Seven Habits of Highly Effective AI Java Coding
summary
Compares models with numbers — DeepSeek 3.1 pass rate ~78% vs GPT 5.2 ~80% but GPT 5.2 emits ~3× the lines of code, so review cost and token spend jump; shows concrete vulnerabilities AI injects (SQL injection, path traversal, hard‑coded creds, typosquatting, lingering Log4j artifacts) and supply‑chain risks from malicious MCP servers. Recommends ownership (human commits, linters, SonarQube CI gates), SDDs and prompt docs, JUnit 5/Quarkus/Java 21 guidance (use virtual threads), and warns tests and k8s manifests from Copilot can be wrong — perceived +20% productivity vs measured −19%.
AI-written summary · about
chapters (13)
- 0:00Introduction and agenda
- 3:42State of AI code generation
- 6:17Model performance evaluation
- 10:22Common AI pitfalls
- 13:20Productivity and the vibe coding trend
- 15:52Habits of effective AI developers
- 27:09Strategic use of documentation
- 33:56Focusing on code simplicity
- 36:58Code guardrails and security
- 39:02Supply chain and verification
- 42:14Infrastructure as code validation
- 43:52Meaningful test generation
- 45:46Conclusion and final remarks
related talks
19mWith great power comes great responsibility in code quality
49mClean Code in the AI era
19mThe AI Code Review Bottleneck: Escaping the Velocity Trap
39mBack to Basics: Cutting Through the AI-Generated Code Noise
26mAl and Software: The Good, The Bad, The Scary
01h 02mPANEL - Unleashing the Future: The Impact of Gen AI on Development Productivity
28mAI-Generated Code: Unmasking the Security Pitfalls
46mCode Quality in the AI era, needed more than ever