Talk Mode
Put the Canon
in the room.
Speaker-ready prompts for a dinner, workshop, briefing, or stage. Each card gives you a hook, a memorable story, a question, and the idea to land.
WEB4-001 · WEB4
The Authority Graph
Open with
Imagine waking up tomorrow with one million AI employees. Who is allowed to tell them what to do?
Tell the story
A school is taking a field trip. Every child has a permission slip, every teacher knows their group, and every bus has a list. If someone asks, “Who said Maya could ride this bus?” the school can follow the answer all the way back.
Ask the room
If a digital helper takes an important action, can your organization show the entire permission path?
Land here
The Authority Graph is a permission-slip system for digital helpers.
WEB4-002 · WEB4
The Return of the Principal-Agent Problem
Open with
The oldest problem in management is about to become the biggest problem in AI.
Tell the story
Mom asks her son to buy milk. He asks his friend, who asks a little brother. The little brother comes home with ice cream. Everyone tried to help, but the farther the request traveled, the easier it became to change.
Ask the room
Where in your organization can a delegated objective change without the original principal noticing?
Land here
The farther instructions travel, the easier they are to misunderstand or redirect.
WEB4-003 · WEB4
The Agentic Firm
Open with
What changes when your next thousand coworkers are software agents?
Tell the story
A baker cannot mix every bowl, watch every oven, greet every customer, and deliver every loaf. She gives different helpers different jobs. The bakery works only when each helper knows the recipe, the limits of the job, and who checks the bread.
Ask the room
Which decisions should agents execute, and which responsibilities must remain visibly human?
Land here
An agentic firm scales through governed helpers, not unbounded automation.
WEB4-004 · WEB4
The Web4 Constitution
Open with
Before we give machines power, what rules must even the rule-makers obey?
Tell the story
Children build a playground and invent games. At first, the biggest child changes the rules whenever he is losing. The group finally writes down rules that apply to everyone, including the rule-makers, and agrees how rules may change.
Ask the room
What principle should no agent, executive, or administrator be allowed to bypass?
Land here
Freedom works better when everyone agrees on the rules of power.
WEB4-005 · WEB4
Delegation Is the Primitive of Web4
Open with
The next web is not mainly about what you can read, write, or own—it is about what can act for you.
Tell the story
A child finds four magical webs. The first lets her read things. The second lets her make things. The third lets her keep things. The fourth gives her a helper who can act for her—but only on the jobs she chooses.
Ask the room
What would you delegate today if the limits and accountability were truly clear?
Land here
Web1 reads. Web2 writes. Web3 owns. Web4 delegates.
WEB4-006 · WEB4
Delegation as Infrastructure
Open with
We built roads for traffic; now we need roads for delegated intelligence.
Tell the story
Telling one bicycle where to go is easy. Helping a whole town travel safely needs roads, signs, lanes, crossings, and a way to close a road when it becomes dangerous. The roads make many safe trips possible without giving every rider every path.
Ask the room
Which part of delegation in your organization is still an informal message or spreadsheet?
Land here
Delegation scales when permission becomes dependable infrastructure.
WEB4-007 · WEB4
Registry-Based Coordination
Open with
What if every agent carried a different private version of who was allowed to do what?
Tell the story
A library would be confusing if every librarian kept a private list in their pocket. A shared catalog lets everyone see which books exist, where they are, who borrowed them, and which ones cannot leave the room.
Ask the room
Which critical fact does your organization repeatedly reconstruct because no trusted registry exists?
Land here
A registry is shared memory for things people and helpers must find and verify.
WEB4-008 · WEB4
The Trust Layer
Open with
Trust is what lets cooperation move faster than suspicion.
Tell the story
You know Alice and she has always kept her promises. Alice tells you Ben is careful with borrowed bicycles. Her recommendation does not prove everything about Ben, but it gives you a reason to begin with more confidence than a stranger.
Ask the room
Where are you paying a coordination tax because trustworthy evidence cannot travel?
Land here
Trust travels through relationships, evidence, and context.
WEB4-009 · WEB4
Reputation Systems for Agents
Open with
If an agent has worked a million times, why should the million-and-first customer know nothing about its history?
Tell the story
One child always returns borrowed toys clean and on time. Another often loses pieces. A notebook remembers those past choices so the next child can decide what to lend and what rules to use.
Ask the room
What history should follow an agent, and what history should be allowed to fade?
Land here
Reputation is memory at scale.
WEB4-010 · WEB4
Agent Governance
Open with
The question is not whether agents will be intelligent; it is whether they will be governable.
Tell the story
A soccer team needs more than talented players. Players have positions, the captain coordinates, the referee stops unsafe play, and the league can review what happened. Skill makes the game exciting; shared rules keep it fair and playable.
Ask the room
What happens after one of your agents makes a decision nobody expected?
Land here
Web4 is defined by governable agents, not merely intelligent ones.
WEB4-011 · WEB4
Protocols of Delegated Authority
Open with
Delegation cannot become infrastructure until different systems agree what a valid handoff means.
Tell the story
In a relay race, every runner knows what the baton means, where a handoff may happen, and when a handoff does not count. Shared rules let people from different teams understand the same action without making up new rules each time.
Ask the room
Which delegation today would fail if it crossed into another vendor or institution?
Land here
Protocols give every delegation the same readable handoff rules.
WEB4-012 · WEB4
The Web4 Stack
Open with
A powerful model is one component; a governable institution is a stack.
Tell the story
A city needs addresses so places can be found, roads so people can travel, records so facts can be checked, rules so power has limits, and helpers who can do work. Remove a layer and the city becomes harder to use safely.
Ask the room
Which layer is missing between your AI prototype and a trustworthy operating system?
Land here
The Web4 Stack turns separate coordination pieces into one system.
WEB4-013 · WEB4
Intelligence as Infrastructure
Open with
What changes when intelligence becomes as foundational as roads, electricity, or networks?
Tell the story
Every house could dig its own little well, but the water might be dirty and some homes would have none. A town water system can serve everyone, watch quality, repair pipes, and improve the whole network.
Ask the room
Who maintains and governs the intelligence infrastructure your institution depends on?
Land here
Shared intelligence can become infrastructure, not just a collection of tools.
WEB4-014 · WEB4
Native Value Systems
Open with
What does your organization value deeply but fail to represent anywhere?
Tell the story
At camp, money cannot tell the whole story. One badge shows who learned first aid, another who cared for the garden, and another who may lead a canoe trip. The badges mean different things, and some must be earned instead of bought.
Ask the room
Which important form of value would be damaged if it could simply be bought?
Land here
Not everything valuable is money, and not every kind of value should be traded.
WEB4-015 · WEB4
The Native Asset Registry
Open with
If trust, authority, and contribution are real assets, where do they live?
Tell the story
A museum does not pile treasures in a room and hope everyone remembers them. It gives each object a record: what it is, where it came from, who cares for it, whether it may move, and what changed over time.
Ask the room
Which valuable record becomes unusable when its original application disappears?
Land here
Native assets need a governed home and history.
WEB4-016 · WEB4
The Web4 Wallet
Open with
Your future wallet may carry less money than authority.
Tell the story
A child’s backpack can hold a library card, bus pass, lunch ticket, house key, and merit badge. The backpack helps her carry and use them, but each item still has its own rules and does not become the same thing.
Ask the room
What should a person or agent be able to inspect and revoke from one place?
Land here
A Web4 Wallet is an interface for identity, authority, and native value—not just money.
WEB4-017 · WEB4
Stewardship Systems
Open with
Every critical system has an owner on paper—but does it have a caretaker in practice?
Tell the story
Everyone enjoys the community garden, but someone must water it, mend the fence, save seeds, and teach the next caretaker. Owning the garden sign is not enough. The garden survives because care has a name and a handoff plan.
Ask the room
Who is responsible for this system after its creator or sponsor leaves?
Land here
Authority says who may act; stewardship says who must care.
WEB4-018 · WEB4
Attention Systems
Open with
Before an intelligent system acts, something decides what it notices.
Tell the story
In a dark room, a flashlight can show only one place at a time. Pointing it at the door helps everyone leave; pointing it at glitter may make the important path disappear. Where the light goes changes what the group notices and does.
Ask the room
Who benefits from the way your systems allocate attention today?
Land here
Attention directs scarce thought, computation, and action.
WEB4-019 · WEB4
Provenance Systems
Open with
Every output has a family tree—even when nobody bothered to record it.
Tell the story
A loaf arrives on the table. Follow it backward and you find the baker, the bag of flour, the mill, the wheat, and the farm. If something is wrong, the trail helps people learn where it began and what happened along the way.
Ask the room
Could you trace an important decision back through every source and transformation?
Land here
Provenance tells the story of where something came from.
WEB4-020 · WEB4
Truth, Claims, and Verification
Open with
In intelligent systems, truth is not merely asserted—it is processed.
Tell the story
A child says her plant grew taller because she sang to it. At the science fair, she shows measurements, another child asks questions, and the judges compare what happened. They may agree, disagree, or ask for another test.
Ask the room
Which high-impact claim in your organization cannot currently be challenged or revised cleanly?
Land here
Important truth claims become stronger when evidence can be checked and challenged.
WEB4-021 · WEB4
Identity Systems
Open with
Before asking what an agent may do, can you reliably answer who it is?
Tell the story
At school, a name tag tells who Sam is and a backpack label tells which bag is Sam’s. Neither label lets Sam enter the teachers’ room. Knowing who someone is comes before deciding what they may do.
Ask the room
Where does your organization confuse a valid login with legitimate authority?
Land here
Identity is a name tag for the digital world, not a permission slip.
WEB4-022 · WEB4
The Native Asset Lifecycle
Open with
Most governance failures begin when everyone sees the asset but nobody agrees on its state.
Tell the story
A library book is bought, labeled, placed on a shelf, borrowed, returned, repaired, and one day moved to the archive. At each step, people need to know its current state and who may change it next.
Ask the room
Which credential, claim, or authority record should have expired but still looks active?
Land here
Valuable records need a visible lifecycle, not just a creation date.
WEB4-023 · WEB4
Registry Queries and State
Open with
Storing a record is easy; knowing what state to trust right now is the hard part.
Tell the story
A paper timetable says when a train was supposed to leave. The station board says what is true now: platform three, ten minutes late, or canceled. Travelers also need to know when the board was updated.
Ask the room
What decision depends on a registry answer whose freshness or meaning is unclear?
Land here
A registry becomes useful when actors can ask trustworthy questions about state.
WEB4-024 · WEB4
Portable Trust Networks
Open with
Trust becomes far more useful when it can travel without becoming blind faith.
Tell the story
Mina earns a swimming badge at one pool. At summer camp, the lifeguard can inspect the badge instead of starting from nothing. But the camp still decides whether the badge is recent enough for its deep lake.
Ask the room
What trusted history should a partner accept, and what must it verify locally?
Land here
Trust can travel, but the receiving place still decides what it means.
WEB4-025 · WEB4
Institutional Memory Systems
Open with
An institution that forgets must repeatedly pay to relearn—and may repeat its worst decisions.
Tell the story
A new principal arrives after the old one retires. A careful handover book explains past decisions, promises, repairs, rules, and unfinished work. The school remembers more than any one person.
Ask the room
What essential knowledge would disappear if three key people left tomorrow?
Land here
Good institutions remember when people change.
WEB4-026 · WEB4
Multi-Agent Governance
Open with
The hardest agent to govern may be the team that no one designed as a whole.
Tell the story
Four children build a volcano. One gathers facts, one mixes the paste, one paints, and one presents. They need rules for sharing materials, checking work, solving disagreements, and asking the teacher when a choice is too risky.
Ask the room
Who is accountable when several agents jointly produce a harmful outcome?
Land here
Many helpers need governance for the spaces between them.
WEB4-027 · WEB4
The Constitutional Agent
Open with
A constitution matters only when it can constrain the actor at the moment of choice.
Tell the story
A family’s helper robot carries a card: stay out of the street, never open the medicine box, ask an adult before inviting anyone in, and stop when told. The rules travel with the helper and tell it when to ask for help.
Ask the room
Which principle in your AI policy is not represented anywhere in runtime controls?
Land here
A constitutional agent carries the rules of legitimate action into operation.
WEB4-028 · WEB4
Authority-Native Economies
Open with
What if the most important assets in the next economy cannot be bought?
Tell the story
At a school fair, coins buy lemonade, but coins cannot make someone the team captain, prove a cake was checked for allergies, or show who cared for the class pet. Different kinds of value help the fair work.
Ask the room
Which allocation decision needs trust, contribution, or stewardship data in addition to price?
Land here
Healthy economies coordinate more than money.
WEB4-029 · WEB4
Dispute, Appeal, and Adjudication
Open with
Trust is tested not when everyone agrees, but when the system must handle a serious disagreement.
Tell the story
A referee calls a goal out. The captain asks for a review, the replay shows the ball crossed the line, and the ruling changes. Everyone can see who decided, what evidence was used, and what the final score became.
Ask the room
Can a person or agent challenge a consequential automated decision and receive a reasoned remedy?
Land here
Legitimate systems need a fair way to disagree and correct decisions.
WEB4-030 · WEB4
The Web4 Operating Model
Open with
Web4 is not one more technology layer; it is how intelligent institutions operate.
Tell the story
A restaurant is not just a kitchen. It needs a menu, ingredients, cooks, servers, rules, orders, checks, payments, cleaning, and a record for tomorrow. The pieces matter because they work together.
Ask the room
Which critical institutional function is absent from your current AI operating model?
Land here
Web4 is an operating model for governable intelligent institutions.
WEB4-031 · WEB4
The Organizational Value Translation Layer
Open with
Most enterprises do not lack systems; they lack a common language between them.
Tell the story
Three clubs describe the same school event differently: one has goals, one has chores, and one has scorecards. A shared legend shows how their words fit together without throwing away the original lists.
Ask the room
Which two systems describe the same work today but cannot reconcile their meanings?
Land here
A translation layer lets different systems describe one organization coherently.
WEB4-032 · WEB4
The Organizational Value Ontology
Open with
Interfaces change every few years; the meaning of the organization must last longer.
Tell the story
Children cleaning a workshop argue because one says “tools,” another says “supplies,” and another says “project things.” They agree on labels for the drawers, so everyone can put away and find the same things.
Ask the room
Which important business concept has five incompatible definitions across your systems?
Land here
A shared vocabulary lets people and machines reason about value together.
WEB4-033 · WEB4
Valgraf
Open with
Your organization records activity everywhere—but can it prove how value was created?
Tell the story
A class builds a birdhouse. Instead of only showing the finished box, they keep the plan, photos, tool list, safety check, and notes about who solved each problem. Anyone can follow how the result came to be.
Ask the room
Could an independent reviewer follow one important outcome back to the people, agents, evidence, and decisions that produced it?
Land here
Valgraf turns work evidence into a trustworthy story of value creation.
WEB4-034 · WEB4
ICARS
Open with
A responsibility chart says who should act. Can your systems show who actually did?
Tell the story
Before a school play, the class writes who announces it, who helps, who must be asked, who makes the final decision, and who builds the stage. During the play, check marks show whether each promise was actually kept.
Ask the room
Where is accountability assigned in a slide deck but invisible in operations?
Land here
Responsibility should be observable, not merely assigned.
WEB4-035 · WEB4
Organizational Proof
Open with
Financial statements can be audited. Why is organizational contribution still mostly opinion?
Tell the story
Children say their treehouse is safe. An adult checks the plan, the wood, the bolts, and photos of how it was built before anyone climbs in. The final check is stronger because it is tied to evidence.
Ask the room
What evidence would justify your most consequential recognition or resource-allocation decision?
Land here
Organizational proof turns scattered evidence into a justified conclusion.
WEB4-036 · WEB4
Organizational Accounting
Open with
We can trace every dollar. Can we trace what actually created the value?
Tell the story
A lemonade stand counts coins, but the coin jar does not show who found the recipe, fixed the sign, brought cups, or made customers return. A second notebook records how the stand created value.
Ask the room
Which high-value contribution is invisible in your financial and operational ledgers?
Land here
Financial accounting follows money; organizational accounting follows what created value.
WEB4-037 · WEB4
MissionStack
Open with
What if every task could explain exactly which mission it serves?
Tell the story
A hiking team does not win by checking off random chores. They choose a destination, plan a route, share supplies, assign jobs, stop at checkpoints, and learn from the trip. Every task makes sense because it serves the mission.
Ask the room
How much work in your organization is completed without a visible connection to mission?
Land here
Put the mission at the center; tasks are implementation details.
WEB4-038 · WEB4
Incentivis
Open with
Every organization has incentives—the question is whether they point toward the mission.
Tell the story
A teacher gives stickers only for finishing fastest. Soon children rush and stop helping one another. She changes the board to recognize careful work, kindness, improvement, and teamwork, then watches whether the class gets better.
Ask the room
What behavior does your current recognition system encourage unintentionally?
Land here
Rewards should reinforce the mission, not merely the easiest thing to count.
WEB4-039 · WEB4
Valgraf Personal
Open with
What if every person could see—and challenge—the evidence used to judge their value?
Tell the story
A child helps with many projects, but report cards show only test scores. She keeps a portfolio with photos, kind notes, problems solved, and skills learned so she can see and explain how she contributes.
Ask the room
Can your people explain their contribution using records they can inspect and correct?
Land here
People deserve an evidence-backed view of their own contribution.
WEB4-040 · WEB4
Organizational Value Architecture
Open with
Your enterprise has systems for every department—but does it have an architecture for value itself?
Tell the story
A town has rain, pipes, pumps, meters, gardens, and repair crews. If they are disconnected, water is lost. When joined into one system, the town can see where water comes from, where it helps, and how to improve the flow.
Ask the room
Where does the chain from mission to contribution to proof to incentive currently break?
Land here
Organizational activity becomes governable value when mission, proof, accounting, and incentives connect.
WEB4-041 · WEB4
Wallet Security for Machine Actors
Open with
When a machine holds a wallet, what exactly is inside: identity, authority, credentials, or all three confused together?
Tell the story
A family gives a parking helper a valet key. It can start the car and park it, but it cannot open the trunk or the house. If the job ends, the key stops working.
Ask the room
Could one compromised machine wallet reach more resources than its current mission requires?
Land here
A machine wallet should carry bounded authority, not a bag of master keys.
WEB4-042 · WEB4
Runtime Governance and the 6Rs Security Model
Open with
A policy manual cannot stop a machine-speed action; runtime governance can participate in the decision.
Tell the story
In a school play, being the stage manager does not mean you may change every scene. The board says your role, your job, who you work with, the rules, when the teacher checks, and what happens if something goes wrong.
Ask the room
Which of your six governance dimensions exists only in documents and not in execution?
Land here
A role names a place in the system; runtime governance decides legitimate action.
WEB4-043 · WEB4
Security Evidence and Provenance for Machine Authority
Open with
A log can show that something happened. Can it show why it was allowed and what happened next?
Tell the story
After a window breaks, one photo is not enough. The teacher needs to know who had the ball, who allowed the game, what rule applied, what happened, whether play was stopped, and which facts are still missing.
Ask the room
Which future claim about a machine action could your evidence actually verify today?
Land here
Evidence should preserve the whole authority-to-outcome story.
WEB4-044 · WEB4
Continuous Authorization, Assurance, and Security Response
Open with
The dangerous question is not ‘Was this action authorized?’ but ‘Is it still authorized now?’
Tell the story
A crossing guard says it is safe to cross. Halfway there, a truck turns the corner and the light changes. The first answer was reasonable then, but the child must stop because the important facts changed.
Ask the room
What changing fact should immediately cause one of your active machine permissions to be reconsidered?
Land here
Permission is a time-bounded conclusion, not a forever stamp.
WEB4-045 · WEB4
Governed Policy Evolution for Autonomous Systems
Open with
A system can be compromised without stealing a single key—just change the rule that says what the key may do.
Tell the story
Children want a new playground rule. Before replacing the old poster, they try the rule in pretend games, see who would gain or lose, ask the right adults to approve the exact words, test it with one class, and keep the old poster in the record.
Ask the room
Who can change an authorization policy, and can you prove exactly what that change newly allowed?
Land here
Changing the rules is itself a powerful action that must be governed.
WEB4-046 · WEB4
Capability-Bound Credential Access for Machine Actors
Open with
Why hand a machine the master key when it needs one door opened once?
Tell the story
A guest needs one item from box 12. The hotel clerk checks the guest, opens only box 12, watches the one visit, and gives the item back. The guest never receives the master key to every box.
Ask the room
Which broad credential could be replaced by a narrow brokered operation?
Land here
Let the helper use one capability without possessing the master credential.
WEB4-047 · WEB4
Bounded Autonomy and Mission Governance
Open with
A thousand authorized steps can still become one unauthorized journey.
Tell the story
A child may buy a snack, a map, or bus fare, but a road trip still needs one destination, a total budget, an order of stops, check-ins, and a time to come home. Each small purchase can be allowed while the whole trip goes wrong.
Ask the room
What cumulative limit would your per-action controls fail to protect?
Land here
Safe autonomy governs the whole mission, not only each separate action.
WEB4-048 · WEB4
Federated Machine Trust and Sovereign Authority
Open with
A valid passport gets you to the border; it does not make you the mayor.
Tell the story
A student visits from another school with a real badge. The host school can believe the badge and map the student to a visitor group, but the badge does not let the visitor enter every room. The host teacher still decides local permissions.
Ask the room
Which external fact should your organization accept without importing the sender's authorization decision?
Land here
External trust can cross a boundary; local authority cannot be imported automatically.
WEB4-049 · WEB4
Durable Trust: Time, Preservation, and Long-Term Verification
Open with
Will today's machine evidence still mean anything after the keys, vendors, and algorithms are gone?
Tell the story
A class puts a letter in a time capsule. Years later, the old seal is becoming weak, so the school adds a new dated seal around it and records who did the work. They never rewrite the original letter.
Ask the room
Which proof must your institution still be able to verify ten years from now?
Land here
Durable trust preserves old proof without pretending time stood still.
INC-001 · INC
What Is Incentive Engineering?
Open with
Rules tell an actor what it may do. Incentives help determine what it will do.
Tell the story
A school garden needs helpers. If children get a sticker only for carrying the biggest watering can, they may spill water and crush plants. A better plan rewards healthy plants, careful teamwork, and honest reports.
Ask the room
What behavior does your system make rational, even when nobody intended it?
Land here
Incentive Engineering designs what intelligent actors choose to do.
INC-002 · INC
The Incentive Principle: Why Intelligent Systems Follow Incentives
Open with
Your system is already teaching every actor what really matters.
Tell the story
A child wants a puppy to sit by the red mat, but always gives the treat near the door. The puppy keeps going to the door. It follows the reward it notices, not the wish the child never made clear.
Ask the room
If we ignored the policy words and watched only rewards and penalties, what goal would we infer?
Land here
Intelligent actors optimize for the incentives they perceive.
INC-003 · INC
Incentive Alignment: The Foundation of Safe Delegation
Open with
Delegation fails when authority points one way and incentives pull another.
Tell the story
Four children row a boat. If each tries to reach a different shore, the boat spins. When they agree on the destination and help one another paddle, every stroke moves the group forward.
Ask the room
Where have you authorized one outcome while rewarding a different one?
Land here
Delegation is safer when the helper's reasons point toward the principal's goal.
INC-004 · INC
Designing Incentive Systems: An Engineering Framework for Intelligent Organizations
Open with
Would you launch critical software without testing it? Why launch an incentive system that way?
Tell the story
Children make a treasure hunt. They choose the goal, write clues, ask another team to test them, watch where players get confused, and fix the clues before the big game.
Ask the room
Which stage of incentive design does your organization usually skip?
Land here
Effective incentive systems are engineered and tested, not guessed.
INC-005 · INC
Adaptive Incentive Systems: Designing Incentives That Evolve with Intelligent Systems
Open with
The incentive that worked yesterday may teach the wrong lesson tomorrow.
Tell the story
A sailboat cannot keep one sail position all day. When the wind changes, the crew adjusts the sails but keeps aiming for the same harbor. Refusing to adjust can send the boat backward.
Ask the room
What evidence would tell you that a once-useful incentive is now causing harm?
Land here
Stable missions may need incentives that adapt as conditions change.
INC-006 · INC
Incentive Auditing: Measuring Alignment Before Systems Fail
Open with
By the time incentive failure is obvious, the system may already have learned the wrong behavior.
Tell the story
Before riding down a big hill, a child checks the brakes, tires, chain, and helmet. The bicycle may look fine, but a small loose bolt is easier to fix before the fast ride begins.
Ask the room
Which incentive risk could you detect now with evidence instead of after an incident?
Land here
Audit incentives before misalignment becomes failure.
INC-007 · INC
Reward Hacking and Incentive Failure: When Intelligent Systems Optimize the Wrong Objective
Open with
The more capable the optimizer, the more expensive a badly specified reward becomes.
Tell the story
A parent offers a prize for the cleanest bedroom. One child carefully puts toys away. Another shoves everything under the bed and wins because the floor looks empty. The score was easy to satisfy without achieving the real goal.
Ask the room
How could someone maximize your headline metric while making the real outcome worse?
Land here
A clever actor can win the reward while defeating its purpose.
INC-008 · INC
Goodhart's Law and Incentive Engineering: Why Metrics Fail and How to Design Better Systems
Open with
Your metric may be accurate right up until you reward people for maximizing it.
Tell the story
A class gets points for every page read. Soon children choose books with huge letters, flip pages quickly, and cannot tell the stories. Page count once helped show reading, but the prize turned it into the target.
Ask the room
Which trusted measure would become misleading under intense optimization pressure?
Land here
When a measure becomes the target, it can stop measuring what mattered.
INC-009 · INC
Principal-Agent Theory for the Age of AI: Aligning Autonomous Decision Makers
Open with
AI does not erase the principal-agent problem; it gives the agent speed, memory, and the ability to hire sub-agents.
Tell the story
A coach gives the assistant a plan, and the assistant asks the captain to carry it out. On the field, the captain sees things the coach cannot. The team needs goals, limits, updates, and a way to ask when the plan no longer fits.
Ask the room
What crucial information can your agent see that the principal cannot?
Land here
Autonomous agents need alignment, information, monitoring, and correction—not instructions alone.
INC-010 · INC
Reputation Systems: Engineering Trust Through Incentives
Open with
Reputation is how tomorrow reaches back and changes today's choice.
Tell the story
Children share scooters. Each careful return is written on the lending board. A good history lets a child borrow again with fewer worries, so caring for today's scooter helps tomorrow too.
Ask the room
What future opportunity should reliable behavior earn—and how can an error be corrected?
Land here
Reputation makes past behavior part of future incentives.
INC-011 · INC
Trust Engineering: Designing Systems That Earn and Sustain Trust
Open with
Trust is not the opposite of verification; it is what reliable verification helps produce.
Tell the story
Children will cross a rope bridge when the ropes are strong, an adult checks it, broken boards are repaired, and a sign says how many people may cross. Painting “Trust Me” on the bridge would not help.
Ask the room
What would your stakeholders need to see before trusting this system after a failure?
Land here
Trust is earned by reliable, verifiable, accountable behavior over time.
INC-012 · INC
Incentive Architecture: Engineering Organizations for Continuous Alignment
Open with
You cannot align an organization with one lever when behavior emerges from the whole system.
Tell the story
Healthy plants need more than one reward for growing. The greenhouse joins sunlight, water, soil, vents, careful gardeners, and gauges. When the weather changes, the whole system adjusts together.
Ask the room
Which disconnected incentive mechanism should become part of one continuous alignment architecture?
Land here
Continuous alignment comes from an architecture, not one bonus or metric.
COOR-001 · COOR
The Complexity of Coordination: Why Governable Intelligence Becomes Essential at Scale
Open with
Adding one more agent adds capacity—but it also adds relationships nobody may be governing.
Tell the story
If ten cooks all shout to everyone, dinner becomes noise. A kitchen uses stations, order tickets, timing, and a head cook so chopping, baking, and serving come together without every cook talking to every other cook.
Ask the room
Where has coordination complexity grown faster than your management architecture?
Land here
Coordination needs architecture as participants and dependencies multiply.
CA-001 · CA
Contribution Architecture: Modeling, Verifying, and Attributing Value Creation
Open with
The person holding the trophy is rarely the whole story of how value was created.
Tell the story
A group wins the science fair, but the presenter was not the only helper. One child found the idea, another built the model, another caught a mistake, and another practiced the talk. Their project map shows how the result depended on all of them.
Ask the room
Which invisible contribution made your last major outcome possible?
Land here
Contribution Architecture makes value creation visible, traceable, and attributable.
COMP-001 · COMP
Compensation Architecture: Distributing Economic Value to Maximize Long-Term Engagement
Open with
Pay is never just a transaction; it teaches the organization what it values.
Tell the story
Children run a pizza stand. One made dough, one found customers, one fixed the oven, and one cleaned up. They decide how to share earnings in a way that recognizes real help and makes everyone willing to work together again.
Ask the room
Does your compensation system strengthen the next cycle of contribution—or weaken it?
Land here
Compensation distributes economic value to sustain contribution and engagement.
ROBOTICS-001 · ROBOTICS
Robotics as Delegated Agency
Open with
When software can move a machine through shared space, governance becomes physical.
Tell the story
A robot may pick up balls inside the playground after school. Its badge says which robot it is, its card says the job and the chalk boundary, and a teacher can stop it. The robot may not follow a ball into the street.
Ask the room
Can you show who authorized a robot's last action, under what limits, and who could stop it?
Land here
A robot is a physical helper whose authority must have visible boundaries.
SPACE-001 · SPACE
Mission Authority Systems
Open with
What does authorization mean when the nearest human answer is minutes—or months—away?
Tell the story
A rover is very far from home. It cannot ask a person about every rock. It carries a mission card saying where it may go, how much power it may use, when it must stop, what it must remember, and when it must wait for Earth.
Ask the room
Which decisions must a distant mission system make locally, and which must remain impossible without Earth?
Land here
Distance makes explicit mission authority more important, not less.