SR-SI accidentally solved the documentation problem too

I arrived at SR-SI to solve context decay. I accidentally solved the documentation problem too.

When you force an AI to maintain a shallow index of architectural decisions, you get comprehensive product documentation as a byproduct, with zero additional effort. Always current. Actually useful.

The constraints are identical:

• For AI self-orientation, the index must be compact, capture why decisions were made, point to relationships between components, and stay current.
• For human comprehension, documentation needs to be concise, explain rationale, show how parts connect, and reflect actual system state.

These are the same constraints.
The documentation that helps AI stay oriented helps humans understand the system.

What you get is layered: the shallow index provides architectural overview and key decisions. In-code comments capture implementation details and rationale. Modular structure keeps components digestible. And because the AI maintains this as it builds (or at key checkpoints), it’s always current.

The traditional documentation problem is real. Writing good docs for a complex system from scratch requires forty to eighty hours. Most teams don’t have that time.

Result: documentation is missing, outdated, or wrong.

SR-SI generates documentation as the AI builds because it needs them to maintain orientation. The kicker? SR-SI methodology generates it automatically as a byproduct.

The quality is sufficient for developer onboarding, stakeholder communication, technical handoff, and understanding the system months later.

Documentation debt kills projects:

• The gap between system complexity and available knowledge creates friction
• New developers can’t onboard
• Stakeholders can’t understand tradeoffs
• And worse: future you can’t remember why that pattern exists

SR-SI makes documentation a natural byproduct of development, not separate overhead. You’re not choosing between fast iteration and good documentation. You get both.

Good AI memory structure equals good human knowledge structure. The AI’s need for clarity creates clarity for everyone.

Next: Why this methodology isn’t for everyone (and that’s okay).

SR-SI: The methodology that gives AI persistent memory across any long-running project