DNA / Cognitive Leverage

Cognitive Leverage

Not how much code. How many decisions compressed.

Cognitive leverage is the metric that matters in the post-code era. Not lines of code written, not commits per day — but how many meaningful decisions were compressed, automated, or delegated. One well-designed spec can replace 40 decisions. One AGENTS.md can replace 200 onboarding questions.

Key Principles

  1. 01Decision compression: a spec that covers 10 edge cases is worth more than 10 separate conversations
  2. 02Automation ROI: measure what decisions are now made automatically vs. manually
  3. 03Context as leverage: well-designed context multiplies every agent's effectiveness
  4. 04Program office thinking: WorkFlow repo exists to prevent decision-making overhead, not to track tasks

Real Examples from Production

Morning ritual → zero cognitive load

Open terminal. Run vcms-scan.js. Open todo.json. Pick task by priority. No ambiguity about "what should I work on today" — the system decides.

Flex-vcms/flex-vcms/README.md

AGENTS.md eliminates onboarding

Any AI agent entering any of the 8 repos has immediate clarity on: what this repo does, what rules apply, how to deploy, what not to touch. Zero discovery time.

flexgrafik-meta/docs/core/agents.md

Handoff docs as knowledge compression

Each handoff doc compresses a session's decisions into 10 lines. The next agent reads 10 lines instead of replaying 2 hours of reasoning.

agent-os/docs/handoffs/

Traditional vs AI Systems Architect

Traditional Developer

Developer spends 3 hours daily deciding what to work on, re-explaining context to collaborators, and recovering from ambiguity.

AI Systems Architect

System pre-decides priorities (todo.json). Context pre-loaded (brain.md). Agent pre-briefed (AGENTS.md). Developer spends 3 hours building.
Back to DNA