Focus
Sacred technology, made buildable
Sacred technology is technology designed to preserve agency, attention, and meaning over time — by making provenance, lineage, and intentional rhythm first-class engineering concerns. Not vague spirituality. A design constraint with formal implementations: provenance graphs, content-addressed immutability, calm UX patterns, and mixed-initiative co-creation protocols.
Sacred persistence and lineage
The central claim: storing information is not the same as transmitting its meaning. A database backup preserves bytes. A lineage-aware system preserves the provenance graph — who created what, derived from what, through what process — so that meaning remains recoverable even when the original context is gone.
This is grounded in the W3C PROV family: a mature standard for representing provenance as a graph of entities (things), activities (what happened), and agents (who or what was responsible). Applied to knowledge systems, it makes attribution, auditing, and forking tractable across time and system migrations.
The four principles
Calm
Technology at the periphery of attention
Calm technology requires the smallest possible amount of attention and should inform without demanding focus. A notification is calm if it can be ignored. A tool is calm if it disappears when not needed. Violation of this principle — technology that constantly competes for attention — is a structural failure, not a design preference.
Lineage
Provenance as infrastructure
Every meaningful artefact has a history: who created it, what it was derived from, and how it changed. Lineage-aware systems make this history first-class — not a comment in a changelog, but a graph of provenance entities, activities, and agents that can be queried, forked, and traversed. Without lineage, storage is not transmission: the meaning decays.
Agency
Human–AI co-creation with preserved control
Mixed-initiative design couples model automation with direct human manipulation. The human sets intent and approves consequential decisions; the model handles scale and synthesis. Good co-creation patterns define the handoff point precisely — when does the model lead, and when does the human override? Systems that get this wrong either remove human agency or make automation pointless.
Ritual
Rhythm and intentional repetition
Ritual in technology is intentional repetition that builds meaning over time: daily notes, weekly retrospectives, versioned commits. These aren't productivity tricks — they're the mechanism by which knowledge compounds. Systems designed for ritual make the rhythm visible and honour it, rather than optimising it away.
Implementation primitives
The concrete building blocks that make sacred technology engineerable.
Provenance
W3C PROV-O for entity–activity–agent graphs. Every artefact knows its ancestry.
Immutability
Git-style content-addressed objects. History is append-only; forks are explicit.
Calm UX
Amber Case's calm technology principles. Minimum attention, peripheral availability.
Co-creation
Mixed-initiative patterns. Model handles scale; human approves meaning.
Human–AI co-creation patterns
Mixed-initiative design defines a precise handoff: what does the model do automatically, and where does the human retain direct control? The failure modes are symmetrical — too much automation removes agency; too little makes the system useless. Good co-creation patterns are explicit about the boundary.
Model leads
Synthesis, pattern-matching, first drafts, search. Scale tasks where human review follows.
Human approves
Consequential decisions, publications, deletions, anything irreversible. Explicit checkpoints.
Human overrides
Any time, without friction. The model must yield immediately and without argument.
Projects and essays
Where these principles show up in concrete work.
Sacred Persistence Protocol (essay)
The core thesis: why storage is not transmission, what lineage-aware knowledge systems look like, and why this matters for AI.
Sacred Codex (project)
A concrete implementation of sacred persistence — a versioned, lineage-aware knowledge protocol.
Agentic AI Landscape 2026: Toolsmith Field Guide
The infrastructure-first view of the agentic ecosystem, including where sacred persistence fits.
About
The engineering philosophy behind this work: why yoga, philosophy, and systems design are the same problem.
Common questions
Isn't 'sacred technology' just spiritual branding for normal software?
Fair challenge. The word 'sacred' here means something specific: technology designed to preserve agency, attention, and meaning over time — not as a marketing claim, but as an engineering constraint. It means lineage is first-class, attention is respected, and human control is structurally preserved. That's different from typical software design, which often treats these as secondary to throughput and engagement.
What is sacred persistence?
Sacred persistence is the thesis that storing information is not the same as transmitting its meaning. A database backup preserves bytes; a lineage-aware knowledge system preserves provenance, context, and the graph of relationships that make meaning recoverable. The Sacred Codex is the implementation: a protocol for knowledge that can survive system migrations, model changes, and the passage of time.
How is this different from PKM (personal knowledge management)?
PKM is typically personal and informal. Sacred technology is about systems-level design: formal provenance (W3C PROV-style entity-activity-agent graphs), content-addressed immutability, and co-creation protocols that preserve human agency in AI-assisted work. It's also explicitly about transmission to others and to future systems — not just personal recall.
What does calm technology mean in practice for AI systems?
Apply Weiser and Brown's principle: an AI system should use the minimum necessary attention and should be available at the periphery without demanding focus. Practically: no unsolicited interruptions, clear signals that don't require interpretation, and the ability to be ignored without consequence. An agent that requires constant supervision is not calm. An agent that surfaces only when its output is ready is.
What is a lineage-aware knowledge system?
A system where every artefact carries a queryable provenance graph: what activities produced it, who (human or agent) performed those activities, and what other entities it was derived from. Modelled formally on W3C PROV-O. This makes attribution, auditing, and knowledge forking tractable — and makes it possible to trace the ancestry of any claim in the system.
Also in Focus
Agentic Workflows →
Patterns for reliable AI agents: orchestration, state, human-in-the-loop, and regression testing.
Also in Focus
AI Tooling →
MCP servers, tool design, context engineering, and the infrastructure that makes agents actually work.
Building lineage-aware systems?
Let's talk sacred infrastructure, knowledge protocols, and systems designed to preserve meaning.