Identifying and fixing Issues in Code using AI based tools
summary
Compares ChatGPT and Claude on real snippets and a 100-person exercise and concludes AI reliably finds bugs faster than humans, but still produces drafty code not fit for production. Demonstrations include a hidden parallel streams causing race conditions that breaks when switched to parallel, null-return that benefits from use Optional instead of null, a Wordle duplicate-letter logic bug, an overload/override type confusion, and a prime-summing sqrt routine where AI explains intent and suggests fixes.
AI-written summary · about
chapters (6)
- 0:00Introduction to AI tools
- 6:24Common code errors overview
- 13:10AI detecting code issues demo
- 28:01Understanding unfamiliar code
- 44:36AI analyzing complex code
- 50:06Q&A and summary
related talks
39mBack to Basics: Cutting Through the AI-Generated Code Noise
29mOne-Click Code Fix: Securing Code Using AI
15mUsing AI effectively in large Java codebases
26mAl and Software: The Good, The Bad, The Scary
58mAI, Determinism, Code Generation and Refactoring
46mClean Code in AI era
01h 06mEndnote: AI-Assisted Software Engineering
01h 28mFuture of Software Programming with AI Bots at Hand