WELCOME TOTHE ATTENTION
CASINO✳
The house always wants one more minute.
Tonight, you get to examine the house.
Walk through seven neon rooms about feeds, rewards, outrage, notifications, approval, defaults and focus. Try a few short games—or just read. Nobody is measuring your mind. Nothing is tracking you.
Your attention is not
a character flaw.
What looks like a personal failure of willpower may also be a very well-designed invitation. A notification can be useful; a feed can teach; a like can connect. The same mechanisms can make stopping harder than starting.
This exhibition refuses two easy stories: “technology is evil” and “just have more discipline.” The interesting question is who designs the choice, what they reward, and whether you can make an informed exit.
The exhibits below are fictional miniatures, not validated cognitive measures. Sources support the discussion; game outcomes describe only what happened here.
Pick your distraction.
Almost-EndINFINITE SCROLL ↗ ROOM 02 / THE SLOT✳Maybe
Next TimeVARIABLE REWARDS ↗ ROOM 03 / THE SPARK⚡Before You
ForwardOUTRAGE ↗ ROOM 04 / THE BELL◌The Important
Little PingNOTIFICATIONS ↗ ROOM 05 / THE MIRROR♡Applause
MachineSOCIAL VALIDATION ↗ ROOM 06 / THE DOOR☑The Choice
Already MadeDEFAULTS & FRICTION ↗ ROOM 07 / THE EXIT↗The Quiet
RoomDELIBERATE FOCUS ↗
Every game is opt-in, explicitly finite, interruptible and entirely local to this browser tab. The reading stays available without JavaScript.
MECHANISM / FRICTIONLESS CONTINUATION
The neverending
almost-end.
A staircase can tell you when you reach the next floor. A feed often declines to tell you when you've arrived anywhere.
Infinite scrolling removes a small decision: whether to ask for the next page. Sometimes that's excellent design. Images, ideas and conversations are easier to browse when exploration is the point. But the absence of an obvious boundary also means a person must invent their own stopping signal, again and again.
The peculiar bargain is this: the next item might be precisely what you came for. Or it might merely renew the promise. When content is served as a continuous stream, continuation feels like the default; stopping becomes the action.
That is not proof that every scrolling product is exploitative. Nielsen Norman Group describes contexts where infinite scrolling works well and others—like task-oriented comparison—where it impairs orientation and refinding. The design question is not “Is scroll bad?” but “Can the user find, finish and leave?” [Research note ↗]
The feed with an end
40 seconds maximum. Decide when enough is enough. Fictional cards, eight maximum, no auto-loading.
Without playing: Ask any feed a simple question: “Where is the natural point at which I can feel done?” A bookmark, page number, end card or deliberate pause can change the answer.
MECHANISM / UNCERTAINTY + FEEDBACK
Maybe
next time.
Certainty can be useful. Uncertainty can be magnetic.
Not every refresh brings something interesting. A new comment, a funny post, a message from someone you like, an update you actually need: all can arrive unpredictably. That occasional reward is part of the pleasure of a living network. The trouble begins when the possibility of reward becomes easier to chase than the reward itself is to enjoy.
Researchers have modelled parts of social posting as reward learning: in observed data and experiments, posting behaviour varied with social feedback. That is meaningful evidence for one mechanism, not a universal explanation of why everyone uses a platform. Friendship, research, identity, boredom and work all complicate the picture. [Original study ↗]
There is a difference between an author surprising readers and an interface that deliberately refuses to clarify what the user is chasing. Unexpected delight is not a crime. When a design obscures the terms of the exchange, however, delight can become an excuse for endless checking.
The almost-jackpot
35 seconds maximum. Reveal up to six fictional envelopes. Nothing at stake; inspect what “maybe” invites.
Without playing: A surprise is a pleasure; compulsive anticipation is something different. Notice when the promise of a future update starts overshadowing the present one.
MECHANISM / EMOTION + AMPLIFICATION
Before you
forward.
A headline can demand a verdict faster than it supplies the facts.
Indignation is not inherently irrational. Anger can notice injustice, protect boundaries and rally help. Yet an interface that rewards immediate reaction may privilege the message that feels certain over the message that is carefully conditional.
One study found that social-media observers overestimated the moral outrage that message authors themselves reported. The researchers connect such overperception to distorted beliefs about how hostile groups are. This is a narrower claim than “social media makes everyone angry,” and the distinction matters. [Study ↗]
A careful pause needn't drain the moral force from a question. It can improve its accuracy. Ask: What happened? What am I inferring? What evidence could change the story? A well-placed “I don't know yet” is sometimes more courageous than a perfectly sharpened insult.
The context desk
40 seconds maximum. Four invented headlines. Share, inspect, or pass—then see what's missing.
Without playing: “I need a fact before I need a reaction” is not indifference. It's a way to care about what is actually happening.
MECHANISM / INTERRUPTION + OPEN LOOPS
The important
little ping.
Someone wants your attention. That does not mean your attention is currently available.
A notification can be merciful: the courier has arrived, a friend needs help, a deadline has moved. But every signal that says “now” competes with whatever “now” already held. The interruption may matter even when you don't open the message, because a question has entered the room: *What was that?*
In a controlled study, researchers found that phone notifications alone disrupted performance on an attention-demanding task, even without interacting with the phone. This experiment generalises neither to every notification nor to every real-life situation. It does illustrate why ignoring a ping isn't necessarily the same as avoiding its cost. [Study ↗]
Being reachable is valuable. Being perpetually interruptible is a different contract. There is no universal correct setting: a caregiver, dispatcher and novelist may need wildly different forms of access. The humane question is whether urgency is chosen according to need or handed to every application by default.
The reading room
40 seconds. Read a short passage while optional fictional on-page pings arrive. Nothing uses system notifications.
Without playing: In a meaningful conversation, ask how it feels when an unrelated notification enters the room. The notification may be tiny; the context it interrupts may not be.
MECHANISM / SOCIAL SIGNALS + SELF-WORTH
Applause
machine.
A like can mean “I saw you.” It cannot tell you who you are.
Numbers compress social life into an elegant little dashboard. A post gets a reaction; a photograph gets a count; a story gets a view. These signals can be useful feedback, especially for artists, organisers and communities trying to understand what resonates.
But a metric is not the thing it measures. A quiet response to an honest piece can coexist with deep private meaning. A viral joke can matter for three minutes and nothing more. Research on social reward learning shows that online feedback can shape behaviour; it does not establish the artistic, moral or personal value of the thing receiving feedback. [Evidence ↗]
There is no shame in enjoying applause. The difficulty begins when applause becomes the only mirror. Different audiences, algorithms, timing and conventions can produce very different numbers for the same work. Sometimes the most important reader will never tap a heart.
The same poem twice
One quick choice, no timer required. See what a like count makes you expect, then reveal the trick.
Without playing: Imagine a beautiful sentence with twelve likes and the same sentence with twelve thousand. Only the surrounding signal changed.
MECHANISM / CHOICE ARCHITECTURE
The choice
already made.
A button doesn't need to lie to steer you. Sometimes it merely needs to be the brightest thing in the room.
Defaults are unavoidable. Someone must decide whether music begins muted, where the keyboard focus lands, whether a box is checked, and how easy it is to go back. Good defaults protect people from tedious work: accessibility settings, backups and security can all benefit from sensible preselection.
Problems emerge when the easiest route serves a business goal rather than the user's informed preference. The US Federal Trade Commission has documented deceptive interface practices including confusing privacy controls, buried terms and cancellation obstacles. Not every nudge is a dark pattern; context, transparency, power and reversibility matter. [FTC report ↗]
Here's the strange thing about a default: it can feel invisible precisely because it works. The way to inspect it is to reverse the question: What would I have to do to say no? Count the steps, the confusion and the emotional pressure. A fair interface makes refusal legible.
The choice architect
Three short fictional interface decisions. Choose the design you'd rather put in the world.
Without playing: A humane opt-in is clear, specific and reversible. Refusal need not be a puzzle.
MECHANISM / DELIBERATE AGENCY
The quiet
room.
The most radical button in a casino is the one that takes you outside.
Focus is not an obligation to be impressive. It is the freedom to give something a little undivided attention: a person, a paragraph, the texture of an ordinary afternoon. Sometimes the attentive choice is to work. Sometimes it is to rest. Sometimes it is to enjoy the ridiculous internet meme and close the tab when the joke is finished.
Behavioural design reminds us that actions depend on more than wishes alone: capability, motivation and prompts matter. That means attention can be shaped by environments, routines and product decisions—not just blamed on personalities. [Fogg model ↗]
This room does not offer an attention score or promise transformation. The timer is merely a piece of furniture. Try thirty quiet seconds, or decline and walk out. The point is that you decided.
And when the world outside remains noisy, that is not a verdict on your virtue. It may be an invitation to redesign a small corner of the world, ask for better tools—or admit that being alive involves more than a tidy mind.
Thirty seconds belong to you
No alerts. No points. No productivity sermon. A simple, stoppable interval.
Without playing: You have already reached the exit. Nothing additional is required.
The attention
ledger.
The only account in this casino is a temporary description of what you did on this page. It is not a measurement of cognitive ability, self-control, addiction or psychological health.
Your visit, not your identity.
These counters exist only in JavaScript memory while this page is open. No analytics, cookies, localStorage, backend or network submission. Refreshing clears everything. Active game time is an approximate timer total, not a measure of concentration.
- No game events yet. Reading is a complete way to visit.
“The house cannot keep what you choose to take back.” — A fictional casino manager, finally making a decent point.BACK TO THE FLOOR ↑
Read past
the neon.
The architecture is satire; the academic claims should stand without it. Below are reputable external readings, with short notes about what they do—and do not—establish. These links were checked in October 2026.
Design trade-offs: refinding, landmarks and contexts where scrolling helps or harms usability. Not evidence that every scroll is manipulative.
OPEN ARTICLE ↗ 02 / REWARD LEARNINGNature Communications — Social Media EngagementEmpirical and computational support for reward-related patterns in social posting. Doesn't explain every individual's motives.
OPEN STUDY ↗ 03 / SOCIAL PERCEPTIONNature Human Behaviour — Overperception of OutrageStudies showing how audiences may overread outrage in posts and then misjudge broader attitudes. Not a universal claim about all anger.
OPEN STUDY ↗ 04 / ATTENTIONAL COSTStothart et al. — Cell Phone NotificationsAn experimental study of task disruption from alerts even without phone interaction. A bounded laboratory finding.
OPEN ABSTRACT ↗ 05 / SOCIAL FEEDBACKReward-Learning Account — Open Full TextOpen access version of reward-learning research; relevant to approval signals, not a measure of a post's inherent merit.
READ FULL TEXT ↗ 06 / CONSUMER DESIGNUS FTC — Bringing Dark Patterns to LightExamples of misleading choices, privacy settings and cancellation hurdles. A regulatory report, not a theory of all defaults.
OPEN REPORT SUMMARY ↗ 07 / BEHAVIOUR DESIGNStanford — Fogg Behavior ModelA framework linking motivation, ability and prompts. A conceptual tool rather than a personal diagnostic instrument.
OPEN MODEL ↗ 08 / FURTHER READINGCenter for Humane Technology — The Attention EconomyAn advocacy-oriented educational perspective on platform incentives; useful for the systemic lens, distinct from primary research.
READ ISSUE GUIDE ↗The best trick
is knowing you
can walk away.
Go back to your people, your art, your curiosity, or your gloriously unoptimized cup of tea. The floor will not send anyone after you.