Problem
We habitually reach for specialized tools - databases for data, UIs for design, diagrams for architecture. These create barriers: tool switching, format conversion, access friction.
Pattern
Text is everything now. Stay in text.
Text is AI's native medium. When you stay in text:
- Everything is at your fingertips - directly editable
- No barriers, no tool switching
- Instant iteration
- Version-controlled by default
- Shared workspace - both human and AI can edit
What can be text:
- Design mockups (ASCII)
- Architecture diagrams (ASCII)
- Process workflows (markdown)
- Data specifications (markdown tables)
- Plans and todos (markdown)
- State descriptions (plain text)
If it can be text, make it text.
Example
UI redesign staying entirely in text:
Designer draws on paper → AI converts to ASCII → edit ASCII together → describe in text → compare old vs new state (both text) → generate todos from diff → implement.
Never left text. No Figma, no drawing tools, no databases. Just editable text files.
Architecture exploration: Ask AI for codebase architecture - get ASCII diagram you can edit and refine together, not a PNG you have to regenerate and open in a different tool.
Process automation: Write processes in markdown. AI reads and follows them. Edit the file, AI adapts immediately.
Text keeps everything lightweight, editable, and accessible.
