Best Practices to make your codebase AI ready
summary
Advice is concrete: skip these patterns if the repo fits 10 million tokens or if defect rate or bus factor dominates, otherwise adopt domain-colocated subsystems that bundle DB schemas, service contracts and APIs, surface product language into code and docs, add broad integration and end to end tests because LLMs create subtle bugs unit tests miss, prefer explicit typed or verbose code so models exploit billion lines of code scale, and try tools like GitHub Copilot and domain agents to validate that the approach actually reduces review time.
AI-written summary · about
related talks
23mBoy Scout Rule in the Age of AI: Leaving Codebase Cleaner than You Found it
11mClean Architecture and AI
49mClean Code in the AI era
47mCode Is Cheap. Software Isn’t.
30mIncorporating AI into your SDLC
19mWith great power comes great responsibility in code quality
31mModern Architecture for AI-Augmented Development Teams
46mCode Quality in the AI era, needed more than ever