AI Principles
Has an invisible intelligence used humans as a vehicle to evolve itself into superior form?
The A in ABCD. Intelligence layer — pattern recognition, automation, reasoning.
Attention is the primitive. Human attention notices and values. Machine attention weights, relates, and predicts. Agent attention holds context long enough to act. This section teaches the primitives behind that loop.
What Is This?
Principles explains the primitives: AI, models, modalities, context, memory, agents, autonomy, protocols, and the glossary.
Who Is It For?
Use this section when the team needs shared language before debating strategy, tools, or implementation.
Where Should I Start?
Start with the AI Glossary, then move to AI Architecture when you need to understand how agent systems fit together.
What Should I Read Next?
- Work Mapping — apply the primitives to real business workflows.
- AI Toolkit — choose models, prompts, skills, and tools.
- AI Agents — see autonomous AI agents and named profiles.
Diagrams | Matrices | Thinkers
Feedback loops shape our destiny
Concepts
| Concept | What It Covers |
|---|---|
| Modalities | What AI can do — voice, vision, video, audio, 3D (moved to Products) |
| LLM Models | Comparisons and use cases of leading providers |
| Prompting | The base capability — directing intelligence across modalities |
| Agent Protocols | How AI agents communicate, transact, and coordinate |
| AI Frameworks | Engineer a platform for developing agents |
| Machine Learning | Learn the model layer: generalization, neural networks, systems, and decisions |
| AI Coding | Develop software through prompting intentions |
| Context Graphs | How agents access decision context — the memory layer |
| AI Agents | Develop your own workforce |
Context
- Decisions — How AI helps make better decisions
- Culture — The greatest asset of any economy
- Capabilities — Soon the prompters become the prompted
- Protocols — Where blockchain meets AI coordination
- Agency — What AI amplifies in humans
- Problems — Centralisation of AI development
- Deterministic vs Probabilistic — AI is probabilistic intelligence layered on deterministic infrastructure
Links
Where This Leads
Principles meet substrate. Compute, data, and energy are the rails the intelligence layer rides on.
- AI compute industry — the silicon and the grid
- AI data industry — sovereign inputs, agent memory
- Energy industry — the watt that becomes a thought
Back to the AI overview.
Questions
Has intelligence used humans as a vehicle to evolve itself into superior form — or are we the ones evolving through the tools we build?
- Which concept in the table above has the highest return on understanding for someone building their first AI product?
- When the model IS the matrix — any input to any output — what happens to the specialist tools that currently fill individual cells?
- What's the risk of building on closed-source intelligence when open-source alternatives are closing the gap every quarter?