiding a loss. And a choice made to avoid a loss that someone else's choice imposed on you is a choice in a much thinner sense than we usually mean by the word.

Step 5 is where the result becomes almost comic in its futility, if it were not costly. At full adoption, the competitive benefit is gone. Everyone improved, so nobody improved relatively. The ranking is roughly what it would have been. What has changed is that the entire population now carries whatever risk the enhancement entails, plus its cost, plus its inconvenience, in exchange for a competitive position they already had. Economists call this dissipation: a collective expenditure that produces no collective gain. It is the same structure as an arms race, as escalating campaign spending, as every situation where individually rational actions aggregate into a collectively irrational outcome.

You may recognize the shape. It is a collective action problem — the family of situations that includes the prisoner's dilemma, the tragedy of the commons, and the arms race. The defining feature of all of them is that the outcome is bad for everyone and no individual acted irrationally. That combination is what makes them so resistant to solution by exhortation. Telling people to behave better does not work, because each person is already behaving sensibly given what everyone else is doing.

Where the argument bites hard, and where it barely bites

The single most common error in enhancement debates is applying this argument uniformly. It is not uniform. It has conditions, and the conditions are checkable.

The argument applies WEAKLY when:

The benefit is largely private and absolute. A person who takes a medication so they can climb the stairs without stopping is not in a contest with anyone. No coercive structure forms, because step 3 never happens.

Performance is not observable or comparable. If nobody can tell who improved, adoption cannot propagate through competitive pressure. It might propagate through belief, as §43.2 noted, but that is a weaker and more fragile mechanism.

The contest is not positional — where a standard is absolute and unlimited numbers of people can meet it, improvement by others does not harm non-adopters.

Adoption is low and likely to stay low, because of cost, access, or route of administration.

The argument applies STRONGLY when:

Competition is explicitly relative and the stakes are large. Ranked contests with concentrated rewards are the natural home of this structure.

Performance is observable, measurable, and continuously compared.

The enhancement is cheap, available, and easy to conceal — concealability prevents the social feedback that might otherwise slow adoption.

The population is young, and the harms are delayed. Discounting future risk against present competitive necessity is something humans do reliably, and enhancement decisions are frequently made at the point in life when that discount is steepest.

Now apply that grid honestly to the peptide compounds this book has covered, and something awkward emerges. The coercion argument is strongest in exactly the domains where the efficacy evidence is weakest. The compounds most discussed in performance contexts are the ones §43.3 rated ❌. If a compound does not actually shift rank, step 2 never produces a real advantage — and the whole cascade runs on perceived advantage rather than real advantage.

Does that dissolve the problem? No — and this is worth being precise about. A coercive structure built on a false belief is still a coercive structure. People still adopt, still spend, still take on risk, still feel the pressure. Everything in steps 3 through 5 happens. The only thing that changes is that at step 5 the population has taken on the full cost and risk in exchange for nothing at all, rather than in exchange for a competitive position they already had. That is worse, not better. It is the same dissipation with the benefit column emptied.

📊 Evidence Rating

Claim: An enhancement adopted widely enough within a positional contest becomes effectively coercive for non-adopters. Rating: ⚠️ Promising but preliminary (as of 2026) — here meaning: structurally well described, sparsely measured. Reason: The mechanism is a well-characterized collective action structure with abundant analogues in economics and game theory, and it has been documented empirically in several non-pharmacological domains; but direct measurement of pressure-driven adoption for peptide enhancement specifically is essentially absent. What would change it: longitudinal studies in a defined competitive population measuring adoption alongside stated reasons — distinguishing seeking advantage from avoiding disadvantage — and tracking whether the stated-reason mix shifts from the former to the latter as prevalence rises. That is the empirically observable signature of step 4, and it is measurable.

I want to flag something about that rating, because it is doing something slightly unusual. A ⚠️ here is not a compliment to the argument, and it is not a hedge against it. It is a statement that a structural argument with excellent theoretical credentials has thin direct empirical support in this particular domain, and that both halves of that sentence are true at once. If you find yourself wanting to round it up because the argument feels right, that is rule 3 (never upgrade with mechanism) applying to a social mechanism rather than a molecular one. The temptation is identical and so is the error.

43.6 The level-playing-field argument, examined

Now the strongest case for permitting and regulating enhancement rather than prohibiting it. I am going to present it at full strength first, without interruption, because arguments presented weakly cannot be evaluated and presenting them weakly is a way of avoiding them.

The argument.

Prohibition does not eliminate use; it relocates it. Where a practice is banned but demanded, the supply moves to an unregulated market, and the unregulated market has properties that are worse on every dimension that matters for safety. Chapter 34 documented these properties directly for peptides: uncertain identity, uncertain concentration, contamination, sterility failures, and no recourse when something is wrong. Chapter 19 documented the consequence — that the actual harms in this space have come predominantly from what was in the vial and how it was used, rather than from exotic pharmacology.

A sanctioned regime, the argument goes, would fix most of that. Supervised use means a clinician sees the person. Pharmaceutical-grade supply means the vial contains what the label says, at the stated concentration, sterile. Documented use means adverse events are reported, aggregated, and eventually studied, which converts a decade of anonymous anecdote into an evidence base. Legal availability means people are not making decisions from the Chapter 6 pipeline in secret, and are not deterred from seeking medical help when something goes wrong for fear of admitting what they did.

And on fairness: an above-board regime is more egalitarian than prohibition, not less. Under prohibition, the people who obtain and use enhancement successfully are disproportionately those with money, connections, private medical access, and the sophistication to navigate a gray market. The prohibition does not stop them; it stops everyone else. A sanctioned regime with equal access at least removes the sophistication premium.

This argument has a serious pedigree. It is the harm-reduction argument, and harm reduction has an extensive record in other domains — needle exchange, opioid substitution, drug checking — where the finding has repeatedly been that regulated availability produced better health outcomes than prohibition. Dismissing it as rationalization is not an answer.

The objections, also at full strength.

It does not address coercion at all. This is the deepest objection and it is usually skipped. Everything in §43.5 runs on adoption, not on legality. A sanctioned regime with excellent safety properties would likely increase adoption — that is much of its point — and higher adoption drives the structure faster. The argument answers a safety question with a safety solution while leaving the freedom question untouched, then presents the whole thing as though the freedom question had been addressed.

It transfers risk to people who did not want to participate. At step 5, non-adopters choose between accepting a disadvantage and accepting a risk they had rejected. Their preference has been overridden by other people's choices, which is a real harm even though nobody did anything to them directly.

Availability is never actually equal. This sounds most like a technicality and is in practice most decisive. §43.3's exemption problem is the demonstration: even inside a system built explicitly for fairness, access to the mechanism tracked access to medical infrastructure. Every real regime has cost, geography, insurance, time, and information barriers, and every one correlates with existing advantage. "Equally available" is a property of the proposal, not of any implemented system.

It assumes the safety question is settled, and Part III showed that for most of these compounds it is not. Supervision manages known risks. It does much less for unknown ones, because you cannot monitor for an effect nobody has characterized. Supervision is not a substitute for evidence, and the argument quietly treats it as one.

And there is a counterfactual problem. The harm-reduction evidence base comes from domains with an existing population of users and an existing pattern of harm. Sanctioned enhancement proposes to make a practice available to people who are not currently doing it. Whether those findings transfer to a regime that changes the size of the exposed population is open, and it is not answered by citing needle exchange.

📊 Evidence Rating

Claim: Sanctioned, supervised enhancement produces better safety outcomes than an unregulated black market for the same compounds. Rating: ⚠️ Promising but preliminary (as of 2026) Reason: General harm-reduction reasoning supports it, and the specific failure modes it would address — contamination, misidentification, wrong concentration, non-sterility, and non-disclosure to treating clinicians — are documented realities of the current market (Chapters 19 and 34); but the model has never been tested for these compounds, supervision cannot manage risks that have not been characterized, and the claim addresses safety while being routinely deployed to answer questions about coercion and fairness that it does not touch. What would change it: outcome data from an actual implemented sanctioned regime — adverse event rates, serious harms, and total exposed population — compared against a defined pre-implementation baseline in the same population. Absent that comparison, the claim rests on analogy to domains with a materially different structure.

I am not going to resolve this. Not because the question is unanswerable, but because resolving it requires weighing autonomy against collective cost, present freedom against delayed harm, and individual benefit against externalized risk — and those weights are values, not findings. You should have them. This book should not assign them to you.

What this book will say is this: the argument is much stronger as a safety argument than as a fairness argument, and it is almost always deployed as a fairness argument. Notice when that substitution happens. It is the single most common move in the debate, and once you can see it, you can see it every time.

43.7 Safety without a physician

Chapter 19 cataloged what medical supervision actually provides. It is worth restating as a list, because "see a doctor" is such a familiar phrase that people stop hearing content in it.

Supervision provides:

Baselines. Values measured before anything is administered. A single abnormal result means very little without a prior value to compare it to; without a baseline, you cannot distinguish a change from a starting point.

Targeted monitoring. Not general wellness screening — monitoring chosen because this compound affects these systems. Knowing what to watch is a function of knowing the pharmacology.

Dose adjustment based on response. Titration to effect, and downward adjustment when a marker moves.

Recognizing tolerance rather than escalating into it. This is the one that most distinguishes supervised from unsupervised use. When an effect diminishes, the untrained response is to increase. The trained response is to ask whether the receptor system has downregulated — which, as Chapter 3 established for peptide signaling generally, is frequently exactly what has happened, and in that situation escalation makes the underlying problem worse while temporarily masking it.

Interaction checking. Against every other medication, supplement, and condition the person has — which requires them to disclose the full list, which requires them to feel able to.

Differential diagnosis. When a symptom appears, someone determines whether it is the compound, an unrelated illness, or a coincidence. People are systematically bad at this about themselves in both directions — attributing everything good to the compound and everything bad to stress.

Someone to call. At two in the morning, when something is wrong and you do not know how wrong.

In enhancement contexts, none of that is present. Not by accident — by construction. There is no diagnosis to monitor, no indication to titrate against, no clinician with a duty, and frequently a positive incentive not to disclose to whatever clinician the person does see. That last point deserves emphasis: non-disclosure is itself a safety failure, because it degrades the information available to every other clinician who treats that person, including in an emergency, and it is produced by exactly the stigma and legal ambiguity the sanctioned-regime argument of §43.6 is designed to address.

Now combine that with §43.1's observation about which compounds are available.

🔬 Read the Study — what happened when growth hormone was actually tested

Chapter 16 covers this in full; this is the compressed version, because it is the empirical anchor under §43.3's rating and readers reach this chapter without always having it to hand.

The question. Does growth hormone improve athletic performance in healthy adults? Not: does it change body composition — does it make people better at things.

The evidence. A systematic review published in a major general internal medicine journal in 2008 pooled the controlled studies of growth hormone administration in healthy participants. Subsequent controlled work has not overturned its central finding.

What it found. Growth hormone administration increased lean body mass, measurably and consistently. It did not produce a demonstrated increase in strength. It did not produce a demonstrated increase in exercise capacity. And adverse effects — notably soft tissue swelling and fatigue — were more common in treated participants than controls.

Why this matters more than it appears to. The gain in lean body mass is real, and a substantial portion of it is understood to be fluid retention rather than contractile tissue. The scale reads higher. The mirror may look different. The function does not change. This is the dissociation Chapter 16 calls mass without function, and it is the most instructive result in the performance-peptide literature, because it demonstrates that a visible, measurable, subjectively convincing effect can coexist with no functional benefit at all.

And it explains why the belief persists. If a compound reliably produces a change the user can see and feel, the absence of functional benefit is nearly invisible to them. Personal experience confirms the effect; personal experience cannot distinguish the effect from the benefit. That is what controlled trials with functional endpoints are for, and it is why "I could tell it was working" is not evidence about the thing people think it is evidence about.

🩺 Safety and Risk — the combination, stated plainly

Put the two halves of this section together and the result is uncomfortable enough to deserve its own statement.

The compounds used in enhancement contexts are, systematically, the ones with the least human safety data. Not by coincidence: §43.1 explained the selection process. Compounds that completed clinical development have indications, prescribers, generic names, and monitoring guidance. Those that did not are cheap, available, and unstudied — and the second group dominates this market.

And enhancement contexts have, by construction, the least medical oversight. No diagnosis, no indication, no duty, no baselines, no targeted monitoring, no differential diagnosis, and often no disclosure to the clinicians who are involved in the person's care.

Highest uncertainty. Lowest oversight. Those two properties share a cause. The regulatory absence that makes a compound available without a prescription is the same absence that means nobody generated the safety data or wrote the monitoring guidance.

This is not an argument that anyone will be harmed. It is a statement about knowledge: in this combination, harm would be difficult to detect, difficult to attribute, and difficult to aggregate into anything anyone could learn from. A rare or delayed effect in an undocumented population using unverified material with no denominator is close to invisible by design.

Whatever you conclude about the ethics, that situation is worth understanding on its own terms — including by people who conclude enhancement should be permitted, since it is exactly what a well-designed sanctioned regime would be trying to fix.

And, as everywhere in this book: decisions about any of this belong with a clinician who knows your history. Chapter 39 is about making that conversation possible — especially when the subject is something you obtained outside the medical system.

43.8 Who bears the risk and who captures the benefit

This section makes an observation from economics rather than from ethics, and I want to be explicit about that, because it means the observation holds regardless of what you concluded in §43.5 and §43.6. It does not require a moral premise. You can accept it while believing enhancement is entirely permissible.

WHERE RISK AND BENEFIT LAND

  THE INDIVIDUAL BEARS:              THE ORGANIZATION CAPTURES:
  ─────────────────────────          ──────────────────────────────
  physiological risk                 output, ranking, revenue
  unknown long-term effects          performance in the decisive window
  financial cost                     lower replacement cost
  legal and reputational exposure    a workforce that competes harder
  the consequences, decades later,   the results, this quarter
    after leaving

  DURATION OF EXPOSURE:              DURATION OF INTEREST:
  a lifetime                         the term of employment

  Risk is borne where it cannot be transferred.
  Benefit is captured where it can be.

In a competitive workplace, a team, or an institution, the individual bears essentially all of the physiological risk. The organization captures a substantial share of the performance benefit — in output, in ranking, in revenue, in whatever the organization is optimizing. Neither party did anything unusual. This is simply where the two things land.

Where risk and benefit separate, the party bearing the risk is systematically under-compensated. That is a standard result, and the reasoning is straightforward: a decision-maker weighing costs and benefits will weigh the ones that fall on them. Costs that fall elsewhere are, from the decision-maker's position, discounted — not maliciously, structurally. Economists call the general phenomenon an externality, and the standard finding is that activities with externalized costs are over-produced relative to what anyone would choose if all costs were counted.

Four features make this asymmetry sharper for pharmacological enhancement than for most workplace risks.

The risk is invisible and cumulative. Unlike an obviously dangerous task, there is no moment at which anyone observes the hazard being taken on. Nothing appears in a safety log.

Attribution is nearly impossible after the fact. A health problem appearing fifteen years later in someone who used an unstudied compound with no documentation, from an unverifiable source, in an undefined population, cannot be traced. This is §43.7's epistemic problem arriving as a compensation problem: you cannot be compensated for a harm nobody can attribute.

The relationship ends before the risk does. Employment is finite; exposure consequences are not. When the harm appears, the beneficiary is usually gone.

Nobody agreed to anything. In occupational risk generally we built compensation structures — hazard pay, workers' compensation, liability, insurance — precisely because we recognized this asymmetry and decided the party capturing the benefit should carry more of the cost. None of those structures exist here. The benefit is captured through ordinary employment, and the risk is carried privately by someone who, in most cases, cannot even discuss it.

Notice how this interacts with §43.5. The coercion structure explains why someone might adopt without wanting to. This section explains why, having adopted, they capture only a fraction of the benefit their risk generates. Together they describe a situation in which a person takes on an uncompensated, unattributable, lifelong risk in order to maintain a competitive position they already held. That is the worst-case shape, it does not require anyone to have behaved badly, and it is reachable through entirely ordinary individual decisions.

📋 Your Evidence Dossier

Before the full exercise below: take one compound from your dossier and answer a single question in one sentence. If this compound harmed you in ten years, who would know, and how?

Not would you be compensated — that comes later. Just: what is the mechanism by which the harm would become known to anyone at all? Trace it. Who records it, who aggregates it, who would ever connect your outcome to your exposure?

For most enhancement contexts the honest answer is that there is no such mechanism, and writing that sentence down is more instructive than any argument in this chapter.

43.9 What would make enhancement defensible

Rather than deliver a verdict, this section builds the criteria. That is a deliberate choice: a verdict tells you what I think, and criteria let you evaluate cases I have never seen — including compounds that do not exist yet, which is most of the ones you will eventually have to assess.

But first, the claim that most often pre-empts this whole discussion.

📊 Evidence Rating

Claim: Enhancement is inevitable. Rating: NOT RATED — this claim falls outside the system's competence. If forced, 🔬. Reason: It is a prediction about aggregate future human behavior, not a claim about a population and an endpoint, and there is no body of evidence that bears on it in the way trial evidence bears on a pharmacological claim. What would change it: nothing available. There is no study design that could establish or refute it, which — by Chapter 5's fifth rule — is precisely what disqualifies it from a rating.

That non-rating is more instructive than a rating would be, because learning where a tool stops working is part of learning the tool.

The rating system was built for claims with a defined population, a defined endpoint, and a body of evidence that could in principle come out either way. "Semaglutide reduces body weight in adults with obesity" has all three. "Enhancement is inevitable" has none. No population — who, exactly? No endpoint — inevitable by when, at what adoption rate, in which domains? And no falsifier: if adoption rises the claim is confirmed, and if it falls the claim becomes "not yet." A claim confirmed by every outcome is not making a prediction; it is expressing an attitude.

That is not a criticism of the people who say it. It is usually said sincerely, and it gestures at something real: technologies do diffuse, and prohibitions often fail. But notice the work inevitable does. It converts a contested choice into a fact about the world, and thereby makes deliberation look naive. A claim that functions to end an argument rather than inform one deserves the same scrutiny as a claim about a molecule — arguably more, because nobody thinks to demand evidence for it.

To make it rateable, sharpen it: "In competitive domain X, adoption of enhancement Y will exceed Z percent by year W." Now it has a population, an endpoint, and a falsifier, and it can be wrong. The sharpening is itself the skill this book has been teaching for forty-two chapters.

The conditions

Five conditions. They are cumulative rather than alternative, and none is sufficient alone.

CONDITIONS FOR DEFENSIBLE ENHANCEMENT

  1. ESTABLISHED SAFETY AT THE RELEVANT EXPOSURE
     Not "safe in patients at therapeutic doses for a defined period."
     Safe in HEALTHY people, at the exposures actually used, for the
     durations actually used. Different population, different question.

  2. GENUINE INFORMED CONSENT
     Real information — the actual evidence, including its absence —
     AND a real alternative. Consent without an alternative is not
     consent; it is acquiescence.

  3. ABSENCE OF COERCIVE STRUCTURE
     Non-adopters must not be made worse off by others' adoption.
     This is the §43.5 condition and it is the hardest to satisfy,
     because it is a property of the CONTEST, not of the compound.

  4. RISK BORNE BY THE BENEFICIARY
     Whoever captures the benefit carries a proportionate share of
     the cost — through compensation, liability, monitoring, or
     insurance. The §43.8 condition.

  5. REVERSIBILITY
     Stopping returns the person approximately to baseline, without
     permanent change and without dependence. Reversible decisions
     tolerate bad information. Irreversible ones do not.

  These are CUMULATIVE. Meeting one is not partial credit toward the rest.

A word on why each earns its place.

Safety at the relevant exposure is first because everything else is conditional on it, and the qualifier is not pedantry — it is the single most common failure of transfer in this entire book. Safety established in a patient population at a therapeutic exposure over a trial duration tells you about that population, that exposure, that duration. Healthy people using a compound for years are a different population under a different exposure for a different duration. Three of the four qualifiers have changed, and the safety claim does not travel.

Genuine informed consent has two components and the second is the one that gets dropped. Real information means the actual state of evidence, including honest statements of what is unknown — precisely what the Chapter 6 pipeline strips out. A real alternative means declining is genuinely available at acceptable cost. §43.4 showed how a hierarchy erodes the second, and §43.5 showed that a sufficiently adopted enhancement erodes it without any hierarchy at all.

Absence of coercive structure is the hardest, and the reason is worth stating precisely: it is not a property of the compound, the user, or the seller. It is a property of the contest. No amount of care by any individual participant can satisfy it.

Risk borne by the beneficiary is the condition we already know how to implement — routinely, for other occupational hazards, through liability, insurance, and compensation — and which is almost entirely absent here.

Reversibility matters most under uncertainty. When evidence is thin, the cost of being wrong depends almost entirely on whether you can undo the decision. A reversible choice made on bad information is a mistake; an irreversible one is a permanent mistake. Effects on developing systems, on endogenous production, and on anything structural are where to ask this hardest.

How many current cases meet these?

Very few. Specifically:

Condition 1 fails for most compounds in Part III, where healthy-population safety data at enhancement exposures over enhancement durations is not thin but absent.

Condition 2 fails wherever information comes through the Chapter 6 pipeline — nearly everywhere in the consumer market — and fails again wherever declining carries a real competitive cost.

Condition 3 fails in every positional contest, which is most of §43.2, §43.3, and §43.4.

Condition 4 fails almost universally, because the compensation structures that exist for other occupational risks have no analogue here.

Condition 5 is the one most often satisfied — many peptides are cleared rapidly and their effects do not obviously persist — but it is satisfied unverifiedly, because demonstrating reversibility requires the long-term follow-up that condition 1 already told us does not exist. We believe these effects are reversible largely because we have no evidence that they are not, which is weaker than it sounds.

So: most current enhancement cases meet at most one or two of five conditions.

Two things that finding is not.

It is not a verdict that enhancement is wrong. It describes the state of evidence and structure as of this writing. A compound with good healthy-population safety data, used in a non-positional context, with honest information, by someone bearing their own risk, reversibly, would meet all five. Nothing prevents that case from existing.

And it is not permanent. Conditions 1 and 5 are empirical and could be satisfied by research. Condition 4 is institutional and could be satisfied by policy. Condition 2 is partly informational, and this book is an attempt at part of it.

Condition 3 is the one that does not yield to any of that. You cannot research your way out of a collective action problem, and you cannot consent your way out of it either. It is solved, when it is solved at all, by changing the structure of the contest — which is what sport attempted, at the cost §43.3 described, with the results §43.3 described. That is the honest end of this argument, and it is why the coercion question, rather than the safety question, is the one this chapter is named after.

📋 Your Evidence Dossier

This chapter fills Field 9 — Risks — for a compound used without a medical indication.

You have filled Field 9 before, in the clinical framing: what adverse effects were seen, at what rates, in which trials, and what monitoring the label specifies. That version of Field 9 assumes a patient, a prescriber, and a surveillance system. Outside the clinic, none of those exist, and the field has to be rebuilt around a different question — not what are the risks, but whose risk is it, and who would ever find out?

Take one compound from your dossier — ideally one you have genuinely considered, or one someone you know uses. Answer these five in writing.

FIELD 9 — RISKS (NON-INDICATED USE)

  1. WHOSE RISK IS IT?
     Who bears the physiological consequence, and who captures the
     benefit? Are they the same party? If not, is there any mechanism
     transferring cost back to the beneficiary — hazard compensation,
     liability, insurance, anything?

  2. MONITORED BY WHOM?
     Is any clinician aware of this use? Does that clinician have a
     duty to this person, or to a seller, a platform, or an employer?
     If nobody is monitoring, name specifically what is going unwatched —
     which organ system, which marker, which interaction.

  3. AGAINST WHAT BASELINE?
     Were pre-use values ever measured? If a marker is abnormal in two
     years, what will it be compared to? Note honestly whether the
     answer is "nothing."

  4. REPORTED TO WHOM IF SOMETHING GOES WRONG?
     Is there any adverse-event pathway? Would a serious event enter any
     database, any registry, any published series? Would anyone else
     considering this compound ever learn of it?

  5. ATTRIBUTABLE HOW?
     If a problem appears in ten years, what evidence would connect it
     to this exposure? Is the product identity documented? The
     concentration? The dates? Is there anything at all to trace?

Worked demonstration — the same compound, two framings

FIELD 9 — A GROWTH-HORMONE-AXIS COMPOUND, CLINICAL USE
  Whose risk        Patient's, and monitored by a prescriber with a duty
                    to the patient
  Monitored by      Treating physician; scheduled labs targeted at the
                    systems this compound affects
  Baseline          Measured before initiation; on file; comparable
  Reported to       Manufacturer pharmacovigilance and the national
                    adverse-event system; aggregated across all users
  Attributable      Product, lot, dates, and indication all documented

FIELD 9 — THE SAME COMPOUND, NON-INDICATED USE
  Whose risk        The user's, entirely. Performance benefit, if any,
                    accrues substantially to an employer, team, or
                    institution that carries none of the cost
  Monitored by      Nobody, typically. Often not disclosed to the
                    person's own physician, which degrades every other
                    clinical decision made about them
  Baseline          None. A future abnormal value has nothing to be
                    compared against
  Reported to       No one. There is no pathway, therefore no
                    aggregation, therefore nothing for the next person
                    to learn from
  Attributable      Not attributable. Unverified product identity,
                    unknown concentration, undocumented dates,
                    no denominator, no cohort

Notice what this exercise reveals that a risk list cannot. The second entry is not more dangerous because the molecule is different — the molecule is identical. It is a different risk situation because every mechanism by which risk is detected, attributed, aggregated, communicated, and compensated has been removed. The compound did not change. The information system around it disappeared.

And notice the consequence of the last line, which outlives any individual decision: because nothing is reported, the evidence base for this entire category is not growing. Ten years of widespread use produces roughly the same usable knowledge as ten years of nobody using anything. That is why Part III's evidence is still thin, and it will remain thin until something in that column changes.

Conclusion

Clinical medicine supplies four things: a diagnosis, an indication, a clinician who owes you a duty, and a risk-benefit judgment made under that duty for you specifically. Enhancement removes all four and keeps the molecule. Everything difficult about this subject follows from that substitution, and almost none of it is pharmacological.

The workplace supplies the pressure, because advancement is frequently positional and positional contests convert other people's improvements into your decline. Sport supplies the only enforced answer anyone has built, at a cost worth studying before proposing another one. Institutions with authority over their members supply a consent problem that persists even when everyone means well, because "voluntary" means something thinner inside a hierarchy.

Underneath all of them is the coercion structure: an enhancement adopted widely enough stops being an advantage and becomes a baseline, at which point the question is no longer whether to gain an edge but whether to accept a disadvantage. Nobody has to compel anything for that to happen. It emerges from adoption alone, which is why it cannot be solved by regulating conduct, and why the level-playing-field argument — genuinely strong as a safety argument — leaves it entirely untouched.

This book has not told you whether enhancement is permissible. That is a values question, and saying so is more useful than pretending that pharmacology settles it. What it insists on is that you evaluate the empirical claims every position smuggles in. Whether supervised enhancement is safer than a black market: ⚠️, plausible, untested for these compounds, and answering a different question than the one it is usually deployed against. Whether these compounds improve performance in healthy trained adults: ❌, because the trials that were run showed mass without function. Whether enhancement is inevitable: not rateable, and understanding why is worth more than any answer to it.

If you take one thing from this chapter, make it the conditions in §43.9 — and specifically that four of them could in principle be satisfied by research, policy, and honest information, and one cannot. Condition 3 is a property of the contest, not of the person, the compound, or the seller. It is the only part of this that no individual can solve by being careful.

Chapter 44 takes the question to population scale and asks what happens to a society when a metabolic drug that actually works becomes widely used. It is the most speculative chapter in this book and says so on nearly every page — which, after this chapter, should read less as a disclaimer than as the correct application of a rule.

Key Terms

Enhancement — use of a medical intervention to improve a capacity in the absence of a diagnosed condition. Distinguished from therapy not by the molecule but by the presence or absence of a patient.

Therapy–enhancement distinction — the line between treating a condition and improving a normal function. Blurry at the margins, load-bearing at the center, since it determines whether an indication, a prescriber, and a surveillance system exist.

Indication — a regulator's statement that a drug is appropriate for a specified condition in a specified population. A compressed summary of trial evidence whose qualifiers do not transfer when the population changes.

Informed consent — agreement to an intervention based on adequate information about risks, benefits, and alternatives. Requires both real information and a real alternative; either one missing makes it something else.

Fiduciary duty — the obligation of a professional to act in the interest of the person they serve, rather than their own or a third party's. The hardest feature of the clinical relationship to replace outside it.

Positional good — something whose value depends on rank rather than absolute level. One person's improvement is functionally identical to everyone else's decline.

Collective action problem — a situation in which individually rational choices aggregate into an outcome worse for everyone. Includes the prisoner's dilemma and the arms race, and is not solvable by persuading individuals to behave better.

Adoption externality — the cost imposed on non-adopters by others' adoption of an enhancement in a positional contest. Arises without anyone intending it and without any rule changing.

Structural coercion — pressure emerging from the shape of a situation rather than from any person's threat. The central concept of §43.5, and the reason "nobody forced anyone" is not a sufficient reply.

Level playing field argument — the position that sanctioned, supervised, equally available enhancement is fairer and safer than prohibition with a black market. Strong as a safety argument, much weaker as a fairness argument.

Therapeutic use exemption — the mechanism by which an athlete with a genuine medical condition may use an otherwise prohibited treatment (Chapter 38). Notable here because it institutionally concedes that the same molecule is therapy or enhancement depending on the person.

Externality — a cost or benefit falling on someone other than the decision-maker. Activities with externalized costs are systematically over-produced.

Risk transfer — the mechanisms by which a party capturing a benefit assumes a proportionate share of the cost: liability, insurance, hazard compensation, mandated monitoring. Well developed for most occupational hazards, effectively absent for pharmacological enhancement.

Reversibility — whether stopping an intervention returns a person approximately to baseline. The property that determines how costly a decision made under uncertainty turns out to be.

Occupational exposure — a hazard encountered as a consequence of work. Named here because the established occupational-health framework is the model against which enhancement's absence of any such structure is most visible.

Harm reduction — reducing the harms of a practice rather than attempting to eliminate the practice. Has a substantial evidence base in several domains; whether it transfers to a regime that would increase the exposed population is open.

Counterfactual — what would have happened otherwise. The comparison any claim about a policy's effect requires, and the thing most often missing from arguments about sanctioned enhancement.

Off-label use — prescription of an approved drug outside its approved indication. Legal, sometimes well supported, and distinct from enhancement in that a prescriber with a duty is still involved.

Spaced Review

1. (Ch 43) State the coercion structure in five steps without looking. Then identify which step changes the moral character of the decision, and explain in one sentence why the person at that step is doing something different from the person who adopted first — even though the action is identical.

2. (Ch 16 + Ch 43) Chapter 16 described a dissociation between mass and function. Explain how that single finding does double duty in this chapter: what it does to §43.3's rating, and what it does to the credibility of a person who says "I could tell it was working."

3. (Ch 19 + Ch 43) Chapter 19 listed what medical supervision provides. Pick the three items you think matter most for a compound with no long-term human safety data, and then explain why supervision is nonetheless a weaker safeguard here than it is for an approved drug — even when a physician is genuinely involved.

4. (Ch 14 + Ch 43) A growth-hormone-axis compound appears in Chapter 14 with real approved uses and in this chapter as a weakly supported enhancement. Explain why this is not a contradiction, using Chapter 5's first rule, and state the population and endpoint that distinguish the two claims.

5. (Ch 43, integrative) A colleague argues: "Everyone's going to do this eventually, so we should legalize it, supervise it, and make it available to everybody equally — that's the fair outcome." Identify the three separate claims in that sentence, rate each one, and say which of the five conditions in §43.9 the proposal would satisfy and which it would not.