1. Applicant & Project Overview
Project Name
Debate Market
Team / Individual Name(s)
Jade MICHEL – Dev full-stack
Links (Website, GitHub, Demo, Socials)
- Landing page : From Noise to Trust
- Whimsical : debate market pres
- Github : Jade-m22 (Jade Michel) · GitHub
100-Word Summary
Imagine an application that looks like a familiar social network, but is built for real debates — structured, weighted, and traceable.
On the surface : a lively feed of opinions, agreements, disagreements and explanations.
Underneath : every contribution becomes a unit of debate, transformed (with your approval) into semantic triples, enriched with signals, and reusable across discussions.
Debates no longer reset or run in circles : they accumulate into a shared reasoning graph, creating the foundation of a true debate market.
An energy system, familiar to Web2 users, gradually introduces the idea that an action sent on-chain has a cost. This system prepare the transition towards a complete Web3 experience : connected wallet, bonding curves, price/share, and in the end reasoning identity in the Intuition ecosystem.
Project Category
- Consumer App
- Agent / Registry
- InfoFi / Knowledge Graph
- Identity / DID / Reputation
- AI Context / MCP
Elevator pitch
Debate Market looks like a simple debate app, but it transforms every opinion into structured, weighted signals that map how collective reasoning evolves. Kialo proved debates can be structured ? We go further : Debate Market is the next evolution of social debates, built on open primitives and open-source code where arguments become interoperable across topics. Ideas no longer disappear : they compound into a shared, living intelligence. This unlocks an entirely new landscape for education and research, and opens the door to huge market opportunities.
Origin story
Today, when people discuss important or everyday topics (telework, best season, productivity, ecology, markets, public policy…) the same arguments loop in ephemeral threads.
Nothing accumulates. Nothing structures itself.
No one is really learning from what has already been said elsewhere.
Intuition now offers precisely what is missing : a knowledge graph, semantic triples, a layer of signal and identity that make these exchanges reusable, interoperable, composable and traceable.
Debate Market is born from the will to :
- Capture opinions, arguments and explanations as structured Atoms, Triples and Nested Triples, instead of letting them disappear in ephemeral feeds.
- Link those together — support, opposition, explanation, causality, proofs — to form a living reasoning graph rather than isolated conversations.
- Assign weight to each argument and signal, so that meaningful contributions rise above noise and collective reasoning becomes visible, navigable and comparable.
- Make this structured knowledge usable by the whole ecosystem, from AI agents to prediction markets, without requiring users to understand the underlying complexity.
- Offer an extremely simple and accessible surface, so anyone can participate just by expressing what they think — the architecture design happens under the hood.
- Open a financial and reputation layer for Web3 users (via TRUST, bonding curves, weighted signals), while providing a smooth on-ramp for Web2 users through an intuitive energy system, “super agree / super disagree” actions, and contextual tooltips.
Notable traction or achievements
Current stage : advanced idea (reasoning, structuring, wireframes, architecture) + landing page.
Validated / worked-through aspects :
- Detailed Web2 → Web3 user journeys, with progressive feature unlock based on engagement and understanding (simple mode, advanced triple mode, market view, reasoning identity, etc.).
- Designing a pipeline for extracting triples from free text, including :
- propositions of assertions,
- validation and correction by the user,
- nested triple generation ([S–P–O] refuted / supported / explained by [S–P–O]).
- Advanced reflection on the non-duplication of Atoms / Triples : one same [S–P–O] used in multiple discussions must be represented only once in Intuition, and simply connected to several contexts.
- Conceptual design of the Reasoning Identity :
- aggregation of contributions by theme,
- flow of support/opposition,
- density of participation,
- emergence of expertise by field.
More or less long-term vision :
- curve annotation (prediction markets, crypto, stock exchange, events) with triples explaining the movements (here too users could respond for/against and say why),
- a DAO of Reasoning to govern the patterns, quality and evolution of the graph,
- exploitation of this data by other Intuition applications and agents (for navigation, recommendation, analysis), without ever replacing the user’s voice,
- …
Current users or early testers
No active users yet (product not developed), but several preliminary exchanges with :
- Web3 builders,
- Web2 users interested in smarter chat spaces,
- profiles attracted by the idea of being recognized not only for “what they say”, but for the quality and consistency of their reasoning.
2. What You’re Building
Problem Statement
The exchange of opinions online today presents three major problems despite clear evidence that people want to debate in structured and meaningful ways.
Platforms like Kialo have demonstrated strong demand for thoughtful, organized debate, gathering large communities who enjoy comparing arguments, voting on their strength, and exploring reasoning.
But even with this demonstrated appetite, today’s mainstream platforms fall short.
1. Ephemerality
Discussions vanish inside infinite feeds. The same topics resurface, the same arguments must be rewritten, and no shared memory accumulates. Debate resets constantly instead of compounding.
2. Lack of structure
Conversations remain unstructured text blobs, not knowledge objects. There is no generalized layer to express:
- “Here is a precise assertion (S–P–O),”
- “Here is what supports or refutes it,”
- “Here is how it connects to other assertions across discussions.”
This is precisely what the graph of Intuition allows to bring : atoms, triples, signal, identity, and a space for the structure.
3. Disconnection of a shared graph and the Web3
Traditional platforms keep these exchanges in silos. Users obtain neither ownership, nor structured signal, nor reasoning identity.
Debate Market wants to solve these three problems by :
- structuring opinions into reusable semantic triples and nested triples,
- linking them into a shared, evolving reasoning graph,
- giving weight and traceability to arguments,
- and gently onboarding Web2 users into Web3 concepts (cost, ownership, signal, markets) through an intuitive, low-friction experience.
Proposed Solution
An open-source debate platform powered by Web3 primitives and the Intuition graph
Users of today’s debate platforms lack what matters most : openness, semantic structure, community ownership, portability of information (and therefore interoperability), meaningful incentives, and durable memory.
Debate Market aims to become the first open-source debate platform built on Web3-native primitives, where every opinion, argument and explanation becomes a structured, reusable, weighted piece of reasoning inside the Intuition graph.
The ambition does not stop there — we could build the first Debate DAO, where structured reasoning fuels innovation, research, markets, and collective intelligence.
The demand already exists (Kialo proved it at scale), but users of today’s platforms lack structure, ownership, portability, incentives, and memory.
Debate Market addresses these long-standing gaps and opens a new frontier where debates become composable knowledge.
A feed of opinions centred on your interests
Users land on a personalized feed, not a chaotic timeline. They see position statements, counterarguments, explanations, and causal links on topics they genuinely care about :
- “AI systems should disclose their training data sources.”
- “Nuclear power is essential to reach net-zero by 2050.”
- “Companies that force full return to office will lose talent in the next 2–3 years.”
- Other users agreeing/disagreeing and explaining why.
On the surface, it’s a straightforward “opinions & replies” app : write, read, react, follow topics. Simple.
Underneath, it is capturing and structuring these exchanges into semantic triples and nested triples.
Non-duplication rule
Non-duplication in Debate Market goes far beyond avoiding exact duplicate text.
It is a core architectural principle that ensures the reasoning graph grows in quality, not in redundancy.
1. Avoiding clones (canonical triples)
Before creating a new triple, the system checks whether an identical [S–P–O] already exists in the Intuition graph.
- If it exists → reuse the canonical triple
- The new contribution only creates a contextual link (discussion, thread, curve annotation, etc.)
- If it does not exist → create a new triple
This prevents graph fragmentation into multiple clones of the same idea.
2. Semantic equivalence, not textual similarity
Two sentences may look different but express the same assertion:
- “Remote work increases developer productivity.”
- “Working from home makes devs more productive.”
Both must map to the same triple. To achieve this, the system uses :
- semantic search (MCP, embeddings, Intuition queries),
- similarity scoring,
- user confirmation in the validation UI.
Humans remain in the loop to ensure correctness.
3. Distinguishing identical / more specific / different
During validation, users classify each assertion:
- Same statement → reuse existing triple
- More specific → create a new triple + link it as a refinement
- Different assertion → create a new triple
4. Knowledge factorization (one triple, many contexts)
If the same [S–P–O] appears in 10 discussion threads, 3 curve annotations, 2 other contexts, we still maintain one canonical triple with multiple contextual links.
This ensures:
- coherent accumulation of support/oppose/stake signals,
- cleaner navigation of reasoning,
- undiluted conviction,
- a denser, more meaningful graph.
5. Hybrid AI + human deduplication
The system combines:
- AI assistance for semantic matching, equivalence suggestions, and relation proposals,
- Human validation (“same meaning / more specific / different”),
- Future gamified curation to merge or classify triples.
This maintains consistency, improves model accuracy, and ensures debates become cumulative knowledge rather than noise.
Two creation modes, progressively unlocked
Mode 1 — Natural Writing (default)
- Write normally
- System suggests assertions
- Approve / edit / discard
- Triples are created or linked
Mode 2 — Triple Mode (unlocked later)
- Pick S / P / O via auto-complete
- Create nodes or reuse existing ones
- Use triples as subjects or objects (nested reasoning)
- Full structured editing power
This ensures the onboarding stays Web2-simple, while power users and Web3 users can go deeper when they are ready.
Milestone 1 will test :
- when to unlock Mode 2,
- how much structure to expose early,
- how to avoid overwhelming Web2 users.
Curves and markets
Alongside the textual feed, the app exposes curves in a dedicated section :
- price curves from prediction markets,
- crypto charts,
- stock charts,
- other relevant datasets.
Users can :
-
click on any point or segment of the curve,
-
add a claim explaining the movement :
“The price jumps here because company X announced a permanent hybrid work policy.”
“The curve is down here because X says Y”
Debate Market then :
- extracts [S–P–O] assertions from that explanation,
- links them to the specific point on the curve and any other related topic,
- lets other users support / oppose / refine them,
- and links them into the Intuition graph.
Over time this creates a structured layer of causal explanations attached to curves — a very useful dataset for InfoFi, agents, and other Intuition applications.
Milestone 1 will validate :
- if the charts must be central or optional,
- the best UX to map an assertion to a point in the market,
- which markets create the best debate loops.
Signal : still in reflection
The signal layer is fully open to experimentation :
- Option A — Classic Upvote / Downvote
Simple, Web2-friendly. - Option B — “Super-Agree / Super-Disagree”
Energy cost, stronger signal.
Web3 side : corresponds to deposits on bonding curves. - Option C — Weighted support with TRUST or staked tokens
Creation of belief markets around affirmations. - Option D — Hybrid model
Web2: simple signals
Web3: financial signals when a wallet is connected.
Milestone 1 will determine :
- what improves quality,
- what reduces the noise,
- what helps the most in growth.
Energy system : teaching on-chain cost without saying “gas”
Every action in the app consumes energy.
For Web2 users, this is a familiar pattern (seen in games and apps). For Debate Market, energy serves as :
- Spam & cost control
- limits the number and type of actions,
- keeps Web2 usage under predictable resource budgets.
- Intuitive bridge to Web3
- instead of explaining gas fees directly, users experience that : “there is a budget behind my actions; some actions cost more.”
- when they later connect a wallet, the idea that on-chain actions have a cost is already internalized.
Energy design (to be defined) :
- Web2 users get X energy per day (e.g. 100 units).
- It is not cumulative and resets every day(?).
- Exact cost of each action to be defined
- [Maybe better if ratio 1:1 for a better comprehension and a better transition ?]
The energy engine in the internal DB :
- decrements energy,
- enforces quotas,
- acts as anti-spam,
- gives full control over how many Web2 actions we can support.
Funding options (to be decided in Milestone 1) :
- initial energy sponsored by the project,
- paid refills via traditional Web2 payments,
- free refills via ads / sponsor content,
- on-chain publishing (see below)
Web2/Web3 writing to Intuition : explore different options
All Web2 actions first land in the internal database :
- table “Every actions” :
- all Web2 actions, status = pending → batched.
- “Energy system” module :
- decrements energy,
- checks quotas.
- “Users data” :
- user ID, interests,
- link to invisible / smart wallet.
There are some main strategies for writing these to Intuition :
1. Sponsorship pool (Web3 funded)
A pool funded by Web3 users periodically batches Web2 actions on-chain.
Pros :
- predictable experience for Web2 (things “just work”),
- good story for Web3 sponsors (“I help grow the graph”).
Cons :
- needs careful parameter tuning (not too generous, not too restrictive).
2. “Web3 pays when interacting” (local until needed)
By default, Web2 actions stay stored locally.
When a Web3 user wants to :
- support/oppose a Web2 claim,
- link it explicitly,
- stake on it,
the UI tells them :
“This claim is not yet on Intuition. To interact with it on-chain, you need to publish it.”
The resulting tx both publishes the triple and records the Web3 action.
This can be combined with a pool (for partial funding) or left purely user-funded.
Both options are on the table ; Milestone 1 explicitly includes deciding and validating which model (or hybrid) is best.
3. Proxy contracts with protocol fee (open design)
A third option is to route some on-chain interactions through a proxy contract (a proxy that calls MultiVault) that takes a small protocol fee on each action.
This fee could be split between :
- funding the Web2 → on-chain sponsorship pool,
- covering part of the app’s infrastructure and development,
- seeding the future Reasoning DAO treasury.
The exact fee rate and distribution are intentionally left open and will be explored with the Intuition team and community. The goal is to align incentives:
- Web3 users interact normally,
- a tiny fraction of value flows back into maintaining the graph, the app, and the reasoning ecosystem.
All three approaches (sponsorship pool, “Web3 pays when interacting”, proxy-based protocol fee) are on the table; Milestone 1 explicitly includes deciding and validating which combination (or hybrid) makes the most sense.
User journey : Web2 → Web3
This progression, and which features unlock when, will be designed up-front and tested via A/B tests (for example: when to show “super-agree”, how much graph UI to reveal early, etc.). That work is part of Milestone 1.
Gamified de-duplication & graph curation (in exploration)
An important challenge is avoiding duplicate triples and improving the quality of links. One idea (still in design exploration) is :
- Create mini-games / quests where users :
- are shown pairs of claims and asked if they are :
- “same meaning”,
- “different”,
- “one refines the scope of the other”, etc.
- are invited to link claims :
- “Does this [S–P–O] explain this other [S–P–O]?”
- “Is this an example / counter-example?”
- are shown pairs of claims and asked if they are :
Web2 users can play these games and in doing so :
- generate high-quality training data for an internal dedupe/structure agent,
- help refine the mapping from natural reasoning patterns to schema patterns.
Web3 users can be invited to take the on-chain side :
- once the community / AI agrees that two triples are equivalent or related,
- Web3 users with stakes or strong signal can confirm this on-chain,
- publishing the final merged triple and relationships into Intuition.
This creates :
- a human-in-the-loop curation layer,
- a dataset for training and improving the “triple suggestion / dedupe agent”,
- and a playful bridge between Web2 and Web3 roles.
For now, this sits under a “Design in exploration” section of the proposal ; it would likely be implemented after the core pipeline works.
Stage of Development (Advanced Idea)
- Detailed design of user journeys
- Description of the triple pipeline
- Advanced reflection on signal modes, energy, financing
- No code, no deployment : the grant aims to transform this specification into a prototype, then into an MVP.
Technical architecture overview (v1)
Next.js / React · CSS/SCSS · TypeScript · Node.js · PostgreSQL · Intuition GraphQL & SDK · Wagmi / viem · MCP / AI Context · Account Abstraction (smart accounts)
Front-end (Web / mobile) :
- feed of opinions,
- text editor + triples proposal validation interface,
- energy bar management (visualization, refills),
- display of curves (markets, crypto prices, etc.) with the possibility to file claims,
- advanced mode (unlocked) for editing triples directly,
- Web3 UI (connected wallet) :
- display of the bonding curves,
- price/share,
- interface for stakes (support/opposition) once the infrastructure is ready.
REST/GraphQL API for :
- web2 account management (simple auth - NextAuth?),
- interaction with MCP for triple extraction,
- energy model (credits consumed/recharged),
- aggregation for the Reasoning Identity dashboard,
- de-duplication logic (search for triples close, use of vectors or Intuition IDs?, choice between creating or linking).
Using MCP / AI context for :
- take an opinion text,
- return a list of [S–P–O] proposal,
- classify potential relationships (support/contradict/explain/nuance…),
- provide suggestions of triples already existing in the graph,
The user remains in the loop to validate, correct or reject.
Intuition Integration (graph + registry + signal)
Intuition API / SDK for :
- create Atoms for the [S-P-O] that do not exist,
- create corresponding Triples and Nested Triples,
- associate signals,
- link the contributions to DIDs (user identity),
- read: existing triples, relations, global signal around a claim, information necessary for Reasoning Identity.
Safe Smart accounts (AA) with 3 signatories : user, backend, backup.
→ With 1/3 signer policy, the user can make decision on their own.
Contracts for sponsorship pools (option) + energy system ?
Analytique Reasoning Identity :
- Aggregating backend :
- number of triples related to a user,
- areas (themes) where they contribute, where they excel,
- density of support/opposition around their assertions,
- reuse of their triples in other discussions.
This identity can then be projected into structures exploitable by the user, Intuition or by external organizations (to be defined, push the idea).
Integrations or dependencies (MCP, A2A, ERC-8004, x402, etc.)
Intuition Graph & APIs
- Atom creation & Triple creation
- Signal APIs (attach support/oppose/“super-agree”/stake information),
- DID / Agent registry for user identities and reasoning profiles,
- read-side queries for :
- searching existing triples,
- retrieving related nodes on the graph,
- pulling signal summaries for the UI.
MCP / AI Context (Intuition)
- take free-form user text + context (topic, previous contributions),
- return candidate assertions as natural-language [S–P–O] proposals,
- suggest likely relations between existing triples (support/refute/explain, etc.),
- suggest candidate matches for de-duplication (triple similarity),
- assist navigation (e.g. “show me related triples I might want to reuse”).
Account Abstraction / Smart Accounts
- AA compatible with Intuition / the underlying chain,
- three signers : user, backend, backup,
- enabling :
- gas sponsorship,
- account recovery,
- gradual transfer of control to the user.
Optional external data sources & markets
- price feeds / charts from prediction markets or DEXs / CEXs for curve annotation,
- these do not need deep integration at first, only read-only charts that users annotate.
Internal storage / infra
- database for :
- pending Web2 actions,
- energy balances,
- local triple IDs and mapping to Intuition IDs,
- Reasoning Identity pre-aggregation.
Analytics / experimentation
- flags / configuration system for :
- A/B testing unlock thresholds,
- comparing different signal variants (upvote vs support/oppose, etc.),
- instrumentation to measure :
- quality of triple suggestions,
- de-duplication effectiveness,
- user understanding and retention.
Security considerations
- Spam : controlled by energy and limits to be set on certain actions.
- AI : the user remains the final filter and the AI does not publish anything alone without relatable validation.
- Non-duplication : particular attention to the logic of search and reconciliation of triples before creation.
- Web3 risk :
- gradual migration towards greater user autonomy (backend & backup signatories can be transferred),
- clear communication via tooltips for each step involving stakes or real costs.
3. Team & Execution Ability
Team backgrounds
Jade MICHEL :
- Associate Degree in Communication (BTS)
- Bachelor’s Degree in Business, Sales, and Digital Marketing
- Master’s Degree in Business Management (specializing in Communication)
- Full-Stack Web Development (The Hacking Project)
Execution proof (past products, repos, shipped work)
- GitHub : Jade-m22 (Jade Michel) · GitHub
INTUITION.BOX :
- Repo : GitHub - intuition-box/intuition.box
- Live : https://intuition.box/
FABLAB.BOX :
- Repo : GitHub - fablab-box/fablab.box
- Live : https://fablab.box/
TRUST CARD :
- Vote :
- Repo : GitHub - intuition-box/TrustCard
- Live : https://trustcard.box/
- Documentation :
- Simulator :
Commitment level (full time, part time)
- Full time for the moment
Optional
- Team structure and roles
- Advisors or collaborators
- Jérémie Olivier
- Saulo
- Expected future hiring
4. Grant Request & Milestones
Overview
| Milestone | Objective | Output |
|---|---|---|
| M1 – Foundation, Intuition Integration & Validation | Build the core prototype, validate extraction pipeline, energy model, Web2→Web3 bridge, and funding models. | Working testnet prototype + validated assumptions. |
| M2 – Advanced Graph, Signals & Funding Model Implementation | Implement dedupe engine v1, curve annotation, advanced signals, and finalize the funding model. | Production-ready reasoning components. |
| M3 – Reasoning Identity, DAO & Curation Layer | Deliver Reasoning Identity v1, graph curation flows, and DAO v1 governance. | Community-driven structured reasoning ecosystem. |
Amount requested (USD and/or TRUST)
This grant request covers Milestone 1, structured into 4 sub-milestones.
Total requested for Milestone 1 :
- 12,000 USD
- 3,000 TRUST
Total duration : 10/12 weeks (≈ 3 months)
Milestone 1 - Step 1 : Core App, Web2 Onboarding & MCP Extraction v1
- Duration : 3 weeks
- Amount : 4,000 USD + 1,000 TRUST
Scope / Tasks :
- Web2 onboarding (email / social login)
- Smart Account (AA) creation
- Opinion Feed v1 (post / reply / react)
- MCP extraction v1 (assertions + relation suggestions)
- Validation UI (approve / edit / reject)
- Basic semantic matching
KPIs & Success Criteria :
-
KPIs (based on ≥100 users)
- ≥ 80% onboarding completion
- ≥ 1 usable assertion per text input
- ≥ 70% comprehension of validation UI
-
Success Criteria
- Feed + onboarding stable end-to-end
- Extraction pipeline functional and testable
Milestone 1 - Step 2 : Intuition Integration, Energy System v1 & Signals v1
- Duration : 3 weeks
- Amount : 4,000 USD + 1,000 TRUST
Scope / Tasks :
- Intuition testnet writes (Atoms, Triples, Nested Triples)
- Web2 energy system (daily budget, action costs, anti-spam)
- Web2 signals (support / oppose)
- Web3 signals (super-agree → bonding curve action)
- Unlock progression (super-agree, triple-mode preview)
KPIs & Success Criteria :
- KPIs (base of ≥300 total actions by ≥30 users)
- ≥ 90% correct energy-consumption events
- ≥ 70% user understanding of the energy model
- ≥ 70% stable and successful Web3 testnet writes
Success Criteria :
- Intuition integration reliable and consistent
- Clear separation between Web2 and Web3 signal flows
Milestone 1 - Step 3 : Funding Models, User Understanding & Prototype Tests
- Duration : 2/3 weeks
- Amount : 3,000 USD + 500 TRUST
Scope / Tasks :
- Implement prototypes for Models A, B, C, D
- UX onboarding for each model
- Funnel analytics + friction identification
- Structured tests with Web2 + Web3 profiles
KPIs & Success Criteria :
-
Success Criteria :
- Comparative A/B/C/D report delivered
- Clear recommended funding model for the next milestones
-
KPIs :
- ≥ 20 test users (mixed Web2 + Web3 profiles)
- ≥ 70% comprehension of cost / publishing flow
- ≥ 60% preference converging toward one funding model
Milestone 1 - Step 4 : Prototype Consolidation & Final Model Selection
- Duration : 2/3 weeks
- Amount : 3,000 USD + 500 TRUST
Scope / Tasks :
- Improve extraction accuracy & semantic matching
- Stabilize onboarding + energy + signals
- Final signal model selection
- Final Web2 → on-chain model selection
- Prepare technical scope for Milestone 2
KPIs & Success Criteria :
-
KPIs
- Prototype validated by ≥10 external testers
Base: minimum 10 new testers (non-internal) completing the full prototype flow. - <15% error rate in extraction validation
Base: evaluation over ≥200 extracted assertions reviewed manually by testers or team. - Clear roadmap validated for M2 & M3
Base: internal + external reviewer approval (min. 2 reviewers), formalized in written document.
- Prototype validated by ≥10 external testers
-
Success Criteria :
- Final prototype delivered
- Final signal and publication models selected + justified
Milestone 2 : Advanced Graph, Signals & Funding Model Implementation
(Exact scope refined after Milestone 1 learnings and user tests)
| **Scope / Tasks ** | Outputs |
|---|---|
| Implement the selected Web2 → on-chain funding model (sponsorship pool, Web3-pays-on-interaction, proxy-fee route, or hybrid). | Final, validated funding model live in the app. |
| Build deduplication engine v1 (semantic matching, confidence scoring, human review). | Consistent triple creation and reduced duplicates across contexts. |
| Implement advanced signals (super-agree mapping to curves, weighted conviction, thresholds). | Stable Web3-native signal layer with bonding curve integration. |
| Curve annotation v1 (attach triples to points/segments on charts). | Structured causal explanations linked to market movements. |
| Improved extraction UX (batch editing, linking suggestions, refinements). | Faster and more accurate user validation flow. |
| Extended progression system (unlock triple mode, graph views, Web3 actions). | Clear onboarding path for Web2 → Web3 transition. |
| Integrate with Intuition production endpoints (if available). | End-to-end Intuition integration at production level. |
Note: All exact features, complexity, and ordering depend on Milestone 1 data, UX tests, and model evaluations.
Milestone 3 : Advanced Graph, Signals & Funding Model Implementation
| Scope / Tasks (High-Level) | Outputs |
|---|---|
| Build Reasoning Identity v1 (expertise, consistency, contribution maps). | First usable version of reasoning profiles for users and agents. |
| Implement identity-based weighting models (expertise by topic, trust scores). | Weighted reasoning signals and more meaningful graph influence. |
| Curation layer v1 (human-in-the-loop merging, classification, relation validation). | Higher-quality graph, fewer duplicates, cleaner reasoning patterns. |
| Optional gamified curation (quests, tasks, rewards). | User-powered graph improvement and dataset generation. |
| Launch Reasoning DAO v1 (governance scope, early proposals, treasury rules). | Community-driven stewardship of signal models, schemas, and funding. |
| Implement proxy-fee or protocol-fee routing (if selected in M2). | Sustainable treasury for DAO + Web2 sponsorship pool. |
| Integration endpoints for external apps, agents, and InfoFi tools. | Reasoning becomes reusable across the ecosystem. |
Note: Milestone 3 intensifies governance, identity, and curation only after the core experience (M1–M2) is validated by real usage.
