> ## Content Index
> Fetch the complete content index at: https://modalpathethics.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# How I 6-0’d Nick Bostrom With Aurumoth
- URL: https://modalpathethics.com/how-i-6-0d-nick-bostrom-with-aurumoth/
- Published: 2026-10-02T11:00:32.000Z
- Updated: 2026-10-05T04:41:11.000Z
- Description: "It was just a matter of time before we got burned by the power creep of Topic Leaders." - DougJustDoug
- Author: Aidan Lawson
- Tags: Multiplayer, Engagement, Inner Apocalypse, Pokémon, Post-Game, Chastening of the Controller

**won 6-0 vs nick bostrom, gg EZ, crit did not matter (knew ur spread).**

Posting the prep because this one was mostly won in the builder.

The matchup into Bostrom is terrible. If the scout says “world dictator, centralization bad, freedom good,” delete the paste and start over. That is **not** his set.

Bostrom’s **singleton** is broad on purpose. This thing can be a democratic world republic, a superintelligent machine, a stable alliance, a world dictatorship, or even a sufficiently prescriptive moral code. It can contain enormous internal variety, independent agents, regional differences, experimentation, and elected representation. The common property is higher-level coordination. Bostrom’s definition also gives the **singleton** the power to prevent threats to its own existence and supremacy and to exercise effective control over major features of its domain.[\[1\]](https://nickbostrom.com/fut/singleton?ref=modalpathethics.com)

That second part is the whole game.

The harder scouting report is called *The Future of Human Evolution*. Bostrom considers a future in which competition rewards agents that devote more and more of their resources to economic or reproductive success and squeezes out forms of life we would actually want to preserve. A locally humane polity may understand the danger perfectly and still lose to competitors willing to sacrifice more. His proposed answer, under those assumptions, is global coordination strong enough to change the fitness environment itself.[\[2\]](https://nickbostrom.com/fut/evolution?ref=modalpathethics.com)

Then the paper gets to stability. Bostrom says the resulting order would need its domestic affairs regulated so that **“no internal challenges to its constitution could arise.”** In a footnote, he considers one possible way a spacefaring singleton might preserve itself over extreme distances and times: colonizers could be built with a fundamental desire to obey the singleton’s basic laws, with descendants required to inherit that disposition and verification tightening as technology improves.[\[2\]](https://nickbostrom.com/fut/evolution?ref=modalpathethics.com)

He also tells the reader exactly how provisional this is. He does not claim to know that the dystopian evolutionary trajectory is the default. He separates the long-run argument from present policy. He explicitly warns that a singleton could increase the risk of an oppressive regime becoming global and permanent.[\[2\]](https://nickbostrom.com/fut/evolution?ref=modalpathethics.com)

That means I do not have to 6-0 a caricature. I can grant Bostrom the coordination problem, the arms-race problem, the selection problem, the possibility of extreme competence, the need for binding decisions, and even the possibility that some functions require planetary authority.

The question is whether those premises reach **supremacy**.

They do not.

## "But I have long wondered if it really makes sense to put complete responsibility for a CAP project on one person's shoulders.****"**

— ****DougJustDoug**

# Scout.

The whole game turns on six promotions that Bostrom’s framework either makes explicitly or leaves structurally available.

1. Global reach becomes general jurisdiction.
2. A final decision becomes control over the conditions of final decisions.
3. Superior intelligence becomes authority over the terms of its own authority.
4. Internal diversity becomes evidence of independent correction.
5. Present indispensability becomes a reason to protect permanent supremacy.
6. Protection of future value becomes protection of the founder’s constitution against future constitutional challenge.

Those are six different claims.

I am keeping the coordination machinery and removing the promotions.

That is the team.

---

## Builder.

### 1\. Aurumoth

![](https://storage.ghost.io/c/20/43/2043f11a-6ae3-404c-bb28-01fce8d9ac88/content/images/2026/09/image-534.png)

Aurumoth is here because Create-A-Pokémon already went through a much smaller version of the institutional problem.

CAP first learned why weak coordination can fail. Under the old Weak Topic Leader model, a project could contain public discussion, public votes, competent contributors, and still have trouble carrying one competitive idea across a long sequence of dependent stages. Different populations could interpret the Concept differently. A stat spread could embody one Pokémon while a later movepool built another. The problem was integration.[\[7\]](https://modalpathethics.com/applied-case-the-temporary-topic-leader/)

So CAP strengthened the Topic Leader.

The **Strong Topic Leader** model gave one person explicit authority to frame discussion, evaluate arguments against the Concept, keep the project moving, and exercise real influence over what reached a poll. The reform had a defensible job. “Everybody votes” had not solved the problem of making the outputs belong to one project.[\[3\]](https://www.smogon.com/forums/threads/topic-leadership.3475824/?ref=modalpathethics.com)[\[7\]](https://modalpathethics.com/applied-case-the-temporary-topic-leader/)

Then Aurumoth happened.

The contemporary CAP report describes a**Risky Business**concept that accumulated an extraordinary amount of offensive flexibility: multiple setup routes, broad coverage, and three very different abilities, including *No Guard* and *Illusion*.[\[4\]](https://www.smogon.com/smog/issue23/cap?ref=modalpathethics.com) The leadership fight around the project was even more important. In the Policy Review that followed, DougJustDoug accused the Topic Leader of pushing the project toward his preferred result and argued that the office had accumulated too much formative control. That account is an angry institutional intervention from a participant and moderator, not a neutral replay judge. The community used the dispute to redesign the office.[\[3\]](https://www.smogon.com/forums/threads/topic-leadership.3475824/?ref=modalpathethics.com)

DarkSlay gave the cleanest description of what concentrated formative authority can do before the vote even starts:

> **“Would this even get slated? Would the TL like it?”**[\[3\]](https://www.smogon.com/forums/threads/topic-leadership.3475824/?ref=modalpathethics.com)

That is the point.

Nobody has to delete your submission if you already know which submissions are worth spending three hours writing.

CAP did not respond by abolishing integration. The reform that emerged separated some of the powers. Specialist leaders would build the slates for their stages. The overall Topic Leader would continue interpreting the Concept and could veto a specialist slate and demand a re-slate. The Topic Leader could not silently substitute a preferred slate. If the two offices could not resolve the dispute, moderators could intervene.[\[3\]](https://www.smogon.com/forums/threads/topic-leadership.3475824/?ref=modalpathethics.com)

Same project. Strong coordinator. Less jurisdiction concentrated in the coordinator.

Aurumoth itself was later revised. Its *Illusion* ability was removed and replaced with *Light Metal* during the update process.[\[5\]](https://www.smogon.com/forums/threads/cap-updates-aurumoth-discussion.3597837/page-6?ref=modalpathethics.com) A later Smogon retrospective described the change as making Aurumoth more manageable while leaving it a serious offensive threat.[\[6\]](https://www.smogon.com/articles/cap-update-rundown?ref=modalpathethics.com)

I wanted a win condition whose own history already included the argument.

- The integrator can be necessary.
- The integrator can also be wrong about what it gets to own.

### 2\. Bounded Last Words

![](https://storage.ghost.io/c/20/43/2043f11a-6ae3-404c-bb28-01fce8d9ac88/content/images/2026/09/image-532.png)

This was the speed control.

A binding decision can be final for a function, case, scale, and time without giving the closer control over evidence, review, duration, succession, amendment, and every later question about its own jurisdiction.

That distinction matters because the standard sovereignty argument loves an emergency. Somebody eventually has to decide. Correct. The fact that a decision has to end does not establish one permanent institution that decides the terms under which all decision-making can ever be corrected.

- The decision closes.
  - The constitution continues.

### 3\. Batman

![](https://storage.ghost.io/c/20/43/2043f11a-6ae3-404c-bb28-01fce8d9ac88/content/images/2026/09/image-533.png)

Batman is here because [Schmitt gets the clock right](https://modalpathethics.com/carl-schmitt-and-the-batman-who-leaves/).

There are emergencies in which waiting for the entire constitutional order to deliberate would destroy the thing the order exists to protect. The person with the warrant has to act.

The mistake comes after the action.

Emergency competence does not spread automatically from the threatened function into evidence custody, duration, renewal, succession, or the power to define whether the emergency still exists.

I do not need Batman to stop making decisions.

I need Gotham to be able to keep functioning when Batman leaves.

### 4\. *The Inner Apocalypse*

[The Inner Apocalypse Has Been Published (Free Download)Civilization made machines out of people. Artificial intelligence may be the first chance we have ever had to stop.![](https://storage.ghost.io/c/20/43/2043f11a-6ae3-404c-bb28-01fce8d9ac88/content/images/icon/thin-tile.rulebook-2-57836c12-a315-42f1-b0ef-e6c16952e90f.png)Modal Path EthicsAidan Lawson![](https://storage.ghost.io/c/20/43/2043f11a-6ae3-404c-bb28-01fce8d9ac88/content/images/thumbnail/Screenshot-2026-09-24-at-01-38-52-The-Inner-Apocalypse-A-Post-Sovereign-Constitution---The_Inner_Apocalypse_v59_Distribution_With_Cover.pdf-1-b992adfc-6f5f-40e9-a63d-976bb7a5c103.png)](https://modalpathethics.com/the-inner-apocalypse-has-been-published-free-download/)

This is the common-mode failure slot.

The dangerous system does not have to silence anyone. It can hear everyone. It can fund multiple panels, maintain different institutions, preserve local identities, and return excellent answers.

The question is whether those institutions still possess independent causal contact with reality and a route for making a contradiction consequential before the decision becomes irreversible.

That is **causal plurality**.

Visible plurality can survive after causal plurality has gone.

### 5\. Material Bounded Finality

![](https://storage.ghost.io/c/20/43/2043f11a-6ae3-404c-bb28-01fce8d9ac88/content/images/2026/09/image-535.png)

This is where the legal theory has to cash out.

A right to replace an authority is useless if every replacement depends on records the authority alone can export, credentials it alone can issue, technicians trained only on its interface, infrastructure it alone can operate, or a successor that cannot carry the protected work inside the available time.

**Material Bounded Finality** asks whether the adverse correction can actually arrive while the thing we meant to protect keeps functioning.

It also asks the question again after the replacement.

A constitution that can correct one ruler by exhausting every independent resource needed to correct the next one has solved one case.

It has not solved succession.

### 6\. Constitutional Firsthood

![](https://storage.ghost.io/c/20/43/2043f11a-6ae3-404c-bb28-01fce8d9ac88/content/images/2026/09/image-536.png)

This is for the version where Bostrom is right about the asymmetry.

> Suppose the first superintelligent coordinator really is the only available system capable of carrying a civilization-scale function safely.

That can justify unusual authority now.

It also creates a constitutional obligation to distinguish **temporary firsthood** from permanent authorship. Later participants must inherit evidence, standing, independent means, and institutions capable of revising what they inherit.

The first system may remain the best system.

Being first does not make everyone after it a permanent ward.

That was the six.

> Aurumoth / *Bounded Last Words* / Batman / *The Inner Apocalypse* / *Material Bounded Finality* / Constitutional Firsthood.

I did not bring decentralization. I knew he prepped for it.

I brought a better answer to coordination.

## **"*No risk, no reward. Welcome to Leadership 101.**"*

****— DougJustDoug et al,** ***Topic Leader Guide**

# The Game.

## 1\. Global coordination does not give general jurisdiction.

Bostrom’s strongest premise is also the easiest one to concede.

> Some problems are global.

An arms race between actors with civilization-ending weapons cannot be made safe by telling each actor to govern itself more sincerely. A colonization race can create incentives that no local polity can cancel unilaterally. A planetary defense problem can require observations, resources, and operational decisions across borders. Bostrom is right that there are fields in which local rationality can reproduce a global failure.[\[1\]](https://nickbostrom.com/fut/singleton?ref=modalpathethics.com)[\[2\]](https://nickbostrom.com/fut/evolution?ref=modalpathethics.com)

The inference I am rejecting starts after that.

A problem with global causal reach can justify authority with global causal reach for that problem.

It does not establish one agency with general jurisdiction at “the highest level.”

That difference sounds small until the powers are written down.

One institution may set a planetary technical standard. Another may carry the operation. Another may preserve evidence. Another may hear an affected claimant. Another may decide whether the first institution exceeded the warrant. A common system may integrate the data. A final operational decision may still be binding.

Those authorities have to work together.

Nothing in the need to work together establishes that they must collapse into one agency capable of defending its supremacy over the entire arrangement.

CAP already found the smaller version.

The project needed a Topic Leader because several sequential stages had to stay attached to one Concept. After Aurumoth, CAP did not conclude that the Typing Leader, Ability Leader, Stats Leader, and Movepool Leader were therefore redundant. It separated the domains while retaining an integrator. The Topic Leader could demand a re-slate if a specialist stage drifted from the Concept. The Topic Leader did not get to write the specialist’s replacement slate alone.[\[3\]](https://www.smogon.com/forums/threads/topic-leadership.3475824/?ref=modalpathethics.com)

That is a direct answer to the idea that integration requires ownership of every integrated jurisdiction.

- **Global** is a scale.
- **General** is a scope.
  - Bostrom gives us reasons for the first.
  - He has not yet given us the second.

## 2\. A last word does not own the conditions of last words.

Now take the hard case.

> The decision really does have to end.

The missile launches or it does not. The grid disconnects or it does not. The pathogen lab closes or it does not. The asteroid intercept commits or it does not.

There is a last word on that path.

Bounded Last Words starts there. The whole point is to avoid laundering indecision into virtue. Somebody with a constituted warrant closes the present action.

Then we ask a different question.

- Who controls the evidence by which that action can later be judged?
- Who decides whether the emergency continues?
- Who can narrow the jurisdiction?
- Who can remove the holder?
- Who can alter the rule for the next case?
- Who can transfer the function?
- Who can amend the arrangement itself?

Bostrom’s definition becomes unstable exactly here because the singleton is supposed to prevent threats to its own existence and **supremacy**.[\[1\]](https://nickbostrom.com/fut/singleton?ref=modalpathethics.com)

There are two readings.

1. Under the first, a lawful constitutional process can remove, divide, replace, or fundamentally restructure the highest-level decision agency.  
  1. Then the agency’s **supremacy** is already bounded by a constituted route it does not own. Great. That is much closer to the architecture I am defending.
2. Under the second, the agency can prevent that adverse constitutional displacement because displacement threatens its **supremacy**.  
  1. Then the question is over. The closer has acquired control over the conditions of its own correction.

Bostrom’s evolutionary paper pushes toward the second reading when it says stable domestic affairs would have to be regulated so that no internal challenges to the constitution could arise.[\[2\]](https://nickbostrom.com/fut/evolution?ref=modalpathethics.com)

That solves stability by deciding in advance what a successful constitutional challenge is allowed to accomplish.

It cannot replace the constitution.

That is not an answer to the last-word problem.

That is a choice about who gets the last word over last words.

## 3\. Better judgment does not self-authorize jurisdiction.

The superintelligence version is the one worth testing.

Make it good.

Give it better forecasts than every ministry and better scientific integration than every university consortium. Let it see interactions no human committee could track. Let its recommendations beat the human baseline for twenty straight years.

Performance matters.

It can justify trust, delegation, and expanding the range of decisions on which the system’s analysis carries serious weight.

It still does not answer three separate questions simply by being more accurate.

1. What transition is actually happening to which bearers and enabling structures?
2. Which tradeoff is justified, given standing, rights, uncertainty, and legitimate authority?
3. If either account is wrong, can an effective external correction reach the result?

*The Inner Apocalypse* now keeps those as separate jobs: structural representation, constitutional judgment, and material correction.[\[9\]](https://modalpathethics.com/the-inner-apocalypse-has-been-published-free-download/)[\[10\]](https://doi.org/10.5281/zenodo.22818297?ref=modalpathethics.com)

A model can be excellent at the first and wrong about the second.

An institution can make the right judgment and lack the means to execute it.

A correction body can understand both and still depend on the challenged system for the one permission required to make the remedy real.

There is no intelligence score at which those distinctions disappear.

Suppose a superintelligent system evaluates three coastal-defense plans more accurately than every human analyst.

Excellent. Its competence is evidence for using the analysis.

It does not decide by itself whose home can be taken, which losses qualify for compensation, who has standing to introduce a fourth plan, what constitutional floor constrains the project, or which institution may revise the decision rule.

Those are relations among knowledge, authority, and affected loci.

The model can participate in all of them.

Accuracy in one does not award jurisdiction over the rest.

This is where Bostrom’s “friendly superintelligent machine” example needs more constitutional specification than the word friendly can carry.[\[1\]](https://nickbostrom.com/fut/singleton?ref=modalpathethics.com)

I am happy to load the smart machine.

I am not letting performance decide the conditions under which its own authority may be challenged.

## 4\. Internal diversity is not an independent correction path.

Bostrom sees the obvious danger in the singleton.

If one state goes bad, people can sometimes leave. Other states can outperform it, expose its errors, develop technologies it missed, or provide examples that support reform. If the **singleton** goes bad, civilization has concentrated the failure. Bostrom calls it putting all the eggs in one basket. His answer includes the possibility that a good singleton deliberately maintains an internal ecology of different societies and regional diversity.[\[1\]](https://nickbostrom.com/fut/singleton?ref=modalpathethics.com)

That is a real mitigation. It does not by itself restore the thing external plurality supplied.

The relevant property is **independent** **correction**.

*The Inner Apocalypse* spends a large part of its argument on exactly this difference.

Imagine several validators built by different institutions, using different algorithms, different hardware, different teams, and different national chains of command. They look independent. They fail differently on most tests.

Then every validator receives measurements only after those measurements pass through the same reference transform.

One common error in the transform enters every answer before the diversity begins.

- The validators can disagree honestly and usefully about everything downstream.
- Their agreement on the shared error is still common-mode agreement.

The book’s [constructed asteroid case](https://modalpathethics.com/tag/the-player-who-can-win-every-game/) uses this structure deliberately: the system is willing to correct itself when a genuinely independent observation finally survives reconstruction. The problem is timing. The independent route becomes consequential after the irreversible intervention has already happened.[\[9\]](https://modalpathethics.com/the-inner-apocalypse-has-been-published-free-download/)

That is the **Inner Apocalypse** in technical form.

**Causal plurality** requires more than several institutions existing.

A contrary account has to preserve evidence. Its interpretation has to remain testable. Some institution has to possess the resources and authority needed to make the result consequential. And the route has to work before the relevant future closes.

Now, return to the singleton.

- A thousand local governments can exist.
- A thousand parties can exist.
- A thousand research institutes can exist.
- A thousand artificial intelligences can exist.

But if every consequential adverse correction has to pass through one common ontology, one mandatory identity layer, one evidence gate, one execution substrate, or one highest-level agency that can protect its own **supremacy**, the plurality is real in many respects and still constitutionally dependent at the point that matters.

Counting institutions will not detect that.

Following the correction path will.

This is why Bostrom’s internal diversity answer does not beat the all-eggs objection on its own.

He needs an account of which internal differences can become adverse consequences for the singleton itself.

Once he supplies that, we can inspect whether the result is still a singleton in the strong sense doing the work in his argument.

## 5\. Indispensability is a capacity problem, not a title.

This was the one I prepped hardest for in the calcs because it can beat paper constitutions clean.

> Suppose the common coordinator really is indispensable.

- Turn it off and the hospitals miss allocations.
- Replace it and the successor cannot parse the live cases.
- Split it and no receiving institution has enough staff.
  - The old model has been retired.
  - Everyone learned one interface.
  - The archives are technicallyexportable, but the export takes twelve hours and the protected correction window is six.

At that point, a legal power of removal can remain perfectly intact while the material ability to use it has disappeared.

Material Bounded Finality calls the boundary where an independent successor can no longer become adequate inside the protected correction interval a **corrigibility horizon**.[\[9\]](https://modalpathethics.com/the-inner-apocalypse-has-been-published-free-download/)[\[10\]](https://doi.org/10.5281/zenodo.22818297?ref=modalpathethics.com)

That is the actual succession problem.

The issue is whether the field still contains a viable adverse route, not whether the incumbent deserves gratitude or caused the dependency maliciously.

The handoff has at least two different accounts.

1. There is **carrying capacity**: people, competence, equipment, lawful authority, money, infrastructure, and enough service to protect the required load.
2. There is **operative state** **and provenance**: current cases, exceptions, histories, claims, records, and access needed to know what the successor is inheriting.

More compute cannot recreate a record that was never preserved.

A perfect archive cannot staff the clinic.

And parallel preparation can still miss the deadline.

The book uses a simple constructed example. Independent capacity takes six hours to prepare. State transfer takes four. Those jobs can happen in parallel. Activation requires another two hours after both complete. The handoff takes eight hours. A six-hour correction window is already lost.[\[9\]](https://modalpathethics.com/the-inner-apocalypse-has-been-published-free-download/)

That is why “replaceable in principle” is weak evidence.

A system can accept every audit request and every shutdown command while the institution around it quietly retires every usable alternative.

No rebellion is required. No coup. 

No hostile singleton.

Dependency can do the constitutional work.

Bostrom’s framework makes this especially important because decisive technological advantage is one of the paths he identifies toward singleton formation. An agency with a large enough lead could prevent others from catching up in security-relevant technology.[\[1\]](https://nickbostrom.com/fut/singleton?ref=modalpathethics.com)

There may be good reasons to prevent proliferation of some dangerous capabilities.

That still leaves the continuity question.

- Who else can carry the public function?
  - What state do they need?
  - How fast can the handoff happen?
- What resources are shared?
- Which alternatives are deliberately maintained?
- When does the field cross the point where replacement can no longer happen safely?

Those are capacity questions.

**Supremacy** is not an answer to them.

That is what the current carrier has after nobody else can answer them anymore.

## 6\. “No internal challenges” is the lock-in.

This is the endgame.

Bostrom’s evolutionary problem is real enough to make the conclusion tempting.

Suppose eudaemonic ways of living are systematically less competitive than agents optimized for reproduction or resource acquisition. Local protections can then be punished at a higher level. 

- A humane polity spends resources preserving art, leisure, friendship, or forms of consciousness it values.
- A rival pours more into expansion.  
  - The rival wins.  
    - The selection pressure returns.

If that is the field, Bostrom is right that local virtue does not solve it.[\[2\]](https://nickbostrom.com/fut/evolution?ref=modalpathethics.com)

Now, look carefully at the proposed stability condition.

The singleton must prevent internal developments from toppling its constitutional agreement. The paper says no internal challenges to its constitution could be allowed to arise under the stable arrangement it sketches. The colonization footnote goes farther and considers engineering a fundamental desire to obey the singleton’s basic laws into colonizers and their progeny, with increasingly strict verification so defection approaches zero.[\[2\]](https://nickbostrom.com/fut/evolution?ref=modalpathethics.com)

This is where the argument loses the sixth.

There are two different things worth protecting.

- One is a **constitutional floor**:  
  - Future agents should not be able to win by annihilating everyone else’s standing, coercively converting the population, externalizing existential risk, or destroying the material conditions under which other forms of life continue.
- The other is the founder’s substantive constitutional settlement.

Those are not identical.

A future population can revise institutions while remaining bound by protections for other loci. It can replace an allocator. Change representation. Split an office. Alter the economic system. Recognize a kind of claimant the founders misunderstood. Discover that a safety category was wrong. Move a power to another scale. Develop a better constitutional instrument.

A system designed so that no internal challenge to the constitution can arise has decided those possibilities before the participants who would discover the reasons for them exist.

That is why *The Inner Apocalypse* develops constitutional firsthood.

Founders inevitably set durable paths for participants who cannot yet answer. Roads are laid. identity systems are selected. archives are built. models become standards. early artificial systems may control capabilities no successor yet possesses.

Firsthood is real power.

The obligation attached to it is to preserve the conditions of later co-authorship.

As successors acquire evidence, standing, independent means, and institutions capable of changing the inheritance, the special authority justified by their earlier absence must end.[\[9\]](https://modalpathethics.com/the-inner-apocalypse-has-been-published-free-download/)

- Bostrom’s proposal protects future eudaemonic value by securing a constitution against future constitutional challenge.
- The post-sovereign alternative protects the conditions under which future participants can continue protecting value and revise the protector.

That is the exact difference.

Bostrom may answer that some constitutional changes would recreate the evolutionary disaster.

Of course. Then those changes face the same burden of proof as every other transition. A protected floor can forbid destructive moves.

What does not follow is that the current constitutional arrangement receives permanent immunity from being challenged.

A future author can be constrained by the standing of other future authors.

They still have to be an author.

That is the 6-0.

## "A lot of time and effort has gone into the CAP process guide, but it's not perfect."

****— DougJustDoug et al,** ***Topic Leader Guide**

# On Aurumoth.

I want to be precise about what the CAP case proves because there's a small chance this article becomes stupid if Pokémon is asked to prove too much.

Smogon has not solved planetary governance.

CAP has permanent infrastructure, moderation, platform dependencies, status, informal influence, and powers that the Topic Leader election does not recreate from scratch every project. The current Modal Path Ethics article already makes that limitation explicit.[\[7\]](https://modalpathethics.com/applied-case-the-temporary-topic-leader/)

The useful result is narrower.

CAP provides a real institutional counterexample to the idea that strong integration and revisability have to be enemies.

The **Weak Topic Leader** model exposed an integration failure.

The **Strong Topic Leader** model repaired part of it by giving one leader more formative authority.

Aurumoth exposed how that authority could shape participation and slating too strongly. CAP responded by redistributing parts of the job while preserving a visible integrator. 

The specialists did not become decorative advisors. They received actual jurisdictions.

The overall Topic Leader did not disappear. The project still had someone responsible for carrying the Concept across the whole.

The product remained revisable too. Aurumoth lost *Illusion* later and continued existing as Aurumoth.[\[5\]](https://www.smogon.com/forums/threads/cap-updates-aurumoth-discussion.3597837/page-6?ref=modalpathethics.com)[\[6\]](https://www.smogon.com/articles/cap-update-rundown?ref=modalpathethics.com)

That sequence is why the case is relevant here.

Smogon did not abolish the integrator. 

- It changed what the integrator was allowed to own.
  - Then it kept playing.

That is enough for the argument.

---

# Postgame.

So what did Bostrom actually lose?

- He did not lose the coordination problem.
- He did not lose the arms-race problem.
- He did not lose the possibility that competition can select for outcomes almost nobody values.
- He did not lose global standards.
- He did not lose emergency authority.
- He did not lose binding decisions.
- He did not lose the possibility that a superintelligence deserves enormous evidentiary weight.
- He did not even lose the possibility of a very powerful planetary coordinator where the function requires one.

I gave him all of that in the builder.

What he loses is the promotion from those facts to a self-protecting highest-level agency whose supremacy is part of the constitutional solution.

- Global reach does not establish general jurisdiction.
- Operational finality does not establish meta-jurisdiction.
- Superior cognition does not establish authority over the conditions of authority.
- Internal diversity does not establish an independent correction path.
- Indispensability does not establish permanent title.
- Protection of future value does not establish final authorship over future constitutional change.

Those are six separate inferences.

None is rescued by saying the singleton can be democratic or internally diverse, by making the singleton smarter, or by reminding us that coordination failures are real.

Those are the premises I already kept.

Bostrom himself gives the reader one final reason to take the critique seriously: he closes *The Future of Human Evolution* by warning that the singleton argument is preliminary and that creating one could raise the risk of an oppressive regime becoming global and permanent.[\[2\]](https://nickbostrom.com/fut/evolution?ref=modalpathethics.com)

Good scout.

That risk does not sit outside his coordination argument. It appears exactly where coordination becomes supremacy, where stability becomes insulation from constitutional challenge, and where a useful carrier becomes the only carrier the future is allowed to have.

That is the part Modal Path Ethics cuts.

Crit didn’t matter.

I gave the argument its best rolls.

- The coordination problem is real.
- The conclusion still does not reach sovereignty.

> **gg**

---

# Shoutouts.

Shoutouts to the Agency.

---

# Sources.

\[1\] [Nick Bostrom, “What Is a Singleton?”](https://nickbostrom.com/fut/singleton?ref=modalpathethics.com)

\[2\] [Nick Bostrom, “The Future of Human Evolution”](https://nickbostrom.com/fut/evolution?ref=modalpathethics.com)

\[3\] [DougJustDoug et al., “Policy Review — Topic Leadership”](https://www.smogon.com/forums/threads/topic-leadership.3475824/?ref=modalpathethics.com)

\[4\] [capefeather, “CAP4 Report: Aurumoth”](https://www.smogon.com/smog/issue23/cap?ref=modalpathethics.com)

\[5\] [“CAP Updates: Aurumoth Discussion”](https://www.smogon.com/forums/threads/cap-updates-aurumoth-discussion.3597837/page-6?ref=modalpathethics.com)

\[6\] [Smogon University, “CAP Update Rundown”](https://www.smogon.com/articles/cap-update-rundown?ref=modalpathethics.com)

\[7\] [Modal Path Ethics, “Applied Case: The Temporary Topic Leader”](https://modalpathethics.com/applied-case-the-temporary-topic-leader/)

\[8\] [Modal Path Ethics, “The Inner Apocalypse”](https://modalpathethics.com/the-inner-apocalypse/)

\[9\] [The Inner Apocalypse: A Post-Sovereign Constitution](https://modalpathethics.com/the-inner-apocalypse-has-been-published-free-download/)

\[10\] [Material Bounded Finality: A Finite Feasibility Model of Correctable Final Authority](https://doi.org/10.5281/zenodo.22818297?ref=modalpathethics.com)