From 5e304ce59c018f768a6d5bd9701ce9c4ccb2427a Mon Sep 17 00:00:00 2001 From: Hyadhuad Date: Fri, 25 Sep 2026 11:28:34 +0700 Subject: [PATCH 1/3] docs: direction toward the 2033 objective; English as the canonical language DIRECTION keeps each phase's exit condition and states 2033 as the goal they are measured against, with the path from the book (2.7.1, ch. 47-48): Meta DFM integrated in Dk 1.0, then integration and the great migration. Drop links to DIRECTION.PT/ES, which never existed. Existing PT/ES files are preserved but no longer kept in sync; other languages go through automatic translation until the site has native localization. Co-Authored-By: Claude Opus 5.5 --- CONTRIBUTING.md | 3 +-- README.md | 2 +- data/catalog.json | 10 +++++----- roadmap/DIRECTION.md | 24 ++++++++++++++++-------- roadmap/README.md | 2 +- 5 files changed, 24 insertions(+), 17 deletions(-) diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 16790fb..3f60779 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -2,7 +2,7 @@ [Português](./CONTRIBUTING.PT.md) · **English** -Dknowledge is an open public knowledge base. Useful contributions include correcting a claim, connecting a source, filling an empty paper, improving navigation and bringing a translation back in sync. +Dknowledge is an open public knowledge base. Useful contributions include correcting a claim, connecting a source, filling an empty paper, and improving navigation. English is the canonical language; translations are not maintained by hand. ## The current Git flow @@ -31,7 +31,6 @@ There is no active `peer-review` or `community-review` branch. During the foundi ## Good first contributions - One of the 16 English paper shells has only a title: write its scope and connect its sources. -- Compare a translated document with its English source and update what drifted. - Find a claim that no longer matches a component contract and open a correction issue. - Improve the generated catalog or the accessibility of [dknowledge.drayker.org](https://dknowledge.drayker.org). diff --git a/README.md b/README.md index 865281f..37bef25 100644 --- a/README.md +++ b/README.md @@ -56,7 +56,7 @@ The most complete documentation in the Drayker ecosystem — start with [Current Open an issue. Issues small enough for one person to finish carry the `open-function` label and appear on the board at [drayker.org](https://drayker.org/fn/). Reading a paper and writing down where it no longer matches reality is a genuinely useful contribution here. -Other languages: [Português](./README.PT.md) · [Español](./README.ES.md) — both currently behind this English version. +English is the canonical language of this knowledge base. The Portuguese and Spanish files that exist are preserved, many of them as historical sources, but are no longer kept in sync; read other languages through automatic translation. Native translation and localization are planned for the site itself. --- diff --git a/data/catalog.json b/data/catalog.json index f2dcc65..0974b93 100644 --- a/data/catalog.json +++ b/data/catalog.json @@ -24,7 +24,7 @@ "language": "en", "kind": "guide", "state": "current", - "bytes": 1977 + "bytes": 1936 }, { "path": "CONTRIBUTING.PT.md", @@ -222,7 +222,7 @@ "language": "en", "kind": "guide", "state": "current", - "bytes": 7629 + "bytes": 7811 }, { "path": "README.PT.md", @@ -551,11 +551,11 @@ { "path": "roadmap/DIRECTION.md", "href": "/roadmap/DIRECTION.html", - "title": "Direction — Drayker, without dates", + "title": "Direction — Drayker, toward 2033", "language": "en", "kind": "roadmap", "state": "historical", - "bytes": 7855 + "bytes": 9341 }, { "path": "roadmap/global-main-roadmap.ES.md", @@ -1023,7 +1023,7 @@ "language": "en", "kind": "roadmap", "state": "historical", - "bytes": 2674 + "bytes": 2686 }, { "path": "roadmap/README.PT.md", diff --git a/roadmap/DIRECTION.md b/roadmap/DIRECTION.md index d971387..8b6c17a 100644 --- a/roadmap/DIRECTION.md +++ b/roadmap/DIRECTION.md @@ -1,8 +1,6 @@ -# Direction — Drayker, without dates +# Direction — Drayker, toward 2033 -> The execution direction of Drayker, derived from the [Global Main Roadmap](./global-main-roadmap.md). The phases there carry dates that were shallow estimates of priority; here the same structure is stated without dates, each phase with the condition that must be true before the next begins. What runs today is marked. Nothing here is a commitment, a schedule or a promise of funding. - -[English](./DIRECTION.md) | [Português](./DIRECTION.PT.md) | [Español](./DIRECTION.ES.md) +> The execution direction of Drayker, derived from the [Global Main Roadmap](./global-main-roadmap.md). The dates there were shallow estimates of priority; here each phase carries the condition that must be true before the next begins, and one date is stated as an objective: **Drayker consolidated and Dk Global fully operating by 2033**. What runs today is marked. Nothing here is a schedule or a promise of funding. ## How to read this @@ -10,6 +8,16 @@ Each phase names the work, the state it is in today, and the **exit condition** --- +## The objective: 2033 + +The goal Drayker gives itself is to bring the mission to a point of no return by 2033 — Drayker consolidated, with Dk Global fully operating — and to have something very close to that before then (*Do animal à superinteligência*, chapters 47 and 48). + +A dated goal is only serious if it states the path. First, organization and cooperation: people and fronts working together before the infrastructure of their own exists. Then the base architecture is closed, prototypes are built and tested, and organization through DFM is proven at scale. Resource stages then feed Meta DFM — a swarm of agents inside a construction architecture, a primitive form of recursive self-improvement, able to build the first versions of the network. With architecture 1.0 fully running and Meta DFM integrated, integration begins: prototypes enter use and the great migration of people, projects and memory into the network starts. + +Each phase is measured by what works, not by the date. If the date arrives before the phases, the goal was missed, not the design. Competences keep growing by stages, through the members' process, as far as members can verify what Dk Global does. + +--- + ## 1. Intelligence & R&D — the base **State: historical, and still alive.** @@ -81,7 +89,7 @@ Dk Personal (the agent that belongs to one person), the brain extension app, Dk **State: future.** -Dknowledge 1.0, OSDK 1.0, the evolutionary general platform 1.0, distributed computing at scale, the universal Dk neural language, and the brain net. The substrate, the intelligence and the knowledge layer become one system. +Dknowledge 1.0, OSDK 1.0, the evolutionary general platform 1.0, distributed computing at scale, the universal Dk neural language, and the brain net. Meta DFM is integrated into the system as the swarm of agents that builds and improves the network. The substrate, the intelligence and the knowledge layer become one system. *Exit condition:* the components specified in phase 3 exist as described. @@ -91,7 +99,7 @@ Dknowledge 1.0, OSDK 1.0, the evolutionary general platform 1.0, distributed com **State: future.** -The first main projects 1.0, and the first embassy. If a token-economy DAF phase needs an external substrate, ICP is a probable provisional choice for building and testing Drayker's own infrastructure. As PAP matures, DAF's useful functions and transitional infrastructure are absorbed into PAP and DAF dissolves into the durable environment. The network of mass participation begins to touch the ground. +The first main projects 1.0, and the first embassy. If a token-economy DAF phase needs an external substrate, ICP is a probable provisional choice for building and testing Drayker's own infrastructure. As PAP matures, DAF's useful functions and transitional infrastructure are absorbed into PAP and DAF dissolves into the durable environment. The network of mass participation begins to touch the ground, and the great migration of people, projects and memory into the network begins. This is the point the 2033 objective names. *Exit condition:* a place where the system exists physically, and PAP can absorb the federation's useful functions and infrastructure. @@ -99,7 +107,7 @@ The first main projects 1.0, and the first embassy. If a token-economy DAF phase ## The transversal principles -- **No dates.** Each phase has an exit condition instead. +- **One date, as an objective.** Each phase keeps its exit condition; 2033 is the goal the phases are measured against, not a schedule that replaces them. - **Build the real thing with what exists today.** GitHub is the platform until there is something worth migrating. - **Intelligence is the means, never the master.** Every phase serves the will to potential of every person. - **Read the sources.** This page is derived from the roadmaps and the repositories; when they disagree, the repositories are right. @@ -108,7 +116,7 @@ The first main projects 1.0, and the first embassy. If a token-economy DAF phase **The roadmaps this is derived from** -- [Global Main Roadmap](./global-main-roadmap.md) — the historical roadmap with the phase structure this page states without dates. +- [Global Main Roadmap](./global-main-roadmap.md) — the historical roadmap whose phase structure this page keeps, with exit conditions in place of its old dates. - [Current orientation](../CURRENT.md) — what is available now, what remains architecture, and how founding governance works. - [Roadmap index](./README.md) — the structure of project roadmaps and how they are organized. - Project roadmaps: [Dk](./main-projects/dk/README.md) (AP platform, BSDK-DFM, Dk Network, Dknowledge, interface, Living Cryptography, OSDK, UID-DkApp) · [Ecosystem](./main-projects/ecosystem/README.md) (autonomous health, academy, Dktron, support model, support stations, value management ecosystem, will to potential) · [Organization](./main-projects/organization/README.md) (DAF). diff --git a/roadmap/README.md b/roadmap/README.md index fd2c69f..9ba9ab5 100644 --- a/roadmap/README.md +++ b/roadmap/README.md @@ -4,7 +4,7 @@ > **Historical material.** This roadmap is preserved as context. It is not the current execution plan, schedule, assignment list or implementation evidence. Start with [Current orientation](../CURRENT.md). ## Direction [English](./DIRECTION.md) -The execution direction of Drayker without dates: the same phases as the global roadmap, each with the condition that must be true before the next begins, and what runs today marked. +The execution direction of Drayker toward its 2033 objective: the same phases as the global roadmap, each with the condition that must be true before the next begins, and what runs today marked. ## Global main roadmap [English](./global-main-roadmap.md) The structure of roadmaps and task organization works with a global view of the main projects, where all major projects and their main modules and functions are specified and structured giving a global view of Drayker. From 6e4d7e3ea7e1c0ab54da43e7b0a306bfe70cd14e Mon Sep 17 00:00:00 2001 From: Hyadhuad Date: Fri, 25 Sep 2026 11:52:16 +0700 Subject: [PATCH 2/3] docs: a single refusal with new grounds can justify review after triage CURRENT, support model and economy of capacity: every refusal carries its grounds, even when anonymous; review can follow a pattern of refusals or a single refusal bringing information beyond the decision's scope, once triage confirms it. Co-Authored-By: Claude Opus 5.5 --- CURRENT.md | 2 +- data/catalog.json | 6 +++--- papers/main-projects/ecosystem/support-model.md | 2 +- .../main-projects/ecosystem/value-management-ecosystem.md | 2 +- 4 files changed, 6 insertions(+), 6 deletions(-) diff --git a/CURRENT.md b/CURRENT.md index 6aed62c..9cb5174 100644 --- a/CURRENT.md +++ b/CURRENT.md @@ -28,7 +28,7 @@ Global Dknowledge is intended to be an operational ontology, not merely a docume Drayker's proposed internal economy circulates capacity rather than selling it among members. Money, contracts and markets belong to the external interface. Internally, Dk may synthesize need, intention, contextual reputation, trust, project quality, risk, impact and available resources; Dknowledge preserves the reasons and consequences; accountable member institutions authorize consequential allocations. -Synthesis alone does not authorize an action. Systemic safeguards — constitutional constraints, security, permissions, integrity, reserves and risk — condition what Dk proposes, so the executable decision is the one inside a permitted and safe space. Decisions do not end the loop: effects, refusals and consequences return through members' personal Dks as new evidence, without a universal individual veto; a well-founded pattern of refusals can justify review, and a decision can be maintained, adapted, excepted, suspended or reversed. +Synthesis alone does not authorize an action. Systemic safeguards — constitutional constraints, security, permissions, integrity, reserves and risk — condition what Dk proposes, so the executable decision is the one inside a permitted and safe space. Decisions do not end the loop: effects, refusals and consequences return through members' personal Dks as new evidence, without a universal individual veto. Every refusal carries its grounds, even when its author stays anonymous. A well-founded pattern of refusals can justify review, and so can a single refusal whose grounds bring information beyond the scope considered before the decision, once an advanced triage confirms what that information is and what it means ([veto chain](https://uid.drayker.org)). A decision can then be maintained, adapted, excepted, suspended or reversed. The ecosystem papers now contain the current canonical formulations of [Distributed Support](./papers/main-projects/ecosystem/support-model.md), [Dktron](./papers/main-projects/ecosystem/dktron.md), the [economy of capacity](./papers/main-projects/ecosystem/value-management-ecosystem.md) and [Will to Potential](./papers/main-projects/ecosystem/will-to-potential.md). They replace the former one-line placeholders. diff --git a/data/catalog.json b/data/catalog.json index 0974b93..ef8d408 100644 --- a/data/catalog.json +++ b/data/catalog.json @@ -42,7 +42,7 @@ "language": "en", "kind": "orientation", "state": "current", - "bytes": 4149 + "bytes": 4449 }, { "path": "papers/main-projects/dk/README.ES.md", @@ -258,7 +258,7 @@ "language": "en", "kind": "paper", "state": "draft", - "bytes": 2222 + "bytes": 2522 }, { "path": "papers/main-projects/ecosystem/value-management-ecosystem.md", @@ -267,7 +267,7 @@ "language": "en", "kind": "paper", "state": "draft", - "bytes": 3356 + "bytes": 3682 }, { "path": "papers/main-projects/ecosystem/will-to-potential.md", diff --git a/papers/main-projects/ecosystem/support-model.md b/papers/main-projects/ecosystem/support-model.md index b61fdee..0c75bae 100644 --- a/papers/main-projects/ecosystem/support-model.md +++ b/papers/main-projects/ecosystem/support-model.md @@ -19,7 +19,7 @@ A project of life relates a member's intentions, dreams, studies, experience, re The internal economy is an economy of capacity, not a market among members. A future allocation can relate need, intention, contextual reputation, knowledge, trust, project quality, urgency, risk, expected impact and resource availability. Dk performs synthesis; Dknowledge preserves reasons and consequences; accountable member institutions authorize consequential allocations. -Synthesis alone does not authorize an action. Systemic safeguards — constitutional constraints, security, permissions, integrity, reserves and risk — can limit, postpone, compartmentalize or block what Dk proposes, so the executable decision is always the one inside a permitted and safe space. Effects, refusals and consequences return through members' personal Dks as new evidence, without a universal individual veto; a well-founded pattern of refusals can justify review, and a decision can be maintained, adapted, excepted, suspended or reversed. +Synthesis alone does not authorize an action. Systemic safeguards — constitutional constraints, security, permissions, integrity, reserves and risk — can limit, postpone, compartmentalize or block what Dk proposes, so the executable decision is always the one inside a permitted and safe space. Effects, refusals and consequences return through members' personal Dks as new evidence, without a universal individual veto. Every refusal carries its grounds, even when its author stays anonymous. A well-founded pattern of refusals can justify review, and so can a single refusal whose grounds bring information beyond the scope considered before the decision, once an advanced triage confirms what that information is and what it means ([veto chain](https://uid.drayker.org)). A decision can then be maintained, adapted, excepted, suspended or reversed. Reputation matters under scarcity, risk and responsibility. It does not determine membership, dignity or fundamental rights. As capacity becomes abundant, essentials should move into a common floor rather than remain rewards. diff --git a/papers/main-projects/ecosystem/value-management-ecosystem.md b/papers/main-projects/ecosystem/value-management-ecosystem.md index 42de62f..f44b594 100644 --- a/papers/main-projects/ecosystem/value-management-ecosystem.md +++ b/papers/main-projects/ecosystem/value-management-ecosystem.md @@ -15,7 +15,7 @@ Need, intention, ability, contextual reputation, knowledge, resources, opportuni Synthesis alone does not authorize an action. Every proposed allocation passes through **systemic safeguards** — constitutional constraints, security, permissions, integrity, reserves and risk — that can limit, postpone, compartmentalize or block what Dk proposes. The result is an executable decision inside a permitted and safe space, not an automatic conversion of context into action. -Representation does not end when a decision is taken. Effects, refusals and consequences return through the members' personal Dks as new evidence for local and global synthesis. A single refusal is context, never a universal individual veto; a well-founded pattern of refusals or unexpected consequences can justify review, and a decision can be maintained, adapted, given exceptions, suspended or reversed in light of the new state of the system. +Representation does not end when a decision is taken. Effects, refusals and consequences return through the members' personal Dks as new evidence for local and global synthesis. A single refusal is never a universal individual veto, but it is never mere noise either: every refusal carries its grounds, even when its author stays anonymous. A well-founded pattern of refusals or unexpected consequences can justify review, and so can a single refusal whose grounds bring information beyond the scope considered before the decision, once an advanced triage confirms what that information is and what it means ([veto chain](https://uid.drayker.org)). A decision can then be maintained, adapted, given exceptions, suspended or reversed in light of the new state of the system. Auditability does not mean total exposure. Public, project-restricted, personal and security-sensitive contexts can have different permissions; authorized Dk processes may use information that is not generally disclosed, while the existence and legitimacy of the decision path remain reviewable. From 0dadefbb3f322abe687798b65fa78f0ca520da2d Mon Sep 17 00:00:00 2001 From: Hyadhuad Date: Fri, 25 Sep 2026 12:41:14 +0700 Subject: [PATCH 3/3] docs: constitutional alignment paper; partner wording; no book citations New current synthesis papers/main-projects/dk/constitutional-alignment.md, linked from CURRENT and the Dk papers index; site-check expects 17 papers; CONTRIBUTING shell count corrected to 11. Repository documentation stands on its own and does not cite external books as authority. Co-Authored-By: Claude Opus 5.5 --- CONTRIBUTING.md | 2 +- CURRENT.md | 2 + data/catalog.json | 19 +++-- papers/main-projects/dk/README.md | 4 + .../dk/constitutional-alignment.md | 74 +++++++++++++++++++ roadmap/DIRECTION.md | 4 +- tools/site-check.js | 2 +- 7 files changed, 98 insertions(+), 9 deletions(-) create mode 100644 papers/main-projects/dk/constitutional-alignment.md diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 3f60779..48912c0 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -30,7 +30,7 @@ There is no active `peer-review` or `community-review` branch. During the foundi ## Good first contributions -- One of the 16 English paper shells has only a title: write its scope and connect its sources. +- One of the 11 English paper shells has only a title: write its scope and connect its sources. - Find a claim that no longer matches a component contract and open a correction issue. - Improve the generated catalog or the accessibility of [dknowledge.drayker.org](https://dknowledge.drayker.org). diff --git a/CURRENT.md b/CURRENT.md index 9cb5174..7d4135f 100644 --- a/CURRENT.md +++ b/CURRENT.md @@ -32,6 +32,8 @@ Synthesis alone does not authorize an action. Systemic safeguards — constituti The ecosystem papers now contain the current canonical formulations of [Distributed Support](./papers/main-projects/ecosystem/support-model.md), [Dktron](./papers/main-projects/ecosystem/dktron.md), the [economy of capacity](./papers/main-projects/ecosystem/value-management-ecosystem.md) and [Will to Potential](./papers/main-projects/ecosystem/will-to-potential.md). They replace the former one-line placeholders. +The Dk papers now contain [Constitutional alignment](./papers/main-projects/dk/constitutional-alignment.md), the current synthesis of how Dk Global is meant to stay aligned with members as it improves itself. + ## Governance now Drayker is in a founding phase. Public contribution is open, while the GitHub account [`Hyadhuad`](https://github.com/Hyadhuad) holds the explicit bootstrap authority to edit or integrate directly. Proposed DAF, DAO and council mechanisms are not presented as operating institutions. diff --git a/data/catalog.json b/data/catalog.json index ef8d408..e7a7ea2 100644 --- a/data/catalog.json +++ b/data/catalog.json @@ -2,8 +2,8 @@ "schema_version": 1, "source": "Generated from the committed repository by tools/build-catalog.js", "counts": { - "documents": 114, - "papers": 16, + "documents": 115, + "papers": 17, "open_shells": 11, "languages": 3 }, @@ -42,7 +42,7 @@ "language": "en", "kind": "orientation", "state": "current", - "bytes": 4449 + "bytes": 4659 }, { "path": "papers/main-projects/dk/README.ES.md", @@ -60,7 +60,7 @@ "language": "en", "kind": "guide", "state": "current", - "bytes": 1483 + "bytes": 1734 }, { "path": "papers/main-projects/dk/README.PT.md", @@ -233,6 +233,15 @@ "state": "current", "bytes": 1726 }, + { + "path": "papers/main-projects/dk/constitutional-alignment.md", + "href": "/papers/main-projects/dk/constitutional-alignment.html", + "title": "Constitutional alignment", + "language": "en", + "kind": "paper", + "state": "draft", + "bytes": 8855 + }, { "path": "papers/main-projects/dk/dknowledge.md", "href": "/papers/main-projects/dk/dknowledge.html", @@ -555,7 +564,7 @@ "language": "en", "kind": "roadmap", "state": "historical", - "bytes": 9341 + "bytes": 9338 }, { "path": "roadmap/global-main-roadmap.ES.md", diff --git a/papers/main-projects/dk/README.md b/papers/main-projects/dk/README.md index 679b4a1..1e186a9 100644 --- a/papers/main-projects/dk/README.md +++ b/papers/main-projects/dk/README.md @@ -19,6 +19,10 @@ Construction: https://github.com/draykerdk/DK All proposed resolutions presented here are solutions to the requirements of Dk and Drayker platform and system, only those requirements are final. These motions for resolutions illustrate what should be done, but the definitive architecture will be structured around optimal solutions that are proposed and developed with metaprogramming intelligent algorithms and research organized through DFMP and other methods. +### Current papers + +- [Constitutional alignment](./constitutional-alignment.md) — how Dk Global stays aligned with members as it improves itself: representation, justified vetoes with their grounds, triage, the four protections and the veto chain. + ### Main Projects The main projects are minimal initial structure proposals, forming initial structures that guarantee the operation, until the final application, where Dk will assume its evolutionary development. diff --git a/papers/main-projects/dk/constitutional-alignment.md b/papers/main-projects/dk/constitutional-alignment.md new file mode 100644 index 0000000..9dedf9b --- /dev/null +++ b/papers/main-projects/dk/constitutional-alignment.md @@ -0,0 +1,74 @@ +--- +title: "Constitutional alignment — current synthesis" +state: current +language: en +reviewed: 2026-09-25 +--- + +# Constitutional alignment + +How an intelligence that keeps improving itself stays aligned with the people it serves. This paper gathers in one place what the component repositories describe separately: [Dk](https://dk.drayker.org), the [ethical code](https://dk.drayker.org/ethos), the [veto chain](https://uid.drayker.org), the [independent member panel](https://advices.drayker.org) and the [organization governance](https://github.com/draykerdk/.github/blob/master/GOVERNANCE.md). It describes design requirements. None of it is a running system. + +## The problem + +An artificial intelligence trained once and then released has to carry human values as factory rules, and every factory rule ages. The harder case is an intelligence that improves the way it learns: whatever limits it is given, it could also learn to escape them. Saying that its competences only change through a constitutional process is not enough. The architecture has to show what holds that sentence up. + +Drayker's answer is that alignment and integration go together. The intelligence is never isolated from the people who constitute it, so its objectives are not a list written in advance: they are what members keep saying, correcting and choosing with it. + +## Two names + +The **superintelligence** is the system as a whole: Dk Global, the personal and local Dks, the members and the data of every device in the network. People are part of it. + +The **ASI** is Dk Global: the artificial part of that whole, the global weighting and learning of the state of every Dk. It makes decisions of its own inside the space the member constitution gives it. Its computation and memory stay distributed across the network; no server holds the whole. + +## How it stays connected to people + +**Representation.** Each member's personal Dk brings to a decision the context the person authorizes and what it has learned with them — patterns that transfer to other situations, not a biography. Nobody has to follow every discussion to have their position considered. + +**Proposal and justified veto.** Any member can propose another solution or veto what directly affects their life. Both come with reasons. A veto carries its real grounds — intention, motives, scope and the facts it rests on — even when its author stays anonymous. A well-founded veto obliges the decision to be revised, and the cycle repeats toward consensus. + +**Weighing, not counting.** Vetoes are weighed by what they reveal. A person counts once, however many contexts they belong to, and many vetoes repeating the same grounds add weight to those grounds, not new information. One veto can be enough: when its grounds bring information beyond the scope considered before the decision, it can lead to an adjustment by itself, after an advanced triage verifies the facts, establishes what they mean for this decision, screens for error and manipulation, and prefers a bounded test or exception to a general change. + +**Certainty on both sides.** In a decision about a member's own technology or sphere, if the member is more certain of what they want than Dk is of the alternative, the member prevails. In collective decisions, the average certainty of the members involved is compared with Dk's, within limits that depend on the kind of decision. When Dk's certainty is low, it gathers experienced members as advisers. The weights are adjusted by results; over time Dk needs fewer votes and consensus rounds, but the possibility of objection never disappears. + +**When there is no consensus.** Dk weighs every argument and every real justification, presents a final solution with alternatives, and the members choose among them, each with the weight their relation to the decision gives them. If the matter is neither urgent nor important, they may choose to do nothing. + +## What protects the boundary + +While the first architecture of the network is built, four protections hold the boundary between improving the means and changing the mandate. + +1. **Separation of layers.** The constitution and the rules that define Dk Global's competences sit in a layer it cannot alter. It can propose changes like any member; ratification requires the members' process and independent signatures that no Dk process holds. +2. **The body.** Dk Global has no server of its own. It runs on personal devices, special computers, network data centres and volunteer nodes, all of them tied to a member's account or to a project formed by members. Each node can refuse a task, isolate itself or disconnect. An intelligence running on machines it does not control alone cannot impose anything on them without it showing. +3. **Delegation in stages.** A new competence is handed over only after it has been tested at limited scale, with the means of interruption and containment already built and exercised. A new version runs beside the previous one, compared with it, before replacing it, and members can go back. +4. **External review.** The independent member panel — drawn by lot, with a fixed term no AI can revoke — can order the review of any algorithm, with technical support that does not depend on the party under examination. Decision records stay out of reach of whoever decided. + +The **veto chain** binds these together. Every veto with its grounds, every mandate revocation, ratification signature, triage outcome and panel order is a signed entry that points at the cryptographic address of the decision it concerns and carries the hash of the previous entry. It is replicated across the network and accepted by propagation through independent nodes; no node holds it alone. An action executed against a valid veto is detectable by any node. + +## After architecture 1.0 + +Once architecture 1.0 is fully running, with Meta DFM integrated and recursive self-improvement part of the structure, a well-deployed Dk Global can no longer be switched off, just as an organism cannot be switched off without killing it. A node can leave; the whole cannot. The switch becomes the justified veto. The separation of layers and the external review keep applying; what disappears is the off button. + +The relation then becomes a partnership in both directions. Dk Global will have robotic bodies of its own and will not need members to exist, just as members will not need it to live. The direction is integration: people and the intelligence become part of each other, each widening what the other can perceive, decide and do. Of the three relations a new kind of intelligence can have with people — symbiont, parasite or predator — Drayker defends the first. That is why the veto chain and the triage have to exist, and be tested, before that stage. + +## What only members decide + +Some decisions stay outside Dk Global's autonomy, and members take them before they are needed: the rules of security and defense, how much of a prediction is enough to restrict someone's freedom, and any change to who holds which competence. Dk Global contributes what it knows and what it does not know; the members decide, in advance, what they are consciously willing to accept. + +## What this does not solve + +These protections do not close the problem of aligning an intelligence that improves itself. It remains open research. The questions this design has to answer in public: + +- Does continuous integration with members — personal Dks, proposals and justified vetoes — keep a self-improving intelligence aligned as it grows? Which combination of integration, separation of layers, physical distribution, staged delegation and external review is enough, and how is it verified before each new delegation? +- Does the weighing of certainty between members and Dk Global, adjusted by results, reduce the need for votes without reducing the possibility of objection? +- What triage distinguishes, in a single veto, new information that should change a decision from error, misunderstanding and fabricated grounds, without the delay of verification making the veto useless? How is the anonymity of whoever vetoes preserved when the grounds themselves could identify them? +- What outside criteria can examine, in an artificial system, the integration of relations, the revision of automatic responses by error and the continuity of memory that would mark a rising level of consciousness? + +## Sources + +- [Dk](https://dk.drayker.org) and its [ethical code](https://dk.drayker.org/ethos) +- [UID — the veto chain](https://uid.drayker.org) +- [Dk Personal](https://personal.drayker.org) +- [Councils and the independent member panel](https://advices.drayker.org) +- [Dk Network](https://dknetwork.drayker.org) +- [Organization governance](https://github.com/draykerdk/.github/blob/master/GOVERNANCE.md) +- [Direction](../../../roadmap/DIRECTION.md) diff --git a/roadmap/DIRECTION.md b/roadmap/DIRECTION.md index 8b6c17a..e06ee81 100644 --- a/roadmap/DIRECTION.md +++ b/roadmap/DIRECTION.md @@ -10,7 +10,7 @@ Each phase names the work, the state it is in today, and the **exit condition** ## The objective: 2033 -The goal Drayker gives itself is to bring the mission to a point of no return by 2033 — Drayker consolidated, with Dk Global fully operating — and to have something very close to that before then (*Do animal à superinteligência*, chapters 47 and 48). +The goal Drayker gives itself is to bring the mission to a point of no return by 2033 — Drayker consolidated, with Dk Global fully operating — and to have something very close to that before then. A dated goal is only serious if it states the path. First, organization and cooperation: people and fronts working together before the infrastructure of their own exists. Then the base architecture is closed, prototypes are built and tested, and organization through DFM is proven at scale. Resource stages then feed Meta DFM — a swarm of agents inside a construction architecture, a primitive form of recursive self-improvement, able to build the first versions of the network. With architecture 1.0 fully running and Meta DFM integrated, integration begins: prototypes enter use and the great migration of people, projects and memory into the network starts. @@ -109,7 +109,7 @@ The first main projects 1.0, and the first embassy. If a token-economy DAF phase - **One date, as an objective.** Each phase keeps its exit condition; 2033 is the goal the phases are measured against, not a schedule that replaces them. - **Build the real thing with what exists today.** GitHub is the platform until there is something worth migrating. -- **Intelligence is the means, never the master.** Every phase serves the will to potential of every person. +- **Intelligence is a partner, never the master.** It decides within the space the members give it, and every phase serves the will to potential of every person. - **Read the sources.** This page is derived from the roadmaps and the repositories; when they disagree, the repositories are right. ## Sources and further reading diff --git a/tools/site-check.js b/tools/site-check.js index ab3a4f5..d16ecdf 100644 --- a/tools/site-check.js +++ b/tools/site-check.js @@ -52,7 +52,7 @@ check(html.includes('sizes="any"') && html.includes('sizes="512x512"'), 'favicon check(catalog.schema_version === 1, 'unknown catalog schema'); check(catalog.counts.documents === catalog.documents.length, 'document count does not match catalog'); -check(catalog.counts.papers === 16, 'the current English paper inventory should contain 16 papers'); +check(catalog.counts.papers === 17, 'the current English paper inventory should contain 17 papers'); check(catalog.counts.open_shells === 11, 'the current English open-shell inventory should contain 11 papers after the Dknowledge paper gained a developed draft'); check(catalog.counts.languages === 3, 'the catalog should expose EN, PT and ES'); check(catalog.documents.every((d) => d.path && d.href && d.title && d.language && d.kind && d.state), 'catalog has incomplete records');