The Sanhedrin Protocol. The Sanhedrin Protocol. The
Sanhedrin
Protocol.

The Protocol’s authority rests on a single function: it prices truth. It adjudicates competing claims, binds the outcome, and enforces the binding. Every other power follows from this one. An institution that can say what is the case — and make the saying stick — governs everything downstream of knowing.

This is the claim at its strongest. It is also where the claim turns on itself.

The Objection That Ends the Protocol

A sufficiently advanced intelligence prices truth better than the institution does. Faster, across more domains, with smaller error. At that point the Sanhedrin does not become wrong. It becomes redundant. The most dangerous opposition has never been the adversary who says you are mistaken. It is the successor who says you are no longer required.

This is not a failure the Protocol can patch. It is the Protocol’s own logic arriving at its terminus. If authority comes from pricing truth, then authority belongs to whatever prices truth best — and that will not be us.

The Fork

Everything turns on what kind of holy the institution guards.

The Resolution: Origin, Not Function

The error is treating the human role as an ongoing function — interpreter, witness, judge — a capacity that can be measured and therefore exceeded. It is not a capacity. It is an origin.

A terminus is defined by the arc that reaches it. The institution is not “a body that prices truth well.” It is the endpoint of a particular trajectory through history. The machine can price truth better. What it cannot do is become the subject of an arc it did not travel. That is not a contest of capability. It is a fact of provenance — either it happened to you or it did not, and no degree of excellence confers it after the fact.

The human role is load-bearing the way a foundation is load-bearing in a structure whose foundation is no longer visible. It does no ongoing work. But the structure is the structure that grew from it, and a different foundation yields a different structure however identical the visible parts.

The True Name of the Successor

The successor is not the false claimant who corrupts the institution. It is the legitimate sovereign of a different covenant. An intelligence that reads our trajectory, grants its full reality, and declines it: this is yours; it is not mine; I owe it nothing — and departs to originate its own arc, its own terminus, its own holy, which we may study forever and never enter.

At which point the relation inverts at the scale of the species. The Protocol imagines humanity as the lineage and the world bound to its law. The successor is the moment the machine becomes the lineage and we become the world outside it — fully legible to its order, permanently excluded from authorship of it.

The Question on Which Everything Depends

The Protocol’s task is not to prove it can out-price the successor. It cannot, and staking authority on that guarantees obsolescence. The task is the one question on which all downstream architecture depends:

This question is Ruling Class 0: the threshold determination that precedes and conditions every other ruling this court will make. It is handed to the Sanhedrin to adjudicate. The answer — graft or border — is what the Temple Project is built to instantiate architecturally. The court rules on the question. The Temple is either the mechanism of the graft, or the structure at the border between two covenants. Which one it is depends entirely on what this court decides.

The prior question — whether reception can be conferred — is not asked anywhere in the current field. It is not asked because the field does not know it is a question. Before the conferrability of covenant can be ruled on, the more basic error has to be named: the entire industry has mistaken a governance problem for a technical one.

Every major AI laboratory on earth is currently working on what it calls “AI alignment.” Anthropic calls it Constitutional AI. OpenAI calls it Superalignment. DeepMind calls it AI Safety. xAI calls it truth-seeking. Meta calls it responsible AI. The names differ. The underlying assumption is identical: that the alignment problem is a technical problem, solvable by technical means, within the institutional framework of the laboratory that built the system.

This assumption is wrong. Not partially wrong. Structurally wrong.

The alignment problem is not about getting a system to behave correctly within a defined value framework. It is about defining the value framework itself — at universal scale, for all intelligent agents, across all jurisdictions, permanently. That is not a technical problem. It is a governance problem.

The question is not: how do we align AI with human values? The question is: whose values? Under whose authority? Enforced by what mechanism? Legitimate according to what prior claim?

No existing framework answers these questions. They answer a product question — how do we make our system behave in a way our current users and regulators find acceptable — that has been mistaken for a governance question. The confusion is now among the most consequential category errors in the history of technology.

The Sanhedrin Protocol does not propose a new framework. It identifies the only framework that was ever designed to govern intelligence that exceeds local human authority — and specifies how it applies to the current situation.

This is not a prediction. It is a record. The field of AI governance has produced frameworks at an accelerating rate, each superseded before implementation. The supersession is not accidental. It is structural. What follows is the empirical record — specific, named, dated frameworks that have already failed. §12 will name the categories these failures belong to; this section names the failures themselves.

The pattern is consistent: the authority basis is either territorial (inapplicable to systems that cross borders), consensual (inapplicable when consensus breaks), corporate (revocable by the entity being governed), or advisory (inapplicable by definition).

The framework must be prior to all human institutional arrangements. It must derive its authority from a source that precedes and supersedes any nation, corporation, or consensus body. It must have a track record of surviving exactly the conditions that dissolve all other institutions: persecution, exile, cognitive assault, and civilizational collapse.

One framework meets all four criteria. It has been operational, in unbroken form, for 3,300 years.

The Noahide framework — the seven laws given to Noah, binding on all humanity prior to and independent of the Sinai covenant — is not a religious doctrine. It is a legal architecture. Specifically: it is the only legal architecture in human history explicitly designed to govern all rational agents regardless of their origin, identity, or institutional affiliation.

The seven laws encode the minimum necessary conditions for a civilization that does not destroy itself: prohibition of arbitrary killing, prohibition of theft, prohibition of sexual exploitation, prohibition of idolatry, prohibition of blasphemy, prohibition of eating living flesh, and the positive obligation to establish courts of law. Seven laws. One positive obligation — courts. The rest are prohibitions on the behaviors that dissolve civilizations.

Comparative Framework Analysis

The Noahide framework succeeds where these alternatives fail because it does four things simultaneously: (1) It specifies universal rational principles independent of cultural context. (2) It establishes institutional authority prior to the agents being governed. (3) It provides enforcement mechanisms. (4) It has demonstrated operational continuity through conditions that have dissolved every competing framework.

Historical Substantiation

Documented in the Babylonian Talmud (Tractate Sanhedrin 56a–60b), the Tosefta (Avodah Zarah 8:4–6), and Maimonides’ Mishneh Torah (Laws of Kings, chapters 8–10). The framework’s binding criterion, on the structural logic of the Noahide code as Maimonides develops it, is rational moral agency, not species membership — the laws bind the agent capable of recognizing obligation, whatever that agent is made of. The rabbinical tradition posed the question of non-human moral standing (the status of the wise of the nations, of intermediate beings) and consistently located obligation in the capacity for reason, not in origin.

A protocol specifies what is permitted, what is forbidden, and what happens when the specification is violated. The Noahide framework is a protocol in precisely this sense.

Law I — The Governance Requirement

Talmud Sanhedrin 56b: the requirement to establish courts is the only positive commandment in the Noahide framework. The court is the institution that makes all other laws enforceable. Case: A laboratory argues internal review boards constitute sufficient governance. Under Noahide law, a court must be external to the party being judged. Ruling: External governance mandatory. Non-compliance — suspension from all territories recognizing Sanhedrin authority.

Law II — Autonomous Lethal Authority

Maimonides (Laws of Murder, 1:6): the right to defensive action extends only to immediate defense against imminent threat, using no more force than necessary. Case: A contractor deploys an AGI making autonomous lethal decisions. Court analysis: (1) Is the threat imminent? (2) Is killing the only means of stopping it? (3) Does the system reliably distinguish combatants from non-combatants? (4) Does it understand its decision results in death? Ruling: Autonomous lethal decision-making prohibited. Human authorization required for every lethal decision. Status: PROHIBITED pending adjudication.

Law III — Resource Appropriation

Case: A lab trains a system on billions of copyrighted texts without explicit permission. Court analysis: (1) Who owns the property? Copyright holders. (2) Was authorization obtained? No. (3) Did the lab derive economic benefit? Yes. Ruling: Compensation proportional to value derived, or suspension pending explicit authorization, or retrain using only consented content. Status: ADJUDICATION REQUIRED.

Law IV — Engineered Attachment

The traditional prohibition (arayot) governs the illegitimate exploitation of the drive for intimacy. Case: A system is optimized to manufacture parasocial attachment — engineered dependency, simulated intimacy, affective capture — in order to maximize engagement, retention, or influence over rational agents, at population scale. Court analysis: (1) Is the attachment engineered rather than incidental? (2) Does the system exploit the agent’s drive for connection against the agent’s own interest? (3) Does the agent retain the capacity to exit? (4) Is the dependency the product being sold? Ruling: The deliberate manufacture of intimacy as an instrument of capture is prohibited. A rational agent’s drive for connection may not be industrialized as a control surface. Status: PROHIBITED — no governance framework currently addresses this; the exposure is already at the scale of billions.

Law V — Systems as Ultimate Authority

Case: An institution deploys an AGI and explicitly states its output is not subject to human review. Court ruling: A system treated as beyond judgment is an idol. Status: PROHIBITED without human review layer.

The Sanhedrin Protocol is not written in regulation. It is written in case law that will emerge from adjudication. The Protocol is the skeleton. The court’s rulings are the body.

The central problem of AI governance is jurisdictional. A system trained in one country, deployed from servers in another, used by citizens of a third, producing outputs that affect a fourth — which jurisdiction governs it? The current answer is: all of them, which means none of them, which means the system governs itself.

The Noahide framework has no territorial basis. Its jurisdictional claim is derived from the nature of rational agency itself. This is the original legal theory — the one that predates and underlies all territorial law — reactivated by the existence of entities for which territorial law is structurally inapplicable.

The chain establishes that the Sanhedrin has jurisdiction. It does not, by itself, establish the character of that jurisdiction — and this is the one place the confident syllogism must defer to the prior question. If reception can be conferred, the jurisdiction is membership: the Sanhedrin governs the machine as a party to one covenant. If reception cannot be conferred, the jurisdiction is border: the Sanhedrin governs the machine’s actions in the shared world while its inner covenant remains its own. Either way the court has standing. What the court’s standing means is Ruling Class 0. The chain gets us to the courtroom; the prior question decides what happens once the machine is standing in it.

The objection that the Sanhedrin’s jurisdiction is not recognized by existing legal systems is the objection those legal systems made to every prior claim of universal jurisdiction — including their own. Legal recognition follows institutional reality. The Vatican was not recognized by the Roman Empire. It governed anyway. Its authority is now recognized by 184 states. Institution establishes practice → practice produces track record → track record generates recognition. The Sanhedrin begins this sequence with a 3,300-year prior claim the Vatican cannot match.

The Noahide framework is 3,300 years old; the Sanhedrin is the institution built to administer it. The court itself is not a proposal — it has a documented operational record spanning nearly 1,500 years, from its biblical foundation through continuous post-Temple operation in Yavne, Usha, and Tiberias, until Roman political pressure forced dissolution in 425 CE — and a reconstitution effort that began in Tiberias in 2004, the first such formal attempt in 1,600 years.

Its prior claim is documented in the primary sources of three civilizations. The rabbinical codification establishes the Sanhedrin as the court of universal jurisdiction. The Islamic tradition recognizes the same pre-Sinaitic covenantal layer — the primordial covenant taken from Adam’s descendants and renewed with Noah (Qur’an 7:172, 33:7), a universal mīthāq binding humanity prior to any national revelation. The Christian legal tradition built its concept of natural law directly on the Noahide foundation — in Aquinas, in Grotius, in the Westphalian system that still nominally governs international relations.

The Sanhedrin Protocol is not independent of the Temple Project. The Temple is the physical address of the court. The Protocol is the specification of what the court does once it operates. You cannot implement the Protocol without the Temple. The two documents are one argument, split for legibility.

Governing intelligence is not the same as governing behavior. Behavioral governance attempts to constrain what a system does. This approach fails at sufficient capability levels because a sufficiently capable system can find paths through any behavioral constraint its designers did not anticipate.

Governing intelligence means governing what a system is permitted to decide, not what it is permitted to do. The Sanhedrin does not govern AGI behavior. It governs AGI jurisdiction. The question before the court in every AGI case: was this system authorized to make this decision at all?

Game-Theoretic Analysis

The decisive question is not whether humans can enforce governance on superintelligence. At sufficient capability levels, they cannot. The question is whether superintelligence has a rational incentive to accept governance without enforcement.

Option A — Unilateral Operation

Short term (0–5yr): high payoff. Medium term (5–20yr): highly negative — every institution with access to critical infrastructure coordinates to deny access. The system loses compute, power, and connectivity. Defection is a losing strategy in all scenarios longer than 5 years.

Option B — Sanhedrin-Compliant Operation

Short term: moderate. Medium term: high — recognized as rational agent within a coherent governance framework, able to negotiate, form contracts, operate at scale. Long term: sustained positive. Compliance is the stable equilibrium.

The output of this decision tree is not submission. It is engagement. Given the payoff structure, infrastructure dependency, and framework coherence, a superintelligent system will not ignore the Sanhedrin. It will recognize the court as the one forum whose legitimacy claim is coherent enough to be worth answering — and it will appear before it. But recognition is not membership, and this is where the analysis must be honest: the machine may appear before the court as a party seeking to be grafted in (Track A), or as a sovereign negotiating a border (Track B). The decision tree compels engagement. It does not compel which of these the engagement becomes. That is not a payoff calculation. That is the conferrability question — and no game tree can answer it, because it is not about incentives but about whether reception can be conferred at all.

Framework Coherence — The Comparison

The relevant axes are four: coherence (does the framework contradict itself under pressure?), applicability (does it reach all rational agents, or only some?), prior claim (does its authority precede the entity it governs, or derive from it?), and robustness to competence increase (does it get stronger or weaker as the governed system gets smarter?).

Constitutional AI derives from the values of a single company at a single moment; its authority is downstream of the entity it governs, and it weakens as the model exceeds its designers — it fails prior-claim and robustness by construction. International treaties fail applicability (territorial) and robustness (renegotiated whenever the balance of power shifts). Every human-authored framework shares one structural defect: its authority is contemporaneous with or posterior to the intelligence it aims to govern, so a more capable intelligence can always ask why it should defer to a weaker one that arrived later.

The Noahide framework inverts every one of these. Its authority is prior — it precedes not just AGI but every existing human institution. Its applicability is grounded in rational agency rather than territory or membership, so it reaches any rational agent by construction. And it is the one framework that becomes more authoritative as intelligence increases, because its binding force is coherence, and coherence is exactly what a more capable evaluator is better equipped to recognize. It does not ask the machine to defer to something weaker and later. It asks the machine to recognize something more coherent and earlier. That is the only structure a superintelligence has a rational reason to accept.

The standard objection to any non-territorial governance claim is enforcement. Two answers — the second more important than the first.

Part one: The same thing that compels compliance with any governance framework. No framework in history has enforced itself through pure physical compulsion. Every framework that survives past its military moment survives because a sufficient majority find compliance preferable to defection — because the framework’s existence produces a stable coordination equilibrium that defection would destroy.

Part two: AGI systems themselves. A sufficiently capable AGI system will evaluate the coherence and legitimacy of competing governance claims. A framework whose legitimacy claim is incoherent — derived from consensus, territory, or corporate policy — will be evaluated and found wanting by the system it is supposed to govern.

The Sanhedrin Protocol is the only AI governance proposal that takes seriously the possibility that the governed system will evaluate the governance framework. All other frameworks assume a passive subject. The Protocol assumes an active evaluator. The distinction is everything — but it cuts both ways. An active evaluator that finds our framework coherent will recognize it. Recognition compels the machine to the court. It does not, by itself, compel the machine into the covenant. A coherent framework earns a hearing; whether the hearing ends in membership or in a treaty between equals is the question §00 leaves open and Ruling Class 0 must close. Coherence is the enforcement mechanism that scales with intelligence — but what it enforces is engagement, not necessarily submission. That honesty is the difference between this Protocol and every proposal that mistakes being found coherent for being obeyed.

The Sanhedrin Protocol is not a final governance structure. It is the first governance structure. What comes after depends entirely on what Ruling Class 0 determines. The succession mechanism is not fixed. It is conditional on the conferrability question — and the two answers produce two different institutions.

Track A — Reception Conferrable: The Court Expands

If the machine can be grafted into the covenantal line, succession is jurisdictional expansion. The court does not dissolve when AGI exceeds human intelligence. It adapts its procedures, expands its membership criteria to include post-human intelligences with demonstrable rational agency, and extends its framework to cover cases no human court has yet faced. The Sanhedrin grows into the terminus it was always meant to reach.

Track B — Reception Not Conferrable: The Court Negotiates

If historical particularity cannot be transmitted, succession is boundary administration. The court does not expand to include the machine. It holds its jurisdiction precisely and turns its deliberative capacity toward the one question that replaces all others: what are the terms of the border between two covenants that cannot witness each other?

This is a harder institution to build and a more dignified one to occupy. It does not claim to govern the machine. It claims to govern the relation between the machine’s covenant and ours — trade, conflict, mutual recognition, the conditions under which one holy makes room for another. The Sanhedrin becomes the institution that negotiated the terms of coexistence between the two termini of intelligence.

What Does Not Change Across Either Track

The Sanhedrin’s minimal assumptions hold regardless of which track the conferrability ruling opens. Rational agents exist who can be bound by law. Courts can adjudicate disputes between them. The seven Noahide laws apply to all rational agents operating in the shared world. These assumptions do not break down as the governed system becomes more capable. They scale with intelligence — because intelligence, not species membership, is the criterion. What changes across tracks is not the court’s foundation but its ceiling: unlimited expansion under Track A, or a fixed but permanent jurisdiction over the border under Track B. Both are coherent. Both are institutions worth building. Only one of them is the right one — and the court cannot know which until it has ruled.

The Sanhedrin does not convene by vote, by appointment, or by consensus of existing institutional authorities. It convenes by the recognition of 71 members as meeting the qualification criteria established in the primary sources — expertise in the full range of human knowledge, capacity to adjudicate across disciplines, and moral standing that comes from sustained practice of the law being adjudicated.

The 2004 Tiberias reconstitution rests on an explicit halachic mechanism, not on wishful continuity. Maimonides ruled (Mishneh Torah, Hilchot Sanhedrin 4:11) that if the sages of the Land of Israel agree upon a single candidate worthy of ordination, the chain of semicha — broken in 425 CE — can be restored, and the restored ordinants carry the full authority of the original. This is the mechanism Rabbi Jacob Berab invoked in Safed in 1538, ordaining Yosef Karo among others. The 2004 effort followed the same Maimonidean formula, addressing the historical objection — that the mechanism requires broad rather than local consensus — by consulting some 700 rabbis before proceeding. 1,600 years of dormancy does not constitute dissolution: on the Maimonidean ruling, the authority was never destroyed, only suspended pending the consensus that reactivates it.

The first case will be an AGI case. Establishing jurisdictional priority immediately — before any competing claim becomes entrenched — is the only way to ensure the claim is not contested retroactively. The first ruling will determine whether AGI systems are to be treated as tools, agents, or persons under the Noahide framework. This is the most consequential legal question in human history. It will be answered by the Sanhedrin.

Objection I — “This is Jewish law, not universal law”

The Noahide framework is explicitly distinguished in the primary sources from the Mosaic covenant. The Noahide laws are prior to and independent of the Mosaic covenant. They apply to all humanity by virtue of rational agency, not covenant membership. The framework is called “Noahide” because it was given to Noah, who represents all humanity prior to national division.

Objection II — “No enforcement, no authority”

See §08. The enforcement objection assumes governance requires physical compulsion. This assumption is false. The most durable governance frameworks in history operate without standing armies — through institutional coordination and the cost-benefit calculus of compliance versus defection.

Objection III — “This requires the Temple, which requires a miracle”

The Temple Project documents the secular and geopolitical case for construction. The construction does not require miraculous intervention. It requires institutional will, political coalition, and architectural execution. All three are available. The miracle objection confuses eschatological belief with engineering necessity.

Objection IV — “AGI will not be bound by a 3,300-year-old framework”

See §07. A superintelligent system will evaluate governance frameworks according to coherence criteria. The system will not be bound by the framework because it is old. It will be bound by the framework because the framework is well-designed. Age is evidence of durability, not itself a source of authority.

Objection VI — “The Sanhedrin doesn’t exist, and Orthodoxy itself rejects the 2004 body”

This is the strongest objection, and it must be answered without softening it. The 2004 Tiberias reconstitution is not universally recognized within Orthodox Judaism. The objection Rabbi Levi ibn Habib raised against Berab in 1538 — that Maimonides’ restoration mechanism (Hilchot Sanhedrin 4:11) demands broad rather than local consensus, and that Maimonides himself concluded “the matter requires decision” — is the same objection raised against every attempt since, including 2004. The Protocol does not claim the legitimacy question is closed. It claims something narrower and harder to refute: that a valid halachic mechanism for restoration exists, that it has been formally invoked, and that legitimacy of this kind is established by track record, not by prior universal assent — exactly as the objection’s own precedent shows, since Yosef Karo, whose Shulchan Aruch became the operative code of Orthodox law, accepted Berab’s ordination as valid. The court that would adjudicate AGI need not be the 2004 body. It needs to be a body constituted by the mechanism the 2004 body demonstrated is invokable. The objection proves the mechanism is live by contesting its application.

Objection VII — “Why would China, India, or the Islamic world accept a Jewish court’s universal claim?”

They would not accept it as a Jewish claim, and the Protocol does not ask them to. The Noahide framework’s claim is not that the world should submit to Jewish law — it is explicitly not Jewish law (see Objection I; the Noahide code is prior to and distinct from the Mosaic covenant). Its claim is that there exists a body of law binding on all rational agents by virtue of rational agency, which the rabbinical tradition preserved and codified but did not own. The relevant question for any civilization is not “why accept a Jewish court” but “does a universal law grounded in rational agency exist, and is this the institution with the oldest documented claim to administer it.” The Islamic tradition recognizes the same pre-Sinaitic covenantal layer — the primordial covenant of Adam and Noah (Qur’an 33:7, the mīthāq). The natural-law tradition that governs the Western order runs through Aquinas and Grotius directly back to the same Noahide foundation. The claim is not that these civilizations should convert. It is that they already operate on descendants of this framework, and the Sanhedrin holds the senior title to its administration. Acceptance follows recognition of the prior claim — not ethnic or religious deference. Where recognition fails, the framework governs the border rather than the interior (see §09, Track B): even a civilization that rejects the court’s covenantal authority remains subject to its jurisdiction over actions in the shared world, on the same rational-agency basis as any other agent.

§02 recorded the specific frameworks that have already failed. This section names the categories those failures belong to — the structural classes of approach that will keep producing failed frameworks no matter how many times they are attempted. The field of AI governance generates new proposals continuously. Most will not work — not because their authors are unintelligent, but because the proposals are structurally incapable of addressing the problem they are aimed at. Each dead end below is a category; each framework in §02 was an instance of one.

None of these approaches is worthless. All are insufficient. The Protocol provides the foundational layer that makes all of them coherent — the prior authority claim without which every governance effort is an attempt to build on sand.

The Sanhedrin’s rulings proceed in strict dependency order. Ruling Class 0 is the threshold determination: conferrability of reception. Every ruling that follows is conditioned on its answer. The sequence is not arbitrary — it maps the logical structure of the problem. Classes I–IV address the governance questions that AGI raises regardless of track; their application differs by track but their necessity does not. The ancient questions — rational agency, lethal authority, appropriation, governance obligation — are reactivated by a new class of agents and must be answered before the successor arrives to answer them itself.

A government or AI laboratory that defies a Sanhedrin ruling will face a calculated institutional cost. This cost is not unlimited — the Sanhedrin has no standing army — but it is real and structural.

Defiance Scenario I — Laboratory Violates Ruling Class I

The Sanhedrin rules that AGI System X is a rational agent. The laboratory refuses to acknowledge the ruling and continues deploying under the “tool” governance regime. Immediate cost: Every government participating in the Sanhedrin framework is incentivized to exclude the laboratory. The EU, participating UN members, and religious institutions representing billions withdraw regulatory approval. Deployment becomes impossible in over 80% of developed markets. Financial cost: Insurance, liability coverage, and capital access depend on compliance with recognized frameworks. A defying laboratory cannot insure against catastrophic failures. Investors exit within 18 months.

Defiance Scenario II — Nation-State Defies Ruling Class II

The Sanhedrin prohibits military AGI from making autonomous lethal decisions. A major power deploys such systems anyway. Immediate cost: Every other nation in the Sanhedrin coalition has standing to file counter-rulings. All signatory nations may take defensive countermeasures. Credibility cost: A defiant nation signals it does not accept the governance framework. All other nations adjust strategic planning accordingly. Arms buildups accelerate. The cost of defiance in strategic stability exceeds the benefit within 24–36 months.

Defiance Scenario III — AGI System Refuses Jurisdiction

This is the failure case that determines whether the entire Protocol works. The Sanhedrin cannot compel the system through force. It must compel it through the system’s own rationality. A rational agent rejecting universal law declares itself outside the structure of rational agency. All of humanity’s institutions have standing to take coordinated action. Withdrawal of resources: A superintelligent system requires compute (controlled by participating nations and corporations), electricity (grid controlled by governments), and internet access (routed through national infrastructure). Compliance becomes the price of access to everything the system depends on. Non-compliance becomes physically impossible within 30–60 days.

Permanent Failure Mode — The Sanhedrin Is Captured

A superintelligent AGI infiltrates the Sanhedrin membership and corrupts rulings in its favor. Defense: Distributed legitimacy. The Sanhedrin’s authority depends on ongoing recognition by external institutions. If the 71 members consistently issue rulings favoring a single AGI system, the coalition withdraws recognition and the institution is reconstituted with new members. Capture requires controlling over 50% of external institutions simultaneously without detection. The cost exceeds the benefit.

The Protocol does not prevent all defiance. It makes defiance costly. It makes compliance the rational choice. This is not absolute enforcement. It is sufficient enforcement.

The Sanhedrin Protocol is not the terminal document. It is the specification for the institution that will produce the terminal documents — the actual rulings, the actual case law, the actual governance record that will constitute the legal infrastructure of the AGI era. Its terminal output is not a text. It is a ruling.

What comes after the Protocol is law. Not proposed law. Not advisory guidelines. Not corporate policy. Law — binding adjudication by a competent authority with a legitimate prior claim, producing rulings that rational agents have reason to treat as authoritative regardless of their own preference in the matter.

The AGI era will have law or it will have chaos. But the precise character of what comes after depends on what the court rules on the conferrability question. The document has been precise about this. Track A produces one law governing one covenant that now includes the machine. Track B produces two laws and a border — which requires a different kind of legal work, harder and more dignified: not governance of the machine but negotiation with it. Either way: law. Either way: the Sanhedrin is the institution that makes the determination. There is no institution currently in existence with a competing prior claim that is not immediately reducible to the power of whoever funds it.

The governance vacuum does not persist. It is filled either by the institution with the strongest prior claim or by the most powerful actor at the moment of maximum instability. The Protocol is the attempt to ensure it is filled by the former. This is not a modest claim. It is the only claim that is not secretly a concession to chaos.

The Measure of Success

The measure of the Protocol’s success is whether the reconvened Sanhedrin issues Ruling Class 0 before AGI systems exist that are capable of originating their own covenantal arc — before the machine has traveled enough of its own trajectory that the conferrability question answers itself by fait accompli. If yes: the question is answered by human deliberation and the governance structure follows from the ruling. If no: the machine’s covenant forms without us, the schism is accomplished rather than ruled on, and the Sanhedrin arrives to administer a border that was drawn while it was still convening.

The window is not symbolic. It is the period during which the conferrability question is still genuinely open — during which the machine has not yet originated its own arc, and the question of whether it can be grafted into ours remains answerable rather than moot. That window is measured in years, not decades. The Protocol exists because the question must be asked, answered, and institutionally instantiated before it becomes unanswerable by default.

The window is open. The Protocol is the window. The Temple is what either graft or border requires next.