[Notice] Balancing AI Efficiency with Engineering Integrity

The mixed use of AI continues to cause persistent issues with history management.

When utilizing AI, please explicitly specify the model and version used.

Please refrain from using AI to draft the Design Doc, ADR (Architecture Decision Record); furthermore, ensure that any accompanying comments or explanations are written by you personally before sharing.

As we integrate AI into our daily tasks, we’ve noticed some challenges in managing our project history and maintaining a consistent technical context. To keep our documentation reliable and easy to follow, I’d like to propose a few simple guidelines:

1. Identify Your Tools: When you use AI to generate code snippets or logic, please explicitly mention the model and version (e.g., GPT-4o, Claude 3.5 Sonnet). Since different models have different strengths and quirks, knowing the source helps us troubleshoot more effectively if issues arise later.

2. Human-First Design Docs & ADRs: Design Docs and Architecture Decision Records (ADRs) represent our team's core thinking. Please refrain from using AI to draft these foundational documents. These records should reflect our specific constraints and long-term vision, which only a human engineer truly understands.

3. Personal Review & Ownership: Before sharing any AI-generated comments or explanations, please rewrite them in your own words. Every word we share with the team must be something we personally verify and stand behind. This prevents "AI hallucinations" from slipping into our shared knowledge base.

Let’s use AI to boost our productivity without losing the personal touch and accountability that make our engineering team great.