Software
The purpose of software is to help people to coodinate in meaningful endeavour. All things start as thought — an intention, a pattern, a class. Software is the discipline that turns thought into thing. Every thing creates a collision surface — the question is whether those collisions compound into meaningful endeavor or extract from it.
Platform
The platform is the product — a business factory commissioned one capability at a time. Platform, products, protocols, and prompts that turn principles into working systems. Each layer replaces a vicious loop with a virtuous one.
| Layer | Replaces | With |
|---|---|---|
| Platform | Rented ground, vendor lock-in | Infrastructure you control |
| Products | Manual process, tribal knowledge | Commissioned capabilities |
| Protocols | Ad-hoc integration, bespoke wiring | Standardised interaction |
| Prompts | Remembering, hoping, guessing | Engineered triggers |
Thought to Thing
| Stage | What | Discipline |
|---|---|---|
| Thought | Intention, picture, class | Pictures — engineer the dream |
| Type | Contract, schema, template | Domain layer — the compiler verifies the class |
| Thing | Artifact, instance, product | Commissioned capability — the class made real |
| Collision | Impact, outcome, compound | Reality — did the thing create positive collisions? |
The same pattern at every scale. A Drizzle schema is a class; each row is an instance. A PRD is a class; each built feature is an instance. A venture is a class; each launched business is an instance. The Legacy Rule closes the loop: every instance improves the class for the next.
The loop back is what makes it virtuous. Every PRD makes predictions — which thoughts are worth turning into things, and what those things will produce. Prioritization is the filter. Credibility is the score: did the thing deliver what the thought predicted? The prediction ledger — correct predictions divided by total, weighted by conviction — is what separates trust earned from trust claimed.
Capability Router
197 features across 23 categories. 16% platform coverage. What to build, buy, or bridge?
| Tier | Categories | Verdict Pattern |
|---|---|---|
| Reserve | Identity, AI, Accounting | Build core data model, buy commodity |
| Primary | CRM, Workflows, Analytics, Marketing | Build what you sell, bridge the rest |
| Secondary | Blockchain, Design, Video, Community, Search | Build on-chain, buy content tools |
| Defer | IoT, Geospatial, Mobile, Field Ops | No demand signal yet |
One table, five answers per row: Capability Router →
Build what touches your data model. Buy commodity. Bridge everything else.
Three Credibility Loops
Software that coordinates intention must earn trust at three levels. Each loop is harder to fake than the last.
| Loop | Question | What Passes | Conviction |
|---|---|---|---|
| Inner | Does it work? | Tests green, benchmarks met, L3 commissioned | LOW — "it works" |
| Story | Does the story match? | Predictions scored, kill criteria honoured | MEDIUM — "it matters" |
| Market | Do others agree? | Adoption, revenue, referral — L4 | HIGH — "others agree" |
"Minimal need for trust" means Loop 3 evidence is verifiable, not narrated. The development journeys pipeline wires all three.
Dig Deeper
- Journeys — Pain to proven value: the full pipeline with three credibility loops
- Platform — Infrastructure you control: AI, blockchain, DePIN, operations
- Products — Commissioned capabilities: design, engineering, AI products
- Protocols — Standardised interaction: agent protocols, smart contracts, standards
- Applications — Build, buy, or bridge: the capability router
- AI Coding — Signal discipline for human-agent orchestration
Context
- Progress — Qualify before you quantify: Pictures, Principles, Priorities, Problem-Solving, Productivity
- Business — The factory's output: ventures from shared mycelium
- Commissioning — What's specified, built, and proven
- Standards — Where proven patterns graduate to
- Navigation System — Value, Belief, Control: the three systems software serves
- Development Journeys — The full pipeline: pain → demand → spec → code → commission → credibility
- Credibility — The prediction ledger: commitments kept / commitments made
Questions
What collision surface does each commissioned capability create?
- Which layer of the stack is the bottleneck — platform, products, protocols, or prompts?
- If the platform is the product, what's the difference between building software and building the factory that builds software?
- What vicious loop is each layer designed to replace — and is it actually replacing it yet?
- If every thing starts as thought, where do the best thoughts come from — and how do you recognise them before the crowd does?