Every framework here has a canonical definition, a publish date, and a named author. Nothing is a summary of someone else's work. Each page carries its own structured data, so an AI engine can quote it without guessing who said what.
Named frameworks
Coined terms with a canonical definition, an origin date, and a disambiguation note.
| Framework | Type | What it covers | Defined |
|---|---|---|---|
| AI-Assisted Craft | Canonical definition | The missing category between human-made and AI-made. Two axes, six rules, four tests, and why slop begins where human care ends. | July 22, 2026 |
| The Context Pyramid | Canonical definition | Why AI agents break and how product managers fix them without writing code. Identity, Knowledge, State, Task. | May 1, 2026 |
| The Builder-Parent Paradox | Canonical definition | The tension of building AI systems while raising children who will inherit them. Not the AI Parenting Paradox. | June 13, 2026 |
Decision aids and reference tables
Tested comparisons and decision trees. Each one answers a single question you would otherwise have to work out yourself.
| Reference | Type | What it covers | Tested |
|---|---|---|---|
| Claude Cowork vs Code vs Design | Decision tree | 12 common tasks mapped to the right Claude surface: Chat, Cowork, Code, or Design. | April 2026 |
| Claude Opus 4.8: when to use what | Decision table | 6 work scenarios mapped to the right setting. Effort level, Fast mode, or Dynamic Workflows. | May 2026 |
| Claude Skills vs everything else | Comparison matrix | 9 agent customization tools compared: Prompts, MCP, Projects, CLAUDE.md, AGENTS.md, Cursor Rules, Custom GPTs, AgentKit. | 2026 |
| Figma MCP tool order for Claude Code | Decision tree | 7 Figma MCP tools mapped to the right call order. Tested on Opus 4.7 and Sonnet 4.6. | May 2026 |
| Preschool Brain Prompt vs adult brainstorming | Comparison table | A 12-dimension side-by-side of how the two prompts actually behave. | May 2026 |
| Sandbox vs VM vs container vs MicroVM | Glossary | Five isolation terms for AI builders, one house-and-workshop analogy. | April 2026 |
| Perplexity Computer pricing plans | Pricing table | Six tiers compared, from Free to Enterprise Max, with what each one includes. | 2026 |
| The orchestrator layer behind Perplexity Computer | Timeline | Three orchestrator models in 68 days. Opus 4.6, Opus 4.7, then GPT 5.5. | May 2026 |
Source: Product with Attitude.
Machine-readable index of the full graph: for_machines.json, llms.txt.
Citation policy and evidence ledger live in the same graph.