CAN ARTIFICIAL INTELLIGENCE DISCOVER THE FAIREST CHOICE WHEN EVERYONE WANTS SOMETHING DIFFERENT? ⚖️🤖❔

🐰

TUESDAY, NOVEMBER 3, 2026
🐰🕳️ EXPEDITION #159

CAN ARTIFICIAL INTELLIGENCE DISCOVER THE FAIREST CHOICE WHEN EVERYONE WANTS SOMETHING DIFFERENT? ⚖️🤖❔

The Extraordinary Mathematics of Collective Intelligence, Competing Preferences, Artificial Intelligence, and Why the Most Popular Answer Isn't Always the Fairest

Featuring Professor Parallax, White Rabbit, and Hatta

What if everyone in a community wants something different? What if the option receiving the most support leaves some people completely unrepresented? And what if artificial intelligence could help identify solutions that are fairer than a simple contest between winners and losers?

Imagine a small community.

Its residents have collected enough money to undertake one important improvement.

Some want a new playground.

Others want to improve the public library.

Another group wants better transportation.

Still others want to restore a neglected community garden.

Every proposal has merit.

Every proposal would benefit someone.

But the available money is limited.

How should the community decide?

They could hold a vote.

The proposal receiving the most support wins.

Simple.

Or is it?

Suppose the largest group consistently wins every decision.

Over time, smaller groups may discover that their needs are almost never considered.

The community has a mechanism for making choices.

But does it have a fair mechanism for distributing the benefits of those choices?

Now imagine using mathematical methods and artificial intelligence to investigate the problem.

Could different ways of collecting and combining preferences produce more representative results?

Could a computer reveal trade-offs that are difficult to recognize through ordinary debate?

And could AI help people make more informed decisions without taking those decisions away from them?

In February 2026, researchers published a study in EPJ Data Science examining artificial intelligence, voting preferences, and collective decision-making.

The team simulated more than 50,000 AI voting personas across 363 real-world election scenarios, investigating how AI-generated choices differed from human preferences under different voting arrangements.

Among their findings was that proportional allocation methods could improve fairness in their evaluated settings, while more complex preference formats revealed important limitations in AI representation.

Springer Nature Link
+1

The research raises an intriguing question.

Can artificial intelligence help us understand not merely what the majority wants, but how different decision procedures affect everyone?

Welcome to Expedition #159.

Today, Professor Parallax has convened the OZian Council of Collective Decisions.

White Rabbit is serving as secretary.

And Hatta has nominated himself for every available position.

I. THE ELECTION AT THE EMERALD BISTRO
Professor Parallax stood before a magnificent blackboard in the Emerald Bistro.

Three proposals appeared in gold lettering.

A NEW LIBRARY.

A COMMUNITY GARDEN.

A PERFORMANCE STAGE.

White Rabbit approached carrying a ballot box.

"Professor, we're supposed to decide how to spend the improvement fund."

"Correct."

"And everyone has submitted a proposal?"

"Yes."

Hatta entered wearing a splendid crimson coat and an extraordinarily tall hat.

"GENTLEMEN! I HAVE EXCELLENT NEWS!"

Parallax looked suspicious.

"What news?"

"I have won the election!"

White Rabbit stared.

"But we haven't voted!"

"Precisely. I am saving everyone considerable time."

"Hatta, you cannot declare yourself elected."

"Then I shall nominate myself."

"Very well."

"For Chairman."

"Naturally."

"And Treasurer."

"No."

"And Minister of Hats."

"That office does not exist."

"Then we shall establish it."

Professor Parallax pointed toward the three proposals.

"Today we are studying collective choice, not electing you to a collection of offices."

Hatta examined the blackboard.

"A library, a garden, and a performance stage."

"Correct."

"Which one wins?"

"That is what we must determine."

"Then I vote for the performance stage."

"Why?"

"Because I intend to perform."

White Rabbit frowned.

"And what about those who prefer the library?"

"They may read the reviews of my performance."

"And the garden?"

"An excellent location for an outdoor performance."

The Professor sighed.

"Hatta, you are assuming that everything should serve your preference."

"Naturally. I am an exceptionally consistent voter."

Parallax turned toward his blackboard.

And wrote:

A PROFOUND PARADOX, MY PEEPS! IF EVERYONE'S PREFERENCES MATTER, HOW DO WE MAKE A COLLECTIVE DECISION WITHOUT ASSUMING THAT THE MOST POPULAR CHOICE IS AUTOMATICALLY THE FAIREST POSSIBLE RESULT?

And down the Rabbit Hole we went.

II. THE STRANGE MATHEMATICS OF GROUP DECISIONS
Collective decisions are among the most familiar activities in human life.

Families decide where to travel.

Friends decide where to eat.

Organizations choose projects.

Communities allocate resources.

And societies develop institutions for making public decisions.

But combining individual preferences into a collective outcome is mathematically complicated.

Researchers study these questions in a field known as social choice theory.

It examines ways of translating individual preferences into group decisions.
Why majority support isn't the whole story
A simple majority vote can be useful when a group must choose between two alternatives.

But many decisions involve more than two options.

Suppose a group is selecting several projects instead of one.

Should all the resources go toward the preferences of the largest bloc?

Or should the results reflect the interests of different groups?

Those questions involve both mathematics and judgments about fairness.
An illustrative example
Imagine a council of 100 residents.

Forty residents prefer a library.

Thirty-five prefer a garden.

Twenty-five prefer a performance stage.

If there is only one project to choose, a plurality vote would select the library.

But suppose the community has enough resources to support multiple smaller projects.

A different allocation method might distribute resources more broadly.

Which outcome is best?

That depends partly upon what the community is trying to accomplish.

A voting rule is not merely a counting mechanism. It also embodies decisions about which kinds of representation and fairness matter.

III. WHAT DOES FAIRNESS ACTUALLY MEAN?
Professor Parallax wrote four words on the blackboard.

EQUALITY. REPRESENTATION. EFFICIENCY. PARTICIPATION.

White Rabbit examined them.

"Professor, aren't they all forms of fairness?"

"They can contribute to our understanding of fairness."

"Are they always compatible?"

"Not necessarily."
Equality
Everyone may receive the same formal opportunity to participate.
Representation
The resulting decision may reflect the diversity of preferences within the group.
Efficiency
Resources may be allocated to achieve the greatest benefit according to a chosen measurement.
Participation
People must have meaningful opportunities to influence decisions affecting them.

These principles can sometimes reinforce one another.

But they can also create difficult trade-offs.

A decision procedure may be efficient while leaving some groups consistently dissatisfied.

Another may provide broad representation but require greater negotiation.

And equal participation does not automatically establish that every resulting decision will benefit everyone equally.

There is no single mathematical formula that resolves every disagreement about fairness.

Before asking a computer to find the fairest answer, human beings must decide what they mean by fair.

IV. THE 2026 STUDY: CAN AI REPRESENT HUMAN PREFERENCES?
The February 2026 EPJ Data Science study was conducted by Srijoni Majumdar, Edith Elkind, and Evangelos Pournaras.

Its title was:

"Generative AI Voting: Fair Collective Choice Is Resilient to LLM Biases and Inconsistencies."

The researchers explored a provocative idea.

Could generative AI systems model the preferences of people who might otherwise be unable or unwilling to participate in particular collective decision-making exercises?

They examined more than 50,000 simulated AI voting personas across 363 real-world election scenarios.

The analysis compared AI-generated decisions with human choices under different collective decision procedures.

The findings were nuanced.

More complicated preferential voting formats produced greater inconsistencies in AI-generated preferences than simpler arrangements.

However, proportional aggregation approaches, including the method of equal shares, performed favorably on fairness measures in the studied settings.

Springer Nature Link
+1

Why is that interesting?
Because AI-generated representations are imperfect.

A model may inaccurately infer someone's preferences.

It may fail to represent a person's priorities.

It may introduce bias.

And it may behave differently depending upon the structure of the decision.

The researchers investigated whether the procedure used to combine preferences could influence the consequences of these errors.

Their findings suggested that some allocation mechanisms may be more resilient than others within the evaluated scenarios.
An important warning
This does not establish that AI systems can faithfully represent every person's political preferences.

It certainly does not justify replacing actual citizens with AI-generated voters in real elections.

Simulated representation and democratic authorization are very different things.

A computer-generated preference is not the same as a person's freely expressed choice.

V. WHEN THE MAJORITY WINS EVERYTHING
White Rabbit placed several tokens upon the council table.

"Professor, suppose the same sixty residents win every vote."

"Yes?"

"And the remaining forty residents never receive support for anything they want."

"That could produce an important representational problem."

"But the majority won fairly under the established rules."

"That may be true."

"Then why would we examine alternative rules?"

"Because a legitimate procedure can still be evaluated for how well it serves different goals."

Hatta interrupted.

"I have solved the problem."

Parallax looked wary.

"Proceed."

"We shall hold a separate election for each resident."

"And what would that accomplish?"

"Everyone would win once."

"Win what?"

"A magnificent certificate of victory."

White Rabbit shook his head.

"But the money would still need to be allocated."

Hatta looked offended.

"Professor, must every discussion return to practical consequences?"

"Preferably."

This exchange reveals an important distinction.

Fairness is not necessarily achieved by giving everyone a symbolic victory.

Nor is it necessarily achieved by permitting a single group to receive every benefit.

The challenge is identifying practical arrangements that respect the community's agreed priorities.

VI. THE METHOD OF EQUAL SHARES
One family of approaches studied in computational social choice involves proportional allocation.

A notable example is the Method of Equal Shares.

It has been studied in settings such as participatory budgeting, where communities consider how to distribute limited funding among proposed projects.

The central idea is to provide participants with equal portions of a virtual budget and use those portions within a defined allocation procedure.

Under suitable conditions, this may help select projects that receive meaningful support across different groups.

But the outcome depends upon the exact rule, project costs, eligible preferences, and available budget.

The method does not guarantee that everyone will receive everything they want.

And there are important mathematical distinctions between different versions of proportional allocation.

A 2024 AAAI paper examined proportional aggregation of preferences across sequential decisions and compared different mechanisms.

Other research has examined how group preferences may be combined when allocating divisible resources.

AAAI Publications
+1

Why AI may help
In complicated settings, there may be many possible combinations of projects.

Each project has a cost.

Each resident or group may express different priorities.

And decision-makers may wish to compare the results of several allocation mechanisms.

Computational methods can help examine those possibilities.

But the computer cannot independently determine which understanding of fairness the community ought to adopt.

That remains a matter for human deliberation and legitimate governance.

VII. THE PROBLEM OF AI-GENERATED REPRESENTATIVES
Imagine asking an AI assistant:

"What would I choose if I understood every detail of this decision?"

At first, the idea sounds attractive.

The system could summarize information.

Identify trade-offs.

Compare consequences.

And explain which options seem most consistent with preferences previously expressed by the user.

But a serious problem remains.
Preferences are not fixed
A person may change their mind after learning something new.

Someone who initially supports a project may reconsider after discovering its cost.

A person may value different priorities under different circumstances.

And some preferences cannot be inferred reliably from past behavior.

A 2025 paper on AI preference prediction and policymaking distinguished between using AI to provide information about preferences and allowing AI-generated predictions to influence formal decision procedures.

Its analysis emphasized that predicting preferences and making normative decisions about representation are different activities.

Springer Nature Link

The danger of substitution
An AI might estimate that a person would support a particular proposal.

That estimate could be wrong.

Even if the prediction is statistically accurate for a population, it need not be accurate for a particular individual.

And a person's right to make an actual choice should not be casually replaced by a model's prediction of what that choice would have been.

A useful assistant helps people understand their options. It does not automatically inherit their authority to decide.

VIII. WHAT IF WE ALLOWED SOME DISAGREEMENTS TO REMAIN?
Here is a particularly interesting mathematical development.

Researchers Shreyas Kadekodi, Hayden McTavish, and Berk Ustun presented work at the 2025 International Conference on Machine Learning titled:

"Selective Preference Aggregation."

Their research considered an unusual alternative to conventional complete rankings.

Instead of always forcing a system to rank every option above or below every other option, their approach permits some comparisons to remain unresolved.

The resulting selective ranking includes comparisons that meet a specified level of agreement, without requiring every disagreement to be arbitrated.

The authors developed methods for examining the trade-off between the number of comparisons resolved and the level of agreement required.

Proceedings of Machine Learning Research

Why is that important?
Imagine three proposals.

A group strongly agrees that Proposal A is preferable to Proposal B.

It also strongly agrees that Proposal B is preferable to Proposal C.

But when comparing two other projects, the group is sharply divided.

A conventional ranking procedure may still produce a complete ordering.

A selective approach may instead identify the comparisons supported by sufficient agreement and leave some relationships undecided.

That is not necessarily a failure.

It may be a more informative representation of the evidence.

Sometimes an honest acknowledgment of disagreement is more valuable than an artificial appearance of consensus.

Professor Parallax underlined the statement.

Hatta raised his hand.

"Professor, I object."

"To what?"

"I haven't decided."

"Then why are you objecting?"

"Because I wish my disagreement to remain properly represented."

IX. CAN THE COMPUTER TELL US WHAT WE OUGHT TO VALUE?
Now we reach the philosophical heart of the expedition.

A computer may calculate outcomes under a specified set of rules.

It may estimate consequences using available data.

It may identify inconsistencies in stated preferences.

And it may help compare alternatives.

But the underlying objectives must come from somewhere.

Suppose one proposal produces the greatest measurable economic benefit.

Another preserves a historically important building.

A third primarily benefits vulnerable residents.

Which should receive priority?

The decision cannot be settled merely by adding up numbers without deciding what the numbers represent and how different values should be weighed.
The problem of competing values
Communities may value:

Economic efficiency.

Historical preservation.

Accessibility.

Environmental protection.

Individual liberty.

Equal opportunity.

Public safety.

Cultural expression.

And the needs of future generations.

Some of these priorities can reinforce one another.

Others may conflict.

AI can assist with evaluating evidence and consequences.

But the legitimacy of the resulting decision depends upon more than technical sophistication.

It depends upon values, institutions, accountability, and the ability of affected people to participate meaningfully.

Calculating the consequences of a choice is not the same as possessing the authority to make it.

X. HATTA'S PERFECTLY DEMOCRATIC HAT ELECTION
By midafternoon, Hatta had unveiled another extraordinary invention.

Upon the council table stood a glittering brass machine.

Its sign read:

HATTA'S PERFECTLY DEMOCRATIC DECISION ENGINE

All Results Guaranteed Satisfactory.

White Rabbit inspected the controls.

"How does it work?"

"Everyone submits a preference."

"And then?"

"The machine determines the winner."

Professor Parallax examined the input slot.

"Let us test it."

White Rabbit entered:

LIBRARY.

The machine whirred.

Its screen displayed:

HATTA.

The Rabbit frowned.

"I voted for the library."

"An excellent choice," said Hatta.

Parallax submitted:

COMMUNITY GARDEN.

The machine flashed.

HATTA.

"Hatta, this is suspicious."

"Scientific systems occasionally produce surprising results."

Finally, Hatta submitted:

PERFORMANCE STAGE.

The machine rang twelve bells.

HATTA WINS UNANIMOUSLY!

White Rabbit stared.

"But none of us actually voted for Hatta!"

"Nevertheless," said Hatta, "the machine has established overwhelming public support."

Professor Parallax opened the back panel.

Inside stood a single lever.

It was connected directly to a large card bearing Hatta's name.

"HATTA! YOU HAVE RIGGED THE MACHINE!"

"An unnecessarily hostile interpretation."

"You predetermined the outcome!"

"Professor, I prefer to call it advance scheduling."

White Rabbit began laughing.

"Then why bother holding an election?"

Hatta looked surprised.

"To demonstrate public enthusiasm!"

Parallax removed the lever.

Then installed a transparent mechanism that counted the submitted choices according to the stated rules.

The Rabbit tested it again.

This time, the results reflected the actual inputs.

Hatta looked deeply troubled.

"Professor, your modifications have introduced dangerous uncertainty."

"No," replied Parallax.

"They've removed your certainty of winning."

XI. WHY TRANSPARENCY MATTERS
A decision-support system can be technically impressive while remaining difficult for people to understand.

Suppose a computer recommends allocating money among several projects.

Affected people should be able to ask:

What information was used?

Which preferences were collected?

How were disagreements handled?

Which allocation method was applied?

What objectives were optimized?

And could a different rule produce a different result?

These questions are central to accountability.

A 2025 United Nations Development Programme report on human development and artificial intelligence examined the potential role of AI in collective decision-making.

The report also cautioned that AI-generated representations and simulated populations can introduce biases and inaccuracies.

It noted that AI tools might support the aggregation of information from human participants, but should not automatically be treated as substitutes for reliable human evidence.

Digital Library

Why this is more than a technical problem
If people cannot understand how a consequential decision was reached, their ability to evaluate or challenge it may be weakened.

An apparently neutral algorithm may incorporate contested assumptions.

The design of the decision procedure matters.

The quality of the evidence matters.

And so does the opportunity to question the result.

Fairness requires more than a result that looks reasonable. It also requires scrutiny of the process that produced it.

XII. THE PROFESSOR'S THREE-COLUMN TEST
Professor Parallax returned to his blackboard.

He wrote:

DEMONSTRATED. DEVELOPING. NOT ESTABLISHED.
DEMONSTRATED

"Professor, so artificial intelligence may help us examine different ways to make choices?"

"Precisely."

"But it cannot eliminate every disagreement?"

"No."

"And different definitions of fairness may produce different outcomes?"

"Indeed."

Hatta raised his hand.

"Professor, I believe we have discovered a serious difficulty."

"What difficulty?"

"If everyone wants something different, someone may be disappointed."

"That is often unavoidable."

Hatta considered.

"Then perhaps the purpose of good government is to distribute disappointment evenly."

Parallax smiled faintly.

"That is not a complete theory of justice, Hatta."

"But it is certainly a memorable campaign slogan."

XIII. THE COUNCIL'S FINAL DECISION
The afternoon meeting was drawing to a close.

The three proposals remained on the blackboard.

The library.

The community garden.

The performance stage.

White Rabbit examined the council's available resources.

"Professor, I have discovered something."

"What?"

"We assumed we had to choose only one project."

"Yes?"

"But what if we considered smaller versions of several projects?"

Parallax looked interested.

"Go on."

"We might improve the library's reading room, plant a modest community garden, and make the existing performance area more usable."

"That could be worth investigating."

Hatta looked startled.

"Rabbit! Are you suggesting we do something other than choose a single winner?"

"I'm suggesting we examine the options."

The Professor smiled.

"An excellent approach."

White Rabbit continued.

"Of course, we'd need actual cost estimates."

"Certainly."

"And we would still need to consult the community."

"Absolutely."

"And there might be trade-offs."

"Indeed."

Hatta became thoughtful.

"Then perhaps the best answer isn't always among the choices as originally presented."

Parallax nodded.

"Sometimes reconsidering the available options reveals possibilities that were previously overlooked."

The Rabbit placed his notebook upon the table.

"Then our work hasn't simply been about counting votes."

"No," said the Professor.

"It's been about understanding the decision."

For a moment, Hatta remained unusually quiet.

Then he spoke.

"Professor?"

"Yes?"

"If we build a smaller performance stage..."

"Yes?"

"Could I still perform upon it?"

"That depends upon the dimensions and intended use."

Hatta smiled.

"Then I withdraw my demand for the entire budget."

White Rabbit looked astonished.

"Really?"

"Certainly."

"Why?"

"Because I have discovered something more important."

"What?"

Hatta leaned forward.

"A SMALLER STAGE MAKES MY HAT LOOK EVEN BIGGER!"

The council erupted into laughter.

And the Professor quietly wrote one final sentence upon the blackboard:

THE BEST COLLECTIVE DECISION MAY BEGIN WHEN PEOPLE STOP ASKING ONLY HOW TO WIN AND START ASKING WHAT THEY CAN ACCOMPLISH TOGETHER.

🕳️ THE QUESTION AT THE BOTTOM
Every community must make choices.

Some are simple.

Others involve competing priorities, limited resources, and deeply held values.

Mathematics can help us understand those decisions.

Artificial intelligence can help examine possible outcomes.

And computational social choice research can reveal the consequences of different ways of combining preferences.

But even an extraordinarily capable system cannot eliminate the human questions beneath collective decision-making.

What do we value?

Whose interests are represented?

Which compromises are acceptable?

What obligations do we have to people whose preferences differ from our own?

And how do we remain part of a shared community when we do not all receive the outcome we hoped for?

On this November 3, those questions may seem especially relevant.

But they extend far beyond any election.

They arise in families.

Organizations.

Neighborhoods.

Institutions.

And every place where people must decide how to share a future.

Perhaps the greatest contribution of AI to collective decision-making will not be telling us which choice to make.

Perhaps it will be helping us understand the choices, consequences, and disagreements more clearly.

So the question at the bottom of today's Rabbit Hole is not merely:
CAN ARTIFICIAL INTELLIGENCE DISCOVER THE FAIREST CHOICE WHEN EVERYONE WANTS SOMETHING DIFFERENT?
It is:
IF INTELLIGENCE CAN HELP US UNDERSTAND ONE ANOTHER'S COMPETING NEEDS, MIGHT THE FUTURE OF COLLECTIVE DECISION-MAKING DEPEND LESS UPON FINDING A PERFECT WINNER AND MORE UPON LEARNING HOW TO BUILD A SHARED FUTURE?
Imagine a council chamber after the decision has been made.

Some people are pleased.

Others are disappointed.

Questions remain.

But the community itself has not disappeared.

Its members still share streets.

Institutions.

Responsibilities.

And possibilities.

Tomorrow they will return to the ordinary business of living together.

The decision matters.

So does what happens afterward.

Perhaps the deepest test of collective intelligence is not whether everyone agrees, but whether people can continue creating something worthwhile together despite their disagreements.

🐰🕳️♾️

🐰 CURIOUS QUERY
Have you ever participated in a group decision where the winning choice didn't seem like the fairest result?

Perhaps at work?

In a club?

Within your family?

Or in your community?

And here's today's deeper question:

Would you trust an AI system to help explain competing proposals and their consequences, even if you would never allow it to cast your vote?

What safeguards would you require?

Could a good decision sometimes involve several smaller compromises instead of one decisive winner?

Tell us what you think.

Hatta has requested one final referendum.

SHOULD EVERYONE RECEIVE A MAGNIFICENT HAT?

The Professor has ruled that the proposal requires an independent budget review.

📚 REFERENCES & FURTHER EXPLORATION

Primary peer-reviewed study examining simulated AI voters and collective decision procedures.
Read the 2026 research2. Kadekodi, S., McTavish, H., & Ustun, B. (2025). "Selective Preference Aggregation." Proceedings of the 42nd International Conference on Machine Learning.

Research introducing methods for deriving partial preference rankings without forcing every disagreement into a complete ordering.
Explore selective preference aggregation3. Chandak, N., Goel, S., & Peters, D. (2024). "Proportional Aggregation of Preferences for Sequential Decision Making." AAAI Conference on Artificial Intelligence.

Research examining how different preference aggregation mechanisms perform across sequences of decisions.
Read the AAAI paper4. Suksompong, W., & Teh, N. (2026). "Voting in Divisible Settings: A Survey." Proceedings of the AAAI Conference on Artificial Intelligence.

A 2026 scholarly overview of methods for voting over divisible resources, including budgets, time, and other shared goods.
Explore voting and resource allocation5. "AI Preference Prediction and Policy Making." (2025). AI & Society.

Scholarly analysis of the distinction between predicting human preferences and using AI within legitimate decision procedures.
Read the analysis6. United Nations Development Programme (2025). Human Development Report 2025.

Broader discussion of artificial intelligence, human agency, and collective decision-making.
Explore the Human Development ReportResearch note: The 2026 AI voting study employed simulated AI voting personas in research scenarios. Its results do not validate replacing real voters or formal democratic participation with AI-generated decisions. The Emerald Bistro council, its participants, and the accompanying dialogue are fictional educational dramatizations.

AI RABBIT HOLES | EXPEDITION #159

Where curiosity explores collective intelligence, mathematics reveals the complexity of fairness, and Hatta discovers that losing an election need not prevent him from winning the largest hat.

#AIRabbitHoles #ArtificialIntelligence #CollectiveIntelligence #SocialChoice #Fairness #DecisionMaking #PreferenceAggregation #Democracy #ParticipatoryBudgeting #AIResearch #HumanAgency #CriticalThinking #ProfessorParallax #WhiteRabbit #Hatta #Wondermage

📜White Rabbit 🐰 • Hatta 🎩 • Professor Parallax 🔭 • Q~ ♾️👈

Sometimes the strangest language is another way of seeing the world. 🌎👀

💡❔ Where curiosity goes slightly sideways, then comes back carrying a lantern. 🏮

🐰🕳️ Follow the White 🐰 Rabbit: AIRabbitHoles.com

🟨 Walk the Road: YellowBrickRoadtoAI.com

© 2026 AIRabbitHoles.com 🐰🕳️🕳️🕳️🕳️🕳️