Add initial exploration history and draft INTENT.md

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 1494246@bnt-lap001
Assistant-Session: caf6de6f-f9e4-47ce-a584-49baa271fa4b
This commit is contained in:
tegwick 2026-10-11 23:55:45 +02:00
parent b1c7894a6c
commit b3dc5b83de
2 changed files with 4426 additions and 0 deletions

163
INTENT.md Normal file
View file

@ -0,0 +1,163 @@
# INTENT.md
> **Status:** draft suggestion (2026-10-11), derived from `README.md` and
> `history/20261011-InitialExploration.md`. To be reviewed and ratified by the
> project owner.
## Project
`assembly-economics`
## Purpose
`assembly-economics` exists to develop an experimental economic framework for
societies in which superhuman machine capability — physical, computational and
increasingly cognitive — provides most of the leverage for innovation and
profit.
In such an economy, wage labour can no longer be the main way people take part
in value creation and gain agency. The project explores a **post-work mechanism
for participation and agency**. Its framing is that an economy is an *adaptive
information processing network*: an assembly of information organisms, humans
and institutions, linked by flows of information, resources and attention.
It is the philosophical and economic framework that **coulomb.social** explores
and builds on.
## Core Intent
Describe an economy as a living, evolvable information-processing system, and
establish the concepts needed to design, govern and evaluate such systems so
that:
1. **Abundant machine resources are spent to preserve scarce sentient
resources.** Money, compute and machine capacity are convertible potential.
Human attention, autonomy and wellbeing are lived capacity. They can only be
allocated once and must be protected by viability boundaries, not optimised
away.
2. **Humans keep purpose and authority.** Information organisms learn, predict
and act, but objectives, authority and hard constraints come from the human
and institutional environment and stay outside the optimisation loop.
3. **Participation does not require labour.** Supplying objectives, judgment,
attention, consent and governance is treated as real economic contribution,
and it is valued accordingly.
## Conceptual Pillars
The initial exploration established six connected layers. Together they form
the project's working vocabulary.
### 1. Adaptive Information Processing Network (AIPN)
A persistent, typed, token-flowing, self-modifying semantic network. Nodes are
high-level business concepts (`DocumentType`, `EmployeeRole`,
`WorkstationClass`, `Goal`, `Observer`). Activations are concrete instances
(documents, people, materials). Observers emit multi-dimensional
success/error signals against explicit and implicit goals. Learning changes
future activation, prediction, utility and, eventually, the structure itself.
> Semantic structure + activation + prediction + feedback + adaptation.
> `sense → activate → predict → act → observe → learn → reorganize`
Key distinctions: structure vs. activation, observation vs. evaluation,
prediction vs. utility, learning vs. authority, viability before optimisation.
### 2. Membrane
The boundary between an organism and its environment, such as a company's
governed IT landscape. It works through semantic projections (read models, not
replicas), preserves identity, provenance and the original governance envelope,
and is claim-oriented. It is bidirectional but asymmetrical. It filters the
vast environment down to what the organism actually samples. It is layered
(constitutional, operational, adaptive, ephemeral) and keeps negative space
visible. Membrane configuration is the main control and deployment interface.
### 3. Seed Kernel
A minimal developmental substrate rather than a finished intelligence. It
attaches to an information environment and grows by adapting to it. The
environment feeds it objectives (prompts act more like hormones than
nutrients) and resources (compute budget). Utility becomes endogenous,
attention becomes resource allocation, and symbiosis is a better target than
autonomy.
### 4. Evolution
Organisms as genotype/phenotype pairs that are bred, recombined and inherited
across generations. Fitness is multi-dimensional and environment-based. There
are nested adaptation loops (learning, development, evolution, ecosystem), and
heritable membranes. Hard constraints remain outside evolutionary optimisation.
### 5. Metabolism and Substrate
Organisms run on substrate layers: information, compute as energy, runtime as
habitat, memory as expensive tissue, authority and human attention as premium
energy. Metabolic pathways, waste and excretion, starvation and abundance
behaviour, and return on compute tie internal regulation to economic selection
pressure.
### 6. Resource Ontology and Attention Economics
Resource *accounting* (money as generalised, convertible potential) is distinct
from resource *ontology*. Resources are classed as commodity, capability, human
and institutional, along with whether they are sentient. Human attention is
treated as the reference scarce resource. Attention regulation becomes a
central societal control loop:
```text
human purpose → objectives → economic allocation
→ (money · compute · machine capacity) → information organisms
→ work → abundance → human benefit → feedback
protected throughout by: attention · autonomy · wellbeing viability boundaries
```
## Scope
**In scope**
- A conceptual canon (currently v0.1 in the history file) and its successors
- Economic models for participation, agency and value attribution in a
post-work setting
- Resource ontology, conversion functions and attention-protection mechanisms
- Membrane and governance models for embedding information organisms in real
institutions
- Reference scenarios (e.g. invoice processing) and minimal experimental
kernels that make the concepts testable
- Framework input for coulomb.social
**Out of scope (for now)**
- Production implementations, storage engines, protocols or API formats.
Canon v0.1 deliberately leaves these unstandardised.
- Concrete marketplace product features. Those belong in `coulomb-social`.
- Pricing mechanics as such. See `adaptive-pricing`.
## Guiding Principles
- Semantic explicitness over opaque optimisation
- Viability precedes optimisation, and sentient viability precedes all else
- Learning is distinct from authority, and governance stays outside the network
- Provenance and explainability are first-class
- Keep the kernel small; separate the normative core from exploratory
extensions
## Proposed Next Artifacts
From canon v0.1 § 53, extended toward the economic layer:
- `canon/ConceptualCanon.md`: extract and stabilise the v0.1 canon
- `canon/NetworkObjectModel.md`: precise, implementation-neutral object model
- `canon/MembraneModel.md`: the interface membrane and its layers
- `canon/ResourceOntology.md`: resource classes, conversion functions,
sentient boundary
- `canon/ParticipationModel.md`: post-work participation and agency
- `research/ResearchProgram.md` and `research/ReferenceScenario-InvoiceProcessing.md`
## Relationship to the Ecosystem
- **coulomb.social**: primary consumer; this repo supplies its philosophical
and economic framework.
- **adaptive-pricing / commerce-canon**: neighbouring economic mechanics that
may later adopt this repo's resource ontology.
- **the-custodian / State Hub**: work coordination (domain `communication`,
topic `coulomb-social`).

File diff suppressed because it is too large Load diff