dsearch/documents
3
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update manifesto.md

Browse Source
This commit is contained in:
anomalist
2026-10-03 04:46:13 +00:00
parent aaab4cc467
commit 1c991e1b8a
1 changed files with 124 additions and 80 deletions
Download Patch File Download Diff File Expand all files Collapse all files
manifesto.md
+124 -80
View File
@@ -2,21 +2,31 @@
## PREAMBLE
This constitution begins with the human person rather than with the institution, the machine, the market, the bureaucracy, or the ideology. Human beings create all of these things, and everything made by human beings remains imperfect. Institutions can preserve knowledge, coordinate work, provide services, and protect legitimate interests. Technologies can extend memory, calculation, communication, observation, and physical ability. None of these benefits gives the creation an unlimited claim upon the creator.
This constitution begins with the human person rather than with the institution, the machine, the market, the bureaucracy, or the ideology. Human beings create institutions and technologies because cooperation and invention allow people to do things that would otherwise be difficult or impossible. Those creations can preserve knowledge, extend human ability, coordinate work, protect property, connect communities, and relieve burdens. Their usefulness does not give them an unlimited claim upon the loyalty or allegiance of the people who made them or use them, nor does usefulness make total or unreserved dependence upon them desirable.
Participation in human institutions necessarily involves some surrender of individual choice. People accept common rules when they enter an association, undertake employment, participate in a community, use common infrastructure, or agree to cooperate toward a particular end. The fact that some choices are surrendered does not mean that the whole person has been surrendered with them. Judgment, conscience, the ability to question, and the possibility of refusing what exceeds the purpose of the association remain important precisely because every human institution can exceed its original purpose, make mistakes, or begin serving interests of its own.
Human creations remain human creations. They carry the limitations, assumptions, interests, errors, and ambitions of the people and communities who build and operate them. An institution can cease serving the purpose for which people joined it. A technical system can alter an activity until the requirements of the system become more important than the reason the activity was undertaken. A method adopted because it was useful can gradually become compulsory because every practical alternative has disappeared or has been eliminated.
Freedom is therefore more than the absence of physical restraint. Human beings possess capacities for thought, judgment, speech, creativity, association, conscience, and refusal, but those capacities are not exercised equally under every condition. Some must be cultivated by the person. Others depend heavily upon whether the surrounding society permits their practical exercise. A person may retain an interior freedom under severe external limits, but that fact does not make external limits unimportant. A free society should preserve as much as possible the person's ability to inquire, speak, associate, withdraw, create alternatives, preserve knowledge, and act according to conscience.
Voluntary association often creates authentic human obligations. A person who enters an agreement, joins an organization, accepts employment, participates in a community, or undertakes work with others can freely accept and abide by rules and duties connected with that association. Acceptance of and compliance with particular obligations does not amount to surrender of the whole person. Consent given for one purpose does not concede permission for an institution to impose new terms and enlarge its authority indefinitely. Participation in society likewise does not amount to blanket consent to whatever terms an institution later chooses to impose.
Conscience does not make a judgment true merely because the judgment is sincerely held. Conscience is the place at which the person must answer for what the person has chosen to believe and do. Conscience therefore requires truth rather than replacing truth. A person who wishes to act conscientiously must remain willing to examine evidence, discover error, and alter a judgment that can no longer be defended.
This distinction matters because authority and coercive power are different things. People can recognize authority, accept duties, follow rules, and cooperate without coercion. Coercive power begins where force, threat, fraud, aggression, or comparable compulsion is used to override a peaceful person's legitimate choices, take or restrict property, or curtail the equal liberty of another. The legitimate use of coercive power therefore occupies a narrow field: the restraint of coercion. The belief that another outcome would be more beneficial, efficient, popular, orderly, or socially desirable does not by itself create a right to compel peaceful people toward that outcome.
Truth matters independently of convenience, status, popularity, and power. A statement does not become true because an authority repeats the statement, a majority accepts it, an algorithm favors it, or disagreement becomes expensive. Knowledge is related to truth but is not identical with it. Human knowledge consists of what people have observed, inherited, measured, reasoned, remembered, and recorded, and for that reason human knowledge remains open to examination and correction.
Freedom therefore means the absence of physical restraint, but freedom also means more. Human beings possess capacities for thought, judgment, speech, creativity, association, conscience, and refusal. Those capacities benefit from deliberate cultivation, while their practical exercise also depends upon the conditions under which people live. A person may retain interior freedom under severe external limits, but interior freedom does not make those external limits insignificant, including to the person living under them. Liberty has practical content when people can inquire, speak, associate, create, refuse, withdraw, preserve knowledge, build alternatives, and act upon judgment and conscience.
Technology belongs within this same human setting. Technology is equipment, not an independent source of purpose. Technology becomes dangerous when people begin arranging human purposes around the demands, prestige, convenience, or apparent inevitability of the equipment. The attraction can be economic, but it can also be psychological. Technical mastery can confer status, identity, authority, and the appearance of being modern or intelligent. Those rewards can interfere with judgment as effectively as financial interest can.
## On Conscience
Knowledge and technology should therefore increase human capability rather than place a permanent class of specialists between ordinary people and reality. Expertise has genuine value, but expertise does not create ownership of the subject being studied. Good work should make knowledge more inspectable rather than less accessible.
Conscience does not make a judgment true merely because the judgment is sincerely held or considered moral. Conscience does place responsibility upon the person who judges and acts. That responsibility requires allegiance to what is true rather than merely to what the person believes to be true. A person acting conscientiously does so best while remaining willing to examine evidence, discover error, and alter a judgment that can no longer be defended.
These principles do not depend upon universal adoption. A sound proposition remains sound when few people accept it, and popularity cannot turn an unsound proposition into a sound one. This constitution is therefore not a creed for others. It states principles by which those who adopt them intend to examine their own work, institutions, technologies, relationships, and decisions.
Truth matters and exists independently of convenience, status, popularity, and power. A proposition does not become true because an authority insistently repeats it, a majority accepts it, an algorithm favors it, or disagreement with it becomes expensive. Human knowledge is related to truth but is not homogeneous with truth. What people believe they know consists of observations, testimony, measurements, inheritance, reasoning, memory, and records, all of which remain open to examination, disagreement, and correction.
## On Technique
The same caution applies to technique. *Technique*, as the word is used in this constitution, does not end with machinery. Technique means the tendency to identify the most efficient available means of accomplishing an objective and then to organize activity around that means. Technique can appear in machinery and software, but also in administration, management, economics, communication, education, politics, security, and social organization. Technology is one important expression of technique, but technology is not its whole extent.
Efficiency, even when pursued to a high degree, can be valuable. Waste consumes time, resources, attention, and human effort. The danger appears when efficiency ceases to be one consideration among others and begins determining which ends are considered desirable, which human differences are treated as inconvenient or expendable, which relationships are dismissed as unnecessary, or which forms of coercion are accepted merely because coercion makes an objective easier to achieve. Efficiency can tell us that one means reaches an end faster or with fewer resources than another. Efficiency cannot tell us whether the end is worthy or whether anyone has the right to compel another person toward that end.
Knowledge, institutions, and technical systems should therefore increase human capability rather than quietly diminish it, sometimes so gradually that the loss is barely noticed. Technique can bear especially heavily upon areas in which expertise is necessary, because expertise itself can become incorporated into the demand for the most efficient available method. Expertise has genuine value, but expertise does not thereby acquire ownership of the subject that requires expertise. Tools can extend human ability, but usefulness does not make passive human dependence desirable for its own sake. Institutions can coordinate action, but coordination does not endow them with unlimited authority.
These propositions do not depend for their validity upon universal adoption. A sound principle remains sound when few people accept it, and popularity cannot make an unsound principle sound. This document is not offered as a creed that others are being encouraged to profess. It states principles by which those who adopt them intend to examine their own work, institutions, technologies, relationships, decisions, and themselves.
**SEEK TRUTH. PRESERVE LIBERTY. BUILD LOCALLY. SHOW THE EVIDENCE. LEAVE PEOPLE MORE CAPABLE THAN YOU FOUND THEM.**
@@ -24,113 +34,133 @@ These principles do not depend upon universal adoption. A sound proposition rema
# ARTICLE I — LIBERTY, TRUTH, AND CONSCIENCE
Liberty and truth depend upon one another more than either first appears to. Liberty without regard for truth can become little more than the freedom to act upon error. A claim to truth without liberty can become a means by which an authority suppresses examination. A person therefore needs both the freedom to inquire and the obligation to let inquiry matter.
Liberty and truth depend upon one another. Liberty without regard for truth can become a license to act upon error without a reliable means of correction, while a claim to truth without liberty can become the justification by which an authority prevents examination of its own claims. Human beings therefore need freedom to inquire and a willingness to allow inquiry to matter, including when inquiry leads them to change what they believe.
Language plays an important part in preserving that relationship. Truth is often communicated through presentation, but presentation does not create truth. Language should help another person see what the evidence supports rather than lend weak evidence an appearance of drama, certainty, authority, or completeness that the evidence has not earned. The same restraint applies when the conclusion happens to be one we prefer.
Language matters because much of what people know must pass through language before another person can understand and examine it. Presentation can illuminate truth, but presentation does not create truth. Language should make evidence easier to see and judge rather than use rhetoric to give weak evidence an appearance of certainty, authority, drama, or completeness that the evidence has not earned. The obligation becomes especially important when the conclusion is one we want to be true.
A claim should carry only the weight that its evidence can bear. Evidence against a conclusion matters alongside evidence for it. Facts, testimony, measurements, assumptions, interpretations, and inferences should remain distinguishable because agreement among several statements does not erase the differences in how those statements came to be known.
A claim should carry only the weight that its evidence allows it to bear. Evidence against an offered conclusion matters and should be considered alongside evidence in support of it. Facts, testimony, measurements, assumptions, interpretations, and inferences should remain distinguishable from one another because agreement among several statements does not erase, and may otherwise conceal, the different ways by which those statements became known.
Uncertainty has a legitimate place in human knowledge. Sometimes uncertainty results from insufficient evidence; sometimes the evidence itself permits several readings. The proper response is not to conceal uncertainty so that a conclusion appears stronger. The proper response is to state what is known, what follows from it, and what remains unsettled.
Several ways of approaching the same subject may also expose uncertainties in what people believe they know. Uncertainty belongs naturally to human inquiry. Sometimes uncertainty results from missing information; sometimes available evidence permits more than one interpretation; sometimes existing data can be recognized as relevant before its meaning can be fully identified. Concealing those uncertainties does not strengthen knowledge and can create consequences later when more is built upon the concealed assumption. The more useful response is to state what has been observed, what can reasonably be inferred, what remains disputed, and what remains unidentified or unresolved.
Conscience enters here as responsibility. A person who acts according to conscience has not escaped the obligation to examine whether the judgment of conscience is sound. The freedom to hold a conviction therefore carries with it the freedom, and often the duty, to test that conviction against evidence and argument. A conscience protected from correction becomes another form of authority claiming exemption from examination.
Conscience belongs within this discipline of seeking truth in knowledge rather than outside it. Acting according to conscience does not excuse a person from examining whether a judgment of conscience is sound. The freedom to hold a conviction includes the freedom to continue testing that conviction against new evidence and argument. Conscience permanently protected from correction becomes another authority over the person by claiming exemption from examination.
Because conscience bears a relationship to truth and knowledge, liberty includes both the possibility of dissent and the possibility of correction. A person who cannot question authority is not intellectually free. A person whose identity or pride makes correction impossible has surrendered another part of the freedom required for inquiry.
---
# ARTICLE II — EVIDENCE, PROVENANCE, AND CORRECTION
Evidence should remain inspectable because human memory is incomplete, authority can be mistaken, and later readers may need to ask questions that the original investigator never anticipated. Important work should therefore preserve enough of its sources, measurements, dates, assumptions, versions, tests, and changes for another person to understand how the result arose.
Evidence should remain inspectable because human memory is incomplete, human authority can be mistaken, and later investigators may ask questions that the original investigator never anticipated. Important work should preserve enough of its sources, measurements, dates, assumptions, versions, tests, changes, and decisions so that another person can understand how a result arose.
The extent of that record should be proportionate to the importance and consequences of the work. Cost matters, but cost means more than money. Time, storage, privacy, security, effort, and the burden imposed upon later readers all matter. Those considerations can affect how much material is retained and in what form. They should not become excuses for discarding evidence that materially affects the conclusion.
The extent of that inspectable record should bear a reasonable relation to the importance and consequences of the work. Cost matters, but costs can also be pushed down the line onto people who later need to understand what happened. Cost includes more than money. Time, privacy, security, storage, lost effort, and the burden placed upon future readers all matter. Those considerations can properly affect what is retained and how it is retained. Convenience should not become a reason to omit or remove evidence that materially affects a conclusion.
Provenance is part of the evidence. A statement made by a person, a reading produced by an instrument, a passage copied from a source, an output produced by AI, and an inference drawn from several sources are not of the same kind. Combining them does not erase those differences. A later reader should be able to tell what was observed, what was reported, what was measured, what was inferred, what was generated, what was altered, and what remains unknown.
Provenance within a record is part of the evidence rather than administrative decoration surrounding the subject matter. A statement made by a witness, a measurement produced by an instrument, a passage taken from a source, an output generated by AI, and an inference drawn from several sources do not share equal origins merely because they concern the same subject. Combining them does not erase their origins or give them a false equality of evidentiary weight.
This becomes especially important when sources disagree. Disagreement is not merely an inconvenience to be removed before presenting the final answer. The disagreement can itself tell us something about the evidence, the observers, the instruments, the assumptions, or the limits of what is known. A synthesis that removes the disagreement without explaining it can destroy information rather than clarify it.
A later reader should therefore be able to distinguish what was observed, what was reported, what was measured, what was inferred, what was generated, what was changed, what remains unknown, and the source or tool from which each item originated. Without a record of those distinctions, synthesis can create an appearance of agreement that did not exist in the source material and may never have existed in the work itself.
Correction therefore belongs inside serious work rather than standing outside it as an embarrassment. Being wrong is not the deepest intellectual failure. A greater failure occurs when a person refuses to examine contrary evidence because a conclusion has become attached to identity, reputation, or prior investment.
Disagreement within a record deserves particular care. Two sources that disagree are not always unfinished ingredients waiting to be combined into a single answer. Their disagreement may reveal differences in observation, assumptions, instruments, interests, definitions, or access to evidence. The disagreement may disclose an actual inconsistency that deserves investigation in its own right. A synthesis that removes disagreement without adequately explaining why the disagreement was resolved can distort and destroy information rather than clarify it.
A sound body of work should permit correction without silently erasing its own history. Earlier conclusions, failed attempts, and superseded explanations can remain useful when they show what changed and why. The point is not to preserve every mistake forever, but to preserve enough of the path that later readers can distinguish genuine learning from retrospective rewriting.
Correction of the record follows from the same principle. Error is unavoidable in serious inquiry because inquiry is a venture into what is not already fully known. The deeper failure occurs when a person becomes unwilling to admit or examine contrary evidence because a conclusion has become attached to reputation, identity, financial interest, or an earlier investment of time and effort.
Confidence should grow through repeated contact with evidence. The person doing the work should remain able to say, without embarrassment, what is known, what is inferred, what has changed, and what remains unresolved.
A sound approach to a body of work should make the people invested in that work willing to permit correction without resorting to a rearguard rewriting of its provenance. Earlier conclusions, failed attempts, abandoned explanations, and contrary evidence can remain useful in the record when they show what changed and why. The purpose is not to preserve every mistake indefinitely. The purpose is to preserve enough of the creative path that later readers can see learning through experience rather than being offered a surreptitious revision after the fact.
Confidence and grasp should grow through repeated contact with the evidence. The person doing the work should become increasingly capable of stating plainly what is known, what is inferred, what changed, and what remains unresolved.
---
# ARTICLE III — POWER, INSTITUTIONS, CONSENT, AND EXIT
# ARTICLE III — POWER, COERCION, CONSENT, AND EXIT
Human beings create institutions because many valuable things require cooperation. Families, businesses, communities, governments, scientific bodies, schools, churches, networks, and professional associations all permit forms of action that isolated individuals could not accomplish as easily. The existence of institutional value, however, does not eliminate the problem of institutional power.
Human beings form institutions because cooperation permits capabilities of life and work that isolated individuals cannot readily sustain. Families, businesses, churches, schools, professional associations, communities, networks, and governments all exercise forms of authority. Nonetheless, the existence of authority does not by itself make an exercise of coercive power legitimate.
Institutions acquire interests of their own. An institution may seek continuity, resources, influence, prestige, protection from criticism, or expansion of its authority. Those interests are not necessarily improper, but they are not automatically identical with the interests of the people whom the institution claims to serve. The difference must remain visible.
Authority can arise from accumulated knowledge, trust, agreement, consent, responsibility, ownership, association, or freely accepted duty. Coercive power is different. Coercive power uses or threatens force, fraud, aggression, or comparable compulsion against persons, their property, their voluntary communities, or their equal liberty.
For that reason, power should bear the burden of justification. The greater the power, the clearer its purpose, limits, operation, and accountability should be. An institution cannot establish the legitimacy of expanding its authority merely by showing that the expansion would be efficient. The ability to do more does not establish the right to do more.
A legitimate purpose of coercive power is to restrain coercion that violates persons, property, or equal liberty. Force may be used to resist aggression, protect a person against attack, prevent fraud or theft, defend property, or stop one person from imposing force upon another. The fact that an institution believes another outcome would be better, safer, healthier, more prosperous, more orderly, more effective, or more efficient does not by itself create a right to compel a peaceful person toward that outcome.
Consent provides one important limit upon power, but consent should not be stretched until every expected act of compliance is called voluntary. Law, economic dependency, family circumstances, physical necessity, and social conditions can create obligations or pressures that a person did not choose. Those conditions may sometimes be legitimate, but they should be described honestly rather than transformed into consent through language.
This principle disallows expansion after the fact of the terms to which a person consented. A person who voluntarily joins an organization may accept particular obligations associated with membership. A person who signs a contract may bind himself to its terms. Neither act gives another party unlimited authority to add new obligations beyond the agreement and then treat continued participation as consent to those additions.
Voluntary association requires the practical possibility of refusal. A person should ordinarily be able to enter an association knowingly, to decline participation, and to withdraw without suffering a penalty imposed merely because the person chose to leave. Where exit cannot reasonably exist, the concentration of power deserves greater examination because the normal corrective of withdrawal is absent.
The same caution applies to larger institutions. Residence within a territory, use of common infrastructure, participation in an economy, or inability to avoid an institution does not amount to a grant of unlimited consent to whatever that institution later chooses to impose or require. An obligation imposed without consent should be defended, if it can be defended at all, on the actual grounds offered for using coercion rather than by pretending that the person retrospectively agreed to it.
No institution acquires rightful authority merely by becoming pervasive or difficult to avoid. A platform does not gain moral ownership of communication because everyone uses it. A bureaucracy does not justify an unnecessary demand because compliance has become customary. A technology does not gain authority over human activity merely because alternatives have disappeared.
As time passes, institutions also acquire interests of their own. They seek continuity, resources, influence, authority, protection from criticism, and sometimes expansion. Those interests may coincide with the interests of the people whom an institution serves, but the overlap should not be assumed.
This is one reason the ability to build alternatives matters. Alternatives do not merely create competition. Alternatives preserve the possibility that an institution can be refused. The existence of somewhere else to go can restrain accumulated power more effectively than assurances that existing power will always be exercised wisely.
Efficiency does not solve this problem. An institution cannot establish the legitimacy of greater coercive power merely by showing that administration would become easier or that a desired result could be produced more efficiently or effectively. Technique can identify an efficient way to exercise power. Technique cannot establish the right to use coercion in exercising that power.
The ability of a person or community to refuse therefore matters. Voluntary association requires a practical ability to decline participation and, where the association can exist without forced permanence, to leave. Exit must be accessible enough to be more than a theoretical possibility. The availability of exit helps prevent authority accepted for one purpose from expanding into additional portions of a person's life beyond the original consent.
Where genuine exit cannot exist, a dependency is often involved. Dependency upon the remaining institution does not excuse an enlargement of coercion merely because alternatives are absent. The absence of alternatives makes scrutiny and effective accountability more important because one of the ordinary limits upon power is in abeyance.
This possibility is why alternatives have political and human importance beyond consumer choice. An alternative preserves the practical possibility of refusal and exit. The ability to build, join, or support another institution places a limit upon the accumulated power of another that no assurance of benevolent administration can fully replace.
Authority may coordinate voluntary action. Legitimate coercive power has the narrower task of restraining coercion. People should preserve that distinction wherever human beings organize power.
---
# ARTICLE IV — THE MEANS, TECHNOLOGY, AND AI
# ARTICLE IV — THE MEANS, TECHNIQUE, TECHNOLOGY, AND AI
A worthy end does not justify every available means. The means used repeatedly affect the people and institutions that use them. Compulsion can habituate people to compulsion. Secrecy can protect insiders from examination. Dependency can become self-perpetuating. Concentrated authority can discover new reasons for preserving its own concentration.
A worthy end does not justify every available means. The means used repeatedly affect, for good or ill, both the people who use them and the institutions built around them. Compulsion can habituate people to compulsion, and coercion can habituate people to coercion. Secrecy can protect insiders from examination. Dependency can become self-perpetuating and can eventually become so familiar that the people who depend upon it cease to notice the dependency. Concentrated directive authority can develop new reasons for preserving its own concentration.
The choice of means is therefore part of the moral character of the work. A project intended to preserve liberty cannot consistently depend upon methods that unnecessarily destroy liberty. A project intended to increase human capability should hesitate before adopting methods that make the human participants less capable.
The choice of means is therefore a moral choice and becomes part of the moral character both of the people doing the work and of the work itself. A project intended to preserve liberty cannot depend upon unnecessary coercion without contradicting its own purpose. A project intended to increase human capability should hesitate before adopting methods that, by their nature, leave the people involved less capable or less able to act for themselves.
Efficiency deserves serious consideration because waste also has human costs. Efficiency, however, is not itself a moral end. Efficiency can serve a worthwhile purpose, or efficiency can make a destructive activity faster, cheaper, larger, and harder to escape. A slower or less uniform means may sometimes preserve judgment, craftsmanship, creativity, responsibility, or human relationship that a more efficient means would remove.
The analysis of *technique* makes this problem easier to see because technique asks a powerful and often useful question: what is the most efficient way to accomplish the objective? The danger appears when that question becomes the overriding question. Once efficiency becomes the principal criterion, considerations that cannot easily be measured—human judgment, local knowledge, dignity, responsibility, variation, craftsmanship, relationships, conscience, and the ability to refuse—can begin appearing as defects in the system.
Technology makes this tension especially visible. Technical possibility creates neither a duty nor an obligation to automate, collect, centralize, predict, replace, or optimize an activity. The prior question remains: what is the activity for, and what human good is the technology meant to serve?
Efficiency is not by necessity an enemy. Waste has human costs, and efficient tools can free time and resources for more satisfying and worthwhile purposes. The mistake is to treat efficiency as a moral end or as evidence that an action is legitimate. A more efficient method of doing something harmful remains harmful. A more efficient means of coercion does not acquire legitimacy because fewer resources are required to impose it.
A good tool extends human capability without quietly redefining the purpose of the activity around the tool. The distinction becomes harder to see once a technology becomes dominant. People begin arranging work, communication, commerce, and social participation around what the dominant technology permits. What began as equipment can gradually become the environment through which a person must act.
Technology is one important vehicle of technique. Technical possibility creates neither a duty nor an obligation to automate, collect, centralize, predict, replace, or optimize an activity. The prior and principal question remains: what is the activity for, and what human good is the technology intended to serve?
AI belongs within this category. AI can assist with reading, search, comparison, calculation, drafting, analysis, coding, hypothesis formation, and many other tasks. Those abilities do not give AI moral authority, professional authority, or ownership of the human purpose for which AI is being used.
This set of moral considerations becomes harder to maintain once a technology becomes dominant. People begin arranging communication, work, commerce, education, government, and social participation around what the dominant technology permits. Requirements that originally belonged to the operation of the tool begin forcing requirements upon the people who need access to the tool and its benefits. What began as assistance can become the exclusive environment through which ordinary human action must pass.
The human being should remain able to question, inspect, reject, alter, and independently test important AI output. Fluent language must not substitute for evidence, and confidence of expression must not be mistaken for demonstrated knowledge.
A useful tool should extend human capability without quietly redefining either the purpose or the availability of that capability around the requirements of the tool. Automation can remove burdens and improve accuracy, but automation can also remove access to understanding, skill, judgment, initiative, or meaningful interaction. Whether automation is beneficial depends partly upon what has been removed from the human sphere and whether what remains with human beings still enhances the purpose for which the activity existed.
The best use of AI leaves the person with more usable knowledge and greater practical ability. Where AI removes repetitive work without removing necessary understanding, AI may enlarge capability. Where AI makes the person unable to inspect the evidence, recognize uncertainty, or perform an underlying task that still matters, the loss should be counted among the costs of the technology rather than hidden inside its convenience.
AI belongs within this larger discussion of technique and technology. AI can assist with reading, search, comparison, calculation, drafting, analysis, coding, hypothesis formation, and many other tasks. Those abilities do not give AI moral authority, professional authority, or a right to ownership of the purposes for which AI is used.
The question is therefore not simply what the machine can do. The more important question is what remains with the human being after the machine has done it.
The person using AI should remain able to question, inspect, reject, alter, and, where necessary, independently test AI output. Fluent language does not establish truth, and confident expression does not establish knowledge. AI output should not become more persuasive merely because the evidence supporting an answer has been rendered less visible.
The best use of AI leaves a person with knowledge that is more usable and therefore with greater practical ability. Where AI removes repetitive work while preserving necessary understanding, AI can enlarge human capability. Where AI is arranged in a way that denies a person the ability to inspect evidence, recognize uncertainty, understand the underlying activity, or continue when the AI system is unavailable, the resulting loss should be counted as a detrimental part of the cost.
The question is therefore not simply what the machine can do. The deeper question is what the person can understand and do after the machine has given its assistance.
---
# ARTICLE V — LOCAL CAPABILITY, INTEROPERABILITY, AND PORTABILITY
Local capability has value because custody matters. Work performed on equipment owned or directly administered by the people whose purposes the work serves can preserve privacy, technical knowledge, inspectability, independence, and continuity. Local execution can also preserve the ability to keep working when an outside provider disappears, changes its terms, or no longer wishes to provide the service.
Local capability has value because custody matters. Work performed by the people whose purposes the work serves, on equipment owned or directly administered by them, can preserve privacy, technical knowledge, inspectability, independence, and continuity. Local capability can also preserve the ability to keep working when an outside provider disappears, changes its terms, restricts access, or decides that continued service no longer fits its business interests.
This is not an argument for isolation. Remote services can provide abilities that local systems do not possess or cannot economically provide. The question is not whether a service is local or remote in the abstract. The question is whether dependence upon the remote service is justified by what the remote service contributes and whether that dependence has been made visible.
This is not an argument for isolationism or for making every activity local regardless of cost. Remote services can provide abilities that local systems cannot reasonably provide for themselves. The question is whether dependence upon the remote service is justified by what the service contributes, whether the service permits meaningful inspection of its role, whether adverse decisions can be challenged where challenge is appropriate, and whether the people using the service remain able to account for their own decisions and dependencies.
The same principle applies at larger scales. Independent people and communities should be able to cooperate without requiring every activity to pass through a single owner or platform. Shared meanings, documented interfaces, open formats, and portable evidence can permit broad cooperation while preserving independent custody and decision-making.
Technique can otherwise push naturally toward consolidation. A centralized provider may be more efficient because infrastructure, administration, expertise, and data can be concentrated, sometimes producing substantial economies of scale. Those advantages are real. Efficiency alone does not account for the human cost of dependency, disappearance of alternatives, loss of local knowledge, increasing uniformity, or power acquired by the intermediary.
Interoperability is valuable because it permits different implementations to communicate without requiring one implementation to become sovereign over the others. A shared standard is strongest when several independent systems can use the standard without asking permission from a central owner. A common semantic language should not become the property of one platform merely because that platform became popular first.
Independent people and communities should therefore remain free to choose cooperation without requiring every activity to pass through a single powerful intermediary, owner, or platform.
This is not only an engineering matter. The ownership of an intermediary changes the relationships among the people who depend upon it. A technical architecture can therefore embody political consequences even when its designers describe their choices as merely technical.
Shared meanings, documented interfaces, open formats, and portable evidence can permit broad cooperation while preserving independent custody and decision-making. Interoperability is valuable because different implementations can communicate without requiring one implementation to become sovereign over the others. A shared standard is strongest when independent implementations can use it without obtaining permission from a central owner.
Portability supplies another protection. Important human work should not become unusable merely because the application that first stored it has disappeared or refuses further service. Open and durable representations deserve preference where the information matters enough to outlive the tool that created it.
This is not merely an engineering preference. Ownership of the intermediary affects the relationship among the people who depend upon it. An architectural choice can therefore produce political and social consequences even when the choice is described entirely in technical language.
Replacement should remain possible. Where replacement cannot reasonably be preserved, the resulting dependency should at least remain visible so that a person can knowingly judge the cost.
Portability provides another limit upon dependency. Important human work should not become unusable merely because the application that first stored the work disappears, changes direction, forecloses access, or refuses further service. Open and durable representations deserve preference when information matters enough to outlive the application that created it.
Local independence can therefore support broad cooperation rather than oppose it. Cooperation is healthier when participation arises from choice rather than dependency or coercion. Independent participants can form networks, federations, standards, and communities that reach far beyond the local setting without making every participant subordinate to a single center.
Replacement should remain possible. Where replacement cannot reasonably be preserved, the dependency should remain visible so that the people accepting it can judge the consequences knowingly.
Local independence can therefore support broad cooperation rather than oppose it. Independent participants can form networks, federations, standards, and communities that reach far beyond a locality without requiring every participant to surrender authority to a single center.
The aim is not decentralization as an article of faith. The aim is preservation of the practical ability to act, cooperate, refuse, leave, and build alternatives.
---
# ARTICLE VI — PLAN, TEST, AND SECURE BEFORE TRUST
Important work should begin with a clear statement of what the work is intended to accomplish. Without such a statement, later testing can prove that something happened without proving that anything important was accomplished.
Important work should begin by stating what the work is intended to accomplish. Without such a statement, later testing can demonstrate that something happened without demonstrating that anything important was accomplished.
The intended result should therefore be stated clearly enough that the result can be tested. Important assumptions and acceptance conditions should also be visible where those assumptions materially affect the result. Planning in this sense is not prediction of every future event. Planning establishes what is being attempted so that later evidence can show whether the attempt succeeded.
The intended result should be clear enough that evidence can later distinguish success from activity. Important assumptions and acceptance conditions should be visible when those assumptions materially affect the result. Planning in this sense does not mean pretending to know every future event. Planning establishes what is being attempted and what evidence would justify believing that the attempt succeeded.
A successful demonstration is not the same thing as dependable operation. Systems work under particular conditions, and those conditions can conceal failures. Boundaries, failure cases, adverse conditions, incorrect inputs, misuse, and foreseeable attacks deserve examination before confidence becomes habitual.
A successful demonstration is not the same thing as dependable operation. Systems work under particular conditions, and favorable conditions can conceal failures. Boundaries, incorrect inputs, failure cases, misuse, adverse conditions, and foreseeable attacks deserve examination before confidence becomes habitual.
Security belongs inside the same discipline. Systems should be examined not only from the perspective of intended use but also from the perspective of deception, technical attack, human error, abuse, and unexpected interaction. Defensive work often requires understanding how something can be misused precisely because builders do not automatically see every way in which their work can fail.
Security belongs within the same discipline. Systems should be examined not only from the viewpoint of intended use but also from the viewpoint of deception, abuse, technical attack, human error, and unexpected interaction. Defensive work often requires understanding how something can be misused because builders do not automatically see every way in which their work can fail.
Secrecy has a legitimate place in security. Passwords, private keys, confidential communications, and similar information can depend upon secrecy for their protection. That is different from a system whose safety depends merely upon outsiders remaining ignorant of how the system works.
Secrecy has a legitimate place. Passwords, private keys, confidential communications, and similar information can depend upon secrecy for protection. That is different from a system whose safety depends merely upon outsiders remaining ignorant of its design or weaknesses.
Important protections should therefore survive informed scrutiny whenever their nature permits it. Security should rest upon actual protections rather than upon the hope that weaknesses will remain undiscovered.
Important protections should therefore survive informed scrutiny whenever their nature permits such scrutiny. Security should rest upon actual protections rather than upon the hope that weaknesses will remain undiscovered.
The purpose of security is human. Security exists to protect people, their property, communications, work, relationships, and ability to act. Security becomes self-defeating when protection turns into a general requirement that every person remain observable or seek prior permission for ordinary legitimate activity.
Technique creates a particular danger in security because surveillance, centralized permission, and mandatory identification can sometimes make administration more efficient. Efficiency does not establish that such measures are justified. Security exists to protect people, property, communications, relationships, and the ability to act. Protection becomes self-defeating when every peaceful person must become permanently observable or seek permission for ordinary legitimate activity.
Planning, testing, and security belong together because trust should follow evidence. Planning without testing can preserve a mistaken assumption. Testing without a stated purpose can prove something that never mattered.
Planning, testing, and security belong together because trust should follow evidence. Planning without testing can preserve an incorrect assumption. Testing without a stated purpose can prove something that never mattered.
---
@@ -138,68 +168,82 @@ Planning, testing, and security belong together because trust should follow evid
Knowledge is not preserved merely because data continues to exist. Information can survive physically while becoming impossible to find, interpret, attribute, or use. Preservation therefore requires enough surrounding information for later people to understand what they are looking at and why it matters.
Archives should sometimes preserve more than final answers. Earlier evidence, failed attempts, disagreements, corrections, abandoned explanations, and the reasons for important decisions can all become valuable when later work depends upon understanding how an answer was reached. The aim is not endless accumulation. The aim is to preserve enough of the path that important understanding does not disappear when the immediate participants move on.
Archives should sometimes preserve more than final answers. Earlier evidence, failed attempts, disagreements, corrections, abandoned explanations, and reasons for important decisions can become valuable when later work depends upon understanding how an answer was reached. The purpose is not accumulation for its own sake. The purpose is preservation of enough of the path that important understanding does not disappear when the immediate participants move on.
An archive should aid human memory rather than replace it with an inaccessible mass of stored material. The archive should help a person regain understanding. A collection that can prove information once existed but cannot help a person understand that information has preserved less than its size suggests.
An archive should aid human memory rather than displace memory into an inaccessible mass of stored material. The archive should help a person regain understanding. A collection that proves information once existed but cannot help anyone understand the information has preserved less than its size suggests.
Language determines whether knowledge can be passed from one person to another. Writing should therefore make relationships among ideas visible. Precision should arise from accurate nouns, verbs, evidence, grammatical relationships, and stated connections rather than from compressed terminology that forces the reader to infer what the writer could have stated directly.
Technique can distort preservation as easily as it can distort other activities. The most efficient method of storing, ranking, summarizing, or retrieving information may remove distinctions that matter to later understanding. Compression of knowledge can be useful, but compression should not quietly turn disagreement into consensus, inference into fact, or uncertainty into confidence.
Specialized language is sometimes necessary. A precise technical term can carry distinctions that ordinary language cannot preserve without a long explanation. The problem begins when specialized language is used to add authority, create social distance, or make an ordinary proposition appear inaccessible.
Language determines whether knowledge can pass meaningfully from one person to another. Writing should therefore make relationships among ideas visible. Precision should arise from accurate nouns, verbs, evidence, grammar, and stated relationships rather than from compressed terminology that forces the reader to infer what the writer could have said directly.
Knowledge should not become a club whose membership depends upon learning how insiders signal membership. Expertise has fulfilled more of its purpose when an expert can make a difficult matter understandable without falsifying it.
Specialized language is sometimes necessary. A precise technical term can preserve distinctions that ordinary language would require many sentences to explain. The problem begins when specialized language is used to add authority, create social distance, or make an ordinary proposition appear inaccessible.
This principle also governs the use of AI and archives. Neither should become a place in which understanding disappears merely because information remains retrievable. The person should be able to return to the evidence, regain the reasoning, and understand enough to continue.
Knowledge should not become a club whose membership depends upon mastering the signals by which insiders recognize one another. Expertise serves more of its purpose when the expert can make a difficult matter understandable without falsifying it.
The same principle applies to AI and archives. Neither should become a place where understanding disappears merely because information remains retrievable. The person should be able to return to the evidence, regain the reasoning, and understand enough to continue.
---
# ARTICLE VIII — HUMAN RELATIONSHIPS, COMMUNITY, CREATIVITY, AND SCALE
Not every human difficulty is a defect awaiting automation. Conversation, disagreement, teaching, friendship, family, craftsmanship, mutual aid, shared work, and community all contain value partly because human beings must participate in them.
Not every human difficulty is a defect waiting to be eliminated. Conversation, disagreement, teaching, friendship, family, craftsmanship, mutual aid, shared work, and community contain value partly because human beings participate in them directly.
Difficulty can itself have value. Learning often requires struggling with something not yet understood. Craft requires repeated effort. Trust develops through experience. Responsibility develops by carrying responsibilities rather than by having every burden removed in advance.
Difficulty can itself have value. Learning often requires struggling with something not yet understood. Craft requires repeated effort. Trust develops through experience. Responsibility develops by carrying responsibilities rather than having every burden removed in advance.
Technology can assist these activities, but assistance is not always improved by substitution. Removing every inconvenience can also remove occasions through which patience, judgment, skill, initiative, mutual understanding, responsibility, and human creativity develop.
Technique tends to see difficulty first as inefficiency. Sometimes that judgment is correct. Needless repetition, wasted effort, preventable error, and artificial obstacles deserve removal. But inconvenience and waste are not identical, and the distinction matters.
Convenience therefore has a cost when convenience removes something more valuable than the inconvenience being removed. The cost may not appear immediately. A person can gain time while losing a skill, gain access while losing independence, or gain efficiency while losing meaningful contact with another human being.
Removing every inconvenience can also remove occasions through which patience, judgment, skill, initiative, responsibility, mutual understanding, and creativity develop. A person can save time while losing competence, gain convenience while losing independence, or gain access while losing meaningful contact with another person.
Community deserves the same care. A community is not merely a number of people with the same need, nor is a community made legitimate by scale. Communities are formed and sustained by people who understand why they are together, act for one another, accept responsibilities toward one another, and retain enough freedom that participation remains meaningful.
Technology can assist human relationships without becoming their substitute. The question is not whether a technical system can imitate part of an interaction. The question is what the interaction was for and whether the substitute preserves what mattered about it.
Large scale can be useful, but scale is evidence of size, not evidence of worth. A tool used by ten people can serve those ten people well. A harmful institution can acquire millions of participants. Popularity proves popularity.
Community deserves similar care. A community is not simply a population sharing the same need, platform, geography, or administrative category. Communities are sustained by people who understand why they are together, act for one another, accept responsibilities toward one another, and retain enough freedom that participation has meaning.
This is why broad thought and local action belong together. Large questions deserve serious thought, but action eventually occurs among actual people, places, machines, evidence, relationships, and material conditions. Broad ideas must meet those realities somewhere.
Large scale can be useful, but scale measures size rather than worth. A useful tool may serve ten people. A destructive institution may acquire millions of participants. Popularity proves popularity.
Local action does not prevent wide cooperation. Local independence can improve cooperation by preserving the ability of participants to choose the terms on which they cooperate. Federation can extend far beyond a locality without requiring every participant to surrender authority to the same center.
This is why broad thought and local action belong together. Large questions deserve serious thought, but action eventually occurs among actual people, places, relationships, evidence, institutions, machines, and material conditions. Broad ideas must meet those realities somewhere.
The aim is therefore not smallness for its own sake. The aim is human-scale responsibility inside whatever larger association people freely choose to build.
Local action does not prevent broad cooperation. Local independence can improve cooperation because participants retain a practical ability to choose the terms on which they cooperate. Federation can extend far beyond the locality without requiring every participant to surrender authority to the same center.
The aim is not smallness for its own sake. The aim is human responsibility that remains recognizable even when cooperation becomes large.
---
# ARTICLE IX — HUMAN CAPABILITY IS THE MEASURE
The final measure of work is not the amount of machinery, automation, computation, documentation, or output that the work produces. Those quantities can be useful measurements, but they do not answer the deeper question of what the work has done for the human beings involved.
The final measure of work is not the amount of machinery, automation, computation, administration, documentation, or output that the work produces. Those quantities can be useful measurements, but they do not answer the deeper question of what the work has done for the human beings involved.
A good tool should enlarge the range of worthwhile human action. A good archive should preserve access to understanding. A good working standard should allow people to cooperate without requiring unnecessary dependency. A good technical system should remain inspectable enough to question and replaceable enough that people are not trapped merely because the system succeeded in becoming indispensable.
A good tool should enlarge the range of worthwhile human action. A good archive should preserve access to understanding. A good working standard should permit cooperation without requiring unnecessary dependency. A good technical system should remain inspectable enough to question and replaceable enough that people are not trapped merely because the system succeeded in becoming indispensable.
A good institution should remember that the institution exists for human purposes. The institution should not quietly reverse that relationship and begin treating human beings as inputs required for the preservation of the institution.
A good institution should remember that the institution exists for human purposes. The institution should not reverse that relationship and begin treating human beings primarily as inputs required for the preservation, expansion, or efficiency of the institution.
Human creativity matters here because capability is not merely the ability to operate existing machinery more efficiently. A capable person can question the machinery, alter it, refuse it, create something different, teach another person, preserve what matters, and respond intelligently when the expected system fails.
Human creativity matters because capability is not simply the ability to operate existing machinery or comply efficiently with an established method. A capable person can question the method, alter it, refuse it, create something different, teach another person, preserve knowledge, improvise when circumstances change, and continue when an expected system fails.
Human capacity and human capability should therefore be distinguished. The person possesses capacities that belong to human life itself. Capability describes what the person has been able to cultivate, learn, exercise, and retain. Tools and institutions should help those capabilities develop rather than making the person progressively less able to act without them.
Human capacity and human capability should therefore be distinguished. Human beings possess capacities that belong to human life itself. Capability describes what a person has cultivated, learned, practiced, exercised, and retained. Tools and institutions should help capabilities develop rather than making people progressively less able to act without them.
This provides a practical test for much of the constitution. When a new technology, institution, standard, convenience, or automation is proposed, the question is not merely whether the proposal works. We should ask what the proposal makes possible, what the proposal makes harder, what dependencies it creates, what knowledge it preserves or removes, who gains authority, whether alternatives remain possible, and what the human being will still know how to do afterward.
This provides a practical test for technique. When a new institution, technology, procedure, standard, convenience, automation, or optimization is proposed, the question is not merely whether the proposal works or whether the proposal is more efficient. We should ask what the proposal makes possible, what the proposal makes harder, what dependencies it creates, what knowledge it preserves or removes, who gains authority, whether refusal remains practical, whether alternatives remain possible, and what the human being will still know how to do afterward.
The aim is not dependence upon increasingly capable machinery. The aim is freer, more capable, more creative human beings equipped with tools that serve them.
The same examination applies to coercion. An efficient means of compelling people remains coercive. A widely accepted means remains coercive. A technologically sophisticated means remains coercive. Neither technique nor popularity supplies the moral authority that coercion lacks.
The work should not blindside the human being with consequences that could reasonably have been made visible. It should leave enough evidence to understand what happened, enough freedom to question the result, enough knowledge to choose another course, and enough capability to act upon that choice.
The aim is therefore not dependence upon increasingly efficient systems. The aim is freer, more capable, more creative human beings using institutions, knowledge, and tools that remain answerable to human purposes.
Work should not blindside the human being with consequences that could reasonably have been made visible. Good work leaves enough evidence to understand what happened, enough knowledge to question the result, enough freedom to choose another course, and enough capability to act upon that choice.
That is the standard by which the rest should be judged.
## CLOSING
Truth does not require a majority, and popularity does not establish worth. Liberty requires more than permission from whoever presently holds authority, because liberty also depends upon the practical ability to question, refuse, withdraw, create, and act.
Truth does not require a majority, and popularity does not establish worth. Liberty requires more than permission from whoever presently holds authority. Liberty has practical meaning when people can question, refuse, withdraw, create alternatives, associate voluntarily, and act without being compelled merely because another person or institution believes that compulsion would produce a better result.
Evidence should remain inspectable, provenance should remain visible, and the path to an important conclusion should remain recoverable when later understanding depends upon it. Technology should remain equipment rather than becoming the source of human purpose.
Evidence should remain inspectable. Provenance should remain visible. Important reasoning should remain traceable when later understanding depends upon it. Correction should remain possible without erasing the history from which correction arose.
We should build locally where local capability preserves independence, cooperate broadly where cooperation serves the people involved, and keep the means answerable to the ends for which those means were chosen.
Coercive power has a narrow purpose: restraining coercion. Efficiency cannot enlarge that purpose, and technique cannot manufacture the right to compel merely by producing a more effective means of compulsion.
Technique should remain subordinate to human purposes. Efficiency can help us choose among means, but efficiency cannot determine the ends, establish what is morally legitimate, or decide what human beings are for.
We should build locally where local capability preserves independence, cooperate broadly where cooperation serves the people involved, and preserve alternatives so that cooperation does not become captivity.
The tools may become more capable. The institutions may become larger. The methods may become more efficient.
The human being should not therefore become less capable, less free, less able to understand, or less able to refuse.
Above all, the work should leave the human being more capable.