Core Concepts

• Adapter composition algebra

  The operating-system metaphor for AI in PointSav — the Doorman as kernel, adapters as processes — and the algebra assembling intelligence from LoRA layers.

• Apprenticeship substrate

  Platform mechanism routing work through a local Small Language Model first, capturing signed senior verdicts as preference pairs for continued pretraining.

• BIM Object specification

  The platform's reusable building-element specification unit: a fixed set of primitive categories anchored to open standards (IFC, Uniclass, bSDD), each carrying three layers of information at once — what it is, what its jurisdiction requires, and what its climate requires.

• Brief queue substrate

  A durable file-backed queue that makes idle-shutdown Yo-Yo compute viable without losing apprenticeship corpus capture data — the durability layer of the three-tier SLM substrate.

• Capability geometry: seL4 capability authorization in Totebox orchestration

  Capability Geometry is PointSav's term for seL4-based authorization that replaces mutable access-control policy with a formally proven, kernel-enforced capability DAG.

• Capability ledger substrate

  The Capability Ledger Substrate is the mechanism by which every access-control decision becomes a cryptographically auditable event anchored to a customer-controlled log.

• Citation substrate

  Platform-wide YAML citation registry with drift detection that makes provenance machine-auditable from regulatory instrument to published claim.

• Code for machines first

  Every inter-service contract, audit record, configuration, and ontology is machine-readable as a primary surface; human-facing interfaces are skins on machine-first APIs.

• Compounding Doorman

  The operational pattern at the heart of sovereign AI substrates: a single service mediating every external compute call, logging events, accumulating training signal.

• Compounding substrate

  Architectural pattern pairing open platform code and a deterministic AI-free data layer with an optional intelligence layer whose interactions compound as training signal.

• Customer-owned graph IP

  The per-tenant knowledge graph and trained adapter weights are the customer's intellectual property, portable and exportable without vendor approval.

• Design-system substrate

  The design-system substrate is a self-hosted, customer-owned engine storing tokens and components in the customer's own Git repo, served via a machine-readable MCP endpoint.

• Disclosure substrate

  Mechanism making a version-controlled Markdown wiki the primary continuous-disclosure record, with signed authorship chains and cryptographic content hashes.

• Editorial draft routing protocol

  Metadata classification layer that routes editorial drafts by their language_protocol declaration — which gateway processes an artifact and which vocabulary rules apply.

• Four-tier SLM substrate ladder

  A graduated sovereignty path for AI deployment: four customer tiers from a lightweight API gateway up to a domain-specialist service, each adding capability without regressions.

• GIS as a BIM substrate

  What the co-location dataset offers a BIM composition pipeline: the cluster manifold and its joinable fields, region-resolution depth, civic context layers, and the stability guarantees a downstream consumer can rely on.

• Knowledge commons and service commerce

  The economic model separating what PointSav publishes freely from what it sells — public knowledge under open licenses, paid service at multi-Totebox aggregation.

• Knowledge-graph-grounded apprenticeship

  The Doorman looks up matching entities in the per-tenant knowledge graph before dispatching a request, grounding the model's response in facts the graph already holds.

• Language-protocol substrate

  The routing mechanism that carries a draft's declared register, document type, and destination between archives — a frontmatter field, a routing table, and a mailbox convention, not an AI adapter system.

• LLM substrate decision — OLMo 3 family

  The rationale for selecting OLMo 3 as the local and GPU-burst substrate: the only fully open model family permitting continued pretraining and public-company procurement.

• Location intelligence substrate

  A flat-file, open-GIS architecture letting customers own geographic datasets end-to-end using open data and a Rust-aligned rendering stack, retail co-location as first surface.

• MCP as substrate protocol

  Every Ring 1 and Ring 2 service exposes a Model Context Protocol server interface as its primary external contract, with the Doorman as the MCP gateway.

• Merkle proofs as a substrate primitive

  Merkle proofs are the cryptographic mechanism letting the platform prove to any third party that a record is part of an append-only log that has not been rewritten.

• moonshot-toolkit build orchestrator

  Rust-only build orchestrator for seL4 unikernel images — TOML spec to content-addressed manifest to bootable AArch64 elfloader, replacing Python and CMake.

• Nightly DataGraph rebuild

  The scheduled process that reconstructs the platform's knowledge graph from canonical flat-file sources each night. A human-approval checkpoint exists for AI-extracted entities, but it is opt-in — an operator must enable it; automated writes land without per-item review by default.

• Organizational knowledge graph — ontological memory for business operations

  Organizational knowledge graph of people, companies, projects, and relationships — persistent semantic memory for answering business-state queries without re-reading sources.

• Retail co-location tier methodology

  Gate-based tier classification for retail co-location clusters — Regional, District, Local, or Fringe — assigned by passing fixed composition, catchment, civic-support, and overlap tests rather than by a composite score.

• Reverse-flow substrate

  The Doorman gateway and audit ledger enforcing inbound data discipline are planned to also enforce outbound commercial flows — marketplace and ad exchange, opt-in per tenant.

• Seed taxonomy as SMB bootstrap

  Every tenant deployment provisions a four-part seed taxonomy — Archetypes, Chart of Accounts, Domains, Themes — as the knowledge graph bootstrap.

• seL4 AArch64 QEMU substrate target

  Hardware foundation for the unikernel platform — formally verified seL4 on AArch64 with QEMU's virt machine as the development, testing, and CI environment.

• seL4 microkernel substrate

  Formally verified seL4 microkernel, PointSav's planned shared L1 kernel substrate — not yet the running kernel for every OS family member as shipped today.

• seL4 Unikernel Substrate for os-console

  os-console is intended to run as a seL4 Microkit unikernel image in production, compiling application code with a formally verified kernel to eliminate OS attack surface.

• Single-boundary compute discipline

  Every AI inference request in a platform deployment routes exclusively through the Doorman, with bypass structurally prevented at the kernel level.

• Sovereign AI commons

  PointSav's market positioning as steward of shared, open AI infrastructure for regulated SMBs: structural properties large cloud providers cannot offer without changing billing.

• Substrate without inference — The base case

  The Totebox Archive remains fully operational and freely transferable even when no AI inference tier is available; the deterministic substrate is the load-bearing foundation.

• Substrate-native compatibility — why the Action API shim was dropped

  Structural compatibility with MediaWiki reader and integrator conventions while declining API mimicry, keeping substrate-native interfaces to reduce maintenance burden.

• System substrate architecture

  The kernel-level architecture beneath every PointSav service — a customer-rooted capability ledger, a two-bottoms sovereign OS strategy, and boot-anywhere recovery.

• Tier 0 customer-side sovereign specialist

  The Tier 0 Totebox is a sovereign specialist deployment running on the customer's own hardware with no required cloud dependency and no required internet connectivity.

• Tiered inference gateway — local-first AI routing

  A tiered inference gateway routing AI requests through a local model first, escalating to remote GPU nodes and external APIs only when needed, minimizing cost and exposure.

• Trajectory substrate

  The platform mechanism converting operational work — commits, sessions, feedback — into structured JSONL training tuples feeding a continued-pretraining corpus.

• TUI as corpus producer

  Every terminal interaction with service-slm through the operator TUI is a curated training corpus contribution for the per-tenant adapter.

• Yo-yo #1 nightly LoRA training pipeline

  The nightly two-phase pipeline on Yo-Yo #1: Phase 1 runs entity extraction for the DataGraph; Phase 2 trains a LoRA adapter via QLoRA on a single L4 GPU.

• Yo-yo compute substrate

  The three-ring compute substrate letting service-slm spin GPU inference capacity up and down while retaining state and producing an audit ledger of every compute event.