{"path":"history.md","content":"# Deliberus History\n\n## Timeline\n\n### ~2008-2010: Genesis\n- Idea conceived around age 20, carried since then as \"the ancient idea\"\n- Early conversations with **Dustin Juliano** (close nerd friend) about collaborative argumentation\n- Influenced by LessWrong rationality community, cognitive science, Kahneman/Tversky\n\n### 2010-2012: Community Building\n- **Facebook group** \"Collaborative Argumentation Analysis\" (`facebook.com/groups/arguable/`) — grew to **250 members** (now 452, archived: 406 posts, 1,956 comments from 134 contributors)\n- Recruitment via LessWrong (\"What are you working on? February 2012\")\n- Stockholm LessWrong meetup participation\n- Feature map created and shared in Facebook group\n- Most active group contributors (by comment count): **James X Nelson** (198), **Curtis SerVaas** (186), **Tomas Kronvall** (111), **Alex Vance** (100), **Emil Alexander** (89), **Frédéric Marcotte Piché** (75), **Dale Langdon** (53), **Daniel Lindsäth** (42)\n\n### 2012: Key Collaborations & Prototype\n- **Curtis SerVaas** — independently discovered identical idea, applied for **Thiel Fellowship** with Deliberus concept\n  - Interned at Leverage Research, Berkeley summer program, SIAI visit\n  - Rich technical discussions (Jan 22, 2012 conversation preserved in `Curtis-jan22.txt`, local archive)\n  - Created detailed planning docs: HOW? (logical structure, UI, open questions), Why?, blog draft, team coordination (preserved on BOHR drive)\n- **Fredrik spent ~3 weeks at Leverage Research** working on Deliberus ideas alongside Curtis and Oliver. Simplenote working notes from this period include the disagreement taxonomy, \"coarse-grained discourse = advent of language\" framing, and earliest \"No Copout Axioms\" formulation (\"start with opaque axioms, drill down carefully with time\"). See `docs/research/simplenote-archive-analysis.md §1`\n- **The \"Consensus software\" board (Proposal FB-01, January 2013)** — a one-page OmniGraffle system proposal Fredrik made during the Leverage stay, rediscovered and provenance-confirmed in July 2026 via a March 2013 email (\"Här är en PDF jag gjorde när jag var hos Leverage Research i januari\"). It anticipates the 2026 architecture in striking detail: \"limit content to atomistic statements\", NLP-assisted extraction of structured arguments from free-form text, concept tracking with merge/fork (\"concepts are defined by all statements using the concept\"), missing/implicit-premise identification, per-statement probability estimates, definitional/constitutional tagging, and \"mitigating psychological barriers to updating\" as a first-class module. One genuine inversion across the 13 years: the board's \"axioms established by users\" became the No Copout Axioms rejection of any permanent bottom. (Artifact kept in the private archive.)\n- **Oliver (Leverage Research)** — suggested collaboration between Leverage and Deliberus team, open to commercialization\n- **Ivan Phillips** — wrote \"A Tool to Industrialize Knowledge\" (Facebook note, Jan 12, 2011): \"each page is a proposition, each link is an inference\" — independently described the AIF architecture. Comment thread with Tom Boniecki (\"spreadsheet for knowledge\"), Ben Hyink (priors/philosophy limits), and Fredrik (prediction market reputation, atomic claim splitting). Full analysis: `docs/research/simplenote-archive-analysis.md §3`\n  - **Tom Boniecki** — early collaborator, contributed \"spreadsheet for knowledge\" analogy and concept tracking / semantics disambiguation idea\n  - **Ben Hyink** — early collaborator, flagged the \"radically different priors\" challenge (fact/value boundary)\n  - Hybrid AGI concept: \"argument map + human society = hybrid AGI\"\n- **CurtisSV-Deliberus** prototype: WorkFlowy-inspired open-source tool (JS/jQuery/HTML, local archive)\n- **Cerebrate** prototype deployed at `arguable.heroku.com/claims`\n\n### 2012-2013: Startup Phase & Divergence\n- Co-founded with **Nils Janse** as a startup. Visions diverged — Nils wanted commercial SaaS, Fredrik wanted societal truth-seeking infrastructure\n- Built a **Meteor.js prototype** with a rich data model: propositions, arguments, evidences, sources, references, voting system (proposition votes, argument-for/against, evidence votes). Deployed to `deliberus.se` with trial instances for Swedish youth political organizations (CUF, Ung Pirat, Gröna Studenter, and others)\n- Nils pivoted the startup → **Delibr** (now **DelibrAI**, he's CEO). Different product, different goals. Non-compete expired ~2014\n- Full details: [history-delibr-era.md](history-delibr-era.md)\n\n### 2013: Public Documentation\n- **P2P Foundation wiki page** published (Nov 3, 2013): `wiki.p2pfoundation.net/Deliberus`\n\n### 2013-2014: Almedalen & Outreach\n- Almedalen sprint — outreach to youth political organizations, educational institutions, and civic tech community\n- `digitaldemokrati.se` interview — overlapping values\n- Full outreach details: [history-delibr-era.md](history-delibr-era.md)\n\n### 2014-2024: Dormancy\n- Competing priorities, technical limitations of the era\n- Idea carried but not actively developed\n- What failed: no LLMs, limited real-time tools, 250 users insufficient for network effects, funding/commercialization challenges\n\n### 2025: AgentLoom Strategy\n- **AgentLoom project** positions Deliberus as \"The Dream\" — the long-term destination\n- AgentLoom = \"the engine\" (force multiplier), Deliberus = \"the vehicle\"\n- Strategy doc: secure funding for Deliberus (Ideal Outcome A) or high-impact AI employment (Ideal Outcome B)\n- Principle: \"Deliberus is the First and Only Customer\"\n\n### 2025: Archaeological Excavation\n- **October 18, 2025**: Comprehensive excavation of old MacBook Pro 2014 backup\n  - 8,084 Safari bookmarks converted to JSON (3.1MB) — ~500 directly relevant\n  - Chat logs recovered (33 logs from 2010-2014)\n  - Full excavation report: `deliberus-excavation-report-2025.md` (local archive)\n- **flowback** project found in archives — Django-based deliberation platform implementing predictive liquid democracy (local archive)\n\n### 2025-2026: Reawakening\n- Facebook posts and AI Safety Gothenburg discussions about reviving the idea\n- ChatGPT and Claude conversations developing modern technical approach\n- Recognition that LLMs + graph databases make the vision feasible\n- Connection to AI safety community provides initial user base\n- Conversations about Lean formal verification, Claimify, computational argumentation\n\n### March 27, 2026: Day Zero\n- First cobblestones. Project repository initialized.\n\n## Key People\n\n| Person | Role | Period | Notes |\n|--------|------|--------|-------|\n| Fredrik Bränström | Founder, visionary | 2008-present | Carried the idea for 18+ years |\n| Nils Janse | Co-founder, business | 2012-2014 | See [history-delibr-era.md](history-delibr-era.md) |\n| Curtis SerVaas | Collaborator | 2012 | Thiel Fellowship applicant, independently had same idea \"down to the web annotation\" |\n| Dustin Juliano | Early conceptualist | 2008-2013 | \"The one who came up with the basic concept\" (Fredrik to Sasha, Nov 2010). Wrote specs for internet protocols, proposed \"Global Authorship Project\" |\n| Oliver | Leverage Research | 2012 | Suggested Leverage/Deliberus collaboration |\n| Ivan Phillips | Theorist | 2011-2012 | Hybrid AGI concept: \"argument map + human society = hybrid AGI\" |\n| Edward Miller | Design contributor | 2011 | Key insights: collaboratively editable meta-rules/ontology, \"Am I Crazy\" belief-web coherence testing, thought experiment visualization |\n| Emanuel Kumlien | Linguistics, concept tracking | 2013 | Discourse-layer insight (semantic agreement masking real disagreement), directly inspired \"concept tracking\" feature sketch. Voice memo transcribed ([voice-memo-emanuel-sofia.md](research/voice-memo-emanuel-sofia.md)) |\n| Edvin | Design contributor | 2011 | Ethical filters as overlays (compute user profile via ethical dilemmas, not -isms), felicific calculus, meritocratic-then-democratic governance |\n| Tobias Wallin | Structured planning | 2011-2012 | Co-created formal goal analysis (målanalys) |\n| Andrew Stein | Organizational | 2010-2012 | Offered project management support |\n| Sasha | Early discussant | 2010 | Received first detailed elevator pitch of the vision |\n| James X Nelson | FB group, top contributor | 2011-2013 | 198 comments — most active non-founder member |\n| Tomas Kronvall | FB group contributor | 2011-2013 | 111 comments — found Icelandic participatory democracy project (betrireykjavik.is) |\n| Emil Alexander | FB group contributor | 2011-2013 | 89 comments |\n| Frédéric Marcotte Piché | FB group contributor | 2011-2013 | 75 comments |\n| Dale Langdon | FB group contributor | 2011-2013 | 53 comments |\n| Daniel Lindsäth | FB group contributor | 2011-2013 | 42 comments |\n| Alex Vance | FB group contributor | 2011-2013 | 100 comments |\n\n## Key Design Insights from Chat Logs (2010-2012)\n\nPreserved from chat logs (local archive) — ideas that emerged in conversation and are NOT in the Meteor prototype:\n\n1. **Arguments as atomic sentences** — single sentences to minimize logical \"jump\" between links (Sasha, Nov 2010). *Independently arrived at same insight as Claimify paper 15 years later*\n2. **\"Logic is universal\"** — no domain-specific validator hierarchy; system validates structure, not domain claims (Curtis, Jan 2012)\n3. **Ethical filters as overlays** — view argument summaries through different ethical lenses; compute user profile via ethical dilemmas, not -isms (Edvin, Jan 2011)\n4. **Gradual AI replacement** — \"as the system grows more intelligent, the people will gradually be made redundant, as we'll replace ourselves by AI modules\" (Sasha, 2010)\n5. **Collaboratively editable ontology** — meta-rules and ontologies as community artifacts, not fixed by developers (Edward, Feb 2011)\n6. **Inductive arguments as the hard unsolved problem** — deductive logic is simpler; probability estimation across all claims \"can get hairy\" (Curtis, Jan 2012)\n7. **\"Am I Crazy\" mode** — anonymous belief-web posting for coherence checking (Edward, Feb 2011)\n8. **Gaming as primary adversarial pressure** — \"enormous incentives to game the system\" (Edvin, Jan 2011)\n9. **Voting on logical strength of links** between argument atoms + tagging (emotional, poetic) (Sasha, Nov 2010)\n10. **Deduplication/clumping** — multiple formulations grouped, best (by vote) visible first, alternatives on double-click (Sasha, Nov 2010)\n\n## Existing Code Assets\n\n| Project | Location | Tech | Status |\n|---------|----------|------|--------|\n| delibr-meteor | Local archive | Meteor.js (106MB) | Archived, DB dump preserved |\n| CurtisSV-Deliberus | Local archive | JS/jQuery/HTML | Archived |\n| flowback | Local archive | Django/Python (422MB) | Archived, related platform |\n\n## Reference Materials\n\n| Resource | Location | Description |\n|----------|----------|-------------|\n| Excavation report | Local archive | 27.4KB comprehensive historical analysis |\n| Safari bookmarks | Local archive (JSON) | 8,084 bookmarks, 3.1MB |\n| Curtis conversation | `Curtis-jan22.txt` (local archive) | 12KB, Jan 2012 technical discussions |\n| Lecture notes PDF | `Deliberus Lecture Notes.pdf` (local) | 13.3MB — \"Online Deliberation: Design, Research, and Practice\" (Davies & Gangadharan, 2009) |\n| Simplenote notes | `Deliberus ideas.txt` (local archive) | \"Conflicting props\" |\n| BSON database dump | delibr-meteor dump (local archive) | MongoDB export of prototype data |\n| Logo files | delibr-meteor assets (local archive) | Original Deliberus logos |\n| Large logo | FERMI drive | High-res logo/wallpaper |\n| Chrome local storage | FERMI drive | Evidence of deployed instances |\n| **PDF Archive** | FERMI drive, Dropbox PDF Archive | 1.8GB, 1,066 PDFs total, **68 directly relevant** to deliberation/argumentation. Includes dedicated `Argument mapping/` folder (15 PDFs), collective intelligence, crowdsourcing, sensemaking, logic/rationality |\n| **LogicalForms.pdf** | BOHR drive, Curtis's Deliberus docs | UI reference from Curtis's planning docs |\n| **Startup Google Docs (BOHR)** | BOHR drive, Google Drive Backup | Shared startup-era docs (Fredrik+Nils): business plan, CUF presentation, data structures v2, samarbetsavtal, Vinnova application. **Google Drive link stubs** — need Google account auth to read actual content |\n| **Curtis's docs (BOHR)** | BOHR drive, Curtis's Deliberus docs | **Curtis SerVaas's** planning materials — HOW? (logical structure, UI, open questions), Why?, blog draft, team coordination. These are HIS notes, not necessarily reflecting Fredrik's vision. Also link stubs. |\n| **Voice memos (BOHR)** | BOHR drive, Google Drive Backup | \"Emanuel + Sofia; Deliberus.mp3\", \"Robin + Martin; Deliberus.mp3\" |\n\n---\n\n**See also**: [Vision](vision.md) · [Object Model](object-model.md) · [Sketches & UI Concepts](sketches.md)\n"}