# The Half-Lives of a Meme

*The joke, the crowd, and the position do not decay together.*

You see a ridiculous image. You laugh. You send it to a friend. Eventually, perhaps, you buy the token attached to it.

The next time you see the image, something has changed. It is still the same joke, but now it also represents a decision you made. A week later, you might be opening its chart without looking at the image at all.

What, exactly, is keeping your attention?

This is the part of memes I find more interesting than the question of which ones become immortal. I want to understand how the same object can move through several different relationships with a person: discovery, amusement, participation, ownership, vigilance, and memory. How something that initially costs three seconds can eventually occupy an afternoon.

My hypothesis is that **a meme has several psychological half-lives, and the interesting behavior happens when they come apart.** Novelty can fade while belonging grows. Enjoyment can decline while a financial position remains open. Selling can end the exposure without ending the argument you are having with yourself.

This is a research essay and a proposed model, not a measured biological history of token traders. Two charts below reproduce published results. The other three make hypotheses visible; they do not measure dopamine, serotonin, cortisol, or anyone’s actual feelings.

## 1. Before asking when a meme dies, ask what is dying

The essay [*Hype and the Hive Mind*](https://hoppa.international/research/HypeAndTheHiveMind/index.html) offers a useful starting distinction: the circulating excitement around an object is not necessarily the same thing as the object’s enduring appeal. Its discussion of digital memes also highlights their ability to mutate through participation.

But we do not need to assume that hype and timelessness are two exclusive categories, or that all hype follows one exponential law. A format can become familiar in one group and newly interesting in another. A derivative can revive attention without reviving enthusiasm for the original.

Nor is quantitative meme research new. Wu and Huberman studied attention to Digg stories in 2007 and modeled novelty decay with a **stretched exponential**, rather than a simple fixed-rate exponential. Leskovec and colleagues subsequently tracked changing phrases through news and blogs. These were studies of information circulation, not a direct account of what token holders feel. [Wu & Huberman](https://pmc.ncbi.nlm.nih.gov/articles/PMC2077036/); [Leskovec, Backstrom & Kleinberg](https://www.cs.cornell.edu/home/kleinber/kdd09-quotes.pdf).

An especially useful empirical result concerns the turnover of popular Twitter hashtags. Lorenz-Spreen and colleagues reported a decline from 17.5 hours in 2013 to 11.9 hours in 2016 in average total time occupying the hourly global top 50 per distinct hashtag. That is about a 32% reduction. It is a historical measure of collective attention allocation—not a claim that people’s individual attention spans fell by 32%. [Study and methods](https://www.nature.com/articles/s41467-019-09311-w).

![Published Twitter results: mean annual total hours in the hourly top 50 per distinct hashtag, 17.5 in 2013 and 11.9 in 2016.](charts/01-attention-evidence.svg)

*Figure 1 · Published summary values. This is neither a continuous trending spell nor a measured half-life of a meme.*

The useful shift is from “How long does hype last?” to “How long does each reason for paying attention last?”

Those are different research questions.

## 2. One image, several relationships

For a tradable meme, I would separate at least six relationships. They are not compulsory stages. People skip them, overlap them, reverse direction, and participate without ever buying.

| Relationship | The question the person may be asking | What could actually be observed |
| --- | --- | --- |
| Discovery | “What is this?” | First exposure, stopping, clicking |
| Recognition and amusement | “Do I get it? Is it good?” | Comprehension, enjoyment ratings, voluntary rewatching |
| Participation | “What can I do with it?” | Sharing, remixing, replying, making something |
| Acquisition | “Do I want exposure to this?” | A purchase, its size, and the person’s stated reason |
| Holding | “What happens next—and what does that mean for me?” | Position changes, checking, forecasts, reported worry |
| Exit and aftermath | “Was that the right decision?” | Selling, subsequent lookups, continued participation, re-entry |

The hook matters because it is the first invitation into this system. A legible object, joke, or action gives someone something they can pass on. But I would distinguish **being easy to understand** from **being easy to participate in**. A meme can succeed at the first and fail at the second.

A repeatable action—a pose, a phrase, a button press—also gives the next participant something to contribute. That is a proposed distribution mechanism, not proof that every participatory meme spreads well.

Once money enters, another distinction becomes unavoidable: **the meme and its token are different objects.** Someone can love the image and sell the asset. Someone else can dislike the image and trade the asset. Buying is not a clean measurement of belief; selling is not a clean measurement of cultural rejection.

“The community” can therefore contain people performing similar visible actions for incompatible reasons.

## 3. The biological machinery, without the chemical fan fiction

It is tempting to label discovery “dopamine,” belonging “serotonin,” and holding a red bag “cortisol.” That would make a beautiful diagram and a poor explanation.

The biology is more interesting when we keep its actual distinctions.

### Dopamine: an outcome is not the same as its surprise

A simplified anatomical route runs from dopamine-producing neurons in the midbrain, including the ventral tegmental area, into regions such as the ventral striatum and prefrontal cortex. This is part of a distributed system involved in learning, motivation, and action—not a pleasure pipe. [Anatomical reference](https://www.ncbi.nlm.nih.gov/books/NBK92777/).

In a classic monkey experiment, Fiorillo and colleagues found that reward responses depended on how expected the reward was. They also identified a sustained anticipatory response that was strongest at a 50% reward probability, where uncertainty about reward delivery was maximal. These were distinct signals, and not every recorded neuron showed the uncertainty response. [Fiorillo, Tobler & Schultz, 2003](https://www.hms.harvard.edu/bss/neuro/bornlab/nb204/final_exam/FiorilloSchultz2003.pdf).

The useful analogy for a trader is that an outcome and its surprise are different quantities. A +10% move against an expectation of +2% is eight percentage points above expectation. The same move against an expectation of +20% is ten points below it. That arithmetic is not a dopamine measurement, but it explains why “green” is an incomplete psychological description.

There is another important separation: pursuing a reward and enjoying it. In one mouse experiment, animals with a dopamine-transporter mutation associated with elevated extracellular dopamine pursued a sweet reward more vigorously, without increased positive taste reactions used to assess liking. The translation to human trading remains a hypothesis. [Peciña and colleagues, 2003](https://pmc.ncbi.nlm.nih.gov/articles/PMC6740586/).

It nevertheless gives us a useful question: **could the desire to check outlast the enjoyment of what is being checked?**

Dopamine receptors do not rescue a simpler story. D1 and D2 name receptor types, not opposing financial decisions. The striatal circuits associated with them cannot be mapped onto buying versus selling; in mice, neurons in both direct and indirect pathways became active before action initiation. [Cui and colleagues, 2013](https://pmc.ncbi.nlm.nih.gov/articles/PMC4039389/).

### Serotonin: patience is not proof of contentment

Serotonin neurons in the brainstem’s raphe nuclei project widely. Their functions cannot be compressed into “the chemical of feeling good about your bag.” [Anatomical reference](https://www.ncbi.nlm.nih.gov/books/NBK11035/).

In a mouse study, activating dorsal-raphe serotonin neurons promoted waiting under particular combinations of reward probability and uncertain timing. The effect was not a general increase in patience under every condition. [Miyazaki and colleagues, 2018](https://www.nature.com/articles/s41467-018-04496-y).

This suggests a distinction worth testing in humans: **“I believe the reward will come, but don’t know when” is different from “I don’t know whether a reward will come.”** Two holders with equally long holding periods could be experiencing very different kinds of waiting.

That does not make holding a serotonin assay. Nor does it establish that community participation releases a predictable amount of serotonin. Belonging needs to be measured as belonging.

### Cortisol: uncertainty is not the same as losing

The relevant anatomical chain is the hypothalamic–pituitary–adrenal axis: hypothalamic CRH prompts the anterior pituitary to release ACTH, which stimulates the adrenal cortex to produce and secrete cortisol. Feedback regulates the system. This hormonal cascade operates on a different timescale from brief neuronal responses; it is not an instantaneous red-candle indicator. [Physiology reference](https://www.ncbi.nlm.nih.gov/books/NBK541120/).

In a small field study of 17 male professional traders over eight working days, Coates and Herbert found associations between cortisol and financial uncertainty, including variability in trading outcomes. Their findings did not reduce to “losses raise cortisol.” This was observational evidence from a specific population, not a universal law. [Coates & Herbert, 2008](https://pmc.ncbi.nlm.nih.gov/articles/PMC2329689/).

Experimental results further resist a one-way story. One study found greater risk aversion after sustained cortisol elevation. Another found increased risky investment after acute administration in a different task. Dose, duration, sample, and task differed, so this is not proof of a neat acute-versus-chronic rule either. [Kandasamy and colleagues, 2014](https://pmc.ncbi.nlm.nih.gov/articles/PMC3948282/); [Cueva and colleagues, 2015](https://www.nature.com/articles/srep11206).

The defensible conclusion is narrower and more useful: a person’s response to financial uncertainty has biological components, but **there is no justified conversion from a price chart into a cortisol chart**.

Excitement and apprehension can coexist as psychological experiences. Dopamine and cortisol should not be drawn as two ends of a single emotional seesaw.

## 4. The half-life of the joke can be shorter than the half-life of the position

Consider four processes: novelty, the pull to participate, financial preoccupation, and identity relevance. For illustration, suppose their intensity after a trigger decays according to:

`remaining intensity = starting intensity × 2^(−time / half-life)`

Assign them different half-lives and their trajectories separate.

![Four illustrative decay curves with different chosen half-lives, not estimated human or biological measurements.](charts/02-four-clocks.svg)

*Figure 2 · A toy model. The half-lives and their ordering are chosen to illustrate the hypothesis. The time units are arbitrary; real trajectories can rise, plateau, stop abruptly, or reactivate.*

Early on, novelty may explain most of a person’s attention. Later, the person could remain because of friends, unfinished creative work, financial exposure, or the significance of having publicly committed.

The visible behavior—opening the same page—might not change. Its purpose could change completely.

An especially revealing counterexample to “attention requires fresh information” comes from research on individual investors. Quispe-Torreblanca and colleagues found that investors revisited known gains, including across successive market-closure days. The paper supports a role for enjoyment of attending to favorable outcomes. It also finds reduced attention to losses in important settings. It does **not** establish that everyone becomes an anxious checker. [*Attention Utility*, 2026](https://academic.oup.com/restud/article/93/1/664/8149267).

So there are at least two possible long tails: returning because the outcome is pleasant to contemplate, and returning because something consequential remains unresolved. There is also avoidance: not opening the page at all.

The strong hypothesis is not that holders necessarily feel worse over time. It is that **ownership can change the reasons they continue—or refuse—to look**.

## 5. Why a rally need not feel safe

Suppose a position rises from $1,000 to $3,000 without a change in token quantity. A hypothetical 10% drawdown now puts $300 at stake instead of $100. That is arithmetic, not a stress measurement. But it shows how gains can enlarge the amount someone feels responsible for protecting.

There can also be a new comparison point: the recent peak, an imagined exit, a friend’s return, or the life the person briefly pictured buying.

We do have experimental evidence that the route matters. In a study involving 150 financial professionals, hypothetical investments ending at the same +10% return received different satisfaction ratings. A fall followed by recovery averaged 1.99; a rise followed by retreat averaged 1.05, on a −4 to +4 scale. The difference was 0.94 scale points—not “90% more happiness,” which would be an inappropriate interpretation of this scale. [Schwaiger and colleagues, 2020, Table 1](https://research-portal.uu.nl/ws/files/188505948/1-s2.0-S016518891930048X-main.pdf).

![Published experiment: the same positive final return elicited different mean satisfaction depending on the preceding price path.](charts/04-same-return.svg)

*Figure 3 · Published laboratory results for the professional subgroup. Participants imagined holdings; these are satisfaction ratings, not hormone measurements or direct measures of stress.*

The study does not prove that rising prices cause anxiety. It does show why current profit is insufficient to describe someone’s experience.

My extension is a testable hypothesis: **during a rally, expected opportunity and perceived responsibility for protecting gains can increase together**. The trader need not choose between being excited and being worried.

This is what a possible lifecycle could look like if we draw psychological hypotheses rather than fictional chemical concentrations:

![A conceptual participant lifecycle showing anticipated reward, perceived threat, and belonging across discovery, participation, buying, holding and exit.](charts/03-lifecycle-waves.svg)

*Figure 4 · One constructed scenario, not a typical or estimated trajectory. Simultaneous peaks show a possibility, not a correlation established by data. Other people would have different curves, including avoidance rather than increased checking.*

The important feature is the overlap. The same event can supply a reason to hope and a reason to monitor. A sell can remove direct exposure while leaving social attachment or comparison with the subsequent price.

There is no natural requirement for these processes to peak together, settle together, or even move in the same direction.

## 6. A second wave might be a different audience

Now move from one person to the market.

A dashboard shows visits to the same token page recovering. It is tempting to say the meme has regained its magic. But at least two different situations could produce that curve:

- Most visits are first visits from newly exposed people.
- Most visits are repeat visits from people who have already been there.

Those compositions can generate identical total page-view counts. Neither identifies the visitor's motive: enjoyment, curiosity, financial worry, or something else. That is a measurement problem, not a matter of choosing the more attractive story.

![Two constructed page-view curves with identical totals but different proportions of first visits and repeat visits.](charts/05-attention-composition.svg)

*Figure 5 · An identifiability example: different compositions can produce the same total. Neither panel is real platform or market data.*

A newcomer encountering the joke today is not at the same psychological age as someone who bought three weeks ago. Plotting both against “days since launch” blurs that difference.

This gives us two clocks even before adding biology: **the age of the meme and the age of the person’s relationship with it**. A revival could reset the first-exposure clock for a fresh audience while doing little for existing holders.

It also means that raw engagement cannot distinguish affection from argument, conviction from coordination, or pleasure from checking. Even a strong relationship between posting and trading would leave open the direction of causation: posts can draw attention to trades, while price moves can give people something to post about.

The useful question becomes: *who is returning, what are they doing, and what has changed for them?*

## 7. Exit does not necessarily end the relationship

Not everyone sells. Some people collect, forget, lose access, or retain a position indefinitely. Others sell part of it. “Eventually everyone sells” is too strong to serve as an assumption.

Among those who remain, holding is not enough to establish enthusiasm. Odean’s classic brokerage study found a stronger propensity to realize gains than losses. The result does not reveal each holder’s emotions, but it warns against interpreting an unsold position as a vote of confidence. [Odean, 1998](https://faculty.haas.berkeley.edu/odean/Papers%20current%20versions/AreInvestorsReluctant.pdf).

And after a sale, the sale price can remain relevant. Research on repurchases found that investors were more inclined to repurchase previously sold stocks when their prices were below, rather than above, the earlier sale price. Regret is one interpretation; the transactions do not directly measure it. Nor do they prove obsessive post-sale checking. [Strahilevitz, Odean & Barber, 2011](https://faculty.haas.berkeley.edu/odean/papers/Repurchases/Once%20burned%20JMR.pdf).

The meme-specific hypothesis is that **capital exposure and psychological involvement have separate exit times**. You can be out of the position and still in the group chat. You can stop liking the group and still own the token.

A lasting artifact complicates the picture further. An image, physical object, or shared ritual could preserve meaning independently of the investment. It could also deepen identification with a financially painful decision. Neither possibility licenses the claim that meaningful art neutralizes losses or lowers cortisol.

The question is not whether the artifact makes someone hold forever. It is whether the reasons for caring become less dependent on price—and how we would tell.

## 8. Michael Levin: a half-life is not a time horizon

There is another dimension missing from the usual hype curve. It tells us how long a reaction persists, but not how far through time the reacting system can think and act.

Michael Levin's **cognitive light cone** offers a useful framework. In his account, the relevant boundary of an agent concerns the scope of events it can measure, model, and try to affect. Its reach includes space and time: what can enter its concerns, and over what scale can it pursue goals? This is a proposal about cognition and agency, not a literal claim that a community emits a physical light cone. [Levin, 2019](https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2019.02688/full).

What I will loosely call *temporal width* is closer to the **temporal depth** discussed explicitly by Levin and coauthors: the reach of recall and planning, including consideration of alternative future actions. I am borrowing that conceptual distinction, not importing the clinical claims of their theoretical paper into meme psychology. [Tolchinsky, Levin and colleagues, 2025](https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2025.1585315/full).

The distinction is simple:

**A half-life asks how long a response lasts. A time horizon asks how much past and future can organize action now.**

A holder can spend months waiting for one number. A small group can spend a week remembering earlier mistakes, dividing work, and completing something none could have completed alone. Duration by itself does not tell us which system is doing the more temporally integrated work. Waiting can be part of a plan; a wallet balance cannot tell us whether there is one.

The same caution applies to scale. A million people watching the same candle are not thereby a million-person mind. In Levin and Resnik's later account, collective agency requires integration sufficient for goal pursuit: coordinating subgoals, responding to disturbances, or changing methods while maintaining an objective. Success is not guaranteed, but there must be more than synchronized attention. [Levin & Resnik, 2026](https://link.springer.com/article/10.1007/s13752-025-00524-5).

That gives us a practical way to connect the framework to memes:

| Capacity | Possible role of a meme community | What would distinguish it from a slogan? |
| --- | --- | --- |
| Usable memory | An archive preserves prior decisions, work, and corrections. | Newcomers use it to avoid repeating mistakes. |
| Future-directed coordination | A shared object organizes a project beyond today's feed. | Different people complete connected tasks toward a specified outcome. |
| Adaptive action | The group can change how it works without losing its purpose. | A handoff or blocked route produces an alternative plan, not paralysis. |

The meme is a possible aid to those capacities: a recognizable reference point, a compact reminder, a way of identifying shared work. Whether the symbol itself should count as an agent is a separate question. No such claim is needed here.

Three less obvious implications follow.

First, **a fast interaction can support a long project**. A two-second ritual and a six-month goal are not opposites. The ritual might give immediate feedback while a shared record makes each contribution part of something cumulative. Equally, frequent checking can coexist with serious long-term planning. Refresh rate is not a measure of temporal depth.

Second, **memory matters only if someone can use it**. An enormous archive can be inert. A short, accurate handoff can change the next decision. More stored material does not automatically mean a larger effective cognitive horizon.

Third, **longer is not automatically healthier or wiser**. A community can organize around a mistaken belief. A distant target can become an excuse to ignore disconfirming evidence. The test is whether the plan can learn, revise, and tolerate people changing their minds—not how effectively it persuades people never to sell.

These are my proposed applications of Levin's framework, not findings about BUY or endorsements from him.

## 9. Why an overnight meme may have years behind it

Some spectacularly fast launches are not starting from zero. They are attaching a new action or instrument to something culture has already made recognizable.

Dogecoin is a clean example of the timing distinction. The project's own history dates the Kabosu photograph to 2010 and Dogecoin's launch to December 2013. The token arrived after the image and the joke. Its launch date was not the culture's birthday. [Dogecoin's history](https://dogecoin.com/dogepedia/articles/history-of-dogecoin/).

This does not prove that inherited familiarity caused its success. It suggests a mechanism: for an audience that already knows the symbol, the new project may need to explain only the new use, rather than teach the whole object from scratch.

The Ice Bucket Challenge illustrates a different ingredient. The ALS Association's account describes a recognizable filmed act, personal nominations, existing relationships, and amplification through athletes and celebrities. People received both something to watch and a way to produce the next instance. [ALS Association retrospective](https://www.als.org/blog/challenge-me-finish-what-was-started).

Those cases suggest three separable advantages: **recognition already available, a contribution easy to reproduce, and paths into audiences beyond the initial group**. They do not tell us how much each ingredient caused, or how many unsuccessful attempts had similar ingredients.

Research makes the network distinction important. Goel and colleagues separate sheer popularity from *structural virality*: a large broadcast is different from a deep chain of people passing something onward. Large cascades can combine both. A celebrity can provide scale without demonstrating that ordinary viewers will independently reproduce the format. [*The Structural Virality of Online Diffusion*](https://www.microsoft.com/en-us/research/publication/the-structural-virality-of-online-diffusion/).

Weng and colleagues, studying Twitter hashtags, found that early spread across communities helped predict later popularity. A group can reinforce a meme while also containing it. Their observational result does not isolate reinforcement from homophily, but it warns against mistaking intense internal repetition for broad adoption. [*Virality Prediction and Community Structure in Social Networks*](https://www.nature.com/articles/srep02522).

My synthesis is that an effective format may offer **quick recognition and room for continued participation**. The visible object stays easy to identify while the setting, participant, or consequence changes. This is related to Shifman's distinction between forwarding content and participating through imitation and remix; it is not a measured growth formula. [Shifman, 2013](https://onlinelibrary.wiley.com/doi/10.1111/jcc4.12013).

There is a useful tension here. A totally closed joke may be easy to understand but leave little to contribute. A lore universe may offer endless depth but demand too much explanation before the first laugh. The design hypothesis is to keep the entrance simple and make the optional second step worthwhile.

The strongest warning against overconfidence comes from an artificial music-market experiment. Salganik, Dodds, and Watts exposed separate groups to the same songs. Social influence increased inequality and unpredictability across those experimental worlds. Appeal mattered, but it did not fix the identity of the winner. [Study](https://www.princeton.edu/~mjs3/salganik_dodds_watts06_full.pdf).

Recognizing useful ingredients is therefore not the same as knowing which meme has the highest probability of winning. Looking only at winners hides the denominator.

## 10. The strongest case for BUY

*Disclosure: I work on BUY. This is an operator's argument for a project I am involved in, not an independent ranking or a conclusion of the studies above. Here, success means broad cultural adoption and continued voluntary participation—not a promised token return.*

The case I find compelling is that BUY can connect the two timescales: an object that is quick to recognize and a story people can continue working on.

### A familiar action with history attached

A buy button is an interface control before it is a crypto narrative. That gives the concept an intelligible entry point for people familiar with purchasing or trading interfaces. A newcomer can grasp the action before understanding the project.

There is also a specific historical layer. Robinhood's own filings record restrictions on purchases of certain securities, including GameStop and AMC, beginning on January 28, 2021. [Robinhood's account of the restrictions](https://investors.robinhood.com/static-files/7ae57af3-0dd6-47fe-9bcc-c5c15389ca51).

BUY's narrative uses the button to recall that experience. The history predates the project, and its meaning is not confined to our community.

My inference is that this offers two possible entrances: *I understand what a button does*, and *I remember why that button mattered*. Recognition need not depend on historical knowledge, while historical knowledge can supply additional meaning. Whether both entrances work with unfamiliar audiences is something to test, not assume.

### A format people can perform

The project's public visual system keeps a stable fluorescent capsule and a simple action: press it. Its grammar includes “button check?” and “still on.” These are existing design choices, not evidence of mass adoption. [Public brand system](https://www.hoodbuybutton.com/brand/).

The opportunity is to let different people supply the scene, reaction, or character while retaining the recognizable object. Someone should be able to contribute a joke instead of merely repeating the founder's explanation.

This is also a way the format could travel between communities without demanding that they abandon their own identity. A character can remain that character while interacting with the button. A collaboration still needs permission and must not imply endorsement that was never given.

The important distinction is between **a symbol with fans and a format with authors**. If outsiders begin making recognizable versions for their own reasons, the project has evidence of a different kind of adoption than a high-view founder post.

### An immediate action with somewhere to go afterward

BUY already has publicly accessible places beyond its market chart. The Witness Archive presents sourced recollections and explicitly offers participation without a wallet. Button Hunt presents a card-collecting game around the object. Their existence is verifiable; durable demand for them is a separate question. [Witness Archive](https://www.hoodbuybutton.com/witness/); [Button Hunt](https://hunt.hoodbuybutton.com/).

The archive offers one concrete application of the temporal-depth argument: a first visit can be brief, while a contributed recollection remains available to later visitors. The game offers another way to interact with the object, though playing it does not by itself establish collective memory or authorship.

That is an opportunity to accumulate cultural work rather than require every generation of participants to restart the same argument about price. It works only if the work is usable, enjoyable, and discoverable. The presence of a game or archive is not evidence that people want it.

The project's stated return of the button on Robinhood Chain supplies a further narrative turn: the symbol reappears in a new setting associated with its history. That is BUY's framing, not institutional validation. BUY is independent and is not endorsed by Robinhood. [Project thesis and affiliation boundary](https://www.hoodbuybutton.com/why-buy/).

### Why I would take the concept seriously—and what could break the thesis

For an audience that understands the interface or remembers the event, BUY begins with available context. Its action is demonstrable, its visual object can remain stable through remixes, and participation need not mean taking financial exposure. **That combination gives me a concrete reason to believe BUY has a strong cultural opportunity.** The opportunity is to turn familiar action into independent authorship.

It does not establish the highest odds across all memes. We have not defined and measured that comparison set. Nor does recognition of the word *buy* automatically flow to this particular project: a generic word can make attribution and discovery harder. Everyone else's buy button is not an advertisement for this token.

The strongest failure modes follow directly from the argument. If the joke only makes sense to existing holders, the entrance is not simple enough. If the same small circle makes all the work, the format has not crossed communities. If every interaction leads back to an investment pitch, non-buyers may reasonably stop participating. If the archive, game, and symbolism grow while the token does poorly, cultural usefulness will not reimburse holders.

The evidence that would strengthen my confidence is therefore specific. I would separately test whether new viewers understand the action, whether they associate the distinctive object with this project, and whether people outside our circle voluntarily make versions of their own. I would then look for repeat participation without a requirement to own BUY, and creative projects that continue through a founder handoff. These are proposed tests, not milestones I am claiming we have reached.

That is the bull case I can defend: **BUY can make entry easy without requiring the relationship to stay shallow.** It starts with something you can recognize and do. Its longer-term promise is to make what people do together accumulate.

## 11. What would an honest test look like?

The first version should measure behavior and reported experience, not pretend to reverse-engineer a nervous system from a wallet.

With explicit participant consent, collect repeated ratings of amusement, anticipated reward, worry, belonging, and the urge to check. Record first exposure, participation, entry, partial sales, and complete exit. Separate content viewing from price checking. Position size relative to someone’s finances is more informative about exposure than a bare token balance, but it is sensitive data and should be optional and protected.

Then align observations to each person’s events, not only the launch date. Keep first-timers, repeat participants, holders, and exited participants distinguishable. People can belong to more than one group.

Four hypotheses would make this essay meaningfully falsifiable:

1. **Attention can outlast amusement because of exposure.** Within people, amusement declines faster than checking while a position remains open. A decline in checking around full exit, after accounting for market events, would be consistent with this. Similar decay rates or no exit-related change would weaken it. Exit is self-selected, so an event study alone would not prove causation.
2. **The comparison point matters beyond today’s return.** At similar current returns, exposure, and volatility, distance below a remembered peak or a previously stated expectation predicts lower satisfaction or greater worry. If these comparisons add no predictive value out of sample, this part of the model needs revision.
3. **Social involvement has its own exit curve.** Conditional on pre-exit participation, stronger pre-existing belonging predicts slower decline in participation during a prespecified period after full financial exit. If belonging adds no predictive value beyond prior activity and exposure history, this part of the model is weakened.
4. **Shared memory supports action beyond the current participants.** In consenting groups doing the same low-stakes creative task, compare an ordinary introduction with a usable record of prior decisions, constraints, and goals. Measure completion and useful independent decisions during a planned handoff. If the record improves recall but not action, the claim of greater practical temporal depth is weakened. Do not manipulate trades or manufacture financial stress for this test.

For decay, begin with baseline-adjusted response curves. Estimate an exponential only where it fits; compare alternatives such as stretched exponentials and flexible curves. Report uncertainty. A response that stays above half its initial level during observation has an unobserved half-decay time, not an invented one. New triggers should be modeled explicitly rather than smoothed into a longer “natural” half-life.

For self-report scales, a literal half-decay time also requires a defensible measurement model: half a rating does not automatically mean half an experience. Without that model, compare time to prespecified response thresholds instead.

Market and engagement data provide context: returns, volatility, liquidity, posting, new contributors, repeat visits. They do not identify hormones, and wallets are not people. Bots, multiple wallets, private chats, recommendation systems, and missing observations can all distort the apparent relationship. Missing should remain missing, not become zero.

If cortisol is genuinely the research target, it requires an appropriate biological study, repeated samples, timing controls, and qualified oversight. Heart rate is not a substitute cortisol assay. Peripheral serotonin measurements cannot, by themselves, establish serotonin release or receptor signaling in the brain circuits of interest, and a public chart cannot reveal dopamine receptor activity. [Evidence on distinct peripheral and central serotonin pools](https://pmc.ncbi.nlm.nih.gov/articles/PMC9105435/).

The payoff would be a **map of changing motives**, with biology supplying constrained hypotheses rather than decorative labels.

## 12. The thing that survives may not be the thing that arrived

The conventional picture is a single wave: discovery, euphoria, decline, death.

I suspect that picture is often a projection of several different processes onto one line. A short-lived joke can initiate a longer-lived relationship. A financial position can retain attention after the original attraction fades. A community can survive a sale. A new audience can make an old format look young again.

Levin's framework adds a second question: can any of that retained attention support usable memory and coordinated action? A community might acquire more of those capacities even while the novelty of its original joke declines. Equally, an old and busy group might never acquire them at all.

For BUY, that turns the ambition into a test rather than a victory declaration: can a very short interaction become an entrance to something people keep making together?

The sharpest question is therefore not “Is the meme still alive?”

It is: **which reason for caring is still alive, in whom?**

That question is less convenient than a universal cortisol curve. It is also much more measurable.

You first opened the image to see what it meant.

Later, you might open the chart to see what your decision meant.

The meme is unchanged. Your position in it is not.

---

### About the figures and evidence

Figures 1 and 3 use the cited papers’ published summary values, not newly collected observations. Figures 2, 4, and 5 are reproducible conceptual constructions. Their shapes, timing, and parameters are not fitted to people, token markets, or biological samples. No shaded area represents statistical uncertainty unless explicitly identified; none is used here.

The supporting package includes editable text, SVG and PNG charts, the chart generator, underlying CSV files, and an evidence/methods note. This is a synthesis and a proposed research agenda, not an established unified theory, a clinical account, or a trading signal. The Levin connections are conceptual applications; the BUY section is explicitly an interested operator's thesis. No probability ranking or return forecast is estimated.