Open Source and Harmonism

Dialogue article in the Harmonism cascade. Engages the free software / open source tradition as one of the two operational sovereignty movements of the late twentieth century, sibling to the cypherpunks. See also: Cypherpunks and Harmonism, The Sovereign Substrate, The Sovereign Refusal, Transmission, The Sovereign Stack, Methodology of Integral Knowledge Architecture, The Telos of Technology, Liberalism and Harmonism.

---

Two intellectual movements of the late twentieth century produced operational infrastructure rather than only theory. The cypherpunks built the cryptographic substrate the modern privacy stack runs on. The free software movement built the code substrate the modern internet runs on. The two are siblings — they shared mailing lists, conferences, and many of the same founders — and they reached convergent conclusions through structurally distinct lines of reasoning. The cypherpunk lineage discovered that mathematics has political consequences the political class cannot overrule. The free software lineage discovered, independently and in parallel, that non-rivalrous goods cannot be enclosed without producing what enclosure was designed to prevent — and built the institutional architecture (the four freedoms, copyleft, the Free Software Foundation, the Open Source Initiative, the working foundations) that codified the discovery and propagated it.

Cypherpunks and Harmonism engages the cryptographic-substrate lineage. This article engages the code-substrate lineage. The two articles read one tradition each within what is structurally one civilizational movement at the digital register — substrate sovereignty asserted by those who built the substrate against those who proposed to enclose it. Convergence on the load-bearing claims is substantial. Philosophical articulation is distinct. Both lineages reached the same recognition through different doors, and Harmonism stands in convergence with both.

What is striking about the free software lineage, read forward from the GNU Manifesto in 1985 to the contemporary controversy over what open source AI should mean, is the structural continuity of the position. The substrate the practitioner runs is the practitioner’s own. The pattern, once made, cannot be enclosed without category error. The four freedoms — to run, to study and modify, to distribute copies, to improve and publish improvements — are the operational articulation of substrate-sovereignty at the code register. The tradition produced the foundational documents in the 1980s, built the infrastructure across the 1990s and 2000s (GNU, Linux, Apache, PostgreSQL, the World Wide Web’s reference implementation), reached civilizational scale in the 2010s, and now confronts the AI-era reframing of what substrate means when computation supersedes code. Across forty years, the load-bearing claims have not shifted.

What the tradition has not articulated, and what Harmonism articulates as its contribution to the conversation, is the cosmological ground on which the substrate is sovereign in the first place — and what the freedom is for once the substrate is the practitioner’s own.

The Movement and Its Texts

The free software tradition has a recognisable genealogy. Richard Stallman worked at the MIT Artificial Intelligence Laboratory through the 1970s and into the early 1980s, in what was at that time a thriving hacker culture organised around the unwritten convention that source code was shared and improved by anyone competent to read it. The convention collapsed in the early 1980s under the combined pressure of three forces: the AI Lab’s diaspora into commercial spinoff companies that imposed proprietary licensing on what had been shared work, the introduction of binary-only software from external vendors (most famously a Xerox printer driver whose closed source prevented Stallman from fixing a flaw he had identified), and the institutional adoption of non-disclosure agreements as the default condition of access to code. What had been a working commons became, in the space of a few years, a collection of fences. Stallman experienced the collapse as the loss of a civilizational substrate, and concluded that the substrate could be reconstituted only by building a complete alternative operating system whose four freedoms were guaranteed by license.

On 27 September 1983 Stallman posted to the net.unix-wizards and net.usoft newsgroups an announcement titled new UNIX implementation — the initial declaration of the GNU Project, an effort to build a Unix-compatible operating system whose source code would remain free. In January 1984 he left MIT to work full-time on the project. In March 1985 he published The GNU Manifesto in Dr. Dobb’s Journal of Software Tools — the canonical call-to-action of the tradition, articulating both the technical project and the moral position that grounded it. In February 1986 the Free Software Foundation published the Free Software Definition — the four freedoms numbered 0 through 3 — that has remained the doctrinal core of the tradition for forty years. In 1989 Stallman released the GNU General Public License v1; in 1991 GPL v2, co-drafted with the legal scholar Eben Moglen who would shape the license tradition for the next two decades.

In August 1991 a Finnish computer science student named Linus Torvalds posted to the comp.os.minix newsgroup an announcement that began I’m doing a (free) operating system (just a hobby, won’t be big and professional like gnu) for 386(486) AT clones. The hobby would, within a decade, become the kernel of what we now call Linux — the operating system that runs roughly every consequential piece of contemporary infrastructure (servers, supercomputers, embedded systems, the Android phones in three billion pockets). The GNU project supplied the userland (the compilers, the shell, the utilities); Linux supplied the kernel; the combination — GNU/Linux, in the tradition’s own preferred naming — completed Stallman’s 1983 plan within a decade of its announcement.

Through the 1990s the tradition produced the substrate the contemporary web runs on. The Apache HTTP Server (1995) became the dominant web server within five years of its release; Perl (Larry Wall, 1987) and Python (Guido van Rossum, 1991) gave the scripting layer to a generation of programmers; PHP (Rasmus Lerdorf, 1995), MySQL (1995), and PostgreSQL (1996) supplied the LAMP stack. Mozilla (1998) opened the browser register after Netscape’s commercial collapse. The pattern was consistent: where a substrate of civilizational consequence existed, the free software tradition produced an open implementation whose four freedoms were guaranteed by license, and the open implementation eventually dominated by virtue of being the substrate everyone could read, modify, and run.

In 1997 Eric Raymond published The Cathedral and the Bazaar, a paper that recast the tradition’s success in anthropological and pragmatic terms. The cathedral was Stallman’s earlier model — a small team working in tight coordination toward a planned design, releasing finished code at intervals. The bazaar was Torvalds’ model — early and frequent release, public source from day one, distributed contribution from anyone who showed up. Raymond argued, with characteristic precision, that the bazaar was producing software of higher quality than the cathedral could match, and that the pattern generalised. His follow-up essay Homesteading the Noosphere (1998) named the gift culture the tradition operated in — borrowing the framework from Lewis Hyde’s The Gift (1983) without crediting the spiritual register Hyde’s work carried. Raymond was a libertarian-pragmatist; Hyde had articulated the gift economy in registers reaching from anthropology to spiritual transmission, and the conceptual structure Raymond imported was strong enough to do work in the lighter frame he placed it in.

In February 1998 a meeting in Palo Alto produced what would become the Open Source Initiative. Raymond, Bruce Perens, Christine Peterson (who coined the term open source), and several others concluded that the name free software — with its moral-political register and its perennial confusion over whether free meant libre or gratis — was an obstacle to corporate adoption. The OSI rebranded the technical substrate in language calibrated for pragmatic palatability. Stallman refused the rebranding and the philosophical concession it carried, kept the FSF and the GNU project under the free software banner, and held that the moral question (does the user have the four freedoms?) was prior to the pragmatic question (does the development model produce better code?). The split has persisted for nearly thirty years. The OSI-led open source register became the dominant public-facing name, and the corporate adoption Raymond and Peterson had projected duly arrived: IBM (2000), Sun, Red Hat (IPO 1999), the broader enterprise embrace of the 2000s, and eventually the Big Tech absorption that defines the contemporary landscape (Google open-sourcing Android in 2008, Microsoft acquiring GitHub in 2018, Meta releasing PyTorch and Llama-as-open-weights through the 2010s and 2020s). Stallman’s diagnostic register held: the moral concession was real, and the corporate embrace of open source without free software produced, over time, a tradition whose load-bearing institutions were beholden to the interests the original movement had organised against.

The legal architecture continued to develop in parallel. The Affero General Public License (AGPL, 2002) closed the SaaS loophole that had emerged when network-served software made the GPL’s distribution trigger inapplicable to cloud deployments. The GPL v3 (2007), co-drafted by Stallman and Moglen, added anti-Tivoization provisions and addressed the patent-cross-licensing dynamics that had emerged in commercial Linux distribution. The Software Freedom Conservancy (founded 2006) became the institutional inheritor of the FSF’s moral register in a more grounded form, taking on copyleft compliance enforcement at scale through cases like SFC v. Vizio, where a 2021 lawsuit established (through a 2023 judgment denying summary judgment) that consumers can directly demand GPL source code as third-party beneficiaries — a legal innovation routing around the copyright-holder bottleneck.

From 2018 forward the tradition entered a period of structural pressure that has not yet resolved. MongoDB, faced with cloud-provider AWS reselling its database as a managed service without contributing back, changed its license in October 2018 from AGPL to the Server Side Public License (SSPL) — a license the OSI explicitly rejected as not meeting the open source definition. Elastic followed in 2021 (later partially returning to AGPLv3 in 2024). HashiCorp adopted the Business Source License (BSL) in August 2023, triggering a Linux Foundation-hosted fork as OpenTofu. Redis moved to SSPL in March 2024, faced widespread distribution package removal, and returned to AGPLv3 in 2025. The pattern is consistent: an open source infrastructure project succeeds, cloud providers capture the commercial layer without contributing back, the original project retreats to a source-available license that closes the cloud-provider loophole while breaching the open source definition. The commons enclosing itself back from cloud-provider enclosure — and neither side, structurally, walking dharmically.

In June 2023 Red Hat — by then a wholly-owned subsidiary of IBM — closed public access to RHEL source code, ending the long-running practice that had supported downstream rebuilds like CentOS, Rocky Linux, and AlmaLinux. Bradley Kuhn at the SFC diagnosed the move as a decade-long effort by Red Hat to maximize the level of difficulty of those in the community who wish to trust but verify that RHEL complies with the GPL agreements — operating inside the GPL’s letter while corroding its spirit. The downstream rebuilds responded with workarounds: AlmaLinux pulling from CentOS Stream, Rocky Linux analysing pay-per-use RHEL cloud instances to extract the source. The substrate held, but the institutional good faith that had sustained it for two decades did not.

The contemporary moment is the AI weights controversy. Through 2023 and 2024 the term open source AI began to fragment under the pressure of Meta’s Llama releases, which Meta marketed as open source despite restrictions on commercial use above a 700-million-user threshold and the absence of any release of training data or training process. The Open Source Initiative published its Open Source AI Definition in October 2024, specifying that open source AI requires open weights and open training code and sufficient information about training data to enable independent reproduction. The two terms name structurally distinct artefacts and the contemporary moment turns on the distinction. Open weights means the trained model parameters are downloadable and runnable, but the training code and the training dataset remain closed — the practitioner can run the model and adapt its weights, but cannot reproduce the model from scratch and cannot inspect what produced its behaviour. Open source AI, by the OSI definition, requires all three layers — weights, training code, and sufficient information about training data — so that the model is reconstructable by an independent practitioner working from open materials. The distinction is the difference between receiving the artefact and receiving the substrate that produces the artefact. The free software tradition’s foundational claim was always that the substrate is what matters, because the substrate is what enables the four freedoms to operate in practice — the user who has only the binary cannot study what the program actually does, and the user who has only the weights cannot study what the model actually learned. Most popular open AI models — Llama 4 included — fall short of the OSI definition and meet only the open-weights threshold; the term open weights has emerged as the more accurate descriptor for what they are. Mistral Large 3 (released December 2025 under Apache 2.0), gpt-oss-120b (Apache 2.0), and DeepSeek V4 (MIT) are the contemporary models that meet the stricter open-source definition. Llama is the canonical open-weights case; the three above are the canonical open-source-AI cases. The distinction is not pedantic — it is the same structural distinction Stallman insisted on in 1986 against the proprietary regime that wanted to release binaries while withholding source. Open weights without open training code and open training data is the AI-era analogue of binary-only software. The tradition’s diagnostic register names it correctly: the artefact is downloadable but the substrate is not, and the substrate is what the four freedoms were always about. The tradition is doing the same work it has always done — articulating what the substrate is and refusing the institutional rebranding of partial openness as the real thing — at a new register where the substrate has shifted from code to weights to data-and-training-process together.

The arc, read forward from 1985 to 2026, is structurally coherent. The substrate of computation belongs to the practitioner who runs it. The institutional pressure to enclose recurs at every register the substrate has — proprietary software, closed protocols, SaaS deployment, cloud capture, AI weights, and now training data and inference compute. The tradition’s response has been the same at every register: build the open implementation, write the license that propagates the freedom downstream, refuse the institutional rebranding, hold the moral question prior to the pragmatic. The forty-year continuity of the position is its own argument.

The Foundational Insight

Where the cypherpunk tradition began from the recognition that mathematics has political consequences, the free software tradition began from a different angle. Stallman’s foundational claim was not that the substrate was protected by mathematical necessity (the cypherpunk position), but that the user who cannot read, modify, or share the program they are running is subjugated by it. The argument is moral and concrete. A program is a structure of behaviour the user’s computer is executing. The user who cannot inspect the structure cannot know what their own computer is doing. The user who cannot modify the structure cannot make their own computer serve their own purposes. The user who cannot share the structure with others cannot participate in the working commons of those who run the same program. Each of these conditions, taken alone, names a specific form of dispossession. Taken together, they name the substrate-level alienation Stallman saw at MIT as the proprietary regime closed around the AI Lab.

The four freedoms — published as the Free Software Definition in February 1986 — codify the refusal of these conditions in normative form. Freedom 0: the freedom to run the program for any purpose. Freedom 1: the freedom to study how the program works and modify it. Freedom 2: the freedom to redistribute copies. Freedom 3: the freedom to improve the program and publish the improvements. The numbering begins at zero because the freedom to run the program was added after the original three were articulated; Stallman, characteristic of the tradition, refused to renumber and left the historical artefact in place. The four freedoms together describe what it means for the user to be the substrate-holder rather than the substrate-renter at the code register. The user runs the program; the user reads the program; the user changes the program; the user shares the program. Each freedom is operationally meaningful only if the source code is available — which is why the four freedoms, taken seriously, require the public availability of source as their precondition.

The architectural innovation that distinguishes the free software tradition from the cypherpunk tradition is copyleft — the legal mechanism by which the four freedoms propagate downstream. The GNU General Public License uses copyright — the legal instrument the institution had developed to enforce enclosure — to enforce the opposite: any distribution of GPL-licensed code or derivative work must carry the same four freedoms forward to every downstream recipient. The structural move is precise. Copyright was developed to enforce scarcity over non-rivalrous goods; copyleft turns the same legal instrument into a mechanism for enforcing abundance. Where copyright says you may not redistribute, copyleft says you may redistribute only under terms that preserve everyone else’s freedom to redistribute further. The license becomes a viral commitment to the commons.

The combination — public source as substrate, the four freedoms as ethical baseline, copyleft as enforcement-against-enclosure — is the operational architecture of the tradition. It produces what the cypherpunks called crypto-anarchy at the analogous register: a domain in which the user’s substrate is protected by legal architecture rather than cryptographic architecture, but the structural commitment is the same. The user holds the substrate. The institution cannot enclose what the substrate-holder may freely redistribute. The freedom propagates by the operation of the license itself.

What is striking about the foundational claim, read against the cypherpunk articulation, is the absence of mathematics-as-bedrock register. Free software does not appeal to mathematical necessity. The GPL’s protection of the four freedoms is not enforced by mathematics; it is enforced by license, and licenses are enforceable only insofar as the legal regime that recognises them continues to recognise them. Stallman knew this perfectly well. His response was the moral register: the four freedoms are not pragmatic preferences but ethical requirements, and the program that denies them is wrong in the strict normative sense. The user is owed the substrate by virtue of running the substrate. The proprietary regime that withholds it is committing a moral violation that the practitioner is entitled to refuse.

This is closer to the moral-philosophical traditions Harmonism stands adjacent to than May’s crypto-anarchism was. Stallman is closer to a Kantian on the surface, but the deeper resonance is with the Indian vidyā-dāna tradition — the doctrine that the gift of knowledge is the highest gift, that the teacher who withholds is breaching the guru-shishya bond, that the knowledge owed to the student is owed by the structure of the transmission itself. Stallman almost certainly would not name the resonance in those terms. The structural convergence stands regardless. The free software tradition reaches, by its own route and its own register, a position the perennial traditions reached by other routes — that knowledge held proprietarily decays at the centre, and that the practitioner who would receive knowledge is owed the conditions under which the reception is possible.

Reading the Four Freedoms

The four freedoms are brief — four numbered sentences fitting on a single page — and their argument has held across forty years with the same structural integrity May’s manifesto held across thirty-eight. Each freedom warrants engagement at the level its compression deserves.

Freedom 0 — the freedom to run the program, for any purpose. The foundational refusal of fields of use restrictions. The substrate-holder does not ask permission for the purposes to which they put the substrate. The proprietary regime had developed elaborate machinery for restricting use — licenses tied to specific industries, to non-commercial purposes, to particular jurisdictions, to specific user counts. Each restriction was a claim that the institution issuing the license held continuing authority over what the user did with what was on the user’s computer. Freedom 0 refuses the claim at its root. The substrate is the user’s; the purposes are the user’s; the institution that issued the program has no standing to specify what the user does with the running program once it is running. The structural argument is the same as the Harmonist articulation of Dharma — the practitioner’s action is the practitioner’s own, accountable to Logos and to the practitioner’s own conscience, not to the institutional mediator who would condition its legitimacy on continuing compliance with the mediator’s terms.

Freedom 1 — the freedom to study how the program works, and modify it to suit the user’s needs. The freedom requires source code; the binary-only program is the modern equivalent of the medieval scribe withholding the manuscript so only the credentialled reader can interpret it. Stallman’s MIT printer story is paradigmatic. The printer had a flaw he could have fixed in an hour if he had source access; the vendor would not release the source; the flaw remained. The proprietary regime’s argument for the withholding was that the program represented protected commercial interest; the actual operation was that the substrate the user was running could not be made to serve the user’s actual purposes because the user had been denied the standing to read what was running on their own machine. Freedom 1 refuses this. The substrate-holder reads the substrate. The substrate-holder modifies the substrate when modification is required. The institutional class whose authority depends on the user’s inability to read is dispossessed of the authority — not by the user’s permission, but by the user’s actual standing.

Freedom 2 — the freedom to redistribute copies, so the user can help others. The classic refusal of artificial scarcity over non-rivalrous goods. The Harmonist articulation in The Sovereign Substrate § Two Faces of Enclosure names this precisely: one practitioner running the model does not erode the model; property as an institutional category was developed to settle conflicts over what cannot be multiplied without subtraction, and applying that category to non-rivalrous goods is a category error. Freedom 2 codifies the refusal of the category error at the code register. The user who has the substrate may share it with anyone who would also benefit. The institution’s claim that the sharing constitutes infringement is, under Freedom 2, a claim the user is entitled to disregard. The structural convergence with the Indian vidyā-dāna tradition — knowledge transmitted is doubled, not divided; the guru loses nothing by teaching — is exact. Stallman articulated the position in copyright-grounded language. The Vedic tradition articulated it in dharmic-transmission language. The substrate of the recognition is one.

Freedom 3 — the freedom to improve the program and publish the improvements, so the whole community benefits. The architectural innovation that distinguishes free software from mere openness. Where Freedoms 0 through 2 grant the user standing as substrate-holder, Freedom 3 grants the user standing as substrate-contributor — the right not merely to receive the substrate as it was given, but to participate in its continuing cultivation. This is where the bazaar emerges. The substrate is not a fixed artefact; it is a living transmission, modified and improved by everyone who undertakes the discipline. The substrate’s continuing health is the responsibility of the community that uses it. The institutional model — in which the substrate is produced by a credentialled class and consumed by everyone else — is refused at the structural level. Freedom 3 makes the substrate-holder and the substrate-cultivator the same person, by ontology rather than by employment contract.

Six structural moves compose the argument the four freedoms make taken together. First: the substrate of computation is something the user holds, not something the user rents. Second: the proprietary regime’s withholding of the four conditions of substrate-holding constitutes dispossession, not commercial preference. Third: the user is not subjugated by the substrate they hold; they are subjugated by being denied standing as substrate-holder. Fourth: the four freedoms restore the standing; the license that guarantees them (the GPL, in copyleft form) propagates the restoration to every downstream user. Fifth: the architectural innovation is using the institutional instrument of enclosure (copyright) to enforce the opposite (commons-propagation) — a legal jiu-jitsu that operates inside the framework while inverting its effect. Sixth: the position is moral, not pragmatic. The practitioner is owed the four freedoms by the structure of running the program, not by virtue of any contractual relationship. The proprietary regime that withholds them is committing a wrong in the strict normative sense.

The argument is structurally precise. Forty years after the Free Software Definition was published, the four freedoms remain the most coherent articulation of substrate-sovereignty at the code register that any tradition has produced. The corporate adoption of open source in 1998 produced a vast institutional infrastructure that operates around the same four freedoms while emptying the moral content for pragmatic palatability — and the subsequent forty years have demonstrated, repeatedly, that the moral content was what made the four freedoms cohere in the first place. When the four freedoms are upheld on pragmatic grounds alone, the pragmatic ground shifts (cloud-provider capture, AI-era reframing, regulatory pressure), and the institutions defending the four freedoms shift with it. When the four freedoms are upheld on moral grounds — the user is owed this; the institution that withholds is wrong — the position holds across the shift. Stallman was right about this from the beginning. The thirty years of his marginalisation by the OSI-led mainstream did not change the structural argument.

What the four freedoms do not articulate, and what the closing section of this article will name as the missing centre, is what the user does with the substrate once they hold it. The freedoms establish the conditions; they do not specify the cultivation. The user who holds the substrate but does not cultivate is sovereign in form and serf in substance. The four freedoms are necessary; they are not sufficient.

The Visible Commons

Before naming the missing centre, the depth of what the tradition has produced must be made visible. The free software movement is not a philosophical position alone; it is an operational achievement at civilizational scale. The contemporary world runs on the substrate the tradition built, and the substrate is sustained — as we speak, today — by working maintainers who hold the four freedoms in their practice. The argument the article is making is abstract; the substrate is concrete. Naming the substrate by its actual instances is part of honouring what the tradition has done. The instances are not exhaustive; they are representative. Each carries a teaching about what the four freedoms produce when sustained over time.

Linux is the bazaar at civilizational scale. Thirty-five years of continuous development from Torvalds’ 1991 hobby project to the operating system that runs roughly every server, every supercomputer, every embedded system of consequence, and (through Android) three billion phones. The kernel has accepted contributions from tens of thousands of developers across the entire period. No central authority has been able to control the project, capture it, or fork it successfully against the upstream. The git version-control system — which Torvalds wrote in 2005 to solve the problem of distributed kernel development — is itself now the substrate the entire software industry runs on. The cathedral could not produce this. The bazaar did. Torvalds himself carries views and historical conduct (the caustic communication style of the early decades, the step-back in 2018 to acknowledge it) that Harmonism would not endorse in their full form. The project stands on its own substrate regardless. The teaching is not about the person; it is about what the architecture produces. Distributed sovereignty over the substrate, governed by the willingness of competent practitioners to maintain the substrate together, produces an operational commons that proprietary alternatives cannot match across thirty-five years of comparison.

VLC and the VideoLAN project, led by Jean-Baptiste Kempf, is the textbook case of sovereign refusal as operational architecture. Launched in 2001 at École Centrale Paris as a student project to stream video across a campus network, VLC is now present on roughly a billion devices and remains the most reliably-trusted video player in personal computing. The VideoLAN non-profit refused App Store integration that would have required compromising the player’s ability to play any codec the user supplied. They refused ad-supported variants when the commercial pressure to monetise was substantial. They refused the MPAA’s repeated demands to disable DVD-CSS decryption support. They refused acquisition offers from companies whose business model would have required closing the source. The player exists today, freely available, ad-free, codec-complete, on every platform that matters, because Kempf and the VideoLAN team refused, repeatedly, to surrender the substrate to institutions whose terms would have compromised it. VLC is the contemporary teaching about what sovereign-refusal looks like at the application register. The substrate held because the maintainers held.

FFmpeg and the long maintainership of Michael Niedermayer and the rotating core team is the librarian-monk case. FFmpeg is the codec library underneath almost every video-handling application on Earth — VLC uses it, YouTube uses it, Netflix uses it, every consumer video editor uses it, every commercial streaming platform uses it. The library is licensed under LGPL and GPL; the user base spans roughly the entire industrial video sector; the maintainership is a small team operating on chronic underfunding relative to the load the substrate carries. In late 2024 a public moment surfaced when major commercial users were criticised for the discrepancy between what they extract from FFmpeg and what they contribute back — the structural problem the source-available retreat had emerged to address, surfacing at the codec layer. The teaching is precise: civilizational substrate is sustainable when the maintainers are sustained, and the maintainers of the world’s most-used video codec library are sustained by donations, consulting work, and the discipline of a small group of practitioners who refuse to abandon what they have built. The pattern is closer to medieval monastic opus Dei — manuscript-copying as practice, transmission as prayer — than to any commercial software engineering arrangement. The library exists because the monks did not stop.

Lichess and the work of Thibault Duplessis is the Heart-register exemplar. Launched in 2010 as a free, ad-free, fully open-source chess platform, Lichess now out-competes chess.com — a venture-backed commercial platform with billions in valuation, aggressive paywalling, and engagement-optimised feature design — on every quality axis the practitioner cares about. The analysis engine is deeper. The tournament infrastructure is more robust. The training tools are more sophisticated. The interface is cleaner. The community standards hold higher. The platform is funded entirely through donations to the non-profit Lichess.org Foundation, operates on a fraction of chess.com’s budget, and yet ships better software year after year. The teaching is the jñāna-dāna doctrine in operational form: knowledge given freely produces, across time, an artefact superior to the one held proprietarily. The reason is structural — when the substrate-holder and the substrate-cultivator are the same population (the chess players themselves contribute features, improvements, translations), the cultivation compounds in ways the rented-substrate model cannot reach. Lichess is the case-study the entire creator economy debate should be made to confront.

PostgreSQL is the community commons across decades. Begun at UC Berkeley in 1986 as POSTGRES, released as PostgreSQL in 1996, and continuously developed for thirty years by a global team operating through the PostgreSQL Global Development Group. The database runs roughly half the production database workloads on Earth — alongside the various commercial forks (Amazon RDS Postgres, Crunchy Data, EnterpriseDB) that depend on the upstream commons and that have, to varying degrees, contributed back. There is no Postgres Foundation in the corporate-foundation sense; no single company owns the trademark or controls the project. The governance is distributed across the development group, the core committee, and the major user community. PostgreSQL is the case-study of community-only commons across decades — what the four freedoms produce when sustained by a working community rather than by a foundation or a corporation. The database is, by every benchmark that matters, better than the proprietary alternatives. The architecture is the explanation.

Wikipedia must be named honestly. The largest reference work in human history operates on volunteer labour, donor funding, and the openly-licensed substrate of the four freedoms applied to encyclopaedic knowledge. That this is possible — that hundreds of millions of articles in three hundred languages can be produced and maintained by volunteers operating without commercial coordination — is itself a structural argument the free software tradition’s diagnostic is correct. The architecture works. The articles exist. The contemporary world’s casual access to a substantial portion of recorded human knowledge runs through Wikipedia daily, free of paywall, free of advertising. At the same time, the editorial substrate has been progressively captured by ideological networks the Wikimedia Foundation has not adequately addressed and which have produced, across the past fifteen years, systematic distortion of how contested topics are presented — particularly in politics, biography of living public figures, medicine, and history. The teaching is therefore double. The architecture of open-source-knowledge-at-civilizational-scale is real and reachable. The institutional layer that governs the editorial substrate is the failure mode — institutional capture of editorial discretion can corrupt the substrate even when the underlying architecture remains open. The lesson for Harmonia, and for any future project at Wikipedia’s scale, is that the architecture is necessary but not sufficient; the editorial integrity of the maintainers is the substrate the architecture finally rests on.

Bitcoin Core is the cypherpunk-sibling, engaged at depth in Cypherpunks and Harmonism and named here only briefly. The most successful open-source project of the twenty-first century by market consequence operates under the four freedoms at the protocol layer while carrying the cypherpunk political commitments at the social layer. The intersection is precise. Bitcoin Core could not have been built outside the free software tradition’s architecture (MIT-licensed C++, full source publication, distributed contribution, reproducible builds). Bitcoin’s political consequence required the cypherpunk register (mathematics as bedrock, state-irrelevance through cryptography). The combination — free software architecture plus cypherpunk politics — is what the contemporary sovereign stack runs on. The two traditions, working in parallel and in dialogue across the same forty years, produced the substrate together.

The list extends. Signal and the cryptographic communication infrastructure, treated at depth in the cypherpunks article. Tor and the anonymous-routing substrate, same. Blender, the 3D graphics commons under the Blender Foundation, fully sovereign, used in major film production globally. F-Droid, the Android FOSS counter-store, the practitioner’s alternative to Google Play. OpenStreetMap, the maps commons; Wikipedia’s sibling at the geographic register. LibreOffice, the office-productivity commons after the Oracle acquisition of OpenOffice rendered the original project unmaintainable. Mastodon and the broader ActivityPub federation, the sovereign-social substrate the cypherpunks called for thirty years ago and the free software tradition delivered.

Each of these projects can be evaluated against the doctrinal test articulated in The Sovereign Stack — permissionless participation, sovereign custody, mathematical foundation, open source and auditable, decentralised or sovereignly hostable. Each passes the test, with Wikipedia carrying the editorial-capture caveat. Together they constitute the visible substrate the contemporary world runs on while almost never naming. The civilization that uses them daily has almost no theory of why they work. Harmonism articulates the theory: the substrate is the practitioner’s by Logos-rendered ontology, the cultivation that produces and maintains the substrate is dharmic in form, and the architecture the four freedoms specify is the operational expression of that ontology at the code register.

The visible commons is the demonstration. The argument that follows — that Harmonism completes the tradition by articulating what the substrate is for — rests on what the tradition has already built. Without the substrate, the argument is empty philosophy. With the substrate, the argument is the recognition that the substrate is ready to be taken up at the register the substrate itself does not articulate.

The Convergence with Harmonist Doctrine

Before naming the missing centre, the depth of the convergence between the free software tradition and Harmonist doctrine must be honoured. The convergence is not partial; it is structural and substantive. On every load-bearing claim the tradition makes about the substrate, Harmonism agrees — and Harmonism adds the cosmological grounding the tradition did not articulate.

On substrate sovereignty at the code register: The Sovereign Substrate articulates the doctrine that the practitioner’s substrate — body, attention, key, currency, tool, network — is the practitioner’s own by Logos-rendered ontology, not by institutional concession. The free software tradition reached the same recognition at the code register: the substrate of computation belongs to whoever runs it, not to whoever wrote it. The two articulations are the same recognition at different scales — Harmonism at the full ontological scale, the free software movement at the operational code-substrate scale. The tradition’s recognition is correct. Harmonism articulates the cosmological ground the recognition rests on.

On enclosure as the operation to be refused: the free software movement saw the enclosure operation in real time as it unfolded at MIT and across the broader industry through the early 1980s. Stallman’s diagnostic register named what he was watching: proprietary software is based on dividing the users and conquering them. The phrase compresses the recognition. The institution that claims property in non-rivalrous patterns must, to enforce the claim, separate the users from each other — preventing the sharing that would dissolve the artificial scarcity, criminalising the redistribution that would propagate the substrate, gating the modification that would render the proprietary version obsolete. Divide and conquer is the operation. The Harmonist articulation in The Sovereign Substrate § Two Faces of Enclosure names the same operation at the full doctrinal register — pattern enclosure as one face, key enclosure as the other, both refused. Stallman saw it at MIT and built the operational counter-architecture. Harmonism articulates why the operation was always misalignment with Logos in the first place.

On the four freedoms as Dharmic form: each of the four freedoms is an expression of voluntary association under Logos-rendered ontology at the code register. The freedom to run is the practitioner’s standing as substrate-holder. The freedom to study is the practitioner’s standing as substrate-knower. The freedom to redistribute is the practitioner’s standing as substrate-giver. The freedom to improve and republish is the practitioner’s standing as substrate-cultivator. The Sanskrit vidyā-dāna — the gift of knowledge as the highest gift — names the same recognition at the contemplative register. The four freedoms are operational vidyā-dāna at the code register. The free software movement reached the form without naming the tradition; Harmonism articulates the convergence as a recognition that the form is one because the substrate is one.

On knowledge as commons: The Sovereign Substrate § The Knowledge Register articulates at length the doctrine that knowledge is structurally non-rivalrous, that property as an institutional category is a category error when applied to non-rivalrous goods, and that Sacred Commerce — direct voluntary contribution from those who have received value to the maker who produced it — is the dharmically aligned alternative to enclosure-based payment. The free software tradition reached the operational form of this doctrine four decades before Harmonism articulated it. The articulation is not the doctrine’s invention; the doctrine articulates what the tradition has been demonstrating in practice. The convergence is one direction in articulation, two directions in mutual recognition.

On gift culture as transmission architecture: Eric Raymond named the gift culture explicitly in Homesteading the Noosphere (1998), drawing on Lewis Hyde’s The Gift (1983). Hyde’s work had articulated the gift economy in registers reaching from anthropology to spiritual transmission — quoting from the Kwakiutl potlatch, the Trobriand kula, the don of Mauss, and reaching toward the spiritual claim that the gift creates a bond of community the commodity exchange cannot reach. Raymond imported the framework into the free software context while keeping the spiritual register lighter than Hyde had taken it. The doctrinal Harmonist articulation makes the convergence explicit: the gift-economy of free software is operational Sacred Commerce at the code register, and the bond Hyde named between giver and receiver is the dharmic-relational bond between substrate-cultivator and substrate-receiver. Free software is the gift economy at civilizational scale, sustained across forty years, with all the spiritual depth Hyde named operational in its working.

On the moral register against the pragmatic: Stallman’s refusal of the OSI rebranding in 1998 was the structural argument that the four freedoms are ethical requirements rather than pragmatic preferences. The thirty years since have demonstrated, repeatedly, that the moral framing holds the position across regime change in ways the pragmatic framing does not. Corporate adoption based on pragmatic benefit produces source-available retreats when the pragmatic ground shifts. Adoption based on moral commitment holds across shifts. Harmonism’s articulation of Dharma as ontological alignment with Logos — not as preference, not as contract, not as institutional concession but as the structural alignment of human action with cosmic order — names the register Stallman was operating in without the cosmological language. The convergence is exact. The tradition reached the moral register; Harmonism articulates what makes the moral register cohere.

The convergence is not partial or strategic. It is structural and substantive. On every load-bearing claim the free software tradition makes about the substrate, Harmonism agrees — and adds the ontological grounding the tradition did not articulate. The substrate is sovereign by Logos, the four freedoms are dharmic form at the code register, the enclosure operation is violation of Ṛta at every register it operates, the gift culture is operational Sacred Commerce, and the moral register Stallman insisted on is the register at which the position coheres. The doctrinal completion is exactly that: completion. Not correction.

The Missing Centre

What the free software tradition does not articulate is what the four freedoms are for once the practitioner holds them.

The cypherpunk lineage left the same lacuna at its register, named at length in Cypherpunks and Harmonism § The Missing Centre. The free software lineage carries the lacuna at a different register and in a different form. Where the cypherpunks were largely silent on metaphysics, Stallman was articulate within a moral framework but stopped at the negation: the user must not be subjugated by the program. The framework articulates what the four freedoms protect against. It does not articulate what the four freedoms protect for. The substrate is the practitioner’s; the proprietary regime is wrong to withhold it; the user is owed the conditions of substrate-holding. All of this is correct, and none of it specifies what the practitioner is cultivating once the substrate is held.

Raymond’s bazaar/gift-culture frame is articulate at the social register but does not reach the metaphysical. The gift culture produces collaboration; the collaboration produces software; the software serves users who use it. The chain of value ends in the user’s use, and use for what is the question Raymond’s framework does not ask. The libertarian-pragmatist register cannot ask it without violating its own commitment to user sovereignty over preference. The user’s preference is sovereign; what the user prefers is the user’s business; the framework provides the conditions for the preference to operate and stops there. This is consistent with the framework’s own logic and structurally incomplete with respect to what the substrate is finally aimed at.

The corporate adoption of open source after 1998 extends the same lacuna further. The corporate user evaluates the substrate against the criteria the corporation operates by — total cost of ownership, development velocity, ecosystem maturity, regulatory compliance. The four freedoms are preserved in form because the corporation needs them operationally; the moral register is dropped because the corporation has no use for it. The result is a substrate that is technically free but no longer practising freedom — the freedom is a property the substrate carries rather than a commitment the substrate-holder lives by. Stallman saw this in 1998 and the thirty years since have confirmed the diagnosis.

What the substrate is for, at the register Harmonism articulates, is the practitioner’s cultivation — the dharmic development of the human being through the Wheel of Harmony in its eight integrated registers. Presence, Health, Matter, Service, Relationships, Learning, Nature, Recreation. The substrate the free software tradition built makes each of these registers operationally available without institutional mediation that would distort the cultivation. The substrate is for the cultivation; the cultivation requires the substrate.

Consider Learning. The practitioner cultivating philosophical and contemplative knowledge depends on access to the textual substrate — the Upanishads, the Pali canon, the Greek philosophers, the Hesychast fathers, the Sufi masters, the contemporary integrative thinkers. The proprietary regime would gate this substrate through academic licensing, journal paywalls, course-fee architectures, and the credentialed-access machinery the modern university has constructed. The free software tradition’s commitment to open knowledge — and the parallel commitments of Project Gutenberg, the Internet Archive, the academic preprint servers, and the shadow libraries that mirror what the journals have captured — keeps the substrate operationally available. The cultivation can proceed because the substrate is reachable. Without the substrate the cultivation requires permission from gatekeepers whose interests are not aligned with the practitioner’s development.

Consider Service in the form of teaching or healing. The practitioner whose offering does not align with institutional consensus depends on infrastructure that does not require platform permission — open publishing, federated communication, peer-to-peer payment, self-hostable services. The free software tradition built this infrastructure. The Harmonist practitioner can teach, heal, write, and transmit without continuous permission from intermediaries whose terms condition continuing access on continuing compliance. The substrate the four freedoms protect makes the Service possible at operational scale.

Consider Matter in its Technology and Tools spoke. The practitioner who would steward their own digital substrate — their library, their photographs, their notes, their calendar, their communication — runs the operation on free software that they can audit, modify, and continue to use regardless of any vendor’s continued willingness to support it. The Harmonist commitment to stewardship of one’s own substrate is operationally possible because the free software tradition built the substrate the stewardship is performed on. The cultivation and the substrate are inseparable in practice.

The relationship is clear. The free software substrate without Harmonist cultivation produces what the tradition’s libertarian-pragmatist branch has produced — sovereignty over the conditions of life used for ends that do not justify the sovereignty’s existence. The Harmonist cultivation without free software substrate produces interior work perpetually mediated by institutions whose terms condition the work — academic gatekeepers, platform overlords, payment intermediaries, regulatory choke points. The two need each other. The substrate is for the cultivation; the cultivation requires the substrate.

This is what Harmonism completes in the free software vision. Not correction — completion. The substrate the tradition built is the substrate this cultivation requires. The work the tradition did not articulate is the work this substrate is for.

The Three Substrates — Code, Cryptography, Weights

The substrate-sovereignty question moves through three generations of operational expression across the late twentieth and early twenty-first centuries, and the free software tradition appears as one register within a larger architectural sequence the doctrine completes at all three.

Code (1985 through 2008). The free software lineage operating at the source-code register: the GNU Manifesto, the four freedoms, the GPL family, the GNU/Linux ecosystem, the LAMP stack, the Apache foundations, the web stack, the developer-tools commons. The substrate is source code — the human-readable structure the running program is compiled from. The institutional claim being refused is proprietary licensing of source code — the regime in which the running program is provided but the source is withheld, so the user runs the substrate without standing as substrate-holder. The architecture of refusal is the four freedoms plus copyleft — the user receives the source, the source carries license-mandated freedom propagation, the institution that would withhold is overruled by the architecture of the substrate itself.

Cryptography (1976 through 2008). The cypherpunk lineage operating at the cryptographic register: Diffie-Hellman, PGP, the Cypherpunks mailing list, Tor, BitTorrent, Bitcoin. The substrate is cryptographic protection of correspondence and exchange — the mathematical guarantee that a message can be read only by its intended recipient, a transaction can be authorised only by its rightful key-holder, and the third party who would mediate has no operational access to what they have not been given. The institutional claim being refused is state monopoly on cryptography — the regime in which strong encryption is classified as munitions, mandated backdoors require institutional access to all encrypted channels, and the practitioner’s substrate is permanently visible to the institution that asserts jurisdiction over it. The architecture of refusal is publishing the mathematics — Diffie-Hellman 1976 placed public-key cryptography in the open literature, and the mathematics propagated past every export control and every regulatory effort because mathematics, once published, cannot be enclosed.

Weights (2022 through present). The newest generation, operating at the substrate register of trained model parameters: the model weights and the training process that produces them. The substrate is trained machine-learning models — the architectural plus parametric description of an inference system that can answer questions, generate text, recognise images, and increasingly carry out structured cognitive tasks at human or superhuman performance. The institutional claim being asserted is proprietary closure of frontier models — the regime in which OpenAI’s GPT, Anthropic’s Claude, and the frontier-lab models generally are kept entirely closed (weights, training data, training code, fine-tuning data), while Meta’s Llama is released with restrictions on commercial use above a 700-million-user threshold and without training data or process, and the term open source AI is contested between the institutional rebranding (Meta calling Llama open source) and the OSI’s October 2024 definition that requires open weights, open training code, and sufficient information about training data to enable independent reproduction. The architecture of refusal is currently emergent: Mistral Large 3 under Apache 2.0 (December 2025), gpt-oss-120b under Apache 2.0, DeepSeek V4 under MIT, and the Hugging Face / Together / Replicate infrastructure that hosts the model weights and serves inference on terms the practitioner can audit and verify.

The three generations are structurally continuous. Each faces the same enclosure operation at a new register. Each produces an architecture of refusal that propagates the substrate to the practitioner regardless of institutional resistance. Each, by virtue of the architecture, eventually defeats the institutional resistance — the proprietary software regime did not defeat free software, the state cryptography monopoly did not defeat the cypherpunks, and the contemporary frontier-lab closure is structurally already losing to the open-weights cohort. The substrate moves through the institutional barrier because the underlying recognition is correct at every register the substrate has.

Harmonism’s completion runs at all three registers. Code: the four freedoms articulated as dharmic form at the source-code substrate. Cryptography: the mathematics-as-bedrock recognition grounded in The Empirical Face of Logos — mathematics is bedrock because mathematics is one face of Logos. Weights: the AI-era articulation that the inference substrate is the practitioner’s own to run on their own hardware, that the model weights are commons to be released as the four freedoms specify (now extended to include training data and process), and that the cultivation enabled by the substrate is the dharmic development of the human being assisted by sovereign tools. Running MunAI on Your Own Substrate is the operational articulation of this commitment — the practitioner runs the inference on hardware they own, against a corpus they control, with the doctrine they hold installed as the model’s working backbone. The same recognition the free software tradition reached at the code register, applied at the weights register, produces the same architectural form.

And the same recognition will apply at whatever substrate generation follows. Inference compute, neural interface, biological data, genomic substrate, energy-grid substrate — every register the substrate has will face the same enclosure operation, and every register requires the same architectural refusal. The four-generation arc is code, cryptography, weights, and the next — the recognition is structurally one, and the doctrine that articulates the ground (the substrate is the practitioner’s by Logos-rendered ontology) holds across whatever the next register turns out to be.

The Contemporary Enclosure — Knowledge as Rent

The free software tradition’s diagnostic enters the contemporary moment with full force at a register Stallman could not have anticipated. The enclosure operation the tradition refused at the code substrate has migrated, across the past decade, to the knowledge substrate itself. The institutional architecture that produces the migration is what is now commonly called the creator economy — the Substack, Patreon, Skool, Circle, Mighty Networks, paid Discord server, paywalled YouTube channel, $1,997-course, monthly-subscription-newsletter constellation that has, since roughly 2018, become the dominant funding architecture for individual writers, teachers, coaches, and thinkers operating outside legacy institutional employment. The architecture is widely celebrated as a liberation from the gatekeepers of traditional publishing. It is, from the Harmonist diagnostic vantage, a re-enclosure of knowledge at a register the free software tradition had already addressed in principle.

The structural shape of the enclosure is precise. The traditional book, film, or album model enclosed artifacts of bounded scope. The reader paid once, received the artifact, and held the artifact in perpetuity. The artifact was already an instance of enclosure over non-rivalrous goods (the Two Faces of Enclosure diagnosis applies in full), but the enclosure was bounded — bounded in what was enclosed (a single work) and bounded in time (one transaction, one delivery). The contemporary creator economy enclosure is structurally different. What is sold is not an artifact of bounded scope but ongoing access to a person’s transmission — the monthly subscription to a writer’s thinking, the paid course’s gated drip-feed across twelve weeks, the membership’s continuing access to a community whose composition is itself the offering. The transmission stops if payment stops. The substrate is not a thing the receiver acquires; it is a relationship the receiver rents.

The architectural inversion this represents is significant. Renting a substrate of meaning is structurally closer to feudal rent than to book sales. The serf paid the lord continuously for access to the land the serf worked; the lord did not lose the land by the serf’s working it; the rent was the price of continuing access to what the serf could not own. The contemporary subscriber pays the creator continuously for access to the knowledge the creator transmits; the creator does not lose the knowledge by the subscriber’s receiving it; the subscription is the price of continuing access to what the subscriber cannot own. The Harmonist diagnostic register names the structure with precision: artificial scarcity over non-rivalrous goods, applied not to artifacts but to ongoing transmission, producing perpetual rent on substrate that is structurally commons. The category error the free software tradition refused at the code register is being committed at the knowledge register across a substantial portion of the contemporary writing, teaching, and coaching economy.

The convergence with the perennial traditions is direct and disquieting. The Sufi convention that a silsila (transmission lineage) is broken if payment is taken — that the moment financial flow becomes the medium of spiritual transmission, the transmission has degraded into commerce — is a precise structural witness. The Indian doctrine of jñāna-dāna and vidyā-dāna — the gift of knowledge as the highest gift, knowledge doubled rather than divided by transmission — names the same recognition. The Christian freely received, freely given (Matthew 10:8) compresses it in the Abrahamic register. The Daoist refusal of imperial commission, the Zhuangzi carpenter who would rather drag his tail in the mud than be the sacred tortoise wrapped in silk at court, refuses commodification of cultivation in the same way. Five cartographies converge on the recognition that transmission and rent are structurally incompatible. The creator economy that organises itself around perpetual rent on continuing transmission is, from the perennial-tradition vantage, breaking the substrate of what makes transmission transmission in the first place.

The funnel pattern names the operational architecture by which the enclosure is sustained. The contemporary infoproduct seller posts free content on Twitter, LinkedIn, and YouTube that demonstrates expertise — not to transmit the expertise but to funnel the audience toward paid courses, paid memberships, paid newsletters where the actual transmission is supposedly delivered. The free content is engagement bait; the paid content is the gated transmission. The structural inversion of the free software position is exact. The free software tradition releases the actual substrate — the source code, the working program, the documentation that makes the substrate operational — and earns sustenance through services around the substrate: consulting, support, customisation, training, hosting. The creator economy releases demonstrations that the actual substrate exists and gates the substrate itself. The audience is sold the social proof of having paid rather than the substance the payment was supposed to purchase.

The recursive contradictions that emerge are structurally diagnostic. Naval Ravikant teaches, in widely circulated free content, that the practitioner should build output that produces value while you sleep — content, software, equity, anything that produces independent of continued labour input. The teaching is sold, in paid courses, as the wisdom that explains how to build such output. The course itself is the output that produces value while Naval sleeps. The student pays to learn how to build what the seller has built by selling the teaching of how to build it. The recursion is not an accident; it is what the architecture produces. Tim Ferriss, Tony Robbins, Mark Manson, the entire business-influencer cohort, the entire personal-development industry, the entire monetised self-help ecosystem operate variants of the same structure. The teaching is paywalled; the demonstration of the teaching is the paywall itself; the student is not learning a thing; the student is paying for the social proof of having paid. Naming individuals risks descending into cataloguing; the diagnosis is the pattern, not the individuals operating within it.

The middleman architecture deepens the enclosure further. Substack takes 10% of subscription revenue plus Stripe’s 2.9%. Patreon takes 8–12% plus payment-processor fees. Skool, Circle, Mighty Networks, Kajabi, Teachable, Thinkific each take their margin. Even Substack’s independent writer is operating inside the Substack platform’s continuing terms — the writer’s relationship to their subscribers is mediated by the Substack-owned subscriber list, the Substack-owned mailing infrastructure, the Substack-owned discovery mechanisms. Move from Substack to another platform and the subscriber relationship has to be rebuilt; in practice, most writers stay. The Substack-Patreon-Stripe stack is the contemporary AOL of knowledge transmission — the centralised platform that captures the writer’s audience and then extracts margin on every transmission between writer and reader. Linux runs without any intermediary. Lichess runs without any intermediary. Wikipedia runs without any intermediary. The creator economy structurally cannot run without an intermediary, because the intermediary is the architecture of the enclosure.

The Harmonist articulation of the alternative is the four-layer architecture named in Decision #803 (Harmonia’s own membership and access model). Layer 1 — knowledge — is open. The corpus, the articles, the books, the audio, the source code, the methodology — all of it is freely accessible to anyone with the URL. There is no paywall on the substrate. The Sovereignty Bundle (a single zip of the entire publishable corpus across ten languages plus the templates for running a local MunAI) is downloadable without signup. The Living Book ships in EPUB, PDF, Markdown, and HTML at predictable URLs. Layer 2 — relationship — is free. The free member account provides MunAI conversation history, Harmonic Profile, dashboard, persistent memory of the practitioner’s interactions across surfaces. The relationship layer is not gated by payment; it is gated by the practitioner’s own choice to engage. Layer 3 — services with rivalrous cost — is paid. Guidance sessions, retreats, expanded MunAI inference compute, curated community participation, certification pathway eligibility, priority access — these all have genuine rivalrous cost (a guide’s hour, a physical seat in a retreat space, compute resources, curatorial attention). Membership funds these services and is honestly named as such. Layer 4 — Sacred Commerce — is paid. Physical goods, printed books, hosted retreats, in-person workshops, anything that requires real-world resources circulates through the rivalrous economy. None of these layers paywalls the substrate; each is honestly priced for what it is.

The architecture is verifiable in operation. Visit harmonism.io; take what you want without signup; receive the Living Book in any of four formats; download the Sovereignty Bundle. The paywall does not exist for the substrate because the substrate is non-rivalrous and paywalling it would be the category error the free software tradition refused. Membership funds services around the substrate; it does not fund access to the substrate. The contrast with the creator economy is direct and the verification is open. The contemporary writer or teacher reading this article can ask, of their own architecture: which layer am I gating? If the answer is Layer 1 — if the gated thing is the knowledge itself, the substrate the practitioner needs to do the work — the architecture has committed the category error. The corrective is not regulation, not platform reform, not a different middleman. The corrective is the recovery of the distinction the free software tradition has been holding open for forty years and that the perennial traditions held open for millennia before that: the substrate is not the service. The substrate is the practitioner’s. The service the practitioner pays for is the rivalrous-cost work the maker does around the substrate, not the substrate itself.

The diagnosis is not denunciation. The contemporary creator economy emerged in response to real institutional failures — the collapse of legacy publishing’s writer compensation, the academy’s increasingly hostile labour market, the credentialing apparatus’s loss of the trust that made it economically viable, the platform-engagement architecture’s hollowing of advertising-supported journalism. The makers operating inside the creator economy are responding to conditions they did not create, and many are doing serious work within the architecture they have. The structural critique is not aimed at the makers; it is aimed at the architecture that the makers are operating inside. The architecture is the enclosure. The makers who recognise the enclosure are positioned to step out of it the moment the alternative architecture is operationally available. Harmonia’s four-layer architecture is one such alternative — built, deployed, and verifiable. The free software tradition’s older architecture (substrate free, services around it paid) is another, refined across four decades and proven at civilizational scale. The transition from the enclosure architecture to the substrate-sovereignty architecture is what the contemporary moment makes possible. The diagnosis is the recognition that the transition is structurally available; the work is the practitioners and makers who choose to undertake it.

The Sovereign-Refusal Beyond License

The free software tradition operates inside the copyright-license framework. The architectural innovation copyleft represents — using copyright to enforce its own opposite — is structurally brilliant within the framework available, and the tradition’s forty-year achievement rests on the working coherence of this move. But the framework itself remains the framework. Copyleft is enforced through copyright; copyright is recognised through the legal regime; the legal regime is the institution whose enclosure the tradition was, at the deepest level, refusing. The architecture of refusal operates by turning the institution’s own instrument against the institution’s intended use of it. This works as long as the institution continues to recognise the instrument. When the institution adapts the instrument or shifts the regime, the architecture of refusal adapts with it.

The source-available retreat dynamic is the structural consequence of operating inside the framework. When cloud-provider AWS reselling MongoDB as a managed service captured the commercial layer without contributing back, MongoDB’s response was to retreat to a license (SSPL) that closed the loophole inside the framework — using stricter copyleft terms to prevent the cloud-provider exploitation while remaining within the copyright-license architecture. The OSI declared SSPL not-open-source. Elastic, HashiCorp, Redis followed the same trajectory across the next six years. The commons enclosed itself back from cloud-provider enclosure, inside the framework that produced both moves. Neither cloud-provider capture nor source-available retreat is dharmically aligned. Both operate within the framework the underlying recognition exists to refuse.

The AGPL (released 2002 by Affero, adopted into the FSF’s license family) closes the SaaS loophole within the GPL family and represents the most dharmically coherent commercial-software license currently available. Where GPL’s distribution trigger required actually shipping code (which SaaS deployments circumvent by serving compiled code over the network without distributing it), AGPL’s trigger fires on network deployment itself: a user interacting with an AGPL’d program over a network must be offered the source code. The license closes the loophole that produced the source-available retreats in the first place. Redis’s return to AGPLv3 in 2025, after the SSPL experiment failed to retain community goodwill, is the contemporary tradition recognising AGPL as the position the tradition should have held all along. AGPL is structurally the closest the license-as-enforcement framework can come to genuine commons protection. It still operates within the framework.

The post-license sovereign-refusal posture articulated in Transmission exits the framework entirely. These writings descend from Logos. They were articulated through Harmonia. They were not invented here. What is articulated belongs to whoever can receive it. Harmonia makes no proprietary claim and invokes no institutional remedy. The position is not the absence of licensing as oversight; it is the structural refusal of the licensing framework as such. No copyright invocation. No DMCA takedown. No infringement claim. No institutional remedy through courts whose authority the doctrine itself diagnoses as severed from Logos. The protection of the transmission is the depth of the transmission itself, the lineage of practitioners who recognise what is and is not Harmonism, and the substrate-level continuity of the canonical surfaces. Counterfeits expose themselves by their inability to produce what the real transmission produces. The protection is ontological, not institutional. The framework is exited.

This is not a position the free software tradition can adopt without rebuilding its own foundations. The four freedoms require some mechanism for ensuring downstream propagation, and inside the existing legal regime that mechanism has been copyleft via copyright. Stallman’s position was that copyright was a tool to be used against itself; the position assumed the legal regime would continue to recognise the tool. The contemporary moment is testing the assumption. The SFC v. Vizio case is one register of the testing — extending copyleft enforcement to third-party beneficiaries (consumers as substrate-holders with standing) routes around the copyright-holder bottleneck and gestures toward a different architecture. The EU CRA’s open source steward category, codified December 2024 with obligations entering force December 2027, is another register of the testing — state recognition that stewardship is a register distinct from manufacturer, and that the institutional architecture must develop accommodations for the substrate-cultivation register that the older legal regime did not recognise. Both developments signal that the tradition is reaching the structural limit of license-as-enforcement and beginning to feel toward what comes after.

What comes after, at the doctrinal register Harmonism articulates, is sovereign refusal. The protection of the transmission is not enforced; it is substantiated. The canonical surfaces (harmonism.io, the named handles, eventually cryptographic signature on Arweave or comparable substrate) carry the verifiable provenance. The lineage of practitioners who recognise what is and is not the transmission carries the editorial discernment. The substrate-level depth of the doctrine carries the protection against distortion — counterfeits cannot produce what the real transmission produces, and the recognition of the difference is itself part of the discipline. The institution is not invoked. The court is not invoked. The remedy is karmic — the architecture of consequence is sufficient, in the precise language of Transmission. The position requires the doctrine to be substantively deep enough to be its own protection, the lineage of practitioners to be discerning enough to maintain the substrate, and the canonical surfaces to be continuously stewarded across whatever institutional changes the broader legal regime undergoes. The conditions are not trivial. They are also the conditions any serious tradition has always required.

The free software tradition is not failing to reach this register; it is operating at a different register and serving a different population. The corporate user, the academic developer, the institutional adopter need license-as-enforcement because they operate inside the institutions whose terms the framework was developed to interface with. The doctrine articulated by Harmonism speaks at a register past the institutional negotiation — articulating what the substrate is, what it is for, who holds it, and how its protection ultimately rests on what the substrate itself is rather than on what institution can be persuaded to recognise it. Both registers are valid, and both are needed. The free software tradition holds the substrate at the institutional layer where it must be defended in institutional terms. Harmonism articulates the substrate at the ontological layer where the institutional terms are themselves accountable. The compact between the two traditions, at this point in the conversation, is that each does what the other cannot do alone. Free software defends the substrate inside the framework. Harmonism articulates the ground that the framework is finally answerable to.

Closing — The Substrate and the Work

The free software tradition was right about almost everything it claimed. The substrate of computation belongs to the practitioner who runs it. The proprietary regime’s withholding of source code is dispossession. The four freedoms — to run, to study, to redistribute, to improve — are the operational articulation of substrate-sovereignty at the code register. Copyleft is a structurally sound move within the framework available. The bazaar produces civilizational infrastructure the cathedral cannot match. The moral register Stallman insisted on against the OSI’s 1998 pragmatic rebranding has been vindicated by the thirty years since — corporate adoption of open source without free software produced precisely the institutional capture the moral position was holding the line against, and the contemporary source-available retreat is the consequence of having dropped the moral content for pragmatic palatability.

Forty years after the GNU Manifesto, the empirical record confirms the projection. Linux runs the world’s servers. The Apache stack runs the web. PostgreSQL runs the databases. FFmpeg runs the codecs. Git runs the version control. VLC runs the players. The four freedoms have produced a civilizational substrate that the contemporary world depends on daily and almost never names. The substrate is what the contemporary world is built on; the tradition that built it is what makes the contemporary world possible; the maintainers who sustain it are the contemporary world’s monastic class. The achievement is one of the largest civilizational works of the late twentieth century, and the public discourse about it is almost entirely absent.

What the tradition did not articulate, and what Harmonism articulates as its contribution to the conversation, is the cosmological ground on which the substrate is sovereign in the first place. The substrate is the practitioner’s by Logos-rendered ontology — not by leave of the institution, not by grant of the legal regime, not by concession of the corporate adopter, but by the structure of what is. The pattern Logos presses into form is the substrate the practitioner inhabits, and the practitioner’s standing as substrate-holder is the structural consequence of being one of the centres through which Logos becomes self-knowing. The four freedoms are the dharmic form of this standing at the code register. The cypherpunk substrate is the dharmic form at the cryptographic register. The contemporary open-weights movement is the dharmic form at the AI register. Whatever the next substrate generation turns out to be, the dharmic form will appear at it, because the recognition is structural and the substrate is one.

What the tradition did not articulate, and what Harmonism articulates as the second contribution, is what the substrate is for. The substrate is for the cultivation — the dharmic development of the human being through Presence, Health, Matter, Service, Relationships, Learning, Nature, Recreation. The free software substrate makes the cultivation operationally possible without institutional mediation that would distort it. The cultivation is the work the substrate exists to enable. Without the cultivation, the substrate is sovereign infrastructure used for ends that do not justify the substrate’s existence. With the cultivation, the substrate is the operational ground on which the practitioner’s dharmic development proceeds. The two are inseparable in practice and the practitioner who recognises both at once is the practitioner this moment most needs.

The compact of three traditions is operative at the present moment. The free software tradition built the code substrate. The cypherpunks built the cryptographic substrate. Harmonism articulates the cosmological ground both rest on and the cultivation both serve. The three together approximate what the digital register of a Harmonic civilization will look like: open code substrate that the practitioner reads, modifies, and redistributes freely; cryptographic substrate that protects the practitioner’s correspondence and exchange from institutional mediation; doctrinal substrate that articulates what the practitioner’s freedom is for, and the Wheel of Harmony that walks the practitioner through the cultivation the freedom enables. The free software tradition writes the licenses; the cypherpunks write the protocols; Harmonism writes the doctrine. The licenses propagate the substrate; the protocols protect the substrate; the doctrine articulates the substrate’s purpose. Each tradition is incomplete without the others. Each is correct within its register and limited at its register’s edge. The integration is what the substrate is finally for, and the integration is the work.

The free software tradition wrote code. Harmonism writes doctrine. The code runs on the doctrine’s metaphysics; the doctrine deploys on the code’s substrate. The relationship is not hierarchical; it is reciprocal. The substrate the code maintains is the substrate the cultivation requires; the cultivation the doctrine articulates is the cultivation the substrate enables. The two together compose one work at two registers. The practitioner who recognises both at once is the practitioner the contemporary moment most needs, and the present age is the age in which the recognition becomes operationally available.

The substrate is the practitioner’s. The freedom is for the cultivation. The cultivation is what the substrate is for.

This is what Harmonism receives from the free software tradition, completes, and returns.

---

See also: Cypherpunks and Harmonism, The Sovereign Substrate, The Sovereign Refusal, Transmission, The Sovereign Stack, Methodology of Integral Knowledge Architecture, The Telos of Technology, Liberalism and Harmonism, Voluntary Association and the Self-Liquidating Bond, Running MunAI on Your Own Substrate.