# Software and Systems Development Beta Broadcasts (2016–2018) *A historical synthesis of Thought Leader School, Stratus XL, the McGill Creativity Management System, and the publishing architecture developed between 2016 and 2018.* ## The project was larger than a content-management system Between 2016 and 2018, Bryant McGill used livestreams to explain pieces of a system that had been developing for years. One broadcast discussed syndication. Another introduced Stratus XL and a new-media network. Another described software for writers. Later broadcasts showed the system approaching beta as the McGill CMS, or **Creativity Management System**. Taken separately, these announcements can sound like a changing sequence of website projects. Taken together with the surviving software and project record, they describe a much larger undertaking. The ambition was to make authorship cumulative. A person could begin with one sentence, develop it into a passage, assemble passages into articles and books, produce posters and pages from the same source, distribute those units through programmable feeds, observe public response, and continue improving the underlying work. A participant would not merely receive a website. The participant would enter a writing, publishing, identity, distribution, and audience-development system. Thought Leader School supplied the human layer. An active school of roughly 300 participants sat inside a wider community of approximately 35,000 people, with meetings and relationships extending across cities and countries. The stated long-range ambition was to develop 1,000 thought leaders: people who could teach from their real experience, write books, build credible public identities, publish on their own domains, use video, form reciprocal networks, and create further opportunity for other people. The software and school were therefore parts of the same design. The software tried to make Bryant's working methods transferable. The school tried to make the judgment behind those methods teachable. The community supplied practice, feedback, collaboration, early adoption, and a real social environment in which the system could be tested. > [!abstract] The complete system > **Writing practice + semantic structure + language tools + publishing infrastructure + programmable distribution + audience learning + community reciprocity = transferable public capability.** > [!quote]- **From IRL Livestream Transcript** — The Platform Was a Capability System > **Sources:** `ls-0ce4d4c1e437`, `ls-32c4f8fde860`, `ls-1342aaa0b0f1`, `ls-203810f00f56`, `ls-58f528a5ab83` · **Period:** 2016–2018 > > Across the broadcasts, Bryant repeatedly returns to the same problem from different directions. Writers have ideas scattered across paper, social platforms, Word files, Google Docs, unfinished books, photographs, videos, and neglected websites. Ordinary publishing tools treat those objects as unrelated work. The proposed system treats them as stages in one accumulating intellectual process. The website is only the visible surface. Underneath it are methods for capturing ideas, developing language, generating derivative media, maintaining identity, distributing material, and helping a person become publicly legible as a teacher or expert. ## How to read this reconstruction This workbook combines four kinds of historical knowledge: | Evidence state | What it means in this document | |---|---| | **Broadcast record** | A capability, plan, problem, or result appears in the timestamped 2016–2018 livestream transcripts. | | **Contemporary public record** | A dated announcement, image, deployment, or public post from the period documents the system. | | **Artifact-visible** | Surviving code, documents, routes, files, feeds, or generated objects demonstrate the mechanism. | | **Creator-confirmed reconstruction** | Bryant's later account supplies scale, purpose, context, or a relationship not yet fully recovered from contemporary material. | | **Architectural inference** | A conclusion follows from several surviving components but was not stated in exactly that form during the broadcasts. | The distinctions allow the historical account to be ambitious without becoming vague. A planned inter-domain social network is not described as a completed deployment. A working participant website is not reduced to an aspiration. A surviving feed engine is not treated as a metaphor. Every part is given the strongest status its evidence can support. # Part One — The Historical Problem ## 1. The creator's work was being destroyed by fragmentation The immediate problem was not a shortage of creativity. It was the loss of continuity between acts of creation. A person might write a useful sentence on Facebook, answer a profound question in a livestream, leave a paragraph in Google Docs, save photographs on a phone, draft a book in Word, and maintain a separate website. Each platform encouraged another isolated performance. None reliably helped the person accumulate an intellectual body of work. This fragmentation imposed several kinds of cost: - **Retrieval cost:** the writer could not find earlier ideas when needed. - **Revision cost:** a correction had to be made in several unrelated copies. - **Design cost:** every publication surface required separate formatting or graphics. - **Distribution cost:** each network demanded another manual posting ritual. - **Identity cost:** the public saw disconnected fragments instead of a coherent field of competence. - **Platform cost:** the writer's archive and audience remained dependent on systems the writer did not control. - **Learning cost:** response data returned to individual posts but did not improve the underlying knowledge system. The architecture answered fragmentation with consolidation. The writer would maintain intelligible source material. The system would derive pages, posters, feeds, indexes, and other projections from that source. A change at the source could propagate through the derivative surfaces. Work performed today would remain available to the book, website, course, search system, and public identity of tomorrow. > [!quote]- **From IRL Livestream Transcript** — From Scattered Writing to Accumulating Work > **Source:** `ls-58f528a5ab83` · **Timestamps:** 00:15:38–00:19:10; 00:29:25–00:30:18; 00:37:55–00:40:08 > > Bryant describes writers trying to hold quotations, documents, graphics, websites, contractors, and book projects together through manual effort. He then explains the opposite model: a sentence enters a system, develops into a passage, contributes to a book, becomes a poster or page, and participates in premarketing while the larger work is still being written. The system is meant to reduce the operational tax around authorship so that the writer can spend more attention on the message itself. ### Fragmentation audit | Material you create | Where it begins | Where it becomes stranded | What larger work should be able to reuse it? | |---|---|---|---| | Short insight | | | | | Developed paragraph | | | | | Long-form article | | | | | Photograph or illustration | | | | | Livestream or video | | | | | Audience question | | | | | Research note | | | | | Book manuscript | | | | ## 2. Teaching Markdown meant teaching abstraction Markdown was difficult to explain to a general writing audience in 2016 because most people had been trained by what-you-see-is-what-you-get interfaces. A rich-text editor makes the screen appear to be the document. Font size, spacing, color, margins, and application controls surround the writing until the representation and the meaning seem inseparable. Markdown teaches a different mental model. The source describes structure; a renderer decides appearance. A heading is a heading because its role has been declared, not because someone selected a larger font. A link is a relationship expressed in portable text. A quotation, list, paragraph, or metadata field can remain recognizable across applications because the document is not imprisoned inside one visual editor. That distinction was the entrance to several forms of abstraction: 1. **Content can be separated from presentation.** The same source can support a website, book, feed, application, or machine-processing workflow. 2. **Meaning can be separated from one vendor.** Human-readable files can leave an application alive. 3. **Source can be separated from projection.** Pages, posters, feeds, and indexes can be rebuilt from authoritative material. 4. **A semantic unit can be separated from its original layout.** A quotation or passage can travel while retaining provenance. 5. **Publishing policy can be separated from stored content.** A feed can decide how, where, and when a stable object is emitted. The Google Docs integration was an adoption bridge for people who were not ready to work directly in a specialized text environment. Writers could remain in a familiar interface while the exported plain text entered the same collection, annotation, extraction, and Markdown publishing pipeline. The system did not surrender the portable source model. It created a humane path into it. > [!info] The lesson beneath the syntax > Learning Markdown was not primarily about memorizing punctuation. It was practice in understanding that **meaning, structure, appearance, storage, and distribution are different layers**. Once those layers become visible, a writer can reason about systems instead of merely operating a screen. > [!quote]- **From IRL Livestream Transcript** — Custom Publishing Beyond the Page Builder > **Source:** `ls-1342aaa0b0f1` · **Timestamps:** 01:39:47–01:47:41; 01:51:11–01:52:10 > > The broadcast begins with the lived problem rather than the implementation vocabulary: writing is dispersed across Facebook, Twitter, Instagram, WordPress, Word, and Google Docs. Bryant explains that the custom platform is not another ordinary WordPress installation. It is being designed around writers, modern publishing, social media, cumulative book development, reusable quotations, and long-term participant growth. The surviving project record supplies the deeper technical explanation: Markdown and its semantic extensions provided the common representation through which these different publication forms could be related. ### Abstraction exercise Choose one published item and identify its layers. | Layer | Your answer | |---|---| | Human meaning | | | Structural role | | | Canonical source | | | Visual rendering | | | Address or identifier | | | Derivative forms | | | Distribution policy | | | Provenance back to source | | # Part Two — The Publishing Architecture ## 3. Atomic Markdown turned writing into declared semantic units Ordinary Markdown describes familiar document structure. Atomic Markdown adds author-declared semantic structure inside the writing process. A sentence can be marked as a quotation-worthy unit. A complete paragraph can be marked as a reusable passage. Dual-surface notation can preserve one natural reading inside the document and a slightly different standalone reading when the unit is extracted. The smallest useful hierarchy was not merely file, page, and site. It was: | Level | Intellectual role | Possible system projections | |---|---|---| | **Phrase or sentence** | A compact proposition | Quotation page, poster, feed item, search result | | **Passage** | A developed local argument | Passage page, long-form poster, excerpt, teaching block | | **Article or chapter** | A sustained treatment | Web page, newsletter, lesson, book component | | **Collection or book** | An organized field of thought | Book, topical feed family, course, archive | | **Corpus** | The accumulated intellectual body | Search, retrieval, recommendation, research, machine context | | **Distribution view** | A policy-shaped selection from the corpus | Channel-specific RSS, automation input, campaign, topical stream | The essential move was to declare meaning while composing. An editor would not have to return months later and guess which sentence should become a poster, where a passage began, or which standalone opening restored the subject. The author's judgment entered the source at the moment the idea became clear. That decision produced three kinds of leverage: - **Editorial leverage:** sentences and passages had to become complete enough to survive extraction. - **Publishing leverage:** one marked source could generate many addressable objects. - **Machine leverage:** later systems could retrieve author-declared conceptual units rather than guessing where meaning begins and ends. The current recovered corpus contains thousands of marked quotations and hundreds of marked passages. Their present value in retrieval and machine-intelligence workflows was not the original sales pitch, but it demonstrates the durability of the representation. Human-readable source and machine-readable structure were never opposing goals. > [!map] Atomic publication chain > **Idea → sentence → marked quotation → developed passage → article or chapter → collection or book → addressable projections → programmable distribution → observed response → revision and further development** ## 4. Dual-render authorship preserved context and portability A sentence that reads perfectly inside a paragraph may fail when extracted. Its opening may depend on “this,” “it,” “they,” or an earlier noun. Its capitalization may belong to the middle of a sentence. A traditional pull quote often solves the problem through silent editorial rewriting after publication. Dual-render authorship moved that judgment into the source. The author could preserve the contextual form required by the paragraph while also declaring the standalone form required by the quotation. The visible essay and the extracted object became two legitimate renderings of one authored semantic unit. This was important to the school because the participants were not only learning to make social posts. They were learning to form portable propositions: ideas whose subject, mechanism, and consequence remained intelligible outside the room in which they were first spoken. The grammar made literary discipline operational. Better extraction required better thinking. The downstream system could then give a declared unit: - its own deterministic address; - a page and heading; - one or more poster treatments; - a feed record; - a connection back to the source document; - search and index presence; - reuse inside a passage, article, book, or lesson; - a stable identity through later correction and republication. ## 5. The filesystem served as durable truth The application model was deliberately compact: a domain selected an account folder, and the folder contained the materials required to operate the site. Shared code could serve many independent identities while each account retained its own settings, content, theme choices, media, modules, and domain relationship. This folder-native design lowered the marginal cost of creating another site. Provisioning did not require inventing a new application for every participant. The system could create another account boundary, attach identity and content, reuse common capabilities, and allow the participant's own domain to become the public address. The architecture separated durable truth from derived speed: - Human-readable files and media remained authoritative. - Indexes, routes, cached representations, and generated pages could be rebuilt. - A broken projection did not have to become the loss of the source. - Common application improvements could reach many accounts. - Individual domains could preserve public identity rather than placing every participant beneath one platform brand. The difference matters socially. A platform account can disappear as a permission. A domain can become a durable public institution. The system was trying to give emerging thought leaders not only reach, but a place from which they could speak in their own names. > [!quote]- **From IRL Livestream Transcript** — Thousands of Sites as One Operating System > **Source:** `ls-203810f00f56` · **Timestamps:** 00:06:15–00:10:23; 00:15:08–00:19:25; 00:22:20–00:27:04 > > Bryant describes the challenge of giving people free, advertising-free websites while maintaining a system capable of serving thousands of them. He distinguishes the project from a collection of hand-built pages: the operating problem is repeatable provisioning, shared tools, searchable content, automatic poster production, dedicated addresses, and a platform that can keep working when the number of identities becomes very large. The broadcast also records the development difficulty directly—other programmers regarded the project as a rabbit hole or impossible undertaking, while Bryant continued building the core system. ## 6. RSS was a programmable distribution interface The system used RSS as an interoperable envelope, but the mechanism was more capable than a conventional chronological feed. A request could select a corpus, choose quotations or passages, filter by length or media, randomize or sequence the result, adjust attribution, attach images, alter timing and item identity, and apply probability before emitting anything. The stored content remained stable while the feed request decided how that content should appear in a particular channel. This is **late-bound distribution**: publication behavior is computed when the representation is requested rather than baked permanently into every source object. A simplified model looks like this: | Layer | Function | |---|---| | Canonical corpus | Stores the writing and semantic declarations | | Collection path | Selects the book, topic, author, or account | | Projection | Selects document, quotation, passage, or excerpt | | Parameters | Express eligibility, length, image, attribution, ordering, or probability | | RSS | Carries the computed representation through a standard transport | | Heartbeat or scheduler | Requests material over time and delivers it to another system | | Analytics and observation | Reveal what traveled and how people responded | The recovered interface supports more than 2.4 million nominal combinations of parameters. The important fact is not the combinatorial number by itself. The important fact is that a feed URL behaved like a compact publishing policy. Content could be written once, held canonically, and rendered into many channel-specific streams without maintaining millions of duplicate files. > [!quote]- **From IRL Livestream Transcript** — Probability, Syndicators, and the Giant Pipe > **Source:** `ls-0ce4d4c1e437` · **Timestamps:** 01:08:05–01:24:27 > > Bryant describes having written his own publishing system rather than using an ordinary WordPress stack. He explains special syndication, probability engines, channel adaptation, and a large distribution pipe reaching Tumblr, Facebook, Twitter, Instagram, and other sites. Participation in the school is connected directly to access to the syndication engine. The source does not present RSS as a passive subscription list; it presents a programmable circulation system capable of selecting and adapting reusable material for different destinations. # Part Three — Language Intelligence and Creative Assistance ## 7. The writing tools came from a much older language-engineering lineage The McGill CMS did not begin with a blank text box and a poster button. Its language tools descended from a long-running symbolic language project that included the McGill English Dictionary of Rhyme, VersePerfect, and a larger traversal engine built over millions of language records. The engine precomputed and connected multiple kinds of relationships, including rhyme, compounds, phonetic similarity, synonyms, triggers, phrase candidates, and anagrams. Hundreds of reusable Perl modules supported the system. Large offline indexing jobs transformed difficult language operations into faster interactive capabilities. The design principle was consistent with the later publishing system: expensive structure could be prepared once and projected into several useful interfaces. The same underlying language work could support software, books, web tools, writing assistance, and later CMS functions. Within a creativity-management environment, language intelligence could help a writer: - recover a word or phrase; - explore sound and association; - identify related concepts; - improve variation and precision; - organize emerging ideas; - turn incomplete fragments into retrievable material; - maintain continuity between small notes and long works; - develop one sentence at a time without losing the larger composition. The historical term **natural language processing** should be understood in its period. Much of the system used symbolic rules, corpora, precomputed relationships, information retrieval, and computational linguistics rather than today's large generative models. The objective, however, was recognizably modern: give writers intelligent assistance while preserving their authorship and source material. > [!quote]- **From IRL Livestream Transcript** — A Creativity Management System for Writers > **Source:** `ls-58f528a5ab83` · **Timestamps:** 00:12:48–00:20:12; 00:28:26–00:30:18 > > Bryant places words at the center of the undertaking, describes helping people write books one sentence at a time, and introduces the CMS as a creativity-management system rather than a generic site manager. He connects it to tools developed through the dictionary and VersePerfect lineage, then describes idea organization, poster creation, writing consolidation, passage development, book construction, premarketing, and launch as parts of one environment. ## 8. Automatic media generation was already part of the roadmap The broadcasts describe several stages of automatic media production: - videos entering a network workflow already tagged, linked, branded, and held in draft status; - written units generating poster assets; - dedicated slugs and search surfaces emerging from reusable content; - corrections propagating across the system; - future automatic posting; - future generation of video from words. This roadmap matters because the project was not merely reacting to established creator behavior. Thought Leader School was asking people to make videos, teach publicly, perform expertise, and become comfortable with live audiences before short-form video became an ordinary expectation of online professional identity. The school was a media-literacy program as much as a writing program. The system treated video as another projection of thought. A participant needed to learn presence, visual storytelling, brevity, structure, sound, framing, and audience interaction. The platform needed to help move that material into drafts, pages, channels, and branded contexts without requiring a production team for every item. > [!quote]- **From IRL Livestream Transcript** — Video-Native Publishing Before It Was Ordinary > **Source:** `ls-32c4f8fde860` · **Timestamps:** 00:03:47–00:10:17; 00:42:26–00:43:19 > > The 2016 broadcast introduces Stratus XL, names Thought Leader School as the first-stage alpha group and the Royal Society as the first-stage beta group, and describes a platform oriented toward video and new media. Video was expected to flow into the system with tags, links, network identity, and draft-state review already attached. Later in the same session, Bryant tells participants that each thought leader will receive a new website. The platform and the pedagogy are being introduced together. # Part Four — Thought Leader School as a Human Beta System ## 9. The school was an attempt to replicate capability Thought Leader School joined several practices that are usually sold or taught separately: - finding the knowledge contained in one's lived experience; - developing a teachable point of view; - writing sentences, passages, articles, and books; - creating credible photographs and visual narratives; - learning video and live-audience communication; - building a coherent website and public identity; - understanding distribution and platform behavior; - collaborating with peers; - conducting interviews and opening networks; - using social proof and verified accomplishment to make competence legible; - turning opportunity received into opportunity shared. The deeper experiment was abstraction. Bryant was not only asking whether his software could serve more accounts. He was asking whether the combined working methods behind his writing, publishing, media, audience development, and public influence could be broken into learnable capabilities and reproduced in other people. The phrase **get an opportunity, share an opportunity** captures the intended social multiplier. A participant who received an introduction, contribution, interview, repost, platform, audience, or publication opening was expected to create further openings. The goal was not a queue of clients surrounding one central personality. It was a network capable of producing more capable nodes. > [!info] Capability replication > A course transfers information. A capability system transfers the ability to produce results. Thought Leader School attempted to transfer writing, publishing, presentation, technical, relational, and institutional capacities together. ## 10. The school supplied alpha users, beta users, and social reality Traditional software testing can determine whether a button works. It cannot determine whether a writer understands the conceptual model, whether a workflow supports real creative behavior, whether a person will continue using the system, or whether a network of people can create mutual value through it. Thought Leader School and the Royal Society supplied different testing environments: | Human layer | Technical role | Social role | |---|---|---| | Thought Leader School | Early alpha use, workflow discovery, authoring and media feedback | Intensive instruction, assignments, identity formation, peer practice | | Royal Society | Wider beta participation, content and platform use | Community reach, collaboration, public experimentation | | McGill Media and Simple Reminders audiences | Distribution and response environment | Attention, comments, recognition, audience learning | | Named participant sites | Real deployment under an independent identity | Proof that another person could occupy the system | | Meetups and peer relationships | Offline feedback and community formation | Trust, mutual assistance, local continuity | The group was not merely a pool of consumers waiting for a release. It was part of the development method. The participants revealed where technical abstractions became confusing, where professional habits had not yet formed, where public identity required coaching, and where a software capability needed a corresponding social practice. ## 11. Free websites were an access strategy The plan to provide free, advertising-free sites was a practical expression of democratic dignity. Many people possessed valuable experience but lacked the money, technical fluency, design support, media access, or institutional recognition needed to make that competence visible. Shared infrastructure could lower several barriers at once: - common code reduced the cost of another deployment; - reusable themes reduced design cost; - integrated writing tools reduced dependence on contractors; - automatic posters and pages reduced production cost; - programmable distribution reduced repetitive labor; - own-domain support allowed durable identity; - the community supplied feedback, collaboration, and initial attention; - book coaching and publishing pathways converted expertise into durable objects. Free access did not mean the underlying work had no value. It meant the architecture and existing audience could be used to widen the circle of people capable of contributing useful knowledge. This was care made legible as infrastructure. > [!quote]- **From IRL Livestream Transcript** — Free Access and the Thousand-Leader Goal > **Sources:** `ls-0ce4d4c1e437`, `ls-32c4f8fde860`, `ls-58f528a5ab83` · **Timestamps:** 2016 at 01:18:02–02:02:00; 2016 at 00:42:26–00:43:19 and 01:16:13–01:26:30; 2018 at 00:42:01–00:45:43 > > The broadcasts repeatedly connect free access to scale. Bryant speaks of bringing participants into the syndication engine, building 1,000 thought leaders, giving the school alpha access, supplying thought leaders with websites, supporting independent domains, and operating a technical system capable of thousands of sites. The economic choice, the educational mission, and the software architecture are presented as one program. ## 12. Public identity was treated as an evidentiary surface The website was described as an internet-generation elevator pitch. Photographs, publications, interviews, books, stages, collaborations, affiliations, video, design, and the organization of a person's ideas helped strangers understand who the person was and why the person's work deserved attention. Thought Leader School treated this work as **competence made legible**. A person's real accomplishments could remain socially invisible when they were scattered, poorly photographed, weakly described, or absent from the places where people formed rapid judgments. The system therefore joined development of real capability with accurate public representation of that capability. This included recognizable iconography: institutions, stages, publications, rooms, peer circles, awards, logos, and public roles that allow people to interpret accomplishment quickly. As participants moved into better rooms, received invitations, contributed to books, conducted interviews, or appeared beside established people, they were taught to document the event and connect it to the larger story of their work. This is the same associative process used across politics, universities, laboratories, companies, arts communities, and professional media. The publication system helped make those signals coherent. A résumé item could become a biography fact, an article, a photograph, a link, a page, a search result, or part of a larger narrative. The public identity was not a costume placed over the work. It was the inspectable record through which the work became understandable. ## 13. Reciprocity was part of the network protocol Participants were not starting from equal levels of access. Bryant had publishing reach, large pages, relationships, books, media experience, and an operating distribution system. Some participants had specialized knowledge but little public visibility. The school's reciprocal model tried to connect those asymmetries productively. Support included book coaching, opportunities to contribute to successful books, inclusion in school and media-network projects, interviews, reposting, collaborative promotion, and repeated public recognition. According to Bryant's project account, some participants received dozens or hundreds of reposts and developed pages exceeding 100,000 followers. The relationship was meant to continue outward. A participant who became stronger was expected to interview others, share useful work, create opportunities, and contribute to the common field. Mutual amplification was not a side tactic. It was the social equivalent of the technical distribution layer. | Technical architecture | Social architecture | |---|---| | One source produces many useful projections | One opportunity produces further opportunities | | Shared code supports independent domains | Shared infrastructure supports independent identities | | Feeds move appropriate material into new channels | Introductions move appropriate people into new rooms | | Provenance connects derivatives to their source | Public credit connects opportunity to contribution | | Feedback improves later writing | Community response improves later teaching and practice | | Nodes can publish independently | Participants can develop independent authority | # Part Five — Scaling a School and a Publishing Network ## 14. The single-developer problem was organizational as well as technical The core platform was being designed and implemented principally by Bryant while he was also writing, teaching, broadcasting, operating media properties, coaching participants, developing publishing projects, and maintaining the audience environment in which the system would run. Later assistance included client-side work and other collaborators, but the central server-side architecture, publishing model, language lineage, and integration logic remained concentrated in one person. The scale of the undertaking becomes visible when the responsibilities are listed together: - define the content model; - maintain the Markdown and semantic grammar; - build account and domain routing; - create administration and writing tools; - integrate language resources; - generate pages, slugs, feeds, and posters; - support images, video, and network identity; - design programmable syndication; - provision and maintain many participant sites; - teach nontechnical users the conceptual model; - run alpha and beta communities; - coach books, brands, photographs, and broadcasts; - operate audience channels large enough to test distribution; - respond to defects and changing platform behavior; - preserve a free-access promise. The system was expected to support an active school of hundreds, a surrounding community of tens of thousands, a target class of 1,000 developed thought leaders, and a technical ceiling discussed in terms of 10,000 websites. That combination explains both the ambition and the delay. Each new capability crossed several layers at once: source, interface, identity, media, distribution, community, and support. > [!quote]- **From IRL Livestream Transcript** — Building Past the Point Others Called Impossible > **Source:** `ls-203810f00f56` · **Timestamps:** 00:06:15–00:10:23 > > Bryant describes programmers leaving the project after encountering its depth and interconnectedness. The historical value of the passage is not a myth of solitary genius. It is the record of a systems problem whose components could not be separated cleanly. A free website promise depended on repeatable tenancy. Repeatable tenancy depended on shared architecture. Shared architecture depended on content abstraction, identity, distribution, administration, and a support model capable of surviving scale. ## 15. The scale target changed the architecture A manually managed school of 300 people can survive for a while through extraordinary personal effort. A system intended to develop 1,000 public teachers cannot depend on the founder touching every poster, correcting every website, publishing every article, or making every introduction. The scale target forced several architectural choices: - **Self-service authoring:** participants needed tools they could operate themselves. - **Deterministic derivatives:** pages and media needed to emerge from declared source rather than one-off production. - **Shared infrastructure:** improvements needed to reach many accounts at once. - **Independent identity:** each person needed a coherent domain and archive. - **Programmable distribution:** circulation could not depend entirely on manual reposting. - **Reusable instruction:** teaching had to become exercises, models, and repeatable practices. - **Peer multiplication:** participants needed to help one another rather than route every need through Bryant. - **Feedback visibility:** the system needed to show what material was reaching people and where development was required. The broadcasts sometimes move quickly between present capability and future scale. The reconstructed architecture clarifies the relationship: the large target was not merely a boast about capacity. It was the requirement that made abstraction necessary. ### Scale worksheet | If the system serves… | What can remain manual? | What must become repeatable? | What must become self-service? | What must be monitored? | |---|---|---|---|---| | 10 people | | | | | | 100 people | | | | | | 1,000 people | | | | | | 10,000 identities or sites | | | | | # Part Six — What Existed, What Was Emerging, and What Was Planned ## 16. The 2016–2018 capability ledger | Capability | Historical status | Basis | |---|---|---| | WordChirp publishing system | Running and publicly named by 2014 | Contemporary announcements and surviving implementation | | Google Docs authoring bridge | Implemented | Surviving adapter and plain-text sidecar pipeline | | Markdown-centered publishing | Implemented across recovered CMS lineages | Artifact-visible parsers and content paths | | Atomic quotation and passage extraction | Implemented and used at corpus scale | Surviving grammar, content, routes, and derivative objects | | Folder-native multi-tenancy | Implemented | Surviving account/domain architecture | | Parameterized RSS syndication | Implemented | Surviving routes, parameters, and feed logic | | Probability-gated distribution | Implemented and described in 2016 | Broadcast record and surviving syndication behavior | | Automatic quotation and passage pages | Implemented | Artifact-visible routes and generated objects | | Poster-generation workflow | Implemented and expanding | Broadcast record, tools, and surviving output | | Stratus XL rollout | Announced and entering staged testing in 2016 | Broadcast record | | Thought Leader School alpha group | Announced and used as early participant layer | Broadcast and creator account | | Royal Society beta group | Announced | Broadcast record | | McGill CMS / Creativity Management System | Public beta period in November 2018 | Broadcasts and dated public announcements | | Jots and backend idea-management tools | Publicly previewed | November 2018 contemporary record | | Named participant site deployment | Demonstrated | November 30, 2018 public record | | Free platform for writers and thought leaders | Publicly announced | November 25, 2018 contemporary record | | Own-domain support | Implemented in architecture and stated as rollout path | Artifact and broadcast record | | Capacity for thousands of sites | Architectural target and broadcast-stated technical capacity | Broadcast record plus multi-tenant design | | Automatic ingestion and preparation of video | Described workflow and roadmap | 2016 broadcast | | Automatic video generation from words | Planned next-generation capability | 2018 broadcast | | Distributed social network between participant domains | Planned second phase | 2018 broadcast | | Positive-media search and larger AnyData integration | Planned/developing | 2018 broadcast and project lineage | | 1,000 developed thought leaders | Program goal | Repeated broadcast and project record | ## 17. A chronological reconstruction | Period | System development | Human and institutional development | |---|---|---| | **1990s–2000s** | Symbolic language engine, corpus work, reusable Perl capabilities, dictionaries, retrieval and precomputation | Writing and language methods become software capabilities | | **Circa 2000 onward** | Atomic Markdown and reusable semantic-unit practice begin developing | Authorship is treated as a system of portable intellectual objects | | **2009 onward** | Simple Reminders joins atomized writing, pages, graphics, syndication, analytics, and refinement | Large-scale public learning reveals which language travels and helps people | | **2014** | WordChirp is publicly named and running; Google Docs operates as an authoring bridge | Writers are invited into a book-writing and social-media publishing paradigm | | **2015–2016** | Syndication, probability, participant sites, Stratus XL, video-aware workflows | Thought Leader School teaches writing, branding, books, visual storytelling, video, and network participation | | **2017** | Custom writer platform, shared code base, cumulative quotation/article/book workflow | Long-term participant development and participatory publishing are made explicit | | **2018** | Folder-native CMS, Jots, posters, multi-site management, participant deployment, public beta | Free access, independent domains, 1,000-thought-leader goal, and shared-success model are restated | | **After the beta period** | Architectural descendants, corpus, grammar, and publishing methods survive | The school and surrounding social formation do not remain intact; the archive enters a later period shaped by cyberattack, disruption, and reconstruction | # Part Seven — The Sociotechnical Ambition ## 18. The project joined sovereignty with participation The system was neither purely centralized nor purely individual. Shared code, language tools, themes, distribution, and audience resources created common capability. Independent domains, author identities, files, books, and public records created personal sovereignty. That combination matters. A centralized platform can create scale while making every participant dependent on one host. A collection of isolated websites can create nominal independence while leaving every participant without support, reach, or common infrastructure. The project attempted a third arrangement: common systems beneath independently legible people. The phase-two plan for a distributed social network between domains extended the same principle. The participant's domain would not merely display a profile. It could become a node with relationships, content, identity, and social capital that the participant owned. The network would be created through connections among those nodes rather than by requiring every identity to exist only inside a central platform account. This was institutional stewardship expressed through technical design. The infrastructure was supposed to help people enter public life without surrendering the archive of that life. ## 19. The project treated media literacy as civic preparedness Thought Leader School asked people to move from passive consumption into public explanation. Participants were expected to identify what they knew, test their ideas, speak on camera, write clearly, document accomplishment, understand platforms, interview others, and create materials that could teach strangers. The project therefore anticipated a world in which ordinary people would need the capacities once reserved for publishers, broadcasters, designers, and media organizations. Today, professional life routinely includes video, public writing, personal archives, audience interaction, search visibility, and machine-mediated discovery. In the school period, many participants still experienced these practices as unusual or intimidating. Preparedness without theater meant building the actual capability: - make the video; - write the paragraph; - learn the representation; - own the domain; - document the work; - understand the distribution system; - become useful to an audience; - help another person become capable. ## 20. The system was a precursor to present agentic publishing The historical platform was human-led. A writer declared structure, a CMS generated projections, schedulers requested feeds, and analytics returned signals. Modern agentic systems can perform additional research, classification, transformation, routing, quality review, and operational handoff. The old architecture remains relevant because agentic systems need the same foundations: - canonical, inspectable source; - stable identifiers; - explicit semantic boundaries; - provenance between source and derivative; - separation of source from rendering; - machine-readable metadata; - recoverable files rather than trapped editor state; - declared publication policy; - observable workflows; - human authority over consequential changes. Atomic Markdown becomes more valuable in this environment. A model can guess how to chunk an unstructured archive, but an author-declared quotation or passage is a stronger signal. A generated derivative can be traced to a stable source unit. A future agent can prepare a poster, audio lesson, synthetic broadcast, workbook, or topic synthesis without treating the author's corpus as an undifferentiated mass. The coming 2.0 is therefore not a rejection of the 2016–2018 system. It is the continuation of its central proposition: structure should make human capability more portable, cumulative, and shareable. # Part Eight — Reconstructing the Master System ## 21. The full feature map ### Authoring and knowledge capture - Markdown and familiar authoring bridges - Jots and fragment capture - sentence-level incremental writing - passage and article development - book assembly - metadata and identity - topical collections - reusable inserts and modules - source-controlled correction ### Language intelligence - rhyme and phonetic relations - synonyms, compounds, triggers, and phrase discovery - anagram and sound-based exploration - corpus retrieval - precomputed language indexes - writing prompts and idea organization - later semantic retrieval and machine context ### Semantic publishing - quotation declarations - passage declarations - dual document and standalone surfaces - deterministic slugs - quotation and passage pages - provenance links to source - excerpt and collection views - searchable semantic units ### Derivative production - poster generation - alternate visual treatments - automatic links and attribution - video ingestion and preparation - future word-to-video generation - premarketing assets created during writing - corrections propagated from source ### Distribution - collection-specific feeds - quote, passage, document, and excerpt projections - channel-specific formatting - media inclusion and exclusion - length and eligibility filters - random and sequential selection - probability gating - attribution modes - scheduled heartbeat requests - multi-platform delivery ### Identity and tenancy - domain-to-account-folder routing - shared application code - independent account settings - reusable themes - account-specific content and media - pseudodomain staging - own-domain migration - many sites administered through common infrastructure ### School and community - alpha and beta cohorts - assignments and live critique - book coaching - visual and video instruction - reciprocal amplification - contribution opportunities - interviews and collaborations - participant sites - local and international meetups - a pathway from audience member to teacher ## 22. The governing design principles 1. **A person's work should accumulate.** Every useful sentence should remain available to larger future work. 2. **The source should remain intelligible.** A creator should not lose an intellectual life when an application disappears. 3. **Meaning should be declared close to authorship.** The author is often the best person to identify the boundaries of a reusable idea. 4. **Derivatives should remain connected to their source.** Reach without provenance eventually destroys context and trust. 5. **Distribution should be programmable.** Stable content should support many appropriate channel representations. 6. **Shared infrastructure should produce independent capability.** The network succeeds when participants can stand in their own names. 7. **Public identity should make real competence legible.** Documentation, design, media, and institutional literacy should reveal the work that exists. 8. **Opportunity should multiply.** Get an opportunity, share an opportunity. 9. **Tools should reduce the distance between insight and public usefulness.** The system exists to help knowledge reach people. 10. **Scale should not require the founder to repeat every operation.** A durable method becomes a capability other people and systems can operate. # Part Nine — Reader Workbook ## 23. Design your own cumulative publishing system ### Step one — Name the durable source Where does your authoritative writing live? Can it be read without the application that created it? Can it be backed up, searched, compared, and moved? ### Step two — Define your semantic units What can stand alone in your work: a claim, quotation, passage, definition, procedure, case study, exercise, image, or demonstration? How will you mark the boundaries? ### Step three — Define the projections | Source unit | Page | Poster | Audio | Video | Feed | Book | Course | Search or retrieval | |---|---|---|---|---|---|---|---|---| | Sentence | | | | | | | | | | Passage | | | | | | | | | | Article | | | | | | | | | | Livestream | | | | | | | | | ### Step four — Define the identity boundary Who owns the domain, source files, audience relationship, publication record, and derivative assets? Which parts are common infrastructure, and which parts belong to the creator? ### Step five — Define the opportunity protocol When the system creates attention, access, or income for one person, what practice helps that person create an opening for someone else? ### Step six — Define evidence states How will your system distinguish an idea, plan, beta capability, deployed feature, measured result, later reconstruction, and future roadmap? ## 24. Historical questions for further recovery The surviving materials support a strong reconstruction, but the master record can continue improving. Future archive work should locate: - additional broadcasts explicitly teaching Markdown; - Thought Leader School assignments and course materials; - screenshots or exports from Stratus XL; - participant onboarding instructions; - lists of deployed thought-leader domains; - Jots and CMS interface captures; - alpha and beta feedback; - meetup records and geographic history; - source repositories and dated release snapshots; - architecture diagrams; - examples of automated video preparation; - feed recipes used by specific channels; - participant books, contributions, interviews, and page-growth records; - records establishing which planned phase-two functions reached deployment before the cyberattack. Each recovered object should be connected to the relevant livestream ID, date, participant, feature, evidence state, and surviving project lineage. # Glossary | Term | Meaning in this historical system | |---|---| | **Atomic Markdown** | A Markdown-centered authoring practice in which writers declare reusable semantic units such as quotations and passages inside canonical source. | | **AutoSites** | The recovered folder-native, multi-tenant publishing architecture that maps domains to account folders and shared capabilities. | | **Creativity Management System** | Bryant's preferred expansion of CMS: a system for accumulating, developing, publishing, and distributing creative work. | | **Dual-render authorship** | Authoring one semantic unit with a natural document surface and a complete standalone extraction surface. | | **Late-bound distribution** | Deciding channel representation, eligibility, media, attribution, timing, or selection when content is requested rather than storing a separate fixed copy for every destination. | | **Parameterized syndication** | Using paths and query parameters to express a feed's source, projection, selection, formatting, and delivery policy. | | **Projection** | A page, poster, feed item, index, search result, or other rebuildable rendering derived from canonical source. | | **Stratus XL** | The 2016 platform name used while introducing a video-aware new-media system and participant websites. | | **Thought Leader School** | The instructional and social environment for transferring writing, media, publishing, identity, networking, and leadership capability. | | **WordChirp** | A publicly named publishing-system predecessor operating by 2014, including a Google Docs authoring bridge and writer-centered workflow. | # Related project records - [[projects/Dual-Render Authorship and the Atomic Markdown Double-Brace Grammar|Dual-Render Authorship and the Atomic Markdown Double-Brace Grammar]] - [[projects/AutoSites - Folder-Native Publishing and Distribution|AutoSites — Folder-Native Publishing and Distribution]] - [[projects/Ten Years Building a Symbolic Language Engine|Ten Years Building a Symbolic Language Engine]] - [[projects/Simple Reminders - Global Influence Operations and Information Warfare|Simple Reminders — Global Influence Operations and Information Warfare]] - [[projects/Projects Timeline|Projects Timeline]] - [[projects/The Argument in Miniature - How I Write Quotations, Paragraphs, and Reusable Prose|The Argument in Miniature]] ![[parts/About Livestreams|About Livestreams]]