Clean Code in the AI era
summary
Concrete stats and tools: static analysis with SonarQube/SonarCloud/SonarLint finds ~1,800 million issues across telemetry with the top 10 categories totaling >500 million; developers lose ~4 days/month to defects. Generative AI usage grows from ~10% today toward a projected 70% by 2027, boosts speed ~55% but Copilot answers are accepted ~40% and only ~29% fully correct. Practical rules: treat AI as an intern and review its code, prefer cognitive complexity over cyclomatic metrics, use Java features (records, toList, virtual threads) and enforce a zero-tolerance policy on new code via CI quality gates.
AI-written summary · about
chapters (9)
- 0:00Introduction and motivation
- 7:35Poor quality code monster
- 12:36AI code generation helper
- 25:01Top code issues revealed
- 30:06Security vulnerabilities explained
- 35:48Cognitive complexity explained
- 41:17Java 21 best practices
- 42:00Tools for clean code
- 45:56Methodology and takeaways
related talks
46mCode Quality in the AI era, needed more than ever
46mClean Code in AI era
19mWith great power comes great responsibility in code quality
34mBuilding Guardrails for AI Coding Systems
40mCode Rot at Machine Speed: Preventing AI-Induced Technical Debt
41mCode Security Reinvented Navigating the era of AI
19mThe AI Code Review Bottleneck: Escaping the Velocity Trap
14mAI-Powered Code Review and Quality Assurance