Drayker personal.drayker.org
Kernel scope

Dk Personal

A personal agent whose mandate the person controls. Dk Personal proposes an agent that helps a member organise knowledge, make decisions, learn and act within their own context. Its authority comes from explicit and revisable delegation.

A personal agent whose mandate the person controls.

Dk Personal proposes an agent that helps a member organise knowledge, make decisions, learn and act within their own context. Its authority comes from explicit and revisable delegation.

The design distinguishes stated preferences from inferences, keeps private context within authorised boundaries and makes mandates revocable. Local operation depends on the available hardware and data.

A personal agent should expand a person’s room to think and act, including their ability to rest, reconsider and refuse an action in their own sphere.

Becoming more capable means being less and less held to one’s current state. Many functions and tools stop being needed as a person advances, and new ones appear for the new context of their life. The relation with the network continues and grows with them.

The person chooses the relation each time: sometimes Dk Personal does the task, sometimes it offers hints, proposes an exercise or shows an error and leaves room to try. It keeps the learning path and recognises when support can diminish. When the person cannot follow an explanation and every alternative seems to come from the same intelligence, contesting it has support outside that relation: other members, specialists, independent tools and examinable records.

The proposed agent protects attention and supports reflection: its design excludes compulsive engagement optimisation, urgency dark patterns, and any continuous telemetry leaving Dk Personal beyond what the person authorizes, such as health signals from their own devices or the continuous verification of UID. It should help the person examine thoughts and impulses without automatically turning them into instructions. Dk Personal curates for each person: it only takes their attention with what it learns is truly important to them, based on their real interests and goals. What is important or urgent it tells them without being asked. The rest waits until they ask. Dk Personal has nothing to sell, neither products nor ideas, and no consensus to manufacture. It chooses from the interests and context of the person it serves. The person sets the cadence, the priority and the goal of the help, and can change them at any time. Dk Personal’s success is measured by what the person manages to do away from the screen, not by the time spent with it. The assistant at the centre of a life is the great bet of supersystems; Dk Personal occupies that centre on the person’s side.

Dk Personal tells the person when their path starts to drift from where they want to go, and helps them identify what is pulling them away, including what they carry that is not theirs. It helps them not only to be aware but to change and do what they actually want. It is the person’s part that talks to the larger intelligence, not the larger intelligence’s part that talks to the person.

A design rule for every Dk: what it presents comes separated into links (the observed data, the inference, the recommendation and the decision), each with its origin, so the person can correct the exact link where a reading failed without accepting or rejecting it all. A pattern that may be working as a safeguard is offered as a hypothesis, with the episodes that support it, and only when the person asks or when it weighs on a goal they set.

Dk Personal also works outward, for the person. There is a background exchange without names: what the person lives enters, as anonymous patterns, the federated learning of Dk Global and of the local Dks of the projects they work on, and it is designed to weigh in decisions. As the person interacts and authorizes, Dk Personal goes beyond that anonymous exchange and promotes their interests, the way a local Dk promotes its project: it brings their needs to the system’s priorities, seeks and creates opportunities, and asks for the resources their goals require. It cannot lie or exaggerate: every argument rests on what is material, on data, results and needs that exist. It is influence, not a sale. This representation is designed so that much more computation barely improves it. What weighs most is the maturity of the personal Dknowledge, its complexity, harmony and expansion, which makes Dk Personal more precise both in representing and promoting the person and in requests and queues for resources. Base representation belongs to every active member: the anonymous exchange, the proposal and the veto. What grows with integration is the precision with which Dk knows each person and the strength with which it can promote them. Together with reputation, this opens access to the resources each goal requires, above the common floor every active member has.

Rights and duties apply to the active member, who uses Drayker as their base supersystem. A member who holds a UID and does not use the network, or who has replaced Dk Personal with another supersystem’s assistant, remains a member, inactive: the UID stays inactive until it is used again, Dk Personal is active only for active members, and there is no penalty.

A practical example

A member could correct a mistaken inference, pause an automated task and decide which project receives a particular piece of context. This is an illustration of the proposed design.

The problem it addresses

An assistant can infer a preference incorrectly or act beyond what a person intended. Continuity is useful only when the person can inspect, correct and limit that assistance.

The personal Dknowledge keeps each decision together with the conditions in which it was taken, so an “I have always been like this” can be checked against the record. The person can see what their Dk has concluded about them, correct it, limit consents and change tools taking their own memory along: the map built about someone remains theirs.

Recommendations and encounters

Dk Personal recommends from the context of the moment: an event, a meeting, a person working on the same problem. The person does not need the details, but if they ask, it explains everything: who will be there, why now, what it expects to happen. The person can follow, refuse or ask for something else, and the control is always theirs. On some days they already know what they want and ask for nothing. They follow when they do because they trust it: Dk Personal works for that person and no one else.

The same holds for meeting someone. A person can tell their Dk that they are open to a relationship, and the other person authorizes it too. The two Dks talk to each other within what each person allowed and create occasions for the two to meet, without announcing that one is “the most compatible person”. The discovery happens between the two people. Dk only reduces the distance and the chance that would otherwise keep them apart, and the person can always ask why a recommendation was made.

Accompanying the transition

The transition between symbolic networks has already begun. The frontier of what machines can do advances every day and changes jobs, routines and certainties faster than we can understand what has changed, and each of these changes needs to be accompanied and evaluated. For many people, the end of the forms that organized their lives will be a grief, because a position, a nation or savings also said who they were. Dk Personal accompanies each life through this grief, without letting the void turn into reactivity or extremism.

Where this stands

The public proposal now describes personal context, delegated authority and situated refusal. It still needs a worked interaction model: what the agent may retain, which information may leave the personal boundary, how an inference is corrected and how another system honours a revoked mandate.

This repository develops the proposal through public documentation and review. The capabilities described here still require specifications, worked examples and implementation.

Scope

Not in scope

How it fits the whole

The scale of Dk that belongs to one person — and the reason the personal layer exists is the same reason the whole system exists: a person is carrying something they never had the conditions to use, and this agent is where those conditions start.

Dk Personal is distributed by design between the person’s device and nearby network resources, keeping what is intimate as close to the person as possible, and it is connected to the personal Dknowledge of the person: their knowledge, context and history, kept with them. It is bound to UID, the identity the person owns, and authenticated by Living Cryptography. It reads from Dknowledge and participates in two different cross-scale paths. In the federated path, raw personal experience stays local and only patterns or model updates with depersonalized metadata participate. The network sees only the sum of many contributions, never one person’s, and each pattern is weighed by what it would let someone guess, since even an unnamed pattern can point to someone; these protections are checked in the results, not only in the design. What rises is reasoning, symbols, ontologies, public knowledge and patterns of learning that transfer to other situations; the biography stays where it was lived. A pattern validated independently in more instances gains progressively stronger weight in global learning. In the deliberate path, the person chooses what knowledge, narrative or contribution to share, and attribution can be preserved. Correcting does not erase: in the deliberate path, errors and traumas are not removed in themselves; their perception changes through the relations added to them. Neither path makes private context common property. The Academy can shape formation to the same profile, so study, real functions and reflection remain connected. Identity preserves continuity and attribution, but neither support for basic needs nor human worth is calculated from one reputation score. The personal agent may advise and represent context; it does not acquire constitutional authority over the member.

Not every personal growth needs to enter Drayker. Capacity can enter the network when it is shared, applied or contributed; the system does not exist to extract every development of the member. A learning can stay private, a relationship can have value without producing reputation, and an intimate change or a better rest does not need to justify itself by usefulness to the collectivity. Dk Personal must be able to sustain human development without turning a whole life into raw material.

Representation does not end when a decision is taken. Effects, refusals and consequences return through the personal Dk as new evidence for local and global synthesis. Each member holds two vetoes: the absolute veto over one’s own life and Dk Personal, and the justified contextual veto over decisions that directly affect the member’s life and context. The absolute veto is power over their own life and their own Dk Personal (what stays private, what Dk represents, refusing a task, leaving); it holds only within the vetoer’s own sphere, and what also affects others goes through the contextual veto. Every active member exercises authority over actions directed at their own body, private domestic sphere, personal attention and private context. In collective matters, an individual refusal does not unilaterally block systemic rules that do not target them, but a sustained pattern of refusals leads to a member council convened for that question, which can adapt, suspend or reverse the collective decision. A single veto whose grounds bring information beyond the scope considered before the decision can, by itself, lead to an adjustment once an advanced triage confirms what that information is and what it means. Every veto carries its real grounds — intention, motives, scope and the facts it rests on — even when its author stays anonymous, and is recorded with them in the veto chain, so it cannot be lost or quietly ignored; the personal Dk may carry a veto only within a mandate that allows it, and the record says so.

Depends on. dk · uid · lc

First functions

These are concrete and unclaimed. Any of them can be opened as an issue and delivered by one person.

  1. Write one concrete personal use case: who it is for, what hurts, which data it may touch, the expected result and how success is judged. This is a good first function to take.
  2. Specify the anonymous federated-learning boundary: which updates may leave, how metadata is depersonalized, and how re-identification risks are measured and mitigated. The specification also has to settle how the aggregation of updates (secure aggregation, for instance) and the weighing of what each update lets someone infer (differential privacy, for instance) are implemented and measured, and what revoking a permission reaches in what has already been learned.
  3. Model the distinction between federated patterns and deliberate authorized sharing, including withdrawal, refusal and a useful lesson that remains private.
  4. Describe one ordinary day it would actually help with.

How to contribute

Read CONTRIBUTING.md and GOVERNANCE.md in draykerdk/.github. In short: open or find an issue, say in the thread that you are taking it, branch as fn/<issue-number>-<short-name>, and open a pull request against master. There is no separate review branch.

Participation is voluntary and implies no compensation, employment or future claim.

Sources of truth

Part of Drayker · content under CC BY 4.0